Common Errors in Hepatic Transporter Parameterization

Why Parameterization Errors Matter

A hepatic transporter model can be structurally plausible and still predict the wrong liver exposure if its parameters mix incompatible concentration bases or count the same process more than once. Four errors are especially consequential: correcting a clearance for binding in the wrong direction, treating uptake as net hepatic elimination, representing passive flux twice, and applying abundance scaling to a value that has already been scaled.

The earlier material on uptake assays, binding corrections, transporter activity scaling, and hepatic transport models provides the relevant measurement context. Here the emphasis is on checking the bookkeeping: what concentration a parameter uses, which flux it represents, and what biological amount it has already been scaled to. The examples use hypothetical values.

Total and Unbound Concentrations Are Not Interchangeable

In a hepatocyte incubation, some compound may bind to proteins, plastic, or other assay components. Let \(C_{\mathrm{tot,inc}}\) be the total incubation concentration and let \(f_{u,\mathrm{inc}}\) be the unbound fraction in that incubation. Under the assumption that binding is at equilibrium and the unbound fraction is constant over the concentration range considered,

If an uptake rate \(v\) is divided by the total concentration, the resulting total-based uptake clearance is \(CL_{\mathrm{uptake,tot}}=v/C_{\mathrm{tot,inc}}\). Dividing the same rate by unbound concentration gives \(CL_{\mathrm{uptake,u}}=v/C_{u,\mathrm{inc}}\). Therefore,

This conversion applies only when the original clearance used total concentration as its denominator and the rate is unchanged. If the rate was already divided by unbound concentration, applying the correction again is an error. If the unbound fraction varies with concentration, a single constant correction may not be adequate; measure or model the relevant unbound concentrations instead. For a saturable uptake model, the concentration basis of the fitted parameters must also be explicit rather than silently switching between total and unbound concentration.

Uptake Is a Flux Into the Cell, Not Automatically Net Clearance

Uptake clearance describes entry from the extracellular or blood-facing space into hepatocytes, relative to the concentration used in its definition. It does not by itself state how much drug is ultimately eliminated from the body. Once inside a hepatocyte, compound may be metabolized, excreted into bile, or returned to blood. Return to blood can reduce net hepatic removal even when uptake is rapid.

A simple linear model makes this distinction precise. Let \(C_p\) be the relevant unbound plasma concentration, \(A\) the intracellular amount, and \(CL_{\mathrm{uptake}}\) the uptake clearance on a matching concentration basis. Let \(k_{\mathrm{met}}\), \(k_{\mathrm{bile}}\), and \(k_{\mathrm{return}}\) be first-order rate constants for intracellular metabolism, biliary loss, and return to blood, respectively. Their units are inverse time. Assume constant parameters, a well-mixed intracellular pool, and steady state for that pool. Then the uptake input rate is \(CL_{\mathrm{uptake}}C_p\), and the intracellular balance is

At steady state, set \(dA/dt=0\) and solve for \(A\). Net irreversible loss is the sum of metabolic and biliary output, not uptake alone. The resulting clearance from plasma into those loss pathways is

In this simplified model, the fraction of uptake that is ultimately lost through metabolism or bile is less than or equal to one. The result depends on the stated steady-state, linear-rate assumptions; more detailed models may include multiple intracellular spaces, saturable processes, or time-varying concentrations. The central bookkeeping point remains: an uptake parameter is one component in a pathway, not automatically the net clearance.

Represent Passive Flux Once

A transporter assay may measure total accumulation that includes both transporter-mediated uptake and passive entry. At a specified concentration and time, write the measured control rate as \(J_{\mathrm{control}}=J_{\mathrm{active}}+J_{\mathrm{passive}}\). If an inhibitor fully and selectively blocks the transporter without changing passive permeability, the inhibited rate estimates \(J_{\mathrm{passive}}\). The transporter-dependent rate is then the difference \(J_{\mathrm{control}}-J_{\mathrm{inhibited}}\).

A model can represent these components either as an active term plus an independently parameterized passive term, or as a measured total term that already includes both. It must not use the total as the active term and then add passive flux again. The same principle applies when passive flux is written explicitly as a permeability term driven by the unbound concentration difference across a membrane: if another parameter already represents that same passive contribution, adding both counts one physical process twice.

The subtraction interpretation requires its assumptions. If the inhibitor is incomplete, the difference underestimates transporter-mediated flux. If it alters passive permeability, the inhibited rate is not a clean passive estimate. Other active pathways may also remain in the inhibited condition. Those limitations should be handled in the assay interpretation, rather than hidden by labeling the residual as pure passive flux.

Scale Transporter Activity Only Once

In vitro transporter activity may be reported per unit assay protein, per unit transporter abundance, or for a reference expression system. The appropriate scaling depends on which of these bases the reported value already uses. For example, if a measured activity is expressed per milligram of membrane protein, scaling to liver may involve the membrane protein per gram of liver and liver mass. If activity is expressed per amount of transporter, it may instead be multiplied by the target liver transporter amount. Do not apply both routes to the same activity unless the quantities are defined to avoid overlap.

Abundance-ratio scaling follows the same rule. If a reference-system activity has already been multiplied by the target-to-reference abundance ratio, the result is already adjusted for that abundance difference. Multiplying the adjusted result by the ratio again applies the same factor twice. Units provide a practical check: multiplying a whole-liver activity by an abundance ratio should not be done merely because the ratio is available; first determine whether the activity already represents the target abundance.

Practical Checks and Common Failure Modes

Before using a transporter parameter in a PBPK model, trace its measurement and transformation from the assay to the model. The following checks help localize errors without mistaking one problem for another.

A smaller corrected unbound clearance than the total-based value is a warning of a reversed correction or another inconsistency; it is not evidence of repeated division by \(f_{u,\mathrm{inc}}\). Repeating that division produces an excessively large value. Keeping these diagnoses separate is important because they point to different corrections.

There are reasonable alternatives to each calculation route. Direct measurement of unbound incubation concentration avoids relying on a binding correction, while a measured \(f_{u,\mathrm{inc}}\) can be useful when direct measurement is impractical and binding is adequately characterized. In uptake assays, inhibitor subtraction can isolate a transporter-dependent component under restrictive assumptions; explicit passive-flux modeling can be preferable when permeability data and intracellular concentrations are available. For scaling, one can use protein-per-liver scaling or transporter-abundance scaling, provided the activity units and abundance basis match. None is automatically superior: the data and assumptions determine which is defensible.

These checks do not establish that a chosen transporter mechanism is correct or that the resulting model is supported by in vivo evidence. They address parameter consistency. A well-scaled, correctly corrected parameter can still be uncertain, poorly identifiable, or unsuitable outside the assay conditions; those limitations should remain visible in model evaluation.