The Canonical Target-Mediated Drug Disposition Model

Why Represent Drug and Target Separately?

A drug that binds its pharmacological target does not remain entirely free in the modeled system. Some drug may be unbound, some may be bound to target, and binding can affect how much free drug is available for other processes. The canonical target-mediated drug disposition (TMDD) model makes these distinctions explicit by representing free drug, free target, and drug–target complex as separate states.

This is a mechanistic model of disposition: its states describe amounts or concentrations of species, and its rates describe processes that move material among or out of those states. The model is deliberately simpler than a full physiological description. Its value is that each modeled process has a clear interpretation and can be checked against its units and stoichiometry.

Definition: In the canonical one-compartment, 1:1 TMDD model, \(L\) is the concentration of free drug, \(R\) is the concentration of free target, and \(LR\) is the concentration of drug–target complex. The complex contains one drug molecule bound to one target molecule. All three concentrations are defined on the same modeled compartment and concentration basis.

The Reaction Network and Its Rate Constants

The model describes reversible binding, alongside processes that remove free drug, produce or degrade free target, and internalize complex. A drug input may also be specified, such as an infusion or an initial concentration after a bolus. The network can be summarized as follows:

L + R \xrightleftharpoons[k_{\mathrm{off}}]{k_{\mathrm{on}}} LR, \qquad L \xrightarrow{k_{\mathrm{el}}} \text{eliminated drug}, \qquad R \xrightarrow{k_{\mathrm{deg}}} \text{degraded target}, \qquad LR \xrightarrow{k_{\mathrm{int}}} \text{internalized complex}.

The target may also be supplied by zero-order synthesis at rate \(k_{\mathrm{syn}}\). “Zero-order” means that this production rate is constant in the model rather than proportional to the amount of free target. The parameters identify distinct processes:

Parameter	Process represented	Units when concentrations are used
\(k_{\mathrm{on}}\)	Association of free drug and free target	concentration\(^{-1}\) time\(^{-1}\)
\(k_{\mathrm{off}}\)	Dissociation of complex into free drug and free target	time\(^{-1}\)
\(k_{\mathrm{el}}\)	First-order elimination of free drug	time\(^{-1}\)
\(k_{\mathrm{syn}}\)	Zero-order synthesis of free target	concentration time\(^{-1}\)
\(k_{\mathrm{deg}}\)	First-order degradation of free target	time\(^{-1}\)
\(k_{\mathrm{int}}\)	First-order internalization and removal of complex	time\(^{-1}\)

For example, if concentrations are in nanomoles per liter and time is in hours, \(k_{\mathrm{on}}\) has units \(\mathrm{nM}^{-1}\mathrm{h}^{-1}\); each first-order rate constant has units \(\mathrm{h}^{-1}\); and \(k_{\mathrm{syn}}\) has units \(\mathrm{nM}\,\mathrm{h}^{-1}\). These units are not interchangeable: the association rate depends on two concentrations, whereas a first-order process depends on one state concentration.

How the Rates Become Fluxes

Under the usual mass-action assumption, the association flux is proportional to both free-drug and free-target concentrations. The dissociation and first-order removal fluxes are proportional to the concentration of the species undergoing that process:

v_{\mathrm{on}} = k_{\mathrm{on}}LR,\qquad v_{\mathrm{off}} = k_{\mathrm{off}}LR,\qquad v_{\mathrm{el}} = k_{\mathrm{el}}L,\qquad v_{\mathrm{deg}} = k_{\mathrm{deg}}R,\qquad v_{\mathrm{int}} = k_{\mathrm{int}}LR.

Each \(v\) is a concentration flux, with units of concentration per time. Association consumes one free-drug molecule and one free-target molecule for every complex formed. Dissociation releases one of each. Elimination acts on free drug in this canonical network; degradation acts on free target; and internalization removes complex from the modeled compartment. With 1:1 binding, each complex contains one drug and one target, so complex internalization removes one equivalent of each from the modeled states.

The mass-action form has a simple justification. If an association event requires one free drug molecule and one free target molecule, doubling either concentration while holding the other fixed doubles the number of potential encounters per unit time. The simplest rate law with both proportionalities is therefore \(v_{\mathrm{on}}=k_{\mathrm{on}}LR\). Its units check the result: \((\mathrm{concentration}^{-1}\mathrm{time}^{-1})(\mathrm{concentration})(\mathrm{concentration})=\mathrm{concentration}/\mathrm{time}\). A first-order process, by contrast, has flux \(kX\), where \(X\) is the affected state and \(k\) has units of inverse time.

Worked Example: Calculating Binding and Complex Fluxes

Suppose a hypothetical modeled compartment has \(L=2.0\ \mathrm{nM}\), \(R=5.0\ \mathrm{nM}\), and \(LR=0.40\ \mathrm{nM}\). Let \(k_{\mathrm{on}}=0.20\ \mathrm{nM}^{-1}\mathrm{h}^{-1}\), \(k_{\mathrm{off}}=0.10\ \mathrm{h}^{-1}\), and \(k_{\mathrm{int}}=0.050\ \mathrm{h}^{-1}\). Calculate the forward association, dissociation, and internalization fluxes.

