Infusion Dose-Exposure Relationships Under TMDD

Why Infusion Rate and Exposure Are Not Proportional Under TMDD

With linear elimination, doubling a constant infusion rate eventually doubles the steady-state concentration. Target-mediated drug disposition (TMDD) can break that relationship: when drug binds a finite target pool, the amount of drug removed through target binding and internalization depends on free drug concentration and target occupancy. As the target becomes saturated, this removal pathway contributes less additional clearance per unit increase in free drug.

This tutorial derives and uses the steady-state infusion relationship for a reduced TMDD model. The central distinction is between free drug, which drives binding and nonspecific elimination in this model, and total drug, which includes drug bound in the drug–target complex. The equations describe a maintained target pool and a steady state; they are not a dosing-interval model.

Concentrations are expressed in nM and \(u\) in nM/h in the examples. Both \(a\) and \(b\) have units of h\(^{-1}\), while \(K\) has units of nM. In a quasi-steady-state reduction, \(K\) commonly represents \((k_{\mathrm{off}}+k_{\mathrm{int}})/k_{\mathrm{on}}\), rather than the equilibrium dissociation constant \(k_{\mathrm{off}}/k_{\mathrm{on}}\). The appropriate value depends on the binding and internalization assumptions used for the reduction.

Steady-State Relationship and Its Assumptions

The infusion rate must balance irreversible drug loss at steady state. Binding itself transfers drug between the free and complex states; it is not a net loss from their sum. In this reduced model, free drug is lost through nonspecific elimination at rate \(aC\), and complexed drug is lost through internalization at rate \(bRC\). Thus the steady-state mass balance is:

Substituting the quasi-steady binding relation gives a single equation for free concentration:

The first term increases linearly with \(C\). The second term increases with \(C\), but approaches the finite maximum \(bR_{\mathrm{tot}}\) as the target becomes saturated. Multiplying through by \(K+C\) gives a quadratic equation:

For \(a>0\), \(K>0\), \(R_{\mathrm{tot}}\geq 0\), and \(u>0\), the constant term is negative and the leading coefficient is positive. The two roots therefore have opposite signs, so exactly one root is positive. That root is the physically meaningful free concentration:

Once \(C\) is found, calculate \(RC=R_{\mathrm{tot}}C/(K+C)\), then \(C_{\mathrm{tot}}=C+RC\). For \(u=0\), the steady-state solution is \(C=RC=C_{\mathrm{tot}}=0\), under these assumptions.

The binding relation follows directly from the reduced complex balance. If formation is \(k_{\mathrm{on}}CR\) and complex disappearance is \((k_{\mathrm{off}}+k_{\mathrm{int}})RC\), the quasi-steady condition sets these rates equal:

Since \(R_{\mathrm{tot}}=R+RC\), substitute \(R=R_{\mathrm{tot}}-RC\) into \(RC=CR/K\). Rearranging gives \(RC(K+C)=R_{\mathrm{tot}}C\), hence \(RC=R_{\mathrm{tot}}C/(K+C)\). This justification depends on the maintained-pool reduction; a full TMDD model with changing target abundance requires its own steady-state target balance.

How Saturation Changes Dose Proportionality

The free concentration response to a change in infusion rate can be examined by differentiating the steady-state input function \(u(C)=aC+bR_{\mathrm{tot}}C/(K+C)\):

The local proportionality of free concentration to infusion rate is summarized by the elasticity \(d\ln C/d\ln u\). Differentiating gives:

For positive \(C\) and a nonzero target-mediated pathway, \(K/(K+C)^2<1/(K+C)\). Therefore the numerator exceeds the denominator and the elasticity is greater than 1: locally, a percentage increase in infusion rate produces a larger percentage increase in free concentration. At low concentrations, binding-mediated removal can be substantial relative to free drug. At high concentrations, the complex approaches its ceiling \(R_{\mathrm{tot}}\), so further increases in input are increasingly balanced by nonspecific elimination, while the free concentration must rise.

This result concerns free concentration. Total concentration also includes a bounded complex pool. Consequently, total concentration need not change by the same factor as free concentration, especially while the bound pool is filling. Dose proportionality should be assessed using the concentration measure relevant to the question or assay.

Clearance Depends on Which Concentration Is Used

At steady state, two useful ratios can be reported, but they answer different questions. The input-to-free-concentration ratio is \(u/C\); from the mass balance, it equals \(a+bRC/C\). It includes the concentration-dependent contribution of target-mediated loss relative to free drug. The input-to-total-concentration ratio is \(u/C_{\mathrm{tot}}\), a total-concentration-based apparent clearance. It is not generally equal to either \(a\) or \(b\), because drug is distributed between free and bound forms and the elimination pathways act on different states.

For the three hypothetical infusion rates, the ratios are approximately:

The ratios change with infusion rate, so a single constant “clearance” cannot summarize every concentration measure and range in this example. Always state which concentration is in the denominator and whether the reported quantity is a model coefficient, an effective ratio, or an apparent clearance based on total measured drug.

Model Choice, Diagnostics, and Common Pitfalls

The reduced steady-state equation is useful when the question is the relationship between a constant input and steady-state exposure, and the binding and target-pool assumptions are defensible. It is more mechanistic than a simple Michaelis–Menten elimination approximation because it explicitly represents the bound pool and its occupancy. A Michaelis–Menten form can be a convenient approximation over a specified concentration range, but its parameters may not map uniquely to binding, target turnover, and internalization processes. A full TMDD model is preferable when transient binding, changing target abundance, or other state-specific processes are important. It requires solving the coupled state equations rather than relying on this single steady-state equation.

Several checks help identify when the reduced relationship is being misused. First, confirm that the input is truly constant and that observations represent steady state; a concentration still changing over time cannot be interpreted using this balance. Second, check that the predicted complex does not exceed \(R_{\mathrm{tot}}\), and that calculated \(C\), \(RC\), and \(C_{\mathrm{tot}}\) are nonnegative. Third, verify the mass balance by substituting the calculated concentrations back into \(u=aC+bRC\). Fourth, compare predictions with the concentration definition used for data: an assay measuring total drug should not be compared directly with predicted free drug without an appropriate observation model.

Systematic disagreement across infusion rates can indicate that the maintained-pool or quasi-steady binding assumptions are inadequate, that a modeled elimination pathway is missing, or that the system has not reached steady state. A good fit at one infusion rate alone does not establish that the predicted saturation pattern is reliable at other rates.

• Mistake: treating binding as drug elimination. Binding moves drug from \(C\) to \(RC\); in the combined free-plus-complex balance, the binding terms cancel. Correction: count irreversible loss pathways, such as the specified complex internalization, in the input balance.
• Mistake: assuming total drug is the free concentration. When a substantial fraction is bound, \(C_{\mathrm{tot}}=C+RC\) can be much larger than \(C\). Correction: calculate and label both quantities.
• Mistake: interpreting an apparent clearance ratio as a fixed model parameter. Ratios such as \(u/C\) vary with occupancy in this model. Correction: state the concentration denominator and report the infusion-rate range and units.
• Mistake: assuming a constant target pool is automatically valid. Target synthesis, degradation, recycling, or abundance changes can alter the relationship. Correction: use the reduced equation only when its target-pool assumption is reasonable.