Cancer Static Shock

When it comes to anticancer drug development, the difficulties of translating in vitro efficacy into clinical success are well known. But what if the metrics scientists use to measure a drug’s effect on cancer cell growth, in vitro, are inherently flawed?

A group of researchers from the Department of Cancer Biology at Vanderbilt University School of Medicine, US, believe that in vitro cell proliferation assays suffer from a number of biases (1). In response, they have developed a new metric, called ‘drug induced proliferation (DIP) rate’. Darren Tyson, Assistant Professor of Cancer Biology and lead author of the study, tells us more.

In what ways are current protocols flawed?

The use of a single measurement of cell number is widely employed across the scientific literature. Since it is based on a single time point measurement, we refer to this type of metric as “static”. Static metrics are flawed in multiple ways. Firstly, because cells grow exponentially, an untreated population will rapidly outgrow a drug-treated population. Perhaps more critically, the ratio of control to drug-treated cells will steadily increase over time, creating the illusion that a drug’s effectiveness is increasing over time. This is an example of what we call “time-dependent bias”.

Another source of time-dependent bias in static metrics is that many drugs exhibit a lag time before their effect stabilizes within a cell population. This stabilization delay can cause drugs to appear more or less potent or effective than they actually are, which means ineffective compounds may be being improperly passed through the drug discovery pipeline or, conversely, effective drugs may be discontinued prematurely.

How does your proposed DIP rate metric differ?

The DIP rate quantifies the growth of a cell population, or more precisely, the rate of change of a cell population size over time. Since the most important characteristic of a cancer drug is whether it can halt or reverse tumor growth, DIP rate is a natural and accurate metric: on a plot of cell population doublings (log2 cell counts) vs time, it appears as the slope of a line. As such, it is independent of time, once any delays in drug action have been accounted for.

When developing the DIP rate metric, our biggest challenge was to determine when, after drug addition, a proliferation rate has stabilized. To support high-throughput drug screens, we had to develop reliable computational methods that could determine, in an automated fashion, when this occurs. The software is written in the R programming language and for academic applications can be obtained as free, open-source software (2).

What impact do you think this will have on the pharma industry?

If in the next few years proliferation rate-based metrics replace static cell counts in cell proliferation assays, we expect increased success rates in drug and biomarker discovery efforts to follow. We, in addition to our colleagues in companion papers, report examples of how DIP rate (and related metrics) can identify novel correlations and eliminate artifactual ones, providing a strong proof-of-concept for converting to time-independent metrics.

What are your next steps?

We want to measure DIP rates in large panels of cancer cell lines and search for novel molecular biomarkers of drug sensitivity, in addition to investigating DIP rate metric predictions of tumor cell responses in vivo. Both are translational tools for precision medicine.

We view the DIP rate as a metric of cell fitness in a particular environment, which extends beyond oncology. The DIP rate can act as a common currency, whether studying, for example, the influence of different drugs across a variety of cell lines, the effects of altering the microenvironment of stem cells, or the variation that exists at the single-cell level within a cell population (clonal heterogeneity, competition, or evolution).