A Quantised Cyclin-Based Cell Cycle Model

Author(s): Chris Emerson, Lindsey Bennie, Niall M. Byrne, Dermot Green, Fred Currell, Jonathan A. Coulter

Background: Computational modelling is an important research tool with applications in predicting the likely outcome from treatment interventions, to estimating tumour growth characteristics. In silico modelling has been ubiquitously used in cancer research investigating processes such as DNA damage and repair, tumour growth, drug/tumour interactions, and mutational status. On a granular scale, modelling can even be used to better understand the interactions between individual proteins on a single cell basis.

Results: Herein, we present a computational model of the eukaryotic cell cycle incorporating the key proteins involved in cell cycle regulation, namely the cyclin family proteins (cyclin A, B, D and E) and cyclin dependent kinases. Quite uniquely, this model provides a fully quantified output based on western blot and flow cytometry data from synchronous HUVEC cells, enabling determination of the absolute number of cyclin protein molecules per cell. Importantly, this quantitative approach confers more realistic control over threshold transitions between cell cycle checkpoints. The results show that the peak values obtained for the four cyclin proteins are comparable, with cyclin B proteins yielding between 5x106 to 9x106 molecules per cell. Comparing this value against the cytoskeletal housekeeping protein actin (5x108 molecules), illustrates the important functional activity of these cyclin proteins despite expression levels approximately two orders of magnitude lower.

Conclusions: This model advances current approaches by determining absolute cyclin concentrations within an individual cell, with progression through each phase of the cell cycle. Using Boolean variables to represent the genetic network (active/inactive), while integrating continuous variables to represent absolute cyclin protein concentrations, this hybrid approach confers computational efficiency permitting rapid calculation of the protein concentrations, while predicting the influence on cell cycle progression. Subsequent iterations could allow for integration of the cell cycle model into larger tumour models, facilitating the tracking of discrete cells within a developing tumour.