Why Genotype and Current Enzyme Activity Can Differ
A pharmacogenetic result describes inherited enzyme function, but it does not by itself describe all of the enzyme activity present when a patient receives a drug. A CYP2D6 genotype can predict a baseline activity score and a genotype-based phenotype. A CYP2D6 inhibitor, inflammation, or another disease-related change can then reduce functional enzyme activity without changing the genotype. This discrepancy is called phenoconversion: the functional phenotype under current conditions differs from the phenotype predicted from genotype alone.
The distinction matters in PBPK because genotype-based activity is not necessarily the activity that governs metabolism during a particular treatment. A patient genetically predicted to have normal CYP2D6 function could have much lower current activity while taking an inhibitor. Conversely, an activity-based assignment is conditional: it can change if the inhibitor is stopped or the disease state resolves. It should not be mistaken for a revised genetic result.
The earlier material on translating diplotypes into enzyme activity provides the starting activity score. Here the task is to adjust that score transparently, state the assumptions behind the adjustment, and interpret the resulting phenotype assignment without treating it as a universal biological measurement.
An Explicit Activity-Scaling Framework
Let \(AS_G\) denote the activity score predicted from genotype. For the examples below, use the following hypothetical allele values: CYP2D6 *1 has score 1, *10 has score 0.25, and *4 has score 0. A diplotype score is the sum of its allele scores, with copy number represented explicitly; for example, *1x2/*1 has score \(2+1=3\). Actual allele-function values and phenotype mappings must be taken from the specified clinical guideline or model version.
Represent each activity-reducing condition by a factor between 0 and 1. Let \(f_I\) be the fraction of activity remaining after inhibitor effects and \(f_D\) the fraction remaining after a disease-related change. Under a multiplicative approximation, the adjusted activity score is:
This equation is an accounting model, not a universal CYP2D6 law. It assumes the factors are suitable summaries of their effects on the same enzyme activity, that they can be combined multiplicatively, and that no relevant compensating process changes the result. A factor of 1 means no reduction in the modeled activity; a factor of 0 means no activity remains in the model. If evidence supports only a qualitative effect, assigning a precise factor would create unjustified precision.
For illustration, this tutorial uses the following phenotype bins: score 0 is poor metabolizer (PM); score greater than 0 and at most 1 is intermediate metabolizer (IM); score greater than 1 and at most 2 is normal metabolizer (NM); score above 2 is ultrarapid metabolizer (UM). These boundaries make the worked calculations unambiguous, but phenotype translation systems can differ, and some use discrete score ranges or special handling of particular values. In practice, use the version specified for the clinical question and record it alongside the assignment.
Estimating an Inhibitor Adjustment
The inhibitor factor should reflect the inhibitor's effect at the relevant concentration and conditions, not merely the fact that a drug is listed as an inhibitor. A simple mechanistic justification is possible for reversible competitive inhibition. For an enzyme following Michaelis–Menten kinetics, let \(S\) be substrate concentration, \(K_m\) the Michaelis constant, \(I_u\) the unbound inhibitor concentration, and \(K_i\) the inhibition constant. With competitive inhibition, the rate in the presence of inhibitor is proportional to \(S/[K_m(1+I_u/K_i)+S]\). Without inhibitor it is proportional to \(S/(K_m+S)\). Taking the ratio cancels the shared rate and substrate terms:
This ratio is the fraction of activity remaining in that simplified setting. It assumes competitive, reversible inhibition, the stated kinetic form, and concentrations appropriate to the enzyme site. It does not automatically account for active metabolites, time-dependent inhibition, changing inhibitor exposure, or other pathways. A PBPK model may represent these mechanisms more explicitly; a simple activity-score adjustment is a summary, not a replacement for mechanistic modeling.
Worked Example: Adjusting a CYP2D6 Score for Competitive Inhibition
Suppose a hypothetical patient has genotype *1/*1, so \(AS_G=1+1=2\). At the relevant condition, assume \(S=K_m\) and \(I_u/K_i=2\), with competitive inhibition as described above. Calculate the remaining activity and assign the illustrative adjusted phenotype.
Substitute the values into the ratio:
The adjusted score is \(AS_{\mathrm{adj}}=2\times0.5=1\). Under the bins defined here, a score greater than 0 and at most 1 is IM. Thus the genotype-based assignment is NM at score 2, whereas the inhibitor-adjusted assignment is IM at score 1. The calculation is conditional on the assumed inhibitor and substrate conditions; it does not establish that every person taking this inhibitor has exactly half of baseline CYP2D6 activity.
Combining Multiple Activity Changes
A disease-related change can affect enzyme abundance or function, while an inhibitor can reduce the activity of the enzyme that remains. If the effects are independently characterized as fractions of the same baseline activity, their factors can be multiplied. This is a modeling assumption to check, not a rule that every disease and drug effect must combine in this way. For example, a disease factor estimated from an activity measurement already made during inhibitor treatment should not be multiplied by a second inhibitor factor that the measurement already includes.
Worked Example: A Strong-Inhibitor Convention
Suppose a hypothetical *1x2/*1 diplotype has genotype score \(AS_G=2+1=3\), which is UM under the stated bins. For a particular analysis, assume the protocol defines exposure to a specified strong CYP2D6 inhibitor as \(f_I=0\), while no disease-related change is present, so \(f_D=1\). Calculate the adjusted assignment.
The score is \(AS_{\mathrm{adj}}=3\times0\times1=0\). The corresponding assignment is PM. This result follows from the protocol's categorical convention; it does not mean the patient's CYP2D6 molecules have been proven to have literally zero activity. If a more mechanistic estimate indicated a small nonzero fraction instead, the adjusted score could be nonzero and would need to be classified using the chosen translation system. The genotype remains *1x2/*1 throughout.
