HeLa Cells from Different Labs Vary in Genetics, Phenotype

The results suggest that HeLa cells have evolved into something slightly different in each lab, says Jallepalli. “What we’re seeing is genetic drift. A starting population is evolving into distinct niches over time.”

In the study, researchers gathered 14 HeLa samples from 13 labs across six countries, and cultured them under the same laboratory conditions. They first quantified gene copy number variation—the number of repeats of a given gene—revealing stark differences between their genomes. This was especially notable between the two most widely used strains of HeLa cells, known as HeLa CCL2—considered to be the “original” variant of the cell line—and HeLa Kyoto, an offshoot of the cell line that has properties that make it useful for specific applications such as imaging.

Further analyses showed that many of these genetic differences translated into changes in mRNA production and, to a lesser extent, changes in protein abundance. The transcriptomic and proteomic profiles of HeLa CCL2 and the Kyoto lineages are as different to one another as are cancer cell lines from two different types of tissue, the researchers report.

The HeLa variants also differed in how fast they grow in culture, with some cell populations taking 17.5 hours to double, whereas others took a little more than 32 hours under the same culture conditions. They also differed in their responses to Salmonella infection: One variant was less susceptible to infection compared to two others, which the researchers attribute to low levels of a protein complex that plays a role in the bacterium’s entry into host cells.

In a separate experiment, the team investigated whether gene expression changed in individual HeLa variants over time by culturing a cell line for three months. The researchers documented a roughly 6 percent difference in gene expression between an early and a later generation of cells.

“It was certainly very dramatic how much these cells differed, and how quickly they changed even in the same lab,” remarks coauthor Ruedi Aebersold, a professor at the Institute of Molecular Systems Biology at the ETH Zurich. He estimates that if a graduate student had done an experiment with a HeLa cell line at the beginning of his or her project and were asked to repeat it after six months, “they might have gotten different results.”

Much of the discussion around the “reproducibility crisis” in research has centered on flaws in experimental design, data analysis, and contaminated or mislabeled cell lines as major drivers. But Aebersold thinks the biological differences in HeLa cells—and cancer cell lines more generally—could play a significant role. At conferences, he has often observed researchers getting into heated arguments over obtaining different results from the same experiment, he notes. “The implication would be one made a mistake,” but another explanation is that “the cells may not be the same cells,” he explains.  

See “The Great Big Clean-Up”

Stanford University cell biologist Tim Stearns sees several possible reasons for why cell lines change under prolonged culture. For one, HeLa cells are cancer cells with known genomic instability and are therefore likely to mutate randomly over time. In addition, they’re subjected to various conditions by growth in the lab that might be pushing the cells to evolve unique characteristics. For instance, he says, culturing mammalian cells involves growing them until they fill a dish, siphoning off a fraction of the cells, and placing them in another dish to grow anew—a process called splitting. “Every person does it a little bit differently,” he says. This “[applies] a selection to the cells in ways that we don’t fully understand.”

Fetal bovine serum—a main ingredient of the growth factor cocktail used to culture mammalian cells—can also vary between labs. “It is not difficult to imagine that based on the source of that material we would create different transcriptional profiles and different selective pressures,” Jallepalli explains. “Even simple, humble things like plastic dishes are likely doing more than we realize.”

The variation between HeLa isolates may worry some life scientists more than others, notes Jallepalli. Molecular biologists and biochemists who use HeLa cells to study universal cellular processes such as DNA replication or vesicle trafficking are less likely to be concerned that their results may not be reproducible, because these processes are unlikely to change in the face of such selective pressures. Developmental and cell biologists who study more-complex traits such as Salmonella infection might have more reason to worry.

Aebersold and his colleagues propose several specific solutions in their paper. For one, researchers ought to use early passages of cancer cell lines and make sure to repeat experiments from one cell line in different samples of the same cell line. Importantly, biologists should clearly report which cell line variants they are using in a given study. “A lot of people aren’t even sure what kind of HeLa cells they have,” notes Stearns, who agrees that more transparency would be a positive change.

Aebersold hopes to identify more solutions in a workshop he is planning later this year with the European Molecular Biology Organization. This will include 25 experts from various fields, such as science policy, publishing, and science in order to come up with recommendations on how to address reproducibility issues in cell line research.

Ultimately, he hopes his work will help enlighten science policy on the causes of irreproducible results in scientific research. The notion that scientists can’t replicate one another’s work is a dangerous one, he says. “The simplistic conclusion is either we don’t know what we’re doing as life scientists, or worse, that things are made up.” This would make it easier for science-averse policymakers to argue that money spent on research is wasted. “I think it is important to provide evidence that it is not so simple, people do not just cheat, people are not incompetent, but it’s more complicated.”