Why Track Cells and Free Virus Separately?
An acute viral infection involves interacting populations that change on different time scales. Susceptible cells are the potential sites of new infection; productively infected cells generate virus; and free virus can infect additional cells or be cleared. Tracking these populations separately makes it possible to express those processes as rates and to examine how their interaction shapes a within-host trajectory.
The basic target-cell-limited model represents infection as depletion of susceptible cells by virus. It is a deliberately compact model: it does not explicitly represent immune responses, an eclipse phase, or target-cell replacement. Its purpose is to describe how infection can grow and then diminish when the supply of susceptible cells is limited and infected cells and free virus are lost.
States, Processes, and Assumptions
The term \(\beta TV\) is the rate at which susceptible cells become infected. It removes cells from \(T\) and adds them to \(I\), so it appears with opposite signs in the first two equations. Infected cells are lost at rate \(\delta I\). They produce free virus at rate \(pI\), while free virus is cleared at rate \(cV\). Thus, the virus equation is a balance between production and clearance, not a direct measure of infection alone.
If cell concentrations are measured in cells/mL, virus concentration in infectious virions/mL, and time in days, then \(\beta\) has units mL/(virion·day), \(\delta\) and \(c\) have units day\(^{-1}\), and \(p\) has units virions/(cell·day). For example, \(\beta TV\) then has units cells/(mL·day), as required for \(dT/dt\). Consistent units are an important check against placing a term in the wrong equation.
The model assumes homogeneous mixing: each susceptible cell experiences infection according to the same mass-action term \(\beta TV\). It also assumes that infection immediately creates a productively infected cell; that infected cells have a constant per-cell loss rate; and that production and clearance are proportional to their respective state variables. There is no baseline target-cell source or loss, and no explicit immune or treatment state. These are modeling assumptions, not claims that those processes are absent in a real host.
Worked Example: Interpreting the Initial Rates
Consider hypothetical initial concentrations \(T(0)=1.00\times10^6\) cells/mL, \(I(0)=100\) cells/mL, and \(V(0)=1{,}000\) virions/mL. Let \(\beta=2.00\times10^{-7}\) mL/(virion·day), \(\delta=0.500\) day\(^{-1}\), \(p=50.0\) virions/(cell·day), and \(c=5.00\) day\(^{-1}\). Calculate the three initial derivatives.
The infection rate is
Therefore, \(dT/dt=-200\) cells/(mL·day). The infected-cell loss rate is \(\delta I(0)=(0.500)(100)=50.0\) cells/(mL·day), so
Virus production is \(pI(0)=(50.0)(100)=5{,}000\) virions/(mL·day), and clearance is \(cV(0)=(5.00)(1{,}000)=5{,}000\) virions/(mL·day). Hence \(dV/dt(0)=0\) virions/(mL·day). At this instant, infected cells are increasing even though free-virus production and clearance balance. A zero virus derivative does not mean that infection has stopped.
Solution Forms and Numerical Trajectories
The equations are coupled: the infection input depends on both \(T\) and \(V\), while changes in \(I\) subsequently alter virus production. In general, the coupled system does not have a simple closed-form solution for all three states. Its trajectory is usually obtained by numerical integration from specified initial conditions and parameter values.
Some exact integral forms clarify what the numerical solution must represent. Dividing the target-cell equation by \(T\), when \(T>0\), gives \(d(\ln T)/dt=-\beta V\). Integrating from 0 to \(t\) yields
Thus \(T\) cannot increase in this model, and it remains positive for finite time if \(T(0)>0\) and the integral is finite. Applying an integrating factor to the infected-cell equation gives
The first term is the surviving contribution from initially infected cells; the integral accumulates new infections, with earlier additions reduced by subsequent infected-cell loss. Similarly, applying an integrating factor to the virus equation gives
These expressions are exact, but they do not decouple the model: \(T(s)\), \(I(s)\), and \(V(s)\) remain interdependent. They help check the meaning of the numerical solution—for instance, virus is supplied by infected cells and removed by clearance—without replacing numerical integration of the coupled equations.