Worked Example: Combining Disease-Related Change and Inhibition
Consider a hypothetical *1/*10 diplotype. Its genotype score is \(1+0.25=1.25\), an NM assignment under the illustrative bins. Suppose independent evidence for this hypothetical case supports a disease-related fraction \(f_D=0.60\), and a moderate inhibitor is represented by \(f_I=0.50\). Calculate the adjusted score and phenotype.
First apply the disease factor: \(1.25\times0.60=0.75\). Then apply the inhibitor factor: \(0.75\times0.50=0.375\). Equivalently, \(AS_{\mathrm{adj}}=1.25\times0.60\times0.50=0.375\). The adjusted assignment is IM because the score is above 0 and at most 1. Both factors are assumed to apply to the same genotype-predicted activity and to be multiplicatively separable. If either factor were uncertain, the result should be presented as conditional or as a range rather than as a definitive individual phenotype.
Assumptions, Diagnostics, and Interpretation
Before reporting an adjusted assignment, check that the calculation matches the clinical and modeling question. The genotype call and copy number must be reliable; the allele-function values and phenotype bins must be named; inhibitor exposure must be relevant to the time period being described; and any disease factor must have a defensible basis. If factors come from different sources or conditions, ask whether they refer to the same baseline and whether multiplying them double-counts an effect.
A categorical label compresses a continuous and sometimes uncertain quantity. Two adjusted scores can fall in the same bin while differing substantially, and a tiny change near a boundary can change the label without a large biological change. Retain the score, factor values, assumptions, and uncertainty where possible; do not report only “IM” or “PM.” In a PBPK context, the score-based assignment can be useful for stratification, but the underlying exposure prediction also depends on drug-specific clearance pathways and other model inputs.
| Approach | What it represents | Main limitation |
|---|---|---|
| Genotype-only assignment | Inherited enzyme-function prediction | Does not reflect current inhibition or disease-related activity change |
| Factor-adjusted score | Genotype score scaled by estimated activity fractions | Depends on the validity and separability of those factors |
| Mechanistic PBPK representation | Enzyme abundance, inhibition, exposure, and relevant drug pathways within a model | Requires additional parameters and evidence; predictions remain conditional on model structure |
Useful diagnostics include checking the arithmetic independently, comparing the predicted direction of change with the known inhibitor or disease mechanism, and examining whether the chosen activity factor is plausible for the relevant concentrations. Where clinical CYP2D6 activity is directly assessed, compare the measurement with the model's conditional prediction while considering what the assay actually measures and when it was performed. Disagreement may indicate an incorrect genotype or copy-number call, an unaccounted inhibitor, an unsuitable scaling factor, a timing mismatch, or model misspecification; it is not automatically evidence that the genotype is wrong.
Common Pitfalls and Limits
- Equating genotype with current activity. The genotype is unchanged by an inhibitor, but current enzyme function may be reduced. Report the genetic prediction and the condition-dependent adjustment distinctly.
- Treating an inhibitor label as a universal numeric factor. Inhibition depends on mechanism, exposure, and conditions. A factor must come from a stated convention, evidence, or model rather than being inferred from the word “inhibitor.”
- Multiplying overlapping effects. If a disease-related estimate already incorporates the effect of an inhibitor, applying the inhibitor factor again double-counts the reduction.
- Assuming category boundaries are universal. The bins in the examples are explicit teaching conventions. Clinical interpretation must follow the selected guideline and its version.
- Interpreting a score of zero as proof of no molecular activity. Zero may be the outcome of a categorical convention or a model approximation; distinguish that from a direct measurement establishing complete absence of activity.
- Ignoring time. An inhibitor-adjusted assignment describes specified conditions. When exposure or disease status changes, the functional activity estimate may change too.
The central limitation is that a single adjusted score cannot capture every determinant of drug disposition. It is a disciplined bridge between genotype-predicted function and condition-dependent activity, not a claim that all biological complexity has been reduced to one exact number. When evidence does not support a precise factor, state the uncertainty rather than presenting a finely calculated but poorly grounded phenotype.
Check Your Understanding
Use the illustrative phenotype bins stated above. Treat each scenario as hypothetical, and state any calculation assumptions.
- A *1/*4 diplotype has genotype score \(1+0=1\). If \(f_I=0.5\) and \(f_D=1\), what is the adjusted score and assignment?
- A *1/*1 diplotype has score 2. If a disease-related factor is 0.8 and there is no inhibitor effect, what is the adjusted score and assignment?
- For competitive inhibition, let \(S=K_m\) and \(I_u/K_i=1\). Calculate \(f_I\). If genotype score is 2, calculate the adjusted score and assignment.
- Why is it incorrect to multiply a disease factor and inhibitor factor if the disease factor was estimated from activity measured while the patient was already receiving that inhibitor?
- A genotype predicts UM, but a current inhibitor-adjusted score is in the IM range. Has the genotype changed? Explain what has changed.
Solutions. (1) \(AS_{\mathrm{adj}}=1\times0.5\times1=0.5\), so the assignment is IM. (2) \(AS_{\mathrm{adj}}=2\times0.8\times1=1.6\), so the assignment is NM. (3) \(f_I=(K_m+K_m)/(K_m(1+1)+K_m)=2K_m/(3K_m)=2/3\), approximately 0.667. The adjusted score is \(2\times2/3=1.33\), so the assignment is NM. (4) The measured disease factor already reflects inhibitor exposure; multiplying by a separate inhibitor factor would count that reduction twice. (5) No. The inherited genotype is unchanged; the current functional activity estimate is lower under the stated inhibitor conditions, producing a different condition-dependent assignment.