Different rates of DNA replication at early versus late S-phase sections: multiscale modeling of stochastic events related to DNA content/EdU (5-ethynyl-2'deoxyuridine) incorporation distributions

Biao Li,Hong Zhao,Paulina Rybak,Jurek W Dobrucki,Zbigniew Darzynkiewicz,Marek Kimmel
DOI: https://doi.org/10.1002/cyto.a.22484
Abstract:Mathematical modeling allows relating molecular events to single-cell characteristics assessed by multiparameter cytometry. In the present study we labeled newly synthesized DNA in A549 human lung carcinoma cells with 15-120 min pulses of EdU. All DNA was stained with DAPI and cellular fluorescence was measured by laser scanning cytometry. The frequency of cells in the ascending (left) side of the "horseshoe"-shaped EdU/DAPI bivariate distributions reports the rate of DNA replication at the time of entrance to S phase while their frequency in the descending (right) side is a marker of DNA replication rate at the time of transition from S to G2 phase. To understand the connection between molecular-scale events and scatterplot asymmetry, we developed a multiscale stochastic model, which simulates DNA replication and cell cycle progression of individual cells and produces in silico EdU/DAPI scatterplots. For each S-phase cell the time points at which replication origins are fired are modeled by a non-homogeneous Poisson Process (NHPP). Shifted gamma distributions are assumed for durations of cell cycle phases (G1, S and G2 M), Depending on the rate of DNA synthesis being an increasing or decreasing function, simulated EdU/DAPI bivariate graphs show predominance of cells in left (early-S) or right (late-S) side of the horseshoe distribution. Assuming NHPP rate estimated from independent experiments, simulated EdU/DAPI graphs are nearly indistinguishable from those experimentally observed. This finding proves consistency between the S-phase DNA-replication rate based on molecular-scale analyses, and cell population kinetics ascertained from EdU/DAPI scatterplots and demonstrates that DNA replication rate at entrance to S is relatively slow compared with its rather abrupt termination during S to G2 transition. Our approach opens a possibility of similar modeling to study the effect of anticancer drugs on DNA replication/cell cycle progression and also to quantify other kinetic events that can be measured during S-phase.
What problem does this paper attempt to address?