Worked Example: Two Explicit Euler Steps
Use the hypothetical parameters and initial values from the preceding example to approximate the trajectory through \(t=0.2\) day with the explicit Euler method and step size \(h=0.1\) day. For a state \(x\) with derivative \(dx/dt=f(x)\), one Euler step is \(x_{\text{new}}=x_{\text{old}}+hf(x_{\text{old}})\).
At \(t=0\), the derivatives are \(-200\), \(150\), and \(0\) in the respective units calculated above. The first step gives
At this updated state, the infection rate is
The infected-cell loss rate is \((0.500)(115)=57.5\) cells/(mL·day), so \(dI/dt=199.996-57.5=142.496\) cells/(mL·day). The virus production rate is \((50.0)(115)=5{,}750\) virions/(mL·day), while clearance remains \((5.00)(1{,}000)=5{,}000\) virions/(mL·day). Thus, \(dV/dt=750\) virions/(mL·day). The second Euler step is
To a sensible precision, this Euler approximation gives \(T(0.2)\approx999{,}960\) cells/mL, \(I(0.2)\approx129.2\) cells/mL, and \(V(0.2)\approx1{,}075\) virions/mL. The result is an approximation, not an exact trajectory. Euler’s method uses the derivative at the start of each step; reducing the step size or using a higher-order solver is appropriate when accurate trajectories are needed.
What the Model Can and Cannot Say
The shape of a simulated trajectory follows from the balances in the equations. Susceptible cells decrease whenever both \(T\) and \(V\) are positive. Infected cells increase when new infections exceed infected-cell loss, meaning \(\beta TV>\delta I\). Free virus increases when production exceeds clearance, meaning \(pI>cV\). Neither infected cells nor virus must rise at every time point: their derivatives depend on the current balance of inflows and outflows.
With no target-cell replenishment, a sustained positive-virus equilibrium is impossible in this closed model. At an equilibrium with \(V>0\) and \(T>0\), the target-cell equation would give \(dT/dt=-\beta TV<0\), contradicting the equilibrium condition \(dT/dt=0\). The model’s only infection-free equilibria have \(I=V=0\), with \(T\) constant. This is a structural implication of the chosen equations, not a general conclusion about every infection.
The model is useful for organizing acute viral dynamics around target-cell availability, infected-cell loss, virus production, and clearance. It is not sufficient when important biology requires additional states or processes. For example, the assumption of immediate productive infection is unsuitable if a delay before virus production must be represented. Likewise, an observed trajectory alone does not guarantee that all four rate parameters can be uniquely estimated; different parameter combinations may produce similar outputs. Model interpretation should therefore distinguish what the equations imply from what the available measurements can identify.
Diagnostics and Common Pitfalls
Check the model implementation before interpreting a simulated curve. Confirm that each rate has the expected units, the infection term removes and adds the same number of cells, and the initial conditions are nonnegative. For nonnegative parameters and nonnegative initial states, the model’s flows should not produce negative concentrations: at \(T=0\), \(dT/dt=0\); at \(I=0\), \(dI/dt=\beta TV\geq0\); and at \(V=0\), \(dV/dt=pI\geq0\). A numerical solver that gives materially negative concentrations may be using an unsuitable step size or settings.
Compare simulations using smaller integration steps or a suitable adaptive solver. If the predicted trajectory changes substantially as the step is reduced, the original numerical result is not yet reliable. Also inspect whether parameter values and units are plausible for the intended biological scale, and whether the measured quantity is actually the modeled state. For example, an assay signal is not automatically identical to infectious free virus concentration.
- Common mistake: treating virus as the infection rate. The model’s infection rate is \(\beta TV\), not \(V\) alone; the same virus concentration can imply different infection rates at different target-cell concentrations.
- Common mistake: interpreting \(dV/dt=0\) as no viral activity. It means production and clearance balance at that instant. Both rates can be large and equal, as in the first worked example.
- Common mistake: assuming the model includes every cause of viral decline. Its equations include infected-cell loss, virus clearance, and target-cell limitation, but omit processes not represented by a state or term.