Association. The mass-action flux is

v_{\mathrm{on}} = k_{\mathrm{on}}LR = (0.20)(2.0)(5.0) = 2.0\ \mathrm{nM}\,\mathrm{h}^{-1}.

Dissociation. Because dissociation is first-order in complex,

v_{\mathrm{off}} = k_{\mathrm{off}}LR = (0.10)(0.40) = 0.040\ \mathrm{nM}\,\mathrm{h}^{-1}.

Internalization. This is also first-order in complex:

v_{\mathrm{int}} = k_{\mathrm{int}}LR = (0.050)(0.40) = 0.020\ \mathrm{nM}\,\mathrm{h}^{-1}.

The calculated association flux is the rate of complex formation. Dissociation and internalization are separate routes by which complex leaves the \(LR\) state. Their values do not mean that association is irreversible or that the complex concentration is at equilibrium; they are instantaneous fluxes at the specified concentrations.

Assumptions Behind the Canonical Representation

The state definitions and flux expressions rely on modeling assumptions, not on a claim that every biological system follows this network exactly. In its basic form, the model assumes:

A single well-mixed modeled compartment. The three states share a concentration basis, and the model does not resolve spatial gradients or separate tissue compartments.
One drug molecule binds one target molecule. The complex is represented as \(LR\), with no additional binding sites or intermediate complexes.
Mass-action association and first-order dissociation. The rates use the flux expressions above, with parameters treated as constant over the modeled conditions unless explicitly varied.
Specified loss and production routes. Free drug is eliminated with rate constant \(k_{\mathrm{el}}\); free target is synthesized at \(k_{\mathrm{syn}}\) and degraded with \(k_{\mathrm{deg}}\); complex is removed by internalization at \(k_{\mathrm{int}}\).
No additional processes unless added explicitly. For example, the basic network does not separately represent distribution, target-mediated recycling, or release of drug from internalized complex back into the compartment.

These assumptions make the model interpretable but also define its limits. If the biology requires multiple target forms, different binding stoichiometry, multiple accessible tissues, or a distinct fate for internalized drug, those mechanisms need explicit states or processes rather than an undocumented reinterpretation of an existing rate constant.

Worked Example: Free-Target Production and Degradation

In a hypothetical model, \(k_{\mathrm{syn}}=0.60\ \mathrm{nM}\,\mathrm{h}^{-1}\), \(k_{\mathrm{deg}}=0.20\ \mathrm{h}^{-1}\), and the current free-target concentration is \(R=2.0\ \mathrm{nM}\). Find the synthesis and degradation fluxes and compare them.

Synthesis is zero-order, so its flux is the stated rate:

v_{\mathrm{syn}} = k_{\mathrm{syn}} = 0.60\ \mathrm{nM}\,\mathrm{h}^{-1}.

Free-target degradation is first-order:

v_{\mathrm{deg}} = k_{\mathrm{deg}}R = (0.20)(2.0) = 0.40\ \mathrm{nM}\,\mathrm{h}^{-1}.

At these values, synthesis exceeds degradation of free target by \(0.20\ \mathrm{nM}\,\mathrm{h}^{-1}\). This comparison concerns those two processes only. If target is also binding drug, association and dissociation contribute additional changes to free \(R\); the difference between synthesis and degradation alone does not establish that the free-target state is increasing overall.

Worked Example: Which Drug Is Removed by Each Route?

Consider hypothetical concentrations \(L=1.2\ \mathrm{nM}\) and \(LR=0.80\ \mathrm{nM}\), with \(k_{\mathrm{el}}=0.050\ \mathrm{h}^{-1}\) and \(k_{\mathrm{int}}=0.40\ \mathrm{h}^{-1}\). Calculate the free-drug elimination flux and complex internalization flux. For this example, define total drug concentration in the modeled compartment as free plus complex-bound drug.

Free-drug elimination is

v_{\mathrm{el}} = k_{\mathrm{el}}L = (0.050)(1.2) = 0.060\ \mathrm{nM}\,\mathrm{h}^{-1}.

Complex internalization is

v_{\mathrm{int}} = k_{\mathrm{int}}LR = (0.40)(0.80) = 0.32\ \mathrm{nM}\,\mathrm{h}^{-1}.

Under the stated 1:1 assumption, the complex contains \(0.80\ \mathrm{nM}\) drug-bound drug equivalents, so internalization removes drug at \(0.32\ \mathrm{nM}\,\mathrm{h}^{-1}\) from the modeled compartment. The \(0.060\ \mathrm{nM}\,\mathrm{h}^{-1}\) elimination flux applies to free drug, not to all drug in the system. The two fluxes represent distinct routes; their numerical comparison does not by itself establish which route dominates over a full dosing interval.

