From Dose to a Time-Varying Pathway Response
A dose does not act directly on a pathway biomarker. Its effect is mediated by the concentration profile in plasma, the amount of drug reaching the unbound target site, and the target-level pharmacology. The biomarker then responds according to its own production and removal rates. Keeping these steps distinct makes it possible to see where assumptions enter and why the biomarker may lag behind drug exposure.
The worked system below is hypothetical. A 100-mg dose is associated with a specified plasma concentration profile; the dose is not itself used as a concentration. The aim is to follow that profile through one target and one downstream biomarker, not to infer a pharmacokinetic profile from dose.
Connecting Plasma Concentration to Target Inhibition
Suppose the total plasma concentration after the hypothetical dose is \(C_{p,\mathrm{tot}}(t)=5.00e^{-0.1733t}\ \mu\mathrm{M}\), where \(t\) is in hours. The plasma unbound fraction is \(f_{u,p}=0.10\). Let \(K_{p,uu}=0.50\) denote the ratio of unbound target-site concentration to unbound plasma concentration. Under the rapid-equilibration assumption,
For this example, fractional inhibition is \(I(t)=C_{T,u}(t)/(IC_{50}+C_{T,u}(t))\), with \(IC_{50}=0.0500\ \mu\mathrm{M}\). This relation assumes a unit Hill coefficient, zero inhibition at zero concentration, and a maximum inhibition of one. It is a simple target-level model, not a universal description of inhibitor pharmacology.
Worked Example: Calculating Target-Site Exposure and Inhibition
Calculate the target-site concentration and fractional inhibition at \(t=2.00\) h. First evaluate the plasma profile:
The unbound plasma concentration is \(0.10(3.535)=0.3535\ \mu\mathrm{M}\). Applying \(K_{p,uu}=0.50\) gives \(C_{T,u}=0.1768\ \mu\mathrm{M}\). Thus,
The model therefore predicts about 78.0% inhibition at 2.00 h. This is a target-level prediction; it is not yet the predicted biomarker value. The plasma profile has a half-life of approximately \(\ln(2)/0.1733=4.00\) h, but the target-site concentration and inhibition are not themselves half-lives.
Turning Inhibition into Biomarker Dynamics
Let \(B(t)\) be the biomarker concentration in \(\mathrm{pg/mL}\), and let its baseline value be \(B_0=100\ \mathrm{pg/mL}\). Assume that inhibition reduces biomarker production in proportion to \(1-I(t)\), while removal is first order with rate constant \(k_{\mathrm{out}}=0.100\ \mathrm{h}^{-1}\). At baseline steady state, production and removal balance. Writing the baseline production rate as \(k_{\mathrm{out}}B_0\), the turnover equation becomes
If inhibition were held constant, the equilibrium biomarker value would be \(B_{\mathrm{eq}}=B_0(1-I)\). When inhibition changes over time, the biomarker generally cannot jump to that equilibrium: it approaches the changing equilibrium at a rate set by \(k_{\mathrm{out}}\). This distinction is the central reason that a pathway biomarker may lag behind exposure.
For a continuously varying inhibition profile, the corresponding solution, with \(B(0)=B_0\), is
The integral weights past production according to how much biomarker removal has occurred since each time point. A practical hand calculation is to divide time into intervals, evaluate inhibition at each interval midpoint, and treat it as constant over that interval. This is a numerical approximation; it becomes more reliable when intervals are short relative to the timescale on which exposure and inhibition change.
Worked Example: Propagating Inhibition Across Four Intervals
Use four-hour intervals from 0 to 16 h. At each midpoint, calculate \(C_{p,\mathrm{tot}}\), then use the target-site and inhibition relations above to find \(B_{\mathrm{eq}}=100(1-I)\). For each interval, update the biomarker using \(e^{-k_{\mathrm{out}}\Delta t}=e^{-0.100(4)}=e^{-0.4}\approx0.6703\). The initial value is \(B(0)=100\ \mathrm{pg/mL}\).
| Interval (h) | Midpoint (h) | \(C_{p,\mathrm{tot}}\) at midpoint (\(\mu\mathrm{M}\)) | \(B_{\mathrm{eq}}\) (\(\mathrm{pg/mL}\)) |
|---|---|---|---|
| 0–4 | 2 | 3.535 | 22.05 |
| 4–8 | 6 | 1.768 | 36.13 |
| 8–12 | 10 | 0.8838 | 53.08 |
| 12–16 | 14 | 0.4419 | 69.35 |
For the first interval, the midpoint inhibition is approximately 0.7795, so \(B_{\mathrm{eq}}=100(1-0.7795)\approx22.05\ \mathrm{pg/mL}\). Starting from 100, the update is
For the second interval, the midpoint equilibrium is about \(36.13\ \mathrm{pg/mL}\). Using \(B(4)\approx74.30\), \(B(8)=36.13+(74.30-36.13)(0.6703)\approx61.72\ \mathrm{pg/mL}\). For the third interval, \(B(12)=53.08+(61.72-53.08)(0.6703)\approx58.87\ \mathrm{pg/mL}\).
For the final interval, use \(B_{\mathrm{eq}}=69.35\ \mathrm{pg/mL}\) and \(B(12)=58.87\ \mathrm{pg/mL}\):
The approximate trajectory is therefore 100, 74.30, 61.72, 58.87, and \(62.33\ \mathrm{pg/mL}\) at 0, 4, 8, 12, and 16 h. Biomarker values are rounded for display; calculations use the shown inputs without treating displayed rounded values as exact. The rebound from 12 to 16 h follows from declining inhibition: the equilibrium value rises above the current biomarker value, so the biomarker begins moving upward.
How Dose Changes Propagate Through the Chain
If pharmacokinetics are dose proportional over the range considered, doubling the dose doubles the plasma concentration at each time point. The target-site concentration also doubles under constant binding and distribution assumptions. Inhibition, however, does not double: the concentration-response relation is nonlinear, and the biomarker response is further shaped by turnover.
Worked Example: Comparing a Doubled Dose
Consider a hypothetical 200-mg dose with the same concentration-time shape and twice the plasma concentrations. At the first interval midpoint, \(t=2\) h, \(C_{p,\mathrm{tot}}\approx7.071\ \mu\mathrm{M}\), so \(C_{T,u}=0.0500(7.071)=0.3535\ \mu\mathrm{M}\). The inhibition and equilibrium biomarker value are
Starting at 100, the first-interval prediction is \(B(4)=12.39+(100-12.39)(0.6703)\approx71.12\ \mathrm{pg/mL}\). At the next midpoint, \(t=6\) h, the doubled-dose plasma concentration is approximately \(3.535\ \mu\mathrm{M}\), giving \(B_{\mathrm{eq}}\approx22.05\ \mathrm{pg/mL}\). Thus \(B(8)=22.05+(71.12-22.05)(0.6703)\approx54.94\ \mathrm{pg/mL}\). The larger dose deepens predicted inhibition and lowers the biomarker, but neither inhibition nor biomarker change is a simple twofold multiple of the single-