The integration of computational models into cancer drug development adds to predicting and understanding cancer drug efficacy and toxicities early on, hence contributing to research efficiency and productivity.
DNA replication and other kinetic events that occur during S-phase of the cell cycle are of great interest to both basic and clinical oncology researchers, and so quantitative analysis of DNA content has been around since the early years of flow cytometry. However, single parameter DNA content analysis cannot detect whether cells with a DNA content equivalent to that of S-phase cells indeed replicate DNA, an important distinction in light of kinetically inactive S-phase cells in tumors . This technical limitation was solved with the introduction of 5-bromodeoxyuridine (BrdU) as a marker of DNA replication, allowing the identification of cells that do progress through S-phase. The technique has been applied extensively in basic research as well as in clinical oncology to evaluate the duration of the DNA synthesis phase . Furthermore, detection of BrdU combined with cyclins expression and DNA content provided information on a time relationship between initiation and termination of DNA replication in relation to cyclin expression .
The efficacy of small molecule, anticancer agents that react with DNA depends significantly on access to target DNA sequences, and timing plays a very prominent role since single-stranded DNA sequences at the telomeric end might only be accessible briefly in the cell cycle during replication . Screening methodologies for early drug discovery efforts using human cell lines by cytometry uncover potential new leads. However, many cell lines present complex genetic backgrounds (cell cycle checkpoint defects) not translatable to normal or other cell cycle progression profiles even in asynchronous, nonperturbed populations . And on the other hand, BrdU (or EdU) labeling is not a high-throughput technique to quickly evaluate candidates and may not truly represent the DNA replication rate .
In this issue of Cytometry A (page 785), Li and coworkers present a novel mathematical model for simulating the process of DNA replication in cycling cell populations. Over the years, mathematical modeling of the cell cycle allowed researchers to quickly connect observation with theory and, more recently, has opened opportunities to design experiments that test predictions in a nonbias fashion, overcoming technical limitations. Cell-cycle models also have an impact on drug discovery. Chassagnole et al.  used a cell-cycle model to quantitatively predict cytotoxicity of a set of kinase inhibitors based on IC50 values, which were measured in vitro. The results allowed them to assess the pharmaceutical value of these inhibitors as anticancer therapeutics. The model was able to predict over three orders of magnitude the cytotoxicity of each compound without model adaptation to specific cancer cell types. In addition, the models have the potential to speed evaluation of cell cycle-directed therapies, as single agents and when used in combination, to assess the optimal sequence of administration for avoiding cell cycle-mediated drug resistance .
Today, our knowledge of the cell cycle regulatory mechanisms continues to advance very much in parallel with the development of cytometric technologies, including laser scanning cytometry and imaging flow cytometry  coupled with the associated growth in computational platforms for robust analysis . Those imaging cytometry methodologies offer the increased value of allowing correlating cytometric data with actual visualization of the individual cells to evaluate changes in morphology and subcellular localization of markers of interest, but in detriment of analytical speed. This has been partially overcome with the introduction of novel approaches like microfluidic image cytometry for cell cycle analysis, which provides higher throughput, while offering additional efficiencies with regard to cost and generation of toxic waste .
Progress in cell cycle analysis and with that its complexity will only dramatically increase with the latest adaptation of mass cytometry to the analysis of the cell cycle as a postfluorescence era requiring additional computational support for data interpretation .
In short, the innovative approach proposed by Li and coworkers describes a multiscale mathematical model simulating DNA replication and cell cycle progression of individual cells and producing in silico EdU/DAPI scatterplots. The model is then compared with experimental data obtained in A549 cells and measured by laser scanning cytometry. The authors took into account the relatively slow DNA replication rate at entrance to S-phase, compared with its rather abrupt termination during the S to G2 transition, and proposed using the observation to simulate the effect of anticancer drugs on DNA replication/cell cycle progression. This approach seems particularly applicable to early drug discovery projects as a nonbias screening tool to identify target lead candidates and to rapidly define the best sequence of administration of combination chemotherapy as a mechanism to overcome cell cycle-mediated resistance.
The effort described here exemplifies progress toward advanced computational approaches in biomedical research, complementing experimental studies and potentially indicating research avenues. One would expect future mathematical models to account for additional imaging data like the subcellular localization of cell cycle proteins as such is more and more recognized as a major factor that regulates cell cycle transitions.