Representing Hepatic Zonation in Transporter Models

Why Represent Hepatic Zonation?

Transporter abundance and other determinants of drug handling need not be uniform throughout the liver. A drug entering a sinusoid encounters a spatial sequence of hepatocyte environments, while different parts of the organ may also receive different fractions of total blood flow. A model that averages these differences into one liver-wide value can predict a different hepatic extraction from a model that preserves the regional structure.

The practical question is not whether zonation exists in general, but whether the available data and the drug’s kinetics justify representing it. A zonated model adds parameters and structure. It is most informative when regional transporter abundance, local intrinsic clearance, or perfusion differences could materially affect the concentration reaching downstream regions.

The earlier material on unbound intracellular concentrations and net intrinsic clearance provides the conceptual basis for constructing a local effective clearance. In this tutorial, \(CL_{\mathrm{int},i}\) is treated as an input to the regional extraction calculation: it summarizes the net local capacity for irreversible removal from blood under the conditions of the model. It should not automatically be equated with a measured uptake clearance or with transporter abundance alone.

Choose the Spatial Structure Before Calculating

Two structures are useful, and they answer different anatomical questions. In a serial-zone model, blood passes through successive regions along a sinusoidal path. If there is no branching or blood loss between regions, the same flow \(Q_H\) passes through each zone. Regional differences affect how much drug remains available to the next zone.

In a parallel-path model, blood is divided among pathways, each with its own flow \(Q_k\). The flow fractions satisfy \(\sum_k Q_k=Q_H\). A pathway can itself contain serial zones. This structure is appropriate when the model represents distinct perfusion territories rather than consecutive locations along one path.

These structures should not be conflated. Assigning independent flows to successive serial zones would count the same blood flow repeatedly. Conversely, averaging parallel pathways before calculating extraction can lose the effect of different flows and clearances.

Serial Zones: Derivation of Extraction

Consider one well-stirred zone \(i\), with inlet concentration \(C_{\mathrm{in},i}\), outlet concentration \(C_{\mathrm{out},i}\), blood flow \(Q_H\), and effective intrinsic clearance \(CL_{\mathrm{int},i}\). Under the stated assumptions, the unbound concentration driving intrinsic removal is \(f_{u,b}C_{\mathrm{out},i}\). At steady state, inflow minus outflow equals removal:

Solving for the outlet-to-inlet concentration ratio gives the fraction that escapes zone \(i\). The zonal extraction is one minus that fraction.

For \(m\) serial zones, the outlet concentration from one zone is the inlet concentration to the next. Multiplying the escape fractions therefore gives total hepatic extraction:

This product is the key consequence of spatial ordering: an upstream zone changes the concentration presented to every downstream zone. The formula assumes each zone is well stirred. A different within-zone assumption, such as plug flow, leads to a different local escape equation and should not be substituted without stating that change.

From Transporter Abundance to Regional Clearance

A regional transporter measurement is not itself an intrinsic clearance. In a linear concentration range, a transporter-mediated clearance component may be scaled in proportion to active transporter abundance if activity per unit abundance, substrate access, and relevant cellular conditions are comparable across regions. For example, if a reference region has abundance \(A_{\mathrm{ref}}\) and a measured activity corresponding to \(CL_{\mathrm{uptake,ref}}\), a simple regional scaling is:

This equation estimates a regional uptake component under its scaling assumptions. It does not establish the region’s net effective \(CL_{\mathrm{int},i}\). The latter can also depend on passive exchange, intracellular binding, metabolism, biliary efflux, and basolateral efflux. The relationship may be nonlinear when uptake or efflux is saturable, when local unbound concentrations differ substantially, or when transporter abundance changes other cellular properties.

Parallel Paths and Regional Perfusion

For parallel pathways \(k=1,\ldots,K\), each receives inlet concentration \(C_{\mathrm{in}}\), has flow \(Q_k\), and has its own total escape fraction \(S_k\). If pathway \(k\) contains serial regions, then:

The total outlet concentration is the flow-weighted average of pathway outlet concentrations. Consequently, overall extraction is the flow-weighted average of pathway extractions:

When Does the Well-Mixed Approximation Suffice?

The well-stirred liver is a compact alternative: it uses one \(Q_H\) and one aggregate \(CL_{\mathrm{int}}\), with extraction \(f_{u,b}CL_{\mathrm{int}}/(Q_H+f_{u,b}CL_{\mathrm{int}})\). It is useful when regional differences are negligible for the question, when only aggregate parameters are identifiable, or when a simpler model is needed for a broader PBPK system. A zonated model is more appropriate when spatial gradients or perfusion heterogeneity are plausible determinants of exposure and can be parameterized or tested.

The difference between predictions is not guaranteed to have one direction in every possible architecture. The serial-zone and parallel-path formulas represent different flow structures, and heterogeneity in clearance or perfusion can affect the aggregate nonlinearly. Comparisons are meaningful only when both models use compatible definitions of clearance, binding, flow, and the measured hepatic output.

Several checks help identify a model that is being used outside its assumptions. Confirm that every clearance has units of volume per time and every flow fraction is dimensionless; check that parallel flows sum to \(Q_H\); and verify that escape fractions lie between zero and one. For serial zones, confirm that the outlet of each zone is the next zone’s inlet. Compare limiting behavior: if all regional clearances are zero, extraction should be zero; if \(f_{u,b}CL_{\mathrm{int},i}\) is small relative to flow, zonal extraction should be small. Large discrepancies from the well-mixed prediction should prompt checks of the assumed structure and scaling, not automatic acceptance of the more complex result.

Common Pitfall and Limitations

A common mistake is to multiply a whole-liver clearance by each zone’s abundance ratio and then sum the resulting values, while also treating each zone as a parallel region with the full hepatic blood flow. The reasoning is that each region has its own transporter abundance, so each should process the whole incoming blood supply. The correction is to define the anatomy first: a serial zone receives the flow that has passed through the preceding zone; a parallel pathway receives only its allocated flow. The regional clearances must also represent net removal consistently, rather than counting uptake as irreversible elimination when drug can return to blood.

Zonation adds biological detail but does not guarantee more accurate predictions. Regional transporter measurements may be uncertain, may not map directly to in vivo activity, and may vary by species or physiological condition. If only systemic concentration data are available, different combinations of zonal abundance, flow, and intrinsic clearance may produce similar predictions. A model should therefore report its structural assumptions, retain a well-mixed comparison where useful, and avoid claiming that a spatial mechanism is established merely because a zonated model can fit observations.