Troubles with VSELs

A stem cell research group at the University of Louisville, Kentucky — famous for apparently discovering an exciting new class of stem cells — could be facing new troubles.

Although the work of Mariusz Ratajczak was supported through large NIH and Vatican grants, no other lab could replicate his findings on very small embryonic-like stem cells (VSELs).

And now, 28 papers from the Ratajczak lab are listed on PubPeer for image duplication and textual similarity concerns.

March 2022 update: No misconduct found

The University of Louisville sent their final decision – dated July 29, 2021 – to me today – March 23, 2022 (shared here with permission).

“The University of Louisville found there was no research misconduct. The institution followed its established, thorough, and robust process and made no findings of research misconduct against Dr. Ratajczak associated with any of the allegations, including all the allegations that continue to persist publicly on the internet. ”

[Link to full letter]

A great discovery backed by the Vatican

Mariusz Z. Ratajczak has received several awards for his work on stem cells. Around 2006, he discovered the presence of stem cells in the bone marrow of adult mice. These adult stem cells, known as VSELs, are primitive, undifferentiated cells with the potential to develop into many different kinds of specialized cells, such as heart, bone, brain, or pancreatic cells.

Ratajczak’s discovery of these VSELs was exciting because it could theoretically lead to all kinds of new treatment options. Patients with heart disease, burnt skin, diabetes, or neurodegenerative disorders, for example, might one day be cured simply by having their own bone marrow cells harvested and programmed to develop into specialized tissues that could restore their health.

This technique, harvesting bone marrow cells from adults, was in particular attractive to people who were opposed to using cells from human embryos. And when the company NeoStem began working with the patents filed by the Ratajczak lab in 2011, the Vatican took the unusual step of investing $1 million in the New York-based company.

NeoStem, now known as Caladrius Biosciences, also received a $1.2 million grant from the National Institute of Health for a project to repair bone defects using VSELs.

VSELs not confirmed by other labs

But other stem cell researchers were simply not convinced that VSELs can be programmed into differentiation, nor that VSELs were even real.

The lab of Irving Weissman at Stanford tried hard to replicate the Ratajczak findings, but was unable to detect these cells at all. In a 2013 paper in Stem Cell Reports, Weissman’s lab concluded that the existence of adult mouse VSELs was at best dubious.

“We tried as hard as we could to replicate the original published results using the methods described and were unable to detect these cells in either the bone marrow or the blood of laboratory mice.”, Dr. Weissman wrote in a Stanford press release.

An accompanying Cell Stem Cell profile article is also highly critical of the work by the Louisville lab. George Daley, back then the Director of Stem Cell Transplantation at Children’s Hospital, Boston said: “I find the work mystifying and lacking in rigor.’’

The headline of a July 2013 RawStory.com article on VSELs was even harsher.

Concerns in 26 papers

Not only have other labs found it impossible to reproduce Ratajczak’s work, his papers also appear to contain irregularities. In an investigation I carried out last year after receiving an anonymous tip, I confirmed image and textual similarity concerns in about two dozen papers from the Louisville lab. All concerns have been posted on PubPeer as of yesterday. There may be a couple more not yet listed on PubPeer, as I’m still going through the large volume of papers and similarities.

Duplicate figures across different papers

Sixteen of the 28 papers have concerns about image reuse. In most cases, it appears as though the images were representing the same experiments. This would be acceptable, so long as the paper acknowledged that the figure had been published before. However, in most of these cases the same figure appeared in multiple papers without citing older publications, while claiming that the results were novel.

In one case, the same figure appears to have been published in four different papers by Ratajczak et al.

Overlapping images

Some papers appear to contain figures with overlapping panels. In these cases, panels within the same figure seem to show areas of overlap, which I have highlighted using boxes of the same color. The panels are labeled differently, so they represent different experiments, and such an overlap would thus be highly unexpected.

Here is a figure from Wysoczynski et al. International Journal of Cancer (2010), DOI: 10.1002/ijc.24732:

An example from Kakar et al., Oncotarget (2017), DOI: 10.18632/oncotarget.20170:

And here is a figure from Libura et al., Blood (2002), DOI: 10.1182/blood-2002-01-0031 :

Duplications within photos

In some papers, the photos themselves appear to contain duplicated features, something that seems hard to explain as the result of an honest error.

Here is an example of such a potential duplication, from the Ratajczak lab’s highly cited Leukemia 2006 paper, DOI: 10.1038/sj.leu.2404171, cited 782 times:

Here is a paper by Shirvaikar et al, Stem Cells and Development (2011), that appears to show duplications within photos of gels and blots:

A 2018 PLOS ONE (2018) paper by Marlicz et al., DOI: 10.1371/journal.pone.0189337 may contain some irregularities in photos of gels:

A paper by Golan et al., Blood (2012), DOI: 10.1182/blood-2011-06-358614, appears to show highly similar — yet slightly different — flow cytometry panels:

And something is definitely not right in this photo from Huang et al., American Journal of Transplantation (2016), DOI: 10.1111/ajt.13511:

Text recycling?

Several papers by Ratajczak et al. appear to contain recycled text, wording that is highly similar in different papers. Text recycling, sometimes called self-plagiarism, is different from plagiarism, in which someone else’s text is copied and passed off as one’s own. Here, text is reused by the same researcher, but passed off as novel. Textual similarities are easy to prove using software such as SimTexter.

Here are some examples, with similarities indicated using highlighted text:

A list of all papers with PubPeer posts

This is a spreadsheet of the papers from the Ratajczak lab with image and textual similarities; 28 as of today. The list could grow, as I may confirm additional findings.