Cell-cycle models
Cell-cycle models describe how a cell moves through growth, DNA replication, preparation for division and mitosis. The aim is to understand how molecular regulation makes these transitions occur in the correct order.
The biological cycle
| Stage | Main biological role |
|---|---|
| \(G_1\) | cell growth and preparation for DNA replication |
| \(S\) | DNA synthesis |
| \(G_2\) | preparation for mitosis |
| \(M\) | chromosome segregation and cell division |
Why ordinary differential equations are useful
Let \(x_i(t)\) represent concentrations or activities of important regulatory proteins. A cell-cycle model can be written generally as
\[\boxed{\frac{dx_i}{dt}=\text{production}+\text{activation}-\text{inactivation}-\text{degradation}.}\]The equations describe how molecular regulation changes continuously through time, while the resulting dynamics can produce sharp biological transitions.
Cyclins and CDKs
Cyclins are proteins whose abundance changes during the cell cycle. Cyclin-dependent kinases, or CDKs, become active when associated with appropriate cyclins and regulatory modifications.
A very simple cyclin model is
\[\boxed{\frac{dC}{dt}=s-k_dC,}\]where \(C\) is cyclin concentration, \(s\) its synthesis rate and \(k_d\) its degradation rate.
If both rates are constant, the system approaches a steady state. That alone cannot generate a repeating cell cycle.
Why feedback is needed
Cell-cycle regulators influence the processes controlling their own activation and destruction. This creates nonlinear feedback.
For example, an active CDK may promote mechanisms that activate still more CDK. This is positive feedback and can create a rapid switch from low to high activity.
A simple activation switch
Let \(X\) denote the fraction of a regulatory protein that is active. A phenomenological positive-feedback model is
\[\frac{dX}{dt}=b+\alpha\frac{X^n}{K^n+X^n}-\delta X.\]The Hill term represents self-reinforcing activation. For sufficiently strong nonlinearity, the production and removal curves can intersect three times.
Why a switch is useful
A gradual molecular input can then trigger a relatively abrupt change in regulatory activity. Once a threshold is crossed, positive feedback rapidly reinforces the transition.
This helps make cell-cycle transitions decisive rather than leaving the cell indefinitely in an intermediate state.
Bistability and memory
If two stable equilibria coexist, the regulatory system can remain in either a low-activity or high-activity state. This is bistability.
The unstable equilibrium between them acts as a threshold. Crossing it can move the system from one stable state to the other.
Hysteresis
In a bistable system, the threshold for switching on can differ from the threshold for switching off. This phenomenon is called hysteresis.
Hysteresis makes a transition more robust to small fluctuations and helps enforce directionality in cell-cycle progression.
But a switch alone does not create a cycle
A bistable switch can change state and remain there. A cell cycle must eventually reset so that the next cycle can begin.
This requires additional regulation, often involving delayed negative feedback or controlled degradation.
A minimal oscillator idea
Consider two variables:
\(C(t)\): a cyclin-like activator that accumulates,
\(D(t)\): a degradation activity induced by the cyclin.
A simple nonlinear model is
\[\frac{dC}{dt}=s-kDC,\]\[\frac{dD}{dt}=\varepsilon\left(\frac{C^n}{K^n+C^n}-D\right).\]Cyclin builds up. High cyclin gradually activates the degradation machinery. Degradation then removes cyclin. Once cyclin falls, degradation activity also declines, allowing cyclin to accumulate again.
The feedback loop
This is delayed negative feedback: cyclin eventually promotes its own removal indirectly.
The delay is important. If the correcting response were instantaneous, the system might simply settle to a steady state instead of producing repeated dynamics.
A model-generated cell-cycle oscillator
To show the basic mechanism clearly, the graph below uses a standard two-variable relaxation oscillator. The variables can be interpreted abstractly as a fast cell-cycle activator and a slower recovery or inhibitory variable.