Reading the Model Correctly

A common mistake is to use “drug concentration” without specifying whether it means free drug \(L\), complex-bound drug \(LR\), or total drug represented by both states. These are different quantities. Under the model’s 1:1 stoichiometry, total drug concentration in the modeled compartment can be defined as \(L+LR\), while total target concentration there can be defined as \(R+LR\). These totals are accounting quantities, not additional states in the basic network.

Another frequent error is to assume that \(k_{\mathrm{int}}\) and \(k_{\mathrm{el}}\) describe the same process because both can reduce drug present in the compartment. They act on different states: \(k_{\mathrm{el}}\) removes free \(L\), while \(k_{\mathrm{int}}\) removes bound \(LR\). Likewise, \(k_{\mathrm{off}}\) is not a loss rate from total drug: dissociation transfers drug from complex to the free state. The drug remains in the compartment unless another process removes it.

Several checks help detect an incorrectly specified or interpreted model:

Confirm that each rate constant has units that produce a flux in concentration per time.
Check that association requires both \(L\) and \(R\), while dissociation and internalization depend on \(LR\).
Check that a process removes or produces the state named by its mechanism; do not apply free-drug elimination to the complex unless the model explicitly includes that route.
Check the measurement definition before comparing data with \(L\), \(LR\), or their sum.
Inspect whether predicted concentrations or fluxes become negative or implausibly large. Such behavior can indicate inconsistent units, an incorrectly specified input, or a model structure that is inadequate for the application.

What the Canonical Model Adds—and What It Does Not

A model with linear drug disposition alone can represent free-drug loss but does not explicitly track binding to a finite target pool. The canonical TMDD network adds target and complex states, allowing binding and complex removal to affect the disposition bookkeeping. In contrast, a purely empirical concentration–response relationship may describe an observed association between exposure and response without specifying these species or processes. The canonical model is useful when that mechanistic detail is relevant and the available data can inform it.

This added structure does not guarantee that every rate constant can be estimated precisely from a particular data set. Sparse measurements, measurements that combine free and bound species, or little variation in exposure can leave multiple parameters compatible with the observations. Model predictions should therefore be checked against the actual assay definitions, dosing information, and biological assumptions. A more detailed network is not automatically a better-supported model.

The canonical formulation also has limits: it assumes a specified accessible compartment and simple binding stoichiometry, and it groups complex internalization into a single removal process. It does not, by itself, establish what happens to internalized drug inside the cell or how target abundance varies among tissues. Those mechanisms should not be inferred merely from the presence of \(k_{\mathrm{int}}\); the parameter represents removal of modeled complex under this formulation.

Key takeaway: \(L\), \(R\), and \(LR\) are distinct states. \(k_{\mathrm{on}}\) and \(k_{\mathrm{off}}\) govern reversible binding; \(k_{\mathrm{el}}\) removes free drug; \(k_{\mathrm{syn}}\) and \(k_{\mathrm{deg}}\) govern free-target production and loss; and \(k_{\mathrm{int}}\) removes complex. Always identify the affected state and units before interpreting a rate.

Check Your Understanding

Answer each question, then compare with the solutions below.

For \(L=3.0\ \mathrm{nM}\), \(R=2.0\ \mathrm{nM}\), and \(k_{\mathrm{on}}=0.10\ \mathrm{nM}^{-1}\mathrm{h}^{-1}\), calculate the association flux and state its units.
If \(LR=0.50\ \mathrm{nM}\) and \(k_{\mathrm{off}}=0.20\ \mathrm{h}^{-1}\), calculate the dissociation flux. Which two free species are released?
Explain why \(k_{\mathrm{el}}L\) and \(k_{\mathrm{int}}LR\) represent different routes, even though both can remove drug from the modeled compartment.
A model has \(k_{\mathrm{syn}}=0.40\ \mathrm{nM}\,\mathrm{h}^{-1}\), \(R=1.5\ \mathrm{nM}\), and \(k_{\mathrm{deg}}=0.10\ \mathrm{h}^{-1}\). Calculate the two target fluxes and their difference, and explain why that difference alone does not determine the net change in free target.

Solutions. (1) \(v_{\mathrm{on}}=(0.10)(3.0)(2.0)=0.60\ \mathrm{nM}\,\mathrm{h}^{-1}\). (2) \(v_{\mathrm{off}}=(0.20)(0.50)=0.10\ \mathrm{nM}\,\mathrm{h}^{-1}\); dissociation releases one free drug and one free target per complex. (3) \(k_{\mathrm{el}}L\) removes free drug, whereas \(k_{\mathrm{int}}LR\) removes drug while it is bound in complex; they act on different states. (4) Synthesis is \(0.40\ \mathrm{nM}\,\mathrm{h}^{-1}\); degradation is \((0.10)(1.5)=0.15\ \mathrm{nM}\,\mathrm{h}^{-1}\); synthesis exceeds degradation by \(0.25\ \mathrm{nM}\,\mathrm{h}^{-1}\). Binding and dissociation also affect free target, so this difference is not the complete free-target balance.