Predicting Intestinal P-Glycoprotein and BCRP Inhibition

Why intestinal efflux inhibition needs a supply check

P-glycoprotein (P-gp) and breast cancer resistance protein (BCRP) are apical efflux transporters in the intestine. For a substrate that enters an enterocyte, efflux can return drug to the lumen instead of allowing it to cross the basolateral side toward portal blood. Inhibiting that efflux can increase the amount of substrate reaching the systemic circulation. The size of the change depends not only on inhibition, but also on how much dissolved substrate is available to the transporter and on what happens after intestinal absorption.

The earlier material on transporter substrates and inhibitors establishes the evidence needed to assign a transporter role; the earlier material on converting inhibition data and estimating inhibitor exposure supplies the relevant in vitro inputs and intestinal-site exposure. Here, the task is to translate those inputs into a prediction for intestinal substrate absorption and, with additional assumptions, systemic exposure. A useful prediction should distinguish a transporter-limited case, where reducing efflux can matter, from a dissolution-limited case, where little dissolved drug is available to be absorbed.

A simplified model of efflux inhibition

For a low substrate concentration relative to the transporter's saturation range, and under reversible competitive inhibition, a simple approximation reduces the efflux rate by a factor \(1+I_{u,g}/K_i\). If both transporters contribute, and their inhibition acts independently, the effective apical efflux rate is

For a model with only one relevant transporter, the other rate is set to zero. This expression assumes the inhibitor concentration at the site is represented by one value, inhibition is reversible, and there is no material change in transporter abundance. It is not a general replacement for a saturable transport model: if substrate or inhibitor concentrations approach relevant saturation ranges, a model that represents that saturation may be needed.

A useful single-pass approximation treats basolateral crossing and apical efflux as competing exits from the enterocyte. The probability of crossing basolaterally after one entry is

This is a probability for the simplified competition, not a complete whole-gut absorption model. In particular, it does not explicitly represent repeated entry after a molecule is returned to the lumen, intestinal transit, enterocyte metabolism, or dissolution. Those processes can be represented in a fuller PBPK model. The approximation is useful for seeing why inhibition has little effect if efflux is already small relative to basolateral crossing, but can have a larger effect when efflux is a substantial competing route.

From crossing probability to oral exposure

A PBPK model may represent intestinal absorption and first-pass escape separately. Oral exposure can be expressed as

where \(CL\) is systemic clearance. This relationship assumes linear pharmacokinetics over the dose range considered and consistent AUC and dose definitions. If an inhibitor changes only the intestinal terms and does not change \(F_h\) or \(CL\), the predicted exposure ratio is

This ratio is not automatically equal to a ratio of enterocyte crossing probabilities. The probabilities are model components; their impact on \(F_a\) and \(F_g\) depends on the model's representation of dissolution, transit, first-pass loss, and repeated transporter cycling. Nor should an intestinal prediction be presented as a systemic exposure prediction without checking the assumption that hepatic first-pass escape and systemic clearance remain unchanged.

Checking whether dissolution limits the effect

Efflux inhibition changes a transport process; it does not directly make an undissolved solid dose dissolve faster. A practical mass-balance view is to track the dissolved amount in the lumen, the rate at which solid drug dissolves, the rate of absorption from dissolved drug, and loss through transit or other processes. If dissolution supplies dissolved substrate more slowly than the intestine can absorb it, increasing the transport capacity may not increase the amount absorbed over the relevant period. If dissolved supply is adequate and efflux competes materially with basolateral crossing, inhibiting efflux can have a larger effect.

This comparison is a diagnostic, not an unconditional rule. Efflux can alter the concentration gradient that drives entry, and returned drug may re-enter the enterocyte; therefore, a coupled gut model may predict some effect even when dissolution is slow. The key question is whether the predicted change survives when dissolution and transit are represented plausibly. A fraction-absorbed estimate based only on transporter parameters can overstate the interaction if it assumes more dissolved substrate is available than the formulation can provide.

Derivation and interpretation of the simplified result

The single-pass expression follows from competing first-order exits. Let the total rate of leaving the enterocyte be \(k_b+k_{\mathrm{eff}}\). For a small time interval \(dt\), the probability of a basolateral exit is approximately \(k_bdt\), while the probability of any exit is approximately \((k_b+k_{\mathrm{eff}})dt\). Equivalently, the time to exit is exponentially distributed with total rate \(k_b+k_{\mathrm{eff}}\), and the fraction of that rate assigned to basolateral crossing is \(k_b/(k_b+k_{\mathrm{eff}})\). Thus the probability that the first exit is basolateral is \(p_{\mathrm{cross}}\) as defined above.

If inhibition lowers \(k_{\mathrm{eff}}\) while \(k_b\) remains fixed, then \(p_{\mathrm{cross}}\) increases: for \(k_{\mathrm{eff,inh}}<k_{\mathrm{eff,control}}\),

This justifies the direction of the effect in the simplified model. It does not establish the magnitude of a whole-body exposure change. That magnitude requires a model linking the transporter process to absorbed fraction, gut escape, and systemic pharmacokinetics.

Diagnostics, alternatives, and limitations

A prediction is more credible when it is consistent with the dose form and the modeled gut processes, rather than being driven by a single inhibition constant. Inspect simulated dissolved drug, luminal drug remaining, intestinal efflux, and the time course of absorption. If almost all drug remains undissolved, a large predicted increase in absorbed amount warrants particular scrutiny. If the predicted effect persists only under a high assumed intestinal inhibitor concentration, revisit the unbound site exposure and inhibition inputs. Sensitivity analysis across plausible values of \(I_{u,g}\), \(K_i\), dissolution, and transit can show whether the conclusion is robust.

The simplified formula offers an interpretable alternative to a full PBPK gut representation, but it cannot resolve spatial variation along the intestine, changing luminal concentrations, transporter saturation, repeated efflux and re-entry, precipitation, or formulation-dependent dissolution. A more detailed model can represent these mechanisms explicitly, at the cost of additional parameters and data requirements. A simple empirical exposure ratio may summarize observed conditions, but on its own it does not separate transporter inhibition from dissolution or other mechanisms. Choose the level of detail that is supported by the available evidence, and avoid implying that added complexity guarantees a better prediction.

A common mistake is to take the inhibitor concentration divided by \(K_i\), apply the resulting efflux reduction, and report that same fold change as the AUC increase. The correction is to treat inhibition as an input to the intestinal transport model, calculate its effect on absorption and gut escape, and then calculate exposure using the relevant disposition terms. A second mistake is to assume every transporter substrate will show a large oral interaction. Substrate recognition is not enough: efflux must be quantitatively important under the dose and formulation conditions, and dissolved drug must be available for the altered transport process to affect absorption.