Separating Transporter Inhibition from Enzyme-Mediated DDI Effects

Why Mechanism Matters

A rise in a victim drug’s plasma exposure during coadministration can be consistent with reduced hepatic uptake, reduced metabolic activity, or both. These mechanisms are not interchangeable. Uptake inhibition reduces delivery of drug from blood into hepatocytes; enzyme inhibition reduces its metabolic loss after it reaches the cell. Because hepatic metabolism depends on intracellular drug, either change can alter metabolite formation and systemic exposure. A plasma AUC ratio by itself may not reveal which process changed.

The useful distinction is therefore between where the disposition pathway is altered and what downstream measurements respond. A coupled liver model makes this explicit by linking plasma drug, hepatocyte drug, transport into and out of the cell, and metabolism. The calculations below use a deliberately simplified linear model to show the logic. Its predictions are mechanistic comparisons, not substitutes for checking the model against clinical data.

A Coupled Uptake–Efflux–Metabolism Model

Assume a well-mixed hepatocyte compartment, linear (unsaturated) transport and metabolism over the concentration range considered, and constant clearances during the comparison. Assume initially that uptake is the only entry route into the hepatocyte and that efflux and metabolism are the only intracellular losses. At steady state, the rate of uptake equals the combined rates of efflux and metabolism:

This balance gives the hepatocyte-to-plasma concentration ratio. Multiplying that ratio by the metabolic clearance gives the metabolic flux per unit plasma concentration, which defines an effective intrinsic clearance from plasma through the hepatic pathway:

The second expression is not simply \(CL_{\mathrm{met}}\): it accounts for the chance that drug entering the cell returns to plasma before metabolism. It is also not a universal transporter formula. It follows from the specified compartment structure and assumptions; additional entry routes, intracellular binding, nonlinear kinetics, or multiple liver compartments require a more detailed model.

From Effective Intrinsic Clearance to Clinical Exposure

A common next step is to use \(CL_{\mathrm{int,eff}}\) in a well-stirred liver model. For this step, assume hepatic blood flow \(Q_h\) is constant, the liver is well stirred, and \(f_{u,b}\) is the fraction unbound in blood. The model treats \(f_{u,b}CL_{\mathrm{int,eff}}\) as the liver’s capacity to remove unbound drug. Under these assumptions, hepatic clearance and the fraction escaping hepatic first-pass extraction are:

The formulas show why the same proportional change in effective intrinsic clearance does not imply the same proportional change in hepatic clearance for every drug. When intrinsic capacity is small relative to flow, clearance is sensitive to changes in that capacity but remains below the flow limit. When intrinsic capacity is large relative to flow, hepatic clearance approaches \(Q_h\), so additional intrinsic capacity has less effect on clearance. First-pass availability \(F_h\) changes in the opposite direction to effective intrinsic clearance in this model.

A short algebraic check explains the oral result in this restricted case. Substituting the well-stirred expressions and cancelling the common terms gives:

Consequently, with hepatic clearance as the only systemic clearance and with all nonhepatic bioavailability terms fixed, oral AUC for a fixed dose is proportional to \(1/CL_{\mathrm{int,eff}}\). This cancellation does not mean an oral AUC ratio identifies whether uptake or metabolism changed: each mechanism can change the same effective intrinsic clearance.

Mechanistic Evidence and Diagnostics

A useful comparison asks whether each candidate mechanism predicts the full pattern of measurements, not just the parent-drug AUC. Consider intracellular concentration, parent drug in plasma, and formation of a metabolite known to be formed by the hepatic enzyme of interest. In the simple model, the fixed-\(C_p\) calculations predict lower intracellular drug and substantially lower metabolic flux with uptake inhibition, but higher intracellular drug with a smaller reduction in metabolic flux after enzyme inhibition. Real measurements are not made at fixed plasma concentration, so a PBPK simulation should compare concentrations and fluxes over the same dosing regimen and sampling times.

In practice, mechanistic separation is stronger when several independent observations agree. Useful evidence can include transporter-selective in vitro uptake data, enzyme activity measurements, a metabolite whose formation pathway is adequately characterized, and clinical data at more than one dose or route. Earlier material on transporter inhibition and CYP interaction modeling provides the relevant parameter-estimation context; here the key requirement is to compare predictions after changing one mechanism at a time and then, if justified, both together.

• Check the predicted intracellular direction. If the model claims selective uptake inhibition but predicts increased \(C_h/C_p\) under the stated assumptions, inspect the equations, parameter mapping, and any additional transport routes.
• Check mass balance. At steady state in the simple model, uptake flux must equal efflux plus metabolism. A simulated liver compartment that does not satisfy this balance may reflect a transient rather than steady-state output, or a coding or unit error.
• Check metabolite specificity. A metabolite is informative about a particular enzyme only if formation through that enzyme is sufficiently characterized and metabolite elimination is considered.
• Check model identifiability. If different transporter and enzyme parameter combinations produce nearly the same observed parent concentration-time profile, parent exposure alone may not estimate their separate effects reliably.

Comparisons, Limitations, and Common Pitfalls

The model is limited by its linear steady-state description. Saturable uptake or metabolism makes clearances concentration-dependent; intracellular binding can make total hepatocyte concentration differ from the concentration available to enzymes; and liver zonation or multiple intracellular compartments can make a single well-mixed compartment inadequate. Changes in protein binding, hepatic blood flow, enzyme abundance, or nonhepatic elimination also violate the assumptions behind the simple clinical exposure calculations.

A common mistake is to infer “enzyme inhibition” from any increase in parent-drug AUC, or to infer “transporter inhibition” from a lower metabolite-to-parent ratio. Both conclusions overstate what those observations establish. The correction is to treat them as patterns to be tested: specify the affected parameter, propagate it through the coupled pathway, and ask whether the predicted parent, intracellular, metabolite, and route-specific observations agree. If both mechanisms fit the available data, report the ambiguity rather than assigning a unique cause.