Patricia F. Fitzpatrick Dimond Ph.D. Technical Editor of Clinical OMICs President of BioInsight Communications

Researchers weigh in on the impact of incorrectly labeled cell lines on cancer research and drug development.

Of 13 human esophageal adenocarcinoma (EAC) cell lines frequently used for cancer research and drug development, three originated from other tumor types, researchers reported in the January 14 online issue of the Journal of the National Cancer Institute (JNCI). University of Rotterdam Medical Center scientists with colleagues at multiple institutions worldwide, found that the SEG-1, BIC-1, and SK-GT-5 cell lines originated from lung, colorectal, and other malignancies. Their analyses, based on data from pathology archives and genotyping assays, confirmed the identities of the 10 other cell lines, however.

“From the scientific point of view it’s not a huge deal, but it’s certainly something you’re glad you found out,” commented Charles Saxe, scientific director of the program in cancer cell biology and metastasis at the American Cancer Society. “This probably doesn’t surprise anybody. The surprise is probably that there were only three.”

Other relatively recent reports have also  found bogus cell lines. In the January 2007 issue of Cancer Letters, Mordechai Liscovitch, Ph.D., and Dana Ravid, Ph.D., from the Weizmann Institute of Science, used DNA fingerprinting to show that the NCI/ADR-RES cell line, which was supposed to be a breast cancer cell line, was actually an ovarian tumor cell line called OVCAR-8. NCI/ADR-RES is included in the NCI panel of 60 cell lines distributed for evaluation of potential anticancer drugs. Dr. Liscovitch and Dr. Ravid estimated that about 300 papers based on the assumption that NCI/ADR-RES was a breast cancer cell line have been published and withdrew their own paper based on this cell line.

The issue of cell-line misidentification has been a huge deal to those involved in basic research. Leading researchers have urged validation of cells either by the organizations that supply them or by individual investigators. In 2007, an open letter to the NIH from Roland Nardone, Ph.D., professor emeritus of the biology department, Catholic University of America, and his colleagues brought the cell-line contamination issue to the attention of Michael Leavitt, then director of the HHS.

By the end of 2007, noting that it was “fundamentally in agreement that misidentified and contaminated cultures constitute a serious, ongoing problem,” the NIH intramural and extramural programs stopped short of mandating cell-line authentication as a requirement for funding. They said that “it would be impractical for the NIH to require application of particular methods in all grant applications,” placing the responsibility on peer reviewers to control their colleagues’ research proposals and manuscripts.

Errant Cell Lines Not So Bad but Validation Is Important

While the use of misidentified cell lines is clearly not the best science, its consequences, at least in cancer drug development, may not be so dire. Scientists point out that cancer is not a heterogeneous disease and that there has been a shift away from defining a cancer type based on its tissue of origin. Thus they assert the need to use multiple cell lines that are continuously validated. If cancer researchers ignore the issue of cell-line contamination and the need for validation, they run the risk of basing their results on one cell line that could turn out to be errant.

“Not only are there many studies that are invalid before they start, but also many others that are successfully published and lead to the dissemination and general acceptance of erroneous data,” noted R. Ian Freshney, Ph.D., of the University of Glasgow and the author of numerous books on cell culture. He cites the story of EVC304, thought to be an endothelial cell line and widely used among scientists working in vascular and angiogenesis research. Even though the cell line was shown to be a T24 bladder carcinoma cell line in 1999, 35 studies using EVC304 for its initial intended purposes were subsequently published between 2007 and 2008.

Similarly, the ongoing clinical trials for Nexavar (sorafenib) in patients with Barrett-related EAC are based on preclinical studies using one of the errant cell lines reported in the January 14 JNCI paper; SEG-1 is actually a lung carcinoma cell line. Nexavar, a small molecule multiple tyrosine kinase inhibitor, targets the Raf/Melk/Erk kinase pathway and is currently marketed by Bayer Healthcare and Onyx Pharmaceuticals for liver and kidney cancers.

In an accompanying editorial in JNCI, Robert Shoemaker, Ph.D., of the screening technologies branch at the NCI’s developmental therapeutics program, said, “To the extent that the induction of mitogen-activated protein kinase pathways and the corresponding high sensitivity to sorafenib are unique to this cell line (SEG-1) or lung cancer, this misidentification may be important.”

He pointed out, however, that cancer is a heterogeneous disease, and the use of only a single cell line in developing the rational for any therapeutic could be questioned. Since mitogen-activated protein kinase pathways are common among many tumor types, “it would not be surprising if using the bona-fide EAC cell lines might provide a preclinical rational for a sorafenib trial.”

The story of the MDA-MB-435 cell line, in use since 1982 as a human breast cancer model and particularly as a drug-screening model, points to the difficulties of cell-line validation, despite the abundance of new characterization tools based on genomic profiling. In 2000, based on DNA microarray assay data, Douglas T. Ross, M.D., Ph.D., who was then at Stanford University School of Medicine, and his team suggested in Nature Genetics that these cells might have originated as melanocytes due to their similarity to melanoma cell lines.

In 2007, James M. Rae, Ph.D., of the University of Michigan Medical Center used karyotyping, comparative genomic hybridization, and microsatellite polymorphism assessments combined with bioinformatic analysis of gene-expression and SNP data. In his paper appearing in Breast Cancer Research and Treatment he concluded that “All currently available stocks of MDA-MB-435 cells are derived from the M14 melanoma cell line and can no longer be considered a model of breast cancer.”

An October 2009, article in Cancer Research, however, proposed an alternative interpretation based on published literature that “both cell lines are indeed identical, but are in fact MDA-MB-435 breast cancer cells.” The author, Ann F. Chambers, Ph.D., of the University of Western Ontario, London, explained, “There is evidence that many cell lines can display lineage infidelity and that assignment to tissue type is unreliably made based on expression patterns.” Dr. Shoemaker explained that this relates back “to one of the real values of the EAC paper, because the investigators were able to match 10 derived cell lines to the original patient sample.” 

In the end, as Dr. Shoemaker commented in his editorial, one answer to the cell-line identity crisis may be the enablement of what he calls a molecular taxonomy to define clinically significant cancer subtypes. He says that in breast cancer, for example, the “multiple subtypes and known heterogeneity would argue for the use of large numbers of cell lines” for cancer research.

“Investigators have the responsibility to attend to this matter and, historically, some have and some haven’t,” Dr. Shoemaker told GEN. “As the horizon expands and people start looking at hundreds of cell lines, it will become more complicated. The good news is that short terminal repeat fingerprinting is done easily and cheaply and is something that people need to do. More and more journals are asking questions about cell-line characterizations and hopefully in the future we will have fewer problems.”

Patricia F. Dimond, Ph.D., is a principal at BioInsight Consulting. Email: [email protected].

Previous articleRoche Expands Agreement with b3 bio to Include Therapeutic Technologies
Next articleResearchers Find Mechanism for Cellular Dysfunction and Apoptosis in Type 2 Diabetes