Tutorials › Nonlinear Mixed-Effects Models (NLME) › Deriving the Quasi-Steady-State Approximation

Home  /  Tutorials  /  Target-Mediated Drug Disposition in NLME

Target-Mediated Drug Disposition in NLME · Tutorial 354 of 797

Deriving the Quasi-Steady-State Approximation

Learn when the TMDD complex can be treated as quasi-steady and how its concentration depends on free drug, free target, and \(K_{SS}\).

Advanced 10 min read

What You'll Learn

  • Derive the quasi-steady-state complex concentration from the TMDD complex equation.
  • Define \(K_{SS}\) and identify its concentration units.
  • Use timescale separation to assess when setting the complex derivative to zero is defensible.
  • Calculate quasi-steady complex concentrations and check formation–loss balance.
  • Recognize initial transients and distinguish quasi-steady state from binding equilibrium.

Why Use a Quasi-Steady-State Approximation?

In a target-mediated drug disposition (TMDD) model, free drug binds free target to form a drug–target complex. The complex can dissociate or be removed through internalization. When complex formation and loss adjust much faster than free drug and free target change, the complex may track their changing concentrations closely. A reduced model represents that tracking relationship algebraically rather than integrating a separate differential equation for the complex.

This is a timescale approximation, not a claim that the system is at thermodynamic equilibrium. The complex may be continually formed and removed while its concentration changes only slowly. The approximation is useful only when the fast adjustment has occurred and the other state variables are changing on a sufficiently slower timescale.

From the Complex State Equation to \(K_{SS}\)

Let \(L\) denote free drug concentration, \(R\) free target concentration, and \(C\) drug–target complex concentration. Let \(k_{on}\) be the association rate constant, \(k_{off}\) the dissociation rate constant, and \(k_{int}\) the complex internalization rate constant. The mass-action equation for the complex is

$$ \frac{dC}{dt}=k_{on}LR-(k_{off}+k_{int})C. $$

The formation term \(k_{on}LR\) has units of concentration per time. The loss term combines dissociation and internalization, because each removes complex from the \(C\) state. Both rate constants in that loss term have units of inverse time.

Definition: The quasi-steady-state (QSS) approximation for the complex replaces its differential equation by the algebraic condition \(dC/dt\approx 0\), when complex adjustment is sufficiently fast relative to changes in free drug and free target. Under this condition, $$ C\approx \frac{k_{on}LR}{k_{off}+k_{int}}=\frac{LR}{K_{SS}}, \qquad K_{SS}=\frac{k_{off}+k_{int}}{k_{on}}. $$ Here \(K_{SS}\) is the quasi-steady-state concentration parameter.

The formula follows by setting the complex formation rate equal to its combined loss rate and solving for \(C\). Since \(k_{off}+k_{int}\) has units of inverse time and \(k_{on}\) has units of inverse concentration per time, \(K_{SS}\) has units of concentration. For example, if \(L\), \(R\), and \(K_{SS}\) are in nM, then \(LR/K_{SS}\) is also in nM.

Why a Fast Complex Can Be Treated as Quasi-Steady

A timescale argument explains what the approximation requires. First hold \(L\) and \(R\) fixed over a short interval. Write \(\lambda=k_{off}+k_{int}\), so the complex equation is \(dC/dt=k_{on}LR-\lambda C\). Its steady value for those fixed inputs is \(C_{QSS}=k_{on}LR/\lambda\). Subtracting this value from \(C\) gives

$$ \frac{d}{dt}(C-C_{QSS})=-\lambda(C-C_{QSS}). $$

Solving this first-order equation yields

$$ C(t)-C_{QSS}=(C(0)-C_{QSS})e^{-\lambda t}. $$

Thus, with the free concentrations held fixed, departures from the quasi-steady value decay exponentially. The corresponding relaxation timescale is \(\tau_C=1/\lambda=1/(k_{off}+k_{int})\). If \(L\) and \(R\) change appreciably only over a much longer timescale, the complex can relax toward the moving value \(k_{on}LR/\lambda\) before that value has changed much. This is the basis of the QSS approximation.

