Roles Belong to an Interaction, Not to a Drug in Isolation
In a drug-drug interaction (DDI), the victim is the drug whose disposition or exposure is changed, and the perpetrator is the drug that causes that change. These labels describe the direction of a specified interaction. They are not permanent classifications of a drug: the same drug can be a victim in one relationship and a perpetrator in another.
For a CYP-mediated interaction, assigning the roles is the first step in organizing the PBPK model. The victim is connected to the substrate pathway and the exposure outcome being evaluated. The perpetrator is connected to its effects on the relevant CYP enzyme and to the concentrations that drive those effects at the enzyme’s site. The enzyme and tissue are part of the interaction description; saying only that one drug “interacts with” another leaves important model components unspecified.
A drug is not a perpetrator merely because a CYP enzyme metabolizes it. Being a CYP substrate connects a drug to a metabolic pathway; being a perpetrator requires that the drug alter a pathway in a way that changes the other drug’s disposition. Conversely, a perpetrator need not be a substrate of the affected CYP. Role assignment therefore requires a causal question: which drug’s exposure is being changed, and which drug is responsible for the change?
Mapping Roles to PBPK Components
A CYP-mediated PBPK interaction can be organized around four linked components: the victim substrate, the CYP enzyme and its tissue location, the perpetrator’s local concentration and enzyme effect, and the resulting victim exposure. The map should identify the particular CYP isoform and site rather than treating CYP activity as a single body-wide quantity.
| Interaction component | Role connection | PBPK question |
|---|---|---|
| Victim substrate | Victim | Which drug’s disposition changes, and which CYP pathway contributes to its metabolism? |
| CYP enzyme and tissue | Shared interaction site | Which CYP isoform is affected, and is the relevant site hepatic, intestinal, or another modeled tissue? |
| Perpetrator effect | Perpetrator | Does the perpetrator inhibit, inactivate, or induce the enzyme, or otherwise alter its activity? |
| Perpetrator concentration | Perpetrator driver | What concentration reaches the relevant enzyme site over time? |
| Victim exposure outcome | Victim response | Which exposure measure, such as AUC or Cmax, changes under coadministration? |
A simple reversible competitive-inhibition expression illustrates how these pieces connect. Let \(C_{P,u,e}(t)\) be the unbound concentration of perpetrator \(P\) at enzyme site \(e\), \(K_i\) its inhibition constant for that CYP, and \(CL_{\mathrm{int},V,\mathrm{CYP},e}^{0}\) the victim \(V\)’s baseline intrinsic clearance through that pathway when enzyme activity is normalized to one. Under the assumptions stated below, the pathway’s effective intrinsic clearance is represented as:
The expression makes the role map explicit: the victim supplies the substrate pathway, the perpetrator supplies the site concentration and inhibition effect, and the CYP isoform and tissue identify where the effect acts. The formula assumes reversible competitive inhibition at rapid equilibrium, a single inhibitor represented by this expression, and concentrations expressed on a consistent unbound basis. It does not by itself predict whole-body AUC or Cmax; those outcomes depend on the full disposition model and the interaction’s sites.
For a brief static illustration, let \(f_m\) denote the fraction of baseline victim clearance attributable to the affected pathway. If other clearance is unchanged, the affected pathway’s relative clearance is \(1/(1+I_u/K_i)\), and total relative clearance is the sum of the unaffected and affected fractions. Under linear pharmacokinetics and unchanged bioavailability, the AUC ratio is the reciprocal of that relative clearance:
This is a simplified static relationship, not a substitute for a dynamic PBPK simulation. It is useful here to show why role labels must connect to specific model inputs: a victim pathway contribution and a perpetrator concentration relative to potency jointly determine the predicted change.
