Why Transporter Location Matters
A transporter-mediated drug interaction cannot be predicted from the word “inhibition” alone. The direction of the exposure change depends on which transporter is inhibited, where it acts, and what happens to the drug after transport. Blocking hepatic uptake can leave more substrate in plasma while reducing its entry into hepatocytes. Blocking apical efflux from intestinal cells can instead allow more drug to reach the circulation. A prediction that ignores site and direction can therefore get even the sign of the interaction wrong.
The focus here is on hepatic uptake by OATP1B transporters and efflux by P-glycoprotein (P-gp) and breast cancer resistance protein (BCRP). Each has more than one clinically relevant location. The same substrate can also be transported by multiple proteins or handled by enzymes and other clearance routes. The predictions below are consequently conditional: each calculation states the simplifying assumptions that make it interpretable.
Map the Transporter to the Site and Direction
OATP1B1 and OATP1B3 are expressed on the blood-facing membrane of hepatocytes. Their uptake can make a circulating drug available for metabolism or biliary excretion. If a drug depends on this uptake for hepatic clearance, inhibition may reduce its removal from plasma and raise systemic exposure. At the same time, less drug may enter the hepatocyte, so plasma exposure and liver exposure need not change in the same direction.
P-gp and BCRP are efflux transporters with important roles at several barriers. In the intestine, apical efflux returns drug from an enterocyte to the gut lumen; inhibiting that process can increase the fraction reaching the blood. In the liver, canalicular efflux moves drug from a hepatocyte into bile. In the kidney, apical efflux can move drug from a tubular cell into urine. Blocking efflux at these latter sites can reduce a route of elimination, but the size of the plasma effect depends on how much total clearance that route contributes and on other transport processes.
| Transporter and site | Normal direction | Possible consequence of inhibition |
|---|---|---|
| OATP1B1/1B3, liver sinusoidal membrane | Blood to hepatocyte | Less hepatic entry; systemic exposure may rise if uptake supports hepatic clearance |
| P-gp/BCRP, intestinal apical membrane | Enterocyte to gut lumen | More drug may reach blood; oral bioavailability may rise |
| P-gp/BCRP, liver canalicular membrane | Hepatocyte to bile | Biliary elimination may fall; intracellular exposure may rise |
| P-gp/BCRP, renal tubular apical membrane | Tubular cell to urine | Secretion may fall; systemic exposure may rise if renal secretion is important |
These are mechanistic possibilities, not guarantees. For example, intestinal efflux inhibition can have little effect if absorption is already nearly complete or if the drug is not a meaningful substrate at clinical concentrations. A rise in plasma exposure after hepatic uptake inhibition also does not imply a rise in hepatocyte exposure.
A Simple Quantitative Framework
For a linear pharmacokinetic comparison after the same oral dose, \(AUC = F \cdot Dose/CL\), where \(F\) is bioavailability and \(CL\) is systemic clearance. Thus, the exposure ratio between an inhibited condition and a control condition is
This identity separates two broad mechanisms: a change in drug reaching the circulation and a change in drug being cleared after it reaches the circulation. It assumes comparable doses, linear pharmacokinetics over the relevant range, and an AUC measured over a suitable interval. It does not identify which transporter caused either change; that requires mechanistic evidence.
For an illustrative hepatic uptake calculation, suppose systemic clearance consists of an uptake-dependent hepatic component \(CL_H\) and all other clearance \(CL_O\). Assume uptake inhibition leaves a fraction \(r\) of the uptake-dependent clearance, while \(CL_O\), bioavailability, binding, and all other processes remain unchanged. If \(f_H=CL_H/(CL_H+CL_O)\) is the baseline fraction of total clearance due to the hepatic component, then:
The derivation uses \(CL_H+CL_O=CL_{\mathrm{ctl}}\), \(CL_{\mathrm{inh}}=rCL_H+CL_O\), and the assumption that \(F\) is unchanged. If hepatic uptake accounts for all baseline clearance, \(f_H=1\), and the ratio simplifies to \(1/r\). If uptake contributes little to total clearance, the same degree of transporter inhibition produces a smaller systemic effect.
The parameter \(r\) is a simplified residual fraction of pathway activity or clearance, not a universal property of a transporter inhibitor. In a simple competitive inhibition assay at low substrate concentration, an idealized fractional inhibition can be related to unbound inhibitor concentration \(I_u\) and inhibition constant \(K_i\) as \(I_u/(K_i+I_u)\). Clinical use of that relation requires the relevant inhibitor concentration at the transporter, an appropriate assay, and consideration of substrate concentration and competing pathways. A plasma concentration alone may not represent inhibitor exposure at the intestinal lumen or inside a tissue.
Worked Example: OATP1B Inhibition of an Uptake-Dependent Substrate
Consider a hypothetical oral drug whose baseline total clearance is 10 L/h. Under the stated model, 8 L/h is OATP1B-dependent hepatic clearance and 2 L/h is other clearance. Assume an inhibitor reduces the hepatic component to 40% of baseline, does not change bioavailability, and does not affect the other clearance.
The baseline hepatic fraction is \(f_H=8/10=0.800\), and the residual fraction is \(r=0.400\). The predicted exposure ratio is:
Checking by direct clearance calculation, inhibited clearance is \(0.400(8)+2=5.2\) L/h, so the ratio is \(10/5.2=1.923\). Under the assumptions, AUC is predicted to increase about 1.92-fold. This is a plasma-exposure prediction; it does not establish that hepatocyte concentration increases.