The fixed-input solution is a justification, not a guarantee for the fully coupled TMDD system. In that system, \(L\), \(R\), and \(C\) affect one another. The approximation is plausible when the complex’s adjustment is fast compared with the timescale of meaningful changes in both free concentrations and in the processes that drive them. A timescale ratio much less than one supports the approximation; there is no universal cutoff that makes it valid in every model.

Worked Example: Calculating \(K_{SS}\) and the Complex Concentration

Suppose a hypothetical model has \(k_{on}=0.20\ \text{nM}^{-1}\text{h}^{-1}\), \(k_{off}=0.30\ \text{h}^{-1}\), and \(k_{int}=0.10\ \text{h}^{-1}\). At a particular time, free drug is \(L=4.0\ \text{nM}\) and free target is \(R=3.0\ \text{nM}\). Calculate \(K_{SS}\) and the QSS complex concentration.

First add the two complex-loss rate constants and divide by the association rate constant:

$$ K_{SS}=\frac{0.30+0.10\ \text{h}^{-1}}{0.20\ \text{nM}^{-1}\text{h}^{-1}} =\frac{0.40}{0.20}\ \text{nM} =2.0\ \text{nM}. $$

Then substitute the free concentrations:

$$ C_{QSS}=\frac{LR}{K_{SS}} =\frac{(4.0\ \text{nM})(3.0\ \text{nM})}{2.0\ \text{nM}} =6.0\ \text{nM}. $$

The units reduce to nM, as required for a concentration. The rate balance also checks: formation is \(k_{on}LR=(0.20)(4.0)(3.0)=2.4\ \text{nM h}^{-1}\), while loss is \((k_{off}+k_{int})C_{QSS}=(0.40)(6.0)=2.4\ \text{nM h}^{-1}\). Under the QSS condition, formation and loss are equal.

Worked Example: Checking the Timescale Separation

Use the rate constants from the previous hypothetical example. Suppose the free drug and free target concentrations each change substantially over a characteristic timescale of about \(30\ \text{h}\). Assess whether the complex’s relaxation is fast enough to make QSS plausible.

The complex relaxation timescale, holding the free concentrations fixed, is

$$ \tau_C=\frac{1}{k_{off}+k_{int}} =\frac{1}{0.30+0.10\ \text{h}^{-1}} =2.5\ \text{h}. $$

The ratio of the complex timescale to the stated free-concentration timescale is \(2.5/30=0.0833\). Equivalently, the free concentrations change over about 12 times the complex relaxation time. This separation supports using QSS as a plausible approximation over that part of the hypothetical trajectory, provided the complex has already adjusted and no faster changes in \(L\) or \(R\) are occurring.

This calculation does not prove that QSS is accurate for every time point. A rapid dose input or another abrupt change could make free drug vary on a much shorter timescale than \(30\ \text{h}\). The relevant comparison is with the actual local changes in the model, not just a typical timescale quoted for the entire treatment period.

Initial Transients and the Limits of the Approximation

Even if the timescale separation is strong after an input, the complex need not start at its QSS value. If \(L\) and \(R\) are held fixed and \(C(0)\ne C_{QSS}\), the exact solution above shows a transient before \(C\) approaches \(C_{QSS}\). During that transient, replacing \(C\) immediately by its QSS value can misrepresent early concentrations and the amount of drug bound in complex. A baseline or initial-condition calculation may inform the starting state, but it does not by itself establish that the QSS assumption holds after dosing.

Worked Example: Quantifying a Frozen-Input Initial Transient

Use \(C_{QSS}=6.0\ \text{nM}\) and \(\lambda=0.40\ \text{h}^{-1}\) from the earlier hypothetical calculations. Suppose, only for this transient calculation, that \(L=4.0\ \text{nM}\) and \(R=3.0\ \text{nM}\) remain fixed and that \(C(0)=0\). Find the exact complex concentration after \(1.0\ \text{h}\), then compare it with the QSS value.

