Transcellular and Paracellular Permeation

Why Separate Epithelial Routes?

Intestinal drug movement across the epithelium can occur through cells or between them. The distinction matters because these routes do not have the same physical barriers. Transcellular permeation crosses the apical membrane, the cell interior, and the basolateral membrane. Paracellular permeation passes through the aqueous spaces and junctional pathway between adjacent cells. Lipophilic compounds may permeate cell membranes readily, whereas small hydrophilic compounds may rely more heavily on the paracellular route, subject to its size and charge selectivity.

A measured or modeled effective permeability can summarize both routes when their fluxes respond linearly to the same concentration difference. But the summary can conceal which route carries the drug, and it can fail when route contributions change with location, concentration, time, or experimental conditions. This tutorial formulates the route contributions and sets out when a single parameter is a defensible simplification.

Flux Through Parallel Routes

Consider a local epithelial region at steady state. Let \(C_L\) be the free, permeating drug concentration at the luminal boundary and \(C_B\) the corresponding concentration at the basolateral boundary. For passive linear transport, the route-specific flux per unit area of that route is proportional to the concentration difference \(C_L-C_B\). Flux is positive from lumen to blood when \(C_L>C_B\).

The two pathways are treated as parallel routes: drug can cross through either route between the same sides of the epithelium. If \(f_t\) and \(f_p\) are the fractions of the chosen reference area assigned to transcellular and paracellular routes, respectively, their sum is one in this two-route representation. The total flux per unit reference area is the sum of the area-weighted route fluxes.

Here \(J_{\mathrm{total}}\) has units of amount per area per time when concentration is expressed as amount per volume. If both routes have the same boundary concentrations and remain linear, the equation can be written using a single effective permeability.

The fractions and route permeabilities must refer to a consistent area convention. In this formulation, \(P_t\) and \(P_p\) are permeabilities per unit area of their respective routes, while \(P_{\mathrm{eff}}\) is normalized to the chosen reference area. The \(f\) terms account for how much of that reference area each route occupies. In practice, route-specific surface areas may be difficult to measure; an empirical \(P_{\mathrm{eff}}\) can therefore summarize geometry and transport together.

What Is Inside the Transcellular Permeability?

The transcellular pathway includes sequential barriers. In a simple steady-state passive model, the same flux passes through the apical membrane, the cell interior, and the basolateral membrane. If each barrier has a linear permeability defined using the concentration difference across that barrier, the total concentration drop is the sum of the individual drops. This makes the permeation resistances add in series.

This expression assumes that there is no net accumulation within the cell at steady state, that the barriers can each be represented by a linear permeability, and that the route is passive. It is a useful conceptual decomposition, not a requirement that every PBPK model resolve three separate transcellular parameters. A fitted transcellular permeability may instead lump these resistances.

The equation also explains why a high permeability of one barrier does not guarantee high permeability of the complete route. A relatively low-permeability membrane or slow intracellular step can dominate the total resistance. In contrast, the paracellular route is commonly represented as a separate pathway with its own effective permeability, reflecting such features as accessible junctional area and pathway restrictions.

When the Route Contributions Differ

The algebraic sum remains useful when route permeabilities differ greatly. The relative contribution of a route is its flux divided by total flux, not simply its area fraction. For a common positive concentration difference, the transcellular fraction is

These fractions sum to one under the parallel, linear model. A route can occupy only a small area yet dominate if its route-specific permeability is sufficiently large. Conversely, an extensive route can carry little flux if its permeability is low.

When One Effective Permeability Is Inadequate

A single \(P_{\mathrm{eff}}\) is a useful reduction when it preserves the flux needed by the model over the conditions of interest. It is inadequate if the reduction cannot reproduce relevant changes in flux or if the route distinction is itself needed for interpretation. Important warning signs include:

• Different concentration dependencies. The linear sum assumes each route flux is proportional to the same concentration difference. If either route has a nonlinear relationship, a constant weighted permeability may fit only a limited concentration range.
• Different driving concentrations. If the routes experience different local free concentrations or concentration differences, one common \((C_L-C_B)\) cannot be factored out of the flux sum. This can occur when the model resolves local boundary layers or pathway-specific conditions.
• Time-dependent intracellular behavior. The simple serial transcellular expression is a steady-state result. If drug temporarily accumulates in cells, the transcellular flux can depend on time and intracellular state rather than only on the current lumen-to-blood concentration difference.
• Changing route contributions. A fitted effective permeability may not transfer if route properties or accessible areas differ across intestinal regions, experimental systems, or conditions. Regional differences in overall permeability do not by themselves identify which route caused them.
• Route-specific questions. A total flux may be predicted adequately while the model still cannot answer whether the flux is transcellular or paracellular. Route-specific data or assumptions are needed to make that distinction.

An effective permeability measured in an assay is therefore not automatically a molecular constant. It may summarize the compound, the epithelial preparation, the available pathways, and the assay conditions. The earlier material on effective permeability and intestinal absorption, and on translating permeability assays to human \(P_{\mathrm{eff}}\), provides the broader context for using such inputs in PBPK models.

Deriving the Parallel-Pathway Result

The single-parameter reduction follows directly from flux conservation across the two parallel routes. It does not require the routes to have equal permeability; it does require them to share the same driving concentration difference and to be linear under the conditions being modeled.

If the two routes instead have distinct driving differences, the correct expression remains \(J_{\mathrm{total}}=f_tP_t\Delta C_t+f_pP_p\Delta C_p\). Unless those differences are equal or have a known fixed relationship, this cannot be represented by the same constant \(P_{\mathrm{eff}}\) multiplying one common concentration difference. This is the precise mathematical reason a single parameter can fail.

Model Choices, Diagnostics, and Limitations

Useful diagnostics include checking whether a fitted permeability remains stable across concentration ranges, assay conditions, and relevant intestinal regions. Systematic residual patterns or a permeability estimate that shifts substantially across these settings may indicate that the lumped model is absorbing changing route contributions or other omitted processes. Such patterns are evidence to investigate, not proof that a particular route is responsible: total-flux data alone generally cannot uniquely separate \(f_tP_t\) from \(f_pP_p\).

A common mistake is to conclude that a pathway contributes little because it occupies little area. That reasoning confuses area with flux. The correction is to compare the weighted conductances \(f_tP_t\) and \(f_pP_p\); the pathway with the larger weighted conductance contributes more under the common-gradient model. Another mistake is to add serial permeabilities as if they were parallel. Serial barriers add as resistances, whereas parallel route fluxes add.