Worked Example: Intestinal P-gp Efflux and Oral Bioavailability
Suppose a hypothetical oral P-gp substrate has \(F_{\mathrm{ctl}}=0.50\) without an inhibitor and \(F_{\mathrm{inh}}=0.75\) with an intestinal P-gp inhibitor. Assume the inhibitor changes neither systemic clearance nor dose proportionality, and that the bioavailability estimates already reflect the net intestinal effects.
The predicted AUC ratio is:
AUC is predicted to rise 1.50-fold, or 50%, because more of the dose reaches the circulation in this hypothetical case. The calculation does not imply that intestinal efflux inhibition always produces this increase: if absorption is already complete, or another absorption barrier dominates, the change in \(F\) may be much smaller.
Worked Example: Renal Efflux as One Component of Total Clearance
A hypothetical drug has total clearance of 12 L/h, including 3 L/h of P-gp-dependent renal secretion. Assume a P-gp inhibitor reduces that secretion to 20% of baseline, with unchanged bioavailability and all other clearance routes. The baseline renal component is \(f_R=3/12=0.250\); its residual fraction is \(r=0.200\).
The inhibited total clearance is \(0.200(3)+9=9.6\) L/h. Therefore:
Under these assumptions, AUC rises 1.25-fold. Even though the renal secretion component falls by 80%, total clearance falls by only 20%, because the other 9 L/h of clearance remains unchanged. This is why a large effect on transporter activity does not by itself establish a large effect on total exposure.
Evaluating Evidence for OATP1B, P-gp, and BCRP
Transporter involvement is best assessed by combining lines of evidence rather than treating one assay result as proof of a clinical interaction. In vitro work should establish that the drug is transported by the relevant protein and should distinguish uptake from efflux using suitable cells, controls, and assay conditions. A transporter being inhibited by a drug is not the same as that drug being a transporter substrate.
A practical evidence review asks:
- Is the drug a substrate? Look for transport that is attributable to the protein, with appropriate comparison systems and concentrations relevant to clinical exposure.
- Can the proposed inhibitor reach the site? Compare relevant unbound concentrations with the inhibition potency, accounting for the intestinal lumen, liver, or kidney as appropriate.
- Does the clinical pattern fit the site? A change in oral exposure may involve intestinal bioavailability, systemic clearance, or both; plasma AUC alone does not uniquely identify the mechanism.
- Are there independent supporting data? Clinical interaction studies and genetic evidence can support a role, but neither automatically isolates one transporter when substrates or inhibitors have multiple actions.
Some substrates illustrate why attribution requires care. Digoxin is a widely used P-gp substrate, but an interaction involving digoxin still needs interpretation in light of study design and other effects of the interacting drug. Rosuvastatin can be transported by OATP1B and BCRP, among other disposition processes. A change in rosuvastatin exposure during inhibition therefore cannot automatically be assigned to only OATP1B or only BCRP. Treating a drug as a single-transporter probe when it is not can lead to a misleading mechanistic conclusion.
Diagnostics, Alternatives, and Limitations
A prediction is becoming unreliable when its assumed mechanism does not match the observed pattern. Warning signs include an effect that is much larger or smaller than expected from the proposed pathway contribution, different changes in AUC and peak concentration that the simple model cannot explain, or inconsistent effects across substrates thought to share the same transporter. These patterns do not prove a particular alternative mechanism, but they are reasons to revisit assumptions about substrate status, site of inhibition, parallel transporters, metabolism, and bioavailability.
For a first-order prediction, the alternatives have different strengths. A clinical exposure ratio directly describes the observed interaction but may not explain its mechanism or transport to a different dose or population. A static calculation makes assumptions and pathway contributions visible, but can be sensitive to uncertain inhibitor concentrations and transporter contributions. A mechanistic physiologically based pharmacokinetic model can represent multiple tissues, transporters, and routes, but adds parameter and structural uncertainty; its complexity does not make weak input evidence strong.
The simple calculations here have important limits. They assume linear behavior, stable dose and bioavailability except where explicitly varied, and no change in other clearance pathways. Transport can be saturable, uptake and efflux may operate in series, and inhibition may change both hepatic and intestinal processes. Total plasma exposure also does not directly report concentrations at the site of pharmacologic action or toxicity. A clinically useful conclusion should therefore distinguish the predicted plasma interaction from what remains uncertain about tissue exposure and mechanism.
Check Your Understanding
Answer each question, then compare with the concise solutions below.
- Why can inhibition of hepatic OATP1B raise plasma exposure while reducing drug entry into hepatocytes?
- A drug’s total clearance is 20 L/h, of which 5 L/h is an uptake-dependent hepatic component. Inhibition leaves 50% of that component. Assuming unchanged bioavailability and other clearance, what is the predicted AUC ratio?
- Why does inhibiting intestinal P-gp not necessarily increase oral AUC?
- Rosuvastatin exposure rises during inhibition of a drug that can affect both OATP1B and BCRP. Why is that observation insufficient to assign the entire interaction to one transporter?
Solutions. (1) Reduced uptake can limit hepatic access to metabolism or biliary elimination, leaving more drug in plasma even as less enters hepatocytes. (2) The inhibited clearance is \(0.50(5)+15=17.5\) L/h, so the AUC ratio is \(20/17.5=1.143\), or about a 14.3% increase under the assumptions. (3) Absorption may already be nearly complete, or another process may limit bioavailability; the net change in \(F\) may therefore be small. (4) The observed change is compatible with effects on either transporter and may also reflect other disposition pathways; the plasma exposure result alone does not isolate the cause.