For fixed free concentrations, the solution is \(C(t)=C_{QSS}+(C(0)-C_{QSS})e^{-\lambda t}\). Substitution gives

$$ C(1.0)=6.0+(0-6.0)e^{-(0.40)(1.0)} =6.0(1-e^{-0.40}) \approx 1.978\ \text{nM}. $$

The instantaneous QSS substitution would give \(6.0\ \text{nM}\), substantially above the exact frozen-input value at \(1.0\ \text{h}\). After three relaxation times, \(t=7.5\ \text{h}\), the exact value is

$$ C(7.5)=6.0(1-e^{-(0.40)(7.5)}) =6.0(1-e^{-3}) \approx 5.701\ \text{nM}. $$

That is about \(95.0\%\) of the QSS value. This example illustrates why an initial transient can matter even when later QSS behavior is reasonable. The fixed-input calculation isolates the complex relaxation; in a full model, \(L\) and \(R\) would generally change during the transient.

Interpretation, Diagnostics, and Alternatives

The QSS relation uses free drug and free target: \(C\approx LR/K_{SS}\). Substituting total drug or total target in their place without a justified transformation changes the relationship. Nor should \(K_{SS}\) be interpreted automatically as a binding affinity measured at equilibrium. The dissociation equilibrium constant is \(K_D=k_{off}/k_{on}\); \(K_{SS}\) also includes internalization. If \(k_{int}=0\), then \(K_{SS}=K_D\), but with positive internalization \(K_{SS}>K_D\).

A QSS model is an algebraic reduction of the full TMDD differential-equation model. The full model explicitly integrates \(C\) and can represent complex transients without setting its derivative to zero; it is a useful reference when those transients may matter, although it requires solving an additional state equation. A rapid-binding equilibrium approximation is a different limiting assumption: it treats binding as equilibrated rapidly enough that internalization does not enter the binding denominator in the same way. The QSS expression here retains \(k_{int}\), so QSS and rapid-binding equilibrium are not interchangeable.

In practice, warning signs include a complex relaxation time comparable to the timescale of changes in free drug or target, sharp changes in dose or concentrations, and systematic disagreement between QSS predictions and a full-model trajectory. Checking only a typical concentration or a single time point is inadequate: timescale separation can vary over the dosing interval. If model predictions are sensitive to the initial complex state or differ materially during early treatment, the algebraic approximation may be unsuitable for that interval.

Common pitfall: “The complex is at equilibrium, so \(K_{SS}=k_{off}/k_{on}\).” This reasoning omits the internalization loss pathway. Under QSS, formation balances both dissociation and internalization, giving \(K_{SS}=(k_{off}+k_{int})/k_{on}\). It equals \(K_D\) only when \(k_{int}=0\).

Check Your Understanding

Answer each question, then compare with the solutions below.

  1. For \(k_{on}=0.50\ \text{nM}^{-1}\text{h}^{-1}\), \(k_{off}=0.20\ \text{h}^{-1}\), and \(k_{int}=0.30\ \text{h}^{-1}\), calculate \(K_{SS}\).
  2. With the rates in Question 1, if \(L=2.0\ \text{nM}\) and \(R=5.0\ \text{nM}\), calculate the QSS complex concentration and verify the rate balance.
  3. If the complex relaxation timescale is \(1.0\ \text{h}\) and free concentrations change substantially over \(2.0\ \text{h}\), does this provide strong timescale separation? Explain.
  4. Why can the QSS prediction be poor immediately after a dose even when it becomes reasonable later?
  5. When does \(K_{SS}\) equal \(K_D\), and why?

Solutions. (1) \(K_{SS}=(0.20+0.30)/0.50=1.0\ \text{nM}\). (2) \(C_{QSS}=LR/K_{SS}=(2.0)(5.0)/1.0=10.0\ \text{nM}\). Formation is \((0.50)(2.0)(5.0)=5.0\ \text{nM h}^{-1}\), and loss is \((0.20+0.30)(10.0)=5.0\ \text{nM h}^{-1}\). (3) The timescale ratio is \(1.0/2.0=0.50\), only a twofold separation, so it does not provide strong support for QSS; accuracy would need careful assessment. (4) The complex may start away from its QSS value and require time to relax, while the dose may also change free drug rapidly. (5) \(K_{SS}=K_D\) when \(k_{int}=0\), since then \(K_{SS}=(k_{off}+0)/k_{on}=k_{off}/k_{on}\). With internalization, QSS includes that additional loss process.

Previous

Next

Back to Tutorials