For this illustrative dynamical system,
\[\frac{dx}{dt}=x-\frac{x^3}{3}-y+I,\]\[\frac{dy}{dt}=\varepsilon(x+a-by).\]With the parameters used in the graph, the nonlinear fast–slow feedback produces sustained oscillation. This is a generic oscillator used here to explain the mathematical principle; detailed biochemical cell-cycle models use mechanistic cyclin–CDK reaction networks.
Fast and slow variables
Many cell-cycle models naturally contain different timescales. Activation of a protein may occur rapidly, while protein synthesis or degradation changes more slowly.
Fast–slow dynamics can produce long periods of gradual change separated by rapid transitions. This resembles biological progression through phases followed by comparatively abrupt checkpoint transitions.
Checkpoints
Cell-cycle checkpoints prevent progression when important conditions are not satisfied. Examples include incomplete DNA replication or DNA damage.
Mathematically, a checkpoint signal can modify an activation rate, inhibit a regulator or move the switching threshold.
For example, if damage signal \(H\) inhibits activation, one might write
\[\frac{dX}{dt}=\frac{\alpha}{1+(H/K_H)^m}\frac{X^n}{K^n+X^n}-\delta X.\]Increasing \(H\) reduces the activation term and can prevent the system from crossing the transition threshold.
Growth and cell size
In some organisms, progression depends partly on cell size. If cell mass \(M\) grows approximately exponentially,
\[\frac{dM}{dt}=\mu M,\qquad M(t)=M_0e^{\mu t}.\]A size-dependent regulator can connect growth to cell-cycle entry.
This illustrates how physiological variables can be coupled to the molecular control network.
Resetting after division
At division, some variables change discontinuously. Cell mass may approximately halve, molecular contents are partitioned between daughter cells, and regulatory states can reset.
A model may therefore combine continuous ODE dynamics between divisions with discrete division events. Such a model is called a hybrid dynamical system.
Population versus single-cell models
A molecular cell-cycle model follows regulatory variables inside one cell. A population model instead follows numbers of cells in different cycle stages.
For example, cells may move through compartments
\[G_1\rightarrow S\rightarrow G_2\rightarrow M.\]These two modelling levels answer different questions and should not be confused.
Age-structured cell-cycle models
Another approach tracks how long cells have spent since division or within a particular phase. This can require age-structured differential equations rather than a small system of ODEs.
Stochastic cell cycles
Real cells do not divide at exactly identical times. Molecular reactions fluctuate, checkpoint durations vary and daughter cells inherit unequal molecular numbers.
Stochastic models can represent random switching, variable phase durations and noisy molecular regulation.
Noise and thresholds
Near a regulatory threshold, molecular noise can determine when a cell crosses from one state to another. Two genetically identical cells can therefore enter a new cell-cycle phase at different times.
Bistability and feedback can nevertheless make the final states relatively robust once the transition has occurred.
Detailed biochemical models
Realistic cell-cycle models may contain cyclins, CDKs, phosphatases, ubiquitin-mediated degradation systems and checkpoint regulators. These are represented by coupled nonlinear reaction equations.
The purpose of simplified models is not to reproduce every molecule, but to isolate the dynamical mechanisms responsible for switching, oscillation and ordered progression.
What mathematics is used?
Cell-cycle modelling brings together biochemical reaction kinetics, nonlinear ODEs, equilibria, nullclines, Jacobian stability, bifurcations, limit cycles, fast–slow systems, stochastic processes and sometimes hybrid or age-structured models.
A practical modelling workflow
Choose the cell-cycle transition or regulator of interest. Identify the essential activating and inhibitory interactions. Translate them into production, activation, degradation and feedback terms. Find equilibria and test their stability. Look for thresholds and bistability. Determine whether delayed negative feedback can generate oscillation. Compare the predicted period and transition timing with experimental data. Add checkpoints, stochasticity or division events when required by the biological question.