How This Model Differs from Common Alternatives
| Modeling choice | What it represents | Consequence for interpretation |
|---|---|---|
| Basic target-cell-limited model | Susceptible cells, productively infected cells, and free virus | Changes arise from infection, infected-cell loss, virus production, and clearance under the stated assumptions. |
| Model with target-cell supply or loss | Additional target-cell turnover or replenishment | Target-cell concentration need not decline monotonically as in the basic model. |
| Model with an infection delay or immune processes | Additional states or processes not present in the basic equations | Changes in viral dynamics can no longer be attributed solely to the basic model’s balances. |
These alternatives are not interchangeable parameterizations of the same assumptions. Adding a process changes the model structure and the interpretation of its parameters. The basic model is most appropriate when its compact representation is adequate for the question and time scale being studied.
Worked Example: A Hypothetical Production-Inhibiting Perturbation
Suppose a hypothetical intervention instantaneously reduces the virus production rate \(p\) from \(50.0\) to \(25.0\) virions/(cell·day), with all other states and parameters initially unchanged from the first example. Calculate the immediate virus derivative and the one-step Euler estimate after \(0.1\) day.
At the initial state, production is now \((25.0)(100)=2{,}500\) virions/(mL·day); clearance remains \((5.00)(1{,}000)=5{,}000\) virions/(mL·day). Therefore,
The one-step Euler estimate is \(V(0.1)=1{,}000+(0.1)(-2{,}500)=750\) virions/mL. At the instant of this modeled change, the equations for \(T\) and \(I\) still have derivatives \(-200\) and \(150\) in their respective units because neither equation contains \(p\) directly. Later, changes in virus and infected cells affect those rates through the model’s coupling. This calculation illustrates an immediate model consequence; it is not evidence that a real intervention produces this response or that the omitted biological processes are negligible.
Check Your Understanding
Answer each question, then compare with the complete solutions below.
- At \(T=2.00\times10^5\) cells/mL and \(V=500\) virions/mL, let \(\beta=1.00\times10^{-6}\) mL/(virion·day), \(I=40.0\) cells/mL, and \(\delta=0.250\) day\(^{-1}\). What are the infection rate and \(dI/dt\)?
- If \(p=20.0\) virions/(cell·day), \(I=300\) cells/mL, \(c=4.00\) day\(^{-1}\), and \(V=1{,}000\) virions/mL, calculate \(dV/dt\) and interpret its sign.
- For the values in Question 1, additionally let \(p=20.0\) virions/(cell·day) and \(c=4.00\) day\(^{-1}\). Use one Euler step of \(h=0.100\) day to approximate \(T\), \(I\), and \(V\) after the step.
- In the basic model, what does \(dV/dt=0\) say about virus production and clearance? What does it not say?
Solutions. (1) The infection rate is \((1.00\times10^{-6})(2.00\times10^5)(500)=100\) cells/(mL·day). Infected-cell loss is \((0.250)(40.0)=10.0\) cells/(mL·day), so \(dI/dt=100-10.0=90.0\) cells/(mL·day). (2) Production is \((20.0)(300)=6{,}000\) virions/(mL·day), and clearance is \((4.00)(1{,}000)=4{,}000\) virions/(mL·day). Thus \(dV/dt=2{,}000\) virions/(mL·day), so free virus is increasing at that instant. (3) The derivatives are \(dT/dt=-100\), \(dI/dt=90.0\), and \(dV/dt=(20.0)(40.0)-(4.00)(500)=800-2{,}000=-1{,}200\), in the corresponding concentration-per-day units. Therefore, after one Euler step, \(T=200{,}000+(0.100)(-100)=199{,}990\) cells/mL, \(I=40.0+(0.100)(90.0)=49.0\) cells/mL, and \(V=500+(0.100)(-1{,}200)=380\) virions/mL. (4) It says production and clearance are equal at that instant, \(pI=cV\). It does not say either process is zero, or that infection and viral replication have stopped.