Worked Example: One Perpetrator Changes a Victim’s Exposure
Suppose Drug P reversibly inhibits CYP2C9, and Drug V is a CYP2C9 substrate. For a hypothetical static screening calculation, let \(f_m=0.70\) for Drug V and let the unbound perpetrator concentration at the relevant enzyme site be \(I_u=2.0\ \mu\mathrm{M}\), with \(K_i=4.0\ \mu\mathrm{M}\). Assume linear pharmacokinetics, unchanged bioavailability, no other interaction mechanism, and an unchanged non-CYP2C9 clearance component.
Drug V is the victim, because its disposition is the outcome that changes. Drug P is the perpetrator, because its concentration inhibits the CYP2C9 pathway. The concentration-to-potency ratio is:
The fraction of baseline CYP2C9 pathway clearance remaining in this simplified model is \(1/(1+0.500)=0.6667\). Thus the relative total clearance is:
With unchanged bioavailability and linear pharmacokinetics, the illustrative AUC ratio is:
The calculation predicts an AUC about 1.30 times the value without Drug P under these assumptions. It does not establish that an observed patient’s AUC will rise by exactly this amount: the simplified calculation omits factors such as time-varying concentrations, other affected sites, and any additional mechanisms.
Reciprocal Roles and Tissue-Specific Effects
A pair of drugs can have reciprocal roles. For example, Drug X can be a victim if Drug Y inhibits the CYP pathway that clears X, while X is also a perpetrator if X inhibits a CYP pathway that clears Y. These are two directional claims, each requiring its own substrate pathway, enzyme effect, and relevant perpetrator concentration. A drug’s role should not be inferred from the direction of the other claim.
Likewise, a perpetrator effect can differ by tissue. Oral dosing may expose intestinal CYP enzymes to concentrations that differ substantially from those at hepatic enzymes. A perpetrator can therefore be a strong inhibitor at one site and a weak inhibitor at another. The same victim may have relevant CYP-mediated metabolism at both sites. The role names remain the same, but the PBPK representation must locate each effect and use the corresponding concentration profile.
Worked Example: A Drug Has Both Roles in a Pair
Consider hypothetical Drugs X and Y, both metabolized by CYP3A. Suppose Y inhibits the CYP3A pathway for X, while X also inhibits CYP3A and changes Y’s exposure. For a simplified, directional screening illustration, let \(f_m=0.80\) for X and \(I_{Y,u}/K_{i,Y}=0.25\); for Y, let \(f_m=0.60\) and \(I_{X,u}/K_{i,X}=0.50\). Assume each calculation is made separately with a fixed perpetrator concentration, linear pharmacokinetics, unchanged bioavailability, and no mechanisms beyond reversible inhibition.
For the direction Y acting on X, X is the victim and Y the perpetrator. X’s CYP3A clearance fraction remaining is \(1/(1+0.25)=0.800\). Its relative total clearance and AUC ratio are:
For the reverse direction, X is the perpetrator and Y the victim. Y’s CYP3A clearance fraction remaining is \(1/(1+0.50)=0.6667\). Therefore:
Thus X is a victim in the Y-to-X direction and a perpetrator in the X-to-Y direction. These two separate calculations demonstrate the role assignment; they are not a joint prediction for simultaneous mutual inhibition. In a joint model, each drug’s time-varying concentration may depend on the other drug’s effect, so the paired changes cannot automatically be combined as if each concentration were fixed.
Worked Example: The Same Perpetrator at Intestinal and Hepatic CYP
Suppose hypothetical oral Drug P inhibits CYP3A, and Drug V is a CYP3A substrate with metabolism in both intestine and liver. Let the unbound concentrations of P at those sites be \(20\ \mu\mathrm{M}\) in the intestine and \(0.50\ \mu\mathrm{M}\) in the liver. Let the relevant \(K_i\) be \(5.0\ \mu\mathrm{M}\) at each site. Under the simple reversible-inhibition expression, calculate the fraction of CYP activity remaining at each site.
