The authors of a 2005 Cancer Research paper that cast some doubt on the safety of a population of adult stem cells used frequently in research have retracted it. According to the retraction, in the August 15 issue of the journal:
Upon review of the data published in this article, the authors have been unable to reproduce some of the reported spontaneous transformation events and suspect the phenomenon is due to a cross-contamination artifact.
Transformation refers to changes in the cells that make them “immortal” — think HeLa cells, made popular by Rebecca Skloot’s book about Henrietta Lacks — and cancer-like.
Only five of the study’s seven authors agreed to the retraction.
When it was originally published, the retracted study garnered some press coverage, and has been cited 284 times, according to Thomson Scientific’s Web of Knowledge — a lot for a five-year-old paper. If adult stem cells researchers hoped to use to treat conditions such as heart disease might actually lead to tumors, that would obviously inject some caution into the field.
Despite this retraction, however, that caution is likely to remain, given a number of other studies with similar findings. In fact, the types of cells described in the retracted study — mesenchymal stem cells — have such affinity for tumors that some have proposed using them to target cancers with gene therapy.
Of note, the retraction follows the publication of a letter to the editor in the August 1 issue of the same journal by the authors of several other studies that found mesenchymal stem cells can become cancer-like. Those authors write:
Inspired by the recent focus on misidentification of cell lines in which cell culture repositories indicate that between 18 to 36% of the cell lines contain misidentified species, we did DNA fingerprinting…analysis comparing the normal [mesenchymal stem cells] with their transformed counterparts. The analysis shows that the transformed mesenchymal stem cells in one laboratory were cross-contaminated with human fibrosarcoma or osteosarcoma cell lines, whereas in the other laboratory cross-contamination was due to two glioma cell lines.
Fibrosarcoma, osteosarcoma, and glioma are all types of cancers, which means that rather than turning into cancer-like cells, the “mesenchymal stem cells” the researchers watched turn into malignant cells may have just been cancer cells to begin with.
The letter, which includes a reference to the now-retracted paper suggesting it is in the same boat, continues with some suggestions for preventing such mishaps in the future:
We believe our findings, although dismal for our research program, convey important messages to the scientific community and to scientific journals. First, because human errors may occur in any laboratory despite stringent working procedures, DNA fingerprinting should be compulsory for all experiments involving cell lines. Scientists should verify the cell lines in their possession and use electronic databases of authenticated DNA profiles against which they can compare their results. Second, scientific journals should require that all cell lines used in an article are verified before publication. Third, funders of grant proposals should encourage expense estimates for cell line verification, in recognition that such verification procedures will increase the costs of research.
We tried reaching the retracted paper’s lead author and the journal’s editor for comment, as well as the author of a major review that cited the study and one of the authors of the letter to the editor. We’ll update if we hear back.
In the meantime, the authors of the August 1 letter about related research say their experience has important lessons:
Subsequent to a learning experience like this, it is important to remember how great thinkers in science have recognized this type of event. In his major disclosure on the scientific method in An Introduction to the Study of Experimental Medicine (1865), Claude Bernard describes what makes a scientific theory good and what makes science important (16). He states “it is always instructive to acknowledge an error. The precept, therefore, is excellent for we are all likely to make mistakes, except those of us who do nothing. But the first requirement in acknowledging a mistake is to prove that there is an error. It is not enough to say: I was mistaken; we must say how we were mistaken; the important point is precisely that.” He further states: “To sum up, we must maintain the conviction that negative facts are determined like positive facts. We posited the principle that all experiments are successful, in that their conditions are determined; in research into the conditions of each of these determinations lie the lessons that teach us the law of a phenomenon; because in this way we learn the conditions necessary to its existence and its nonexistence.”
Update, 9:20 a.m. Eastern, 8/17/10: A few minutes after this post went live, Rolf Bjerkvig, of the University of Bergan, Norway — and one of the authors of the August 1 letter to the editor — responded by email to our emailed questions:
It is highly likely that this is due to our letter. I will suggest that you read our letter carefully, since I truly hope that this will have impact in the field. This relates to the fact that I now believe journals should implement stringent criteria for cell line identification. A topic that is very important and also highlighted by a recent article in Nature Reviews Cancer.
I also believe that our conclusions may be positive for people working on MSCs as therapeutic vehicles, even though it is not clear to what extent MSCs contribute to cancer development.
Update, 9:30 Eastern, 9/9/10: The editor of Cancer Research forwarded our questions to the journal’s publisher, the American Association for Cancer Research, which declined comment. Also see an update with comment from author Javier Garcia-Castro.
Thanks to Marilyn Mann, who tipped us off to this retraction.