Representing Perpetrator Concentrations at CYP Enzyme Sites

Why the Concentration at the Enzyme Site Matters

A CYP perpetrator can inhibit enzyme activity or increase enzyme abundance only to the extent that it reaches the relevant enzyme environment. The concentration measured in plasma is useful, but it is not automatically the concentration at hepatic or intestinal CYP enzymes. Differences in protein binding, tissue partitioning, uptake, efflux, and intracellular binding can make these concentrations substantially different.

The practical task is to select a concentration metric that corresponds to the concentration scale used to estimate the perpetrator's potency. For example, an unbound concentration should be compared with an unbound inhibition constant, not with a nominal assay concentration that has not been reconciled for binding. A mismatch can produce a mathematically valid model with a systematically incorrect interaction effect.

CYP3A, CYP2C, and other drug-metabolizing CYP enzymes are associated with the endoplasmic reticulum in cells. For hepatic CYP effects, a liver intracellular concentration is therefore a more direct conceptual target than plasma concentration. For intestinal CYP effects, the relevant target is intracellular exposure in enterocytes. These targets may be approximated from systemic plasma exposure when evidence and assumptions support that approximation, but they should not be treated as identical by default.

Total and Unbound Concentrations

Total concentration includes both bound and unbound drug. If \(C_{\mathrm{total}}\) is the total concentration in a particular matrix and \(f_u\) is the unbound fraction in that matrix, the corresponding unbound concentration is

Here \(f_u\) is dimensionless and lies between zero and one. The matrix matters: plasma \(f_u\) is not generally interchangeable with liver-tissue or enterocyte unbound fraction. The equation also assumes that the fraction is applicable to the concentration and conditions being represented. Concentration-dependent binding, changing matrix composition, or an inappropriate binding measurement can undermine that assumption.

Potency parameters must use a compatible concentration scale. If inhibition potency is expressed as an unbound \(K_{i,u}\), use an unbound concentration at the enzyme site in the ratio \(C_u/K_{i,u}\). If the potency value is a nominal concentration from an in vitro assay, determine whether assay binding has already been addressed before translating it. Where the assay unbound fraction \(f_{u,\mathrm{inc}}\) is known and the assay conditions justify the correction, an approximate conversion is \(K_{i,u}=f_{u,\mathrm{inc}}K_{i,\mathrm{nom}}\). Applying this correction twice would be just as problematic as not applying it when needed.

Representing Hepatic CYP Exposure

A useful way to connect plasma and liver exposure is the unbound liver-to-plasma partition coefficient:

If this ratio is applicable to the modeled situation, then \(C_{u,\mathrm{liver}}=K_{p,uu,\mathrm{liver}}C_{u,p}\). A value near one is consistent with similar unbound concentrations in the two matrices under the relevant conditions; a value above or below one indicates a difference. It is not safe to assume a value of one merely because passive diffusion is plausible. Active uptake or efflux, intracellular binding, pH gradients, metabolism, and non-equilibrium conditions can alter tissue exposure.

If total liver concentration is available, it can also be converted using \(C_{u,\mathrm{liver}}=f_{u,\mathrm{liver}}C_{\mathrm{total,liver}}\), provided the tissue unbound fraction is applicable. When total liver concentration and liver binding are used, the result should be consistent with the unbound partition relationship if all quantities refer to compatible conditions. A total liver-to-plasma ratio alone does not determine the unbound ratio unless binding in both matrices is also accounted for.

Representing Enterocyte CYP Exposure

For intestinal CYP effects, a model needs an enterocyte concentration or a defensible proxy for it. Oral dose or gut-lumen concentration alone is not the same as intracellular enterocyte concentration: only drug that is available to enter the cells, and remains there in an unbound form, contributes directly to intracellular CYP exposure. Permeability, intestinal transport, intracellular binding, local metabolism, and transit can all matter.

If an enterocyte model supplies total intracellular concentration and a suitable enterocyte unbound fraction, the site concentration can be calculated as \(C_{u,\mathrm{ent}}=f_{u,\mathrm{ent}}C_{\mathrm{total,ent}}\). Alternatively, a validated model may provide unbound enterocyte concentration directly. In either case, the inhibition or induction potency used with this input must correspond to the chosen intracellular concentration scale. A plasma-derived potency value may need careful interpretation before it can be paired with an enterocyte prediction.

A Practical Selection and Implementation Sequence

The concentration choice should be made separately for each relevant tissue and effect. When the same perpetrator has both inhibitory and inductive effects, the concentration metric used for each effect should be matched to its own potency evidence; the two effects need not be supported by identical assays or identical exposure surrogates.

For example, if an induction model uses an \(EC_{50}\) estimated against unbound intracellular exposure, its concentration input should be an estimate of unbound exposure at the relevant site. Substituting total plasma concentration changes the scale of the exposure-to-potency ratio and therefore changes the implied induction. The same compatibility principle applies to inhibition parameters such as \(K_i\), although the functional equation depends on the selected inhibition model.

Derivation, Diagnostics, and Limits

The concentration conversion follows directly from the definition of unbound fraction: \(f_u=C_u/C_{\mathrm{total}}\). Multiplying both sides by \(C_{\mathrm{total}}\) gives \(C_u=f_uC_{\mathrm{total}}\). Likewise, the definition of \(K_{p,uu,\mathrm{liver}}\) gives \(C_{u,\mathrm{liver}}=K_{p,uu,\mathrm{liver}}C_{u,p}\). These are accounting relationships, not guarantees that a measured fraction or partition coefficient remains constant over dose, time, or physiological conditions.

A practical diagnostic is to trace every concentration-to-potency ratio back to its units and matrix. If a model compares total plasma concentration in \(\mathrm{\mu M}\) with an unbound \(K_i\) in \(\mathrm{\mu M}\), matching units alone does not make the comparison valid. Check also whether a liver concentration was calculated from total liver content using an appropriate liver binding estimate, or from plasma using a supported \(K_{p,uu}\); do not combine both approaches as though they were independent corrections.

The main alternatives are not simply “right concentration” and “wrong concentration.” A model may use measured tissue concentration, a PBPK-predicted intracellular concentration, an unbound plasma concentration as a proxy, or a total concentration under a potency convention that is explicitly compatible with it. These choices differ in data requirements and assumptions. A tissue-specific estimate is more direct in concept but may be difficult to measure; a plasma proxy is convenient but can conceal tissue partitioning; and a total-concentration approach can be defensible only when the potency parameter and model formulation use the same convention.

Important limitations include uncertain tissue binding, concentration-dependent transport, lack of equilibrium between plasma and tissue, and imperfect correspondence between in vitro assay conditions and in vivo enzyme environments. A single partition coefficient may not capture changing exposure over time or across doses. Enterocyte concentration estimates can be especially sensitive to assumptions about absorption and intracellular disposition. These uncertainties should be represented or explored rather than hidden by reporting a single precise site concentration.

Common Mistake

• Wrong reasoning: “The perpetrator has a total plasma concentration of 6 \(\mathrm{\mu M}\), and its unbound \(K_i\) is 0.12 \(\mathrm{\mu M}\), so the inhibition ratio is \(6/0.12=50\).”
• Correction: The concentration and potency are on different binding scales. If plasma unbound fraction is 0.080 and plasma is the justified site proxy, then \(C_{u,p}=0.48\ \mathrm{\mu M}\) and the ratio is \(0.48/0.12=4.0\). If the target is liver or enterocyte, estimate that site's unbound concentration instead, using supported tissue information.