NEW YORK (GenomeWeb) – Last week, researchers from the Human Toxome project reported in Nature's Scientific Reports that they observed genetic variability between samples from the same cell line, which resulted in research delays and lost funds.
The cell line in question, Michigan Cancer Foundation-7 (MCF-7) was first isolated in 1970 and has been held as a standard model for in vitro cancer for over 40 years. According to the paper, "almost 23,000 articles using MCF-7 can be retrieved in PubMed." Considering its widespread use, these findings could have significant implications for the research community.
The Human Toxome project is a collaborative research initiative that has received $6 million in funding from the National Institutes of Health and involves researchers from Johns Hopkins University, Brown University, Georgetown University, the Hamner Institutes for Health Sciences, and Agilent Technologies. They are attempting to understand how the human body creates toxicity pathways when it's dealing with different toxins; something that is currently poorly understood.
"[The project is an] attempt to revamp safety sciences," Thomas Hartung, the corresponding author on the study and a researcher from Johns Hopkins, told GenomeWeb. "You want to be safe with the chemicals that you encounter. There are nearly 83,000 chemicals [and toxicants that humans encounter], and we have safety data for about 3 percent of them."
In conducting their experiment, researchers at Johns Hopkins and Brown each received MCF-7 samples from the same batch and practiced the same protocols. "We were very meticulous to ensure the quality of our work," Hartung said. Working in two separate labs served two purposes: it ensured the quality of the data and helped split the workload, he added.
However, even though Hartung and his colleagues followed current validation protocols and procedures in both labs, they weren't getting the same results. "Differences persisted even when the two groups exchanged their cell cultures," Heidi Kijenski, global marketing director for Agilent's diagnostics and genomics division told GenomeWeb via email.
"It became clear that the cause was in the cells themselves — not in the lab protocols, reagents, or handling by the groups," Kijenski noted. "This was particularly surprising since the cells were not only obtained from the same cell bank, they had the same lot number."
Fortunately, the researchers did have two unused vials of the MCF-7 cells left over, one from each lab. They noted differences after they performed experiments using Agilent gene expression microarrays and found that the cells were not identical. "We saw significant differences between the cells in the expression of certain estrogen-responsive genes at the baseline level," Kijenski said.
The researchers confirmed their findings by performing comparisons between the relative gene abundances between the cells at each lab site using Agilent's GeneSpring GX software.
Using Agilent's comparative genomic hybridization (CGH) array technology, they found that "it was not a simple case of cellular contamination; the cells were indeed MCF-7s," Kijenski told GenomeWeb.
The genetic differences didn't show up using standard validation protocols, demonstrating that current quality assurance methods aren't effective enough, Hartung suggested.
Researchers have noted in the past that the MCF-7 cell line was variable, and though it's not the only cell line with such genetic variability, there hasn't been another cell line that's been identified to have variability to this extent. "It's only one of many unstable cell lines," according to Kijenski. "The second most common breast cancer cell line, T47D, is also known to have genetic and phenotypic heterogeneity."
HeLa cells, the first human tumor cells isolated in 1952 at Johns Hopkins, have also been noted to have some variability problems. This cell line has been widely used in research and appears in over 70,000 published studies. But as recent as last year researchers were pointing out to the scientific community that people working with HeLa cells "are not [all] working with the same stuff," Hartung said.
Hartung and his colleagues have already had major setbacks because of the genetic variability issues in the MCF-7 cell line, and a primary issue is that it undermines reproducibility.
In this case it caused huge delays in the research and set the research team back about a million dollars in research costs. "We certainly lost two years of time," Hartung said. "It broke my heart."
One possible solution to the genetic variability issues in the cell lines themselves is for scientists to encourage action from the big cell banks, Hartung said. These banks are normally very responsive to the problems encountered in the scientific community and have been responsible for setting scientific standards in the past, including many current validation standards, he added.
"Cell banks and labs which use their strains would not need to adopt [these protocols] themselves," Kijenski clarified. "Running a test can be as easy as sending DNA to a provider."
However, that is only one smaller aspect of a larger problem. "This study indicates that cell authenticity is unlikely the major source of irreproducibility in in vitro research," Kijenski said. "It is well known that genetic drift occurs in cultures and that this problem is exacerbated in tumor cell lines with high degrees of genomic instability. The new and unexpected lesson we learned from this study was that acquiring cells from the same lot of a tissues bank [and] minimizing [the number] of passages required to expand the cells for the study is not always sufficient to get reproducible results."
For Agilent, the only company involved in the Human Toxome project, the research has allowed it to develop better technology for the toxicology testing market. "Most in vitro assays for environmental toxicity testing [used] today were developed 40 to 50 years ago," Kijenski said. These tests "pre-date modern mechanistic understanding of molecular and cell biology."
One of the major beneficiaries of the firm's participation in the Human Toxome project has been Agilent's bioinformatics software. "Numerous new functionalities and workflows were added to Agilent GeneSpring/MPP and several other software products based on in-depth understanding of the project data and its analyses needs," Kijenski said.