Why Evidence Must Converge
A drug may interact with a transporter in a laboratory assay without being meaningfully transported or inhibited in patients. Conversely, a transporter can contribute to disposition even when a single assay gives an ambiguous result. The practical task is therefore not to label a compound from one signal, but to decide whether multiple lines of evidence support a specific, biologically plausible role.
Two claims must be kept distinct. A substrate claim is that a transporter carries the drug across a membrane under relevant conditions. An inhibitor claim is that the drug reduces transport of another molecule through that transporter. A drug can be a substrate, an inhibitor, both, or neither at the concentrations and conditions tested. Neither claim alone establishes the magnitude of a clinical drug–drug interaction (DDI).
What Makes a Substrate Assay Persuasive?
Transport assays answer different questions. In an uptake assay, drug accumulation is compared between cells expressing the transporter of interest and matched control cells. A higher initial uptake rate in transporter-expressing cells is supportive only if the difference is reproducible and is reduced by a suitable transporter inhibitor or by another specific perturbation, such as loss of transporter expression. Initial rates are particularly useful because later measurements can be affected by metabolism, binding, or movement back out of cells.
For an efflux transporter, polarized cells can measure apparent permeability in both directions across a cell layer. The efflux ratio is defined as permeability from the basolateral to the apical side divided by permeability from the apical to the basolateral side. A ratio above one may indicate directional efflux, but it is not by itself transporter-specific: cell polarity, other transporters, and experimental artifacts can also produce directional differences. A reduction in the ratio with a selective inhibitor or transporter-deficient cells strengthens the inference.
Before interpreting either design, check that the assay can detect transport at all. Useful controls include a known substrate, matched cells without the transporter, a selective inhibitor where available, cell viability, and appropriate measurement of nonspecific binding. In polarized-cell studies, confirm monolayer integrity and assess recovery of the compound. Low or inconsistent recovery can make permeability comparisons unreliable, while a damaged monolayer can allow passive leakage to overwhelm a transporter signal.
Worked Example: Uptake Supports OATP1B1 Transport
In a hypothetical initial-rate assay, uptake of compound A is measured in matched cells expressing OATP1B1 and in control cells. Uptake is 0.18 and 0.06 pmol/min/mg protein, respectively. A selective OATP1B1 inhibitor reduces uptake in the expressing cells to 0.07 pmol/min/mg. A known OATP1B1 substrate responds as expected, and cell viability is similar between conditions.
The expression-associated uptake difference is \(0.18-0.06=0.12\) pmol/min/mg. Relative to control-cell uptake, the expressing cells show \(0.18/0.06=3.00\)-fold uptake. The inhibitor-sensitive component in the expressing cells is \(0.18-0.07=0.11\) pmol/min/mg, or \(0.11/0.18=0.611\), approximately 61.1% of uptake in that condition.
The higher uptake in expressing cells, together with the inhibitor-sensitive component and functioning positive control, supports OATP1B1-mediated uptake of compound A in this assay. It does not establish that OATP1B1 is the only uptake route in humans, or that it controls the compound’s clinical exposure.
Worked Example: Interpreting an Efflux Ratio
In a hypothetical polarized-cell experiment, compound B has apical-to-basolateral permeability of \(1.2\times10^{-6}\) cm/s and basolateral-to-apical permeability of \(7.2\times10^{-6}\) cm/s. With a selective P-glycoprotein inhibitor, the corresponding values are \(2.4\times10^{-6}\) and \(3.6\times10^{-6}\) cm/s. Recovery is 88% without inhibitor and 91% with inhibitor, and an integrity marker indicates intact monolayers in both conditions.
Without inhibitor, the efflux ratio is \((7.2\times10^{-6})/(1.2\times10^{-6})=6.00\). With inhibitor, it is \((3.6\times10^{-6})/(2.4\times10^{-6})=1.50\). The ratio falls by \((6.00-1.50)/6.00=0.750\), or 75.0%.
