Dr Begley was disappointed to note, therefore, that the current study was not randomized, and the investigators were not “blinded to allocation during experiments and outcome assessment.”
“Based on our experience, failure to perform experiments in a blinded manner is always a cause for concern,” Dr Begley said.
The paper highlights the shortcomings of relying on a few preclinical cell line studies in drug development and shows that scientists who are developing new therapies need to plan additional corroborating cell line and animal model experiments, and “more harmonization” and standardization of lab methods and data collection, Dr Bettegowda said.
“This study raises a likely contributing factor influencing replication results, and there are at least 2 major ways to view the implications on replicability,” said Tim Errington, PhD, of the Center for Open Science in Charlottesville, Virginia. “When using a cell line that could evolve over time, if a similar result is achieved in a replication, it indicates whatever alterations occurred to the cell line are not responsible for the original result and that the result is robust to these changes. That’s useful to know [because] it informs us how robust the result might be.”
For example, as part of the Reproducibility Project: Cancer Biology (RP:CB), scientists confirmed that a BET inhibitor, JQ1, selectively downregulated MYC transcription in multiple myeloma cells when they used cells shared by the authors of an earlier study, Dr Errington noted.3
“Conversely, if a different result is achieved when using a cancer cell line, it is likely the cell line has drifted, as this recent paper implicates,” Dr Errington said. “That’s also useful to know, as it suggests limitations, and potential mechanisms, for a given response.” For example, as part of RP:CB, a replication that used transformed melanocytes shared by the authors of an earlier study failed to produce the same effect on tumorigenesis that was seen in the original study. This was likely a result of cell adaptation during extended cultivation in vitro, Dr Errington said.4
Thus, it is key for researchers to retain and share cell lines, Dr Errington noted.
The emerging field of evolutionary tumor biology has focused largely on measuring and predicting cancer cells’ responses to artificial selection pressures from chemotherapy and targeted cancer treatments, rather than the evolution of cancer cell lines in response to culture media or other contact materials.
“People tend to assume that cell lines are homogeneous and stable, but the fact is, all cell lines are living materials,” said Amanda Capes-Davis, PhD, of the International Cell Line Authentication Committee (ICLAC) in Sydney, Australia. “Cell lines are typically made up of many different clonal populations that change and evolve based on their handling and living conditions.”
The study authors looked at genomic variation in commonly used cancer cell lines and compared lines that had undergone more handling with those that have spent more time in the freezer.
“The answer is, there is a lot of impact” from all that handling, Dr Capes-Davis said.
“Your choice of growth medium, the flasks or plates you use, where you obtained your cells, and how long you keep them in culture — these choices can all affect a cell line’s genetic makeup, clonality, and behavior,” she told Cancer Therapy Advisor. “It’s important to document all of these things and to think about how they affect your experimental data and its reproducibility.”
However, the current study does not render the use of cancer cell lines obsolete, Dr Errington emphasized.
“[It] rather illustrates the need to be aware of [multiple] factors and how they can impact the research results and interpretation,” he said.
“There have been a few prior reports regarding cell line genetic drift, including variations in growth conditions and the impact they can have on compound potency, [including] impacts of atmospheric oxygen and variation in cell division rates, but it is great to see a large, systematic approach to evaluate this,” Dr Errington noted.
The findings bolster the case for implementing good cell culture practice standards, Dr Capes-Davis said.5
Good cell culture practice includes authentication testing to spot cell lines that were identified incorrectly, she noted. With genome-wide sequencing costs declining sharply over time, tracking cell line evolution is increasingly feasible. The US National Institutes of Health (NIH) already requires grant recipients to confirm cell line identities as a routine part of research.
“I was intrigued to see that the authors included a known-misidentified cell line, KPL-1, and confirmed the previous finding that it is actually MCF-7,” Dr Capes-Davis said. “[Single-nucleotide polymorphism]-based DNA fingerprinting was used here to authenticate cell lines. This can be a very effective approach.”
The study shows that labs need to do more replication, with multiple in vitro conditions controlled or tested using multiple strains and cell lines to validate conclusions, Dr Bettegowda said. “There need to be multiple lines of evidence that are independent of each other to support published claims,” he explained. That can be done in parallel instead of sequentially to avoid slowing research efforts, he added.
“The thing that most intrigues me here is that the deeper characterization of these clonal populations might be used in a positive way,” noted Dr Capes-Davis. “We tend to see heterogeneity as a problem for [research] reproducibility. But the authors comment that you can use this information for good — for example, by looking for particular phenotypes and exploring the genetic variations that are associated with those populations.”
“That’s an exciting possibility to explore,” she concluded.
- Ben-David U, Siranosian B, Ha G, et al. Genetic and transcriptional evolution alters cancer cell line drug response [published online August 8, 2018]. Nature.doi: 10.1038/s41586-018-0409-3
- Aird F, Kandela I, Mantis C, et al. Replication study: BET bromodomain inhibition as a therapeutic strategy to target c-Myc. eLife. 2017;6:e21253.
- Horrigan SK, Courville P, Sampey D, Zhou F, Cai S; Reproducibility Project: Cancer Biology. Replication study: melanoma genome sequencing reveals frequent PREX2 mutations. eLife. 2017;6:e21634.
- Geraghty RJ, Capes-Davis A, Davis JM, et al. Guidelines for the use of cell lines in biomedical research. Br J Cancer. 2014;111(6):1021-1046.
- Begley CG. Reproducibility: six red flags for suspect work. Nature. 2013;497:433-434.