For intestine, the concentration ratio is \(20/5.0=4.0\), so the remaining activity fraction is:
For liver, the ratio is \(0.50/5.0=0.100\), giving:
P is the perpetrator at both sites, while V is the victim whose CYP3A-mediated disposition may change. The calculation suggests much stronger local inhibition in the intestine than in the liver under the stated assumptions. These activity fractions are not an overall AUCR: calculating that outcome requires the PBPK model to represent the victim’s disposition and the contributions of both sites.
Checking the Role Map
Before implementing an interaction, write the direction in words: “Drug P changes Drug V’s exposure by affecting CYP2C9 in the liver.” This short statement identifies a candidate perpetrator, victim, enzyme, site, and outcome. Then check that the model has corresponding quantities: the victim’s substrate pathway, the perpetrator’s mechanism and site concentration, and an exposure measure for the victim.
A dynamic PBPK model can represent time-varying concentrations and effects in tissues, whereas a static calculation summarizes an interaction using selected concentrations and simplifying assumptions. Neither approach changes the definitions of victim and perpetrator. The choice affects how much temporal and spatial detail is represented, not which drug has which role.
Useful diagnostics follow directly from the map. If the predicted victim exposure changes but the perpetrator concentration at the implicated site is negligible relative to the modeled potency, check the concentration input, units, and tissue assignment. If the perpetrator concentration is substantial but the predicted victim response is absent, check that the affected CYP pathway is connected to the victim’s disposition and that the modeled endpoint reflects that pathway. If an interaction is attributed to CYP3A without identifying intestine versus liver, the model may be missing a meaningful site distinction.
- Reversing the labels: the drug whose AUC is reported is usually the victim for that endpoint; it is not automatically the perpetrator.
- Equating substrate status with perpetrator status: a CYP substrate does not necessarily alter the enzyme or another drug’s disposition.
- Using plasma concentration as if it were automatically the site concentration: the relationship between systemic concentration and concentration at the enzyme site depends on the model and route.
- Ignoring direction: reciprocal interactions require two explicit directional role assignments, not one ambiguous label for the pair.
- Collapsing sites or enzymes: an effect on one CYP isoform or tissue does not establish the same effect on another.
Role assignment is an organizing step, not evidence that a proposed mechanism is correct. The available evidence may be incomplete, the relevant concentration may be uncertain, or more than one pathway may contribute. The map should preserve those uncertainties rather than imply that a single role label resolves them. This keeps the distinction clear between identifying the model components and establishing their values or adequacy.
Check Your Understanding
Answer each question, then compare with the brief solutions below.
- Drug A is a CYP2C19 substrate. Drug B inhibits CYP2C19 and increases A’s AUC. Which drug is the victim, which is the perpetrator, and what additional model location should be identified?
- Does evidence that Drug C is metabolized by CYP3A establish that C is a perpetrator? Explain.
- At a relevant enzyme site, a perpetrator has an unbound concentration of \(3.0\ \mu\mathrm{M}\) and \(K_i=6.0\ \mu\mathrm{M}\). Under the simple reversible-inhibition expression, what fraction of enzyme activity remains?
- Drug D changes Drug E’s exposure, and E also inhibits the CYP pathway that clears D. How should the roles be described?
- Why should a high intestinal perpetrator concentration not automatically be used as the hepatic enzyme-site concentration?
Solutions. (1) A is the victim and B the perpetrator; identify the CYP2C19 site, such as intestine or liver, relevant to the modeled pathway. (2) No. Metabolism by CYP3A makes C a substrate, but perpetrator status requires an effect of C that changes another drug’s disposition. (3) \(1/(1+3.0/6.0)=1/1.5=0.667\), so 66.7% of activity remains under the stated assumptions. (4) D is the perpetrator and E the victim in the D-to-E direction; E is the perpetrator and D the victim in the E-to-D direction. (5) Intestinal and hepatic concentrations can differ because tissue exposure is site- and route-dependent, so each enzyme site needs an appropriate concentration representation.