The marked, inhibitor-sensitive directional transport, supported by acceptable recovery and integrity checks, is consistent with P-glycoprotein-mediated efflux in this assay. The residual ratio of 1.50 and the possibility of other transport processes mean the result should not be described as proof of exclusive P-glycoprotein transport.
Evaluating Inhibitor Evidence
Inhibition assays typically measure transport of a probe substrate with and without the suspected inhibitor. Convincing evidence is stronger when inhibition is concentration-dependent, reproducible, observed with an appropriate positive control, and not explained by cytotoxicity, precipitation, assay interference, or generalized disruption of membranes. Testing more than one relevant transporter can reveal whether the effect is selective or broad. A nominal concentration that inhibits several unrelated transport systems only at very high levels is a weak basis for assigning a specific clinical mechanism.
An in vitro potency estimate, such as an IC50, describes the concentration producing a specified degree of inhibition under the assay conditions used. It is not automatically the concentration that will produce the same effect in vivo. Protein binding, assay composition, substrate concentration, active metabolites, and concentrations near the transporter can all matter. Thus, comparison with clinical exposure is a plausibility check, not a stand-alone prediction or a substitute for evaluating exposure at the relevant site.
Worked Example: Potent In Vitro Inhibition but Uncertain Clinical Relevance
In a hypothetical study, compound C inhibits OATP1B1-mediated probe uptake with an IC50 of 0.80 µM. Its measured maximum unbound plasma concentration at the tested clinical dose is 0.050 µM. The compound also inhibits CYP3A, and a clinical probe used in a DDI study is a substrate of both CYP3A and OATP1B1.
The simple exposure-to-potency comparison is \(0.80/0.050=16.0\): the in vitro IC50 is 16-fold higher than the unbound plasma concentration. This comparison makes substantial hepatic OATP1B1 inhibition less immediately plausible based on plasma exposure alone, but it does not rule it out. Concentrations at the transporter site, metabolites, and differences between the assay and clinical conditions have not been resolved.
Suppose the clinical probe’s area under the concentration–time curve rises 1.25-fold with compound C. That observation establishes a clinical exposure change in the hypothetical study, but does not identify OATP1B1 as its cause: CYP3A inhibition is a competing explanation. The defensible conclusion is that compound C inhibits OATP1B1 in vitro, while its clinical transporter-inhibitor relevance remains uncertain without evidence that separates transporter effects from CYP3A effects and supports plausible exposure at the transporter site.
Clinical Evidence and Causal Attribution
A clinical DDI can support transporter involvement when the victim drug is a well-characterized substrate, the perpetrator has credible inhibitory activity at relevant exposure, and alternative mechanisms are limited or independently assessed. The interpretation is more persuasive when results align across evidence types—for example, transporter-dependent uptake in vitro, a clinical exposure change consistent with reduced uptake, and no strong competing metabolic or renal explanation.
Clinical observations are rarely transporter-specific by themselves. A perpetrator may affect enzymes and several transporters; the victim may have multiple clearance routes; and changes in absorption, renal function, or protein binding can alter exposure. A change in total plasma concentration also need not identify which tissue or transport step changed. Clinical pharmacogenetic or expression differences can provide supporting evidence, but associations can be confounded and do not replace mechanistic assay validation.
Worked Example: A Clinical Signal with a Competing Mechanism
In a hypothetical study, exposure to victim drug D increases from 100 to 150 mg·h/L when coadministered with perpetrator E. Drug D is transported by OATP1B1 in a validated uptake assay. Perpetrator E inhibits OATP1B1 in vitro, but it also inhibits an enzyme responsible for part of drug D’s metabolism. No independent clinical measurement separates the two pathways.
The exposure ratio is \(150/100=1.50\), so drug D exposure is 1.50-fold higher during coadministration in this hypothetical comparison. The result is consistent with reduced OATP1B1-mediated uptake if that uptake normally contributes to drug D disposition. It is also consistent with reduced metabolism due to enzyme inhibition.
Because both mechanisms predict an exposure increase and both are plausible here, the clinical result alone cannot assign the increase to OATP1B1. The justified conclusion is that the study demonstrates a DDI and that transporter inhibition is one candidate mechanism, not that it has been isolated as the cause.
A Practical Evidence Review
A structured review prevents a vivid but weak result from dominating the conclusion. For substrate status, ask whether the assay demonstrates transporter-dependent movement, whether controls establish specificity, and whether the result is reproducible at interpretable concentrations. For inhibitor relevance, ask whether inhibition is concentration-dependent and sufficiently selective, whether the experimental conditions are credible, and whether relevant clinical exposure could plausibly reach the transporter. Then ask whether clinical observations support the same mechanism or leave strong alternatives.
Specify the transporter, the drug’s proposed role as substrate or inhibitor, and the tissue or direction of transport at issue.
Check matched controls, positive controls, selective perturbation, viability, recovery, and monolayer integrity as applicable.
Consider other transporters, nonspecific effects, assay concentrations, and whether clinical exposure could support the proposed role.
Compare in vitro findings with clinical evidence, identify competing pathways, and state what is supported versus unresolved.
Diagnostics, Alternatives, and Limitations
Signals deserve caution when transporter-expressing and control cells behave similarly, the positive control fails, inhibition appears only near cytotoxic concentrations, or measured recovery is poor. A high efflux ratio without inhibitor sensitivity or an orthogonal confirmation is also less persuasive. If a result depends strongly on one cell line, one probe concentration, or one inhibitor, test whether expression level, substrate concentration, or inhibitor selectivity explains the finding.
No single assay is universally superior. Uptake systems can isolate an influx process but may differ from tissue architecture and transporter expression in vivo. Polarized monolayers capture directional movement across a barrier but may express multiple transporters and can be sensitive to monolayer integrity. Vesicle assays can characterize transport activity in a controlled membrane preparation, but they do not reproduce whole-cell disposition. Genetic or expression-based evidence can support attribution but may be confounded by correlated biological differences. The choice should match the transport question, and independent methods are valuable when a conclusion has substantial consequences.
The main limitation is that evidence establishes varying degrees of plausibility, not automatic clinical importance. Assays simplify biological systems; clinical exposure at the transporter site may be uncertain; and observed DDIs can reflect several pathways at once. The appropriate outcome may therefore be “supported in vitro, clinical relevance unresolved,” rather than a forced yes-or-no classification.
Common Mistake
Wrong reasoning: “The drug increased a victim’s AUC in a clinical study, so it must inhibit the transporter that carries that victim.” The conclusion does not follow: the perpetrator could inhibit an enzyme, another transporter, or more than one pathway.
Correction: Treat the clinical change as evidence of an interaction, then evaluate transporter attribution separately. Require a plausible transporter-specific assay signal, exposure plausibility, and consideration of competing mechanisms. Keep the certainty of the claim proportional to the evidence.
Check Your Understanding
Answer each question, then compare your reasoning with the solutions.
- In an uptake assay, transporter-expressing cells show greater uptake than control cells, but a selective inhibitor does not reduce uptake. What does this support, and what remains uncertain?
- A polarized-cell assay reports apical-to-basolateral permeability of \(2.0\times10^{-6}\) cm/s and basolateral-to-apical permeability of \(8.0\times10^{-6}\) cm/s. Calculate the efflux ratio.
- Why does a clinical exposure increase not, by itself, prove transporter inhibition?
- An inhibitor’s in vitro IC50 is 2.0 µM and its measured unbound plasma concentration is 0.10 µM. Calculate the ratio and state one limitation of interpreting it.
Solutions. (1) The higher uptake is consistent with an expression-associated difference, but without inhibitor sensitivity or another specific perturbation, transporter attribution remains uncertain. (2) The efflux ratio is \((8.0\times10^{-6})/(2.0\times10^{-6})=4.00\). It indicates directional transport in that assay but does not identify the responsible transporter without additional evidence. (3) Enzyme inhibition, other transporters, or changes in absorption or elimination may also alter exposure; the clinical signal does not isolate a mechanism. (4) The ratio is \(2.0/0.10=20.0\). It is only a plausibility screen: plasma concentration may not represent concentration at the transporter site, and assay conditions may differ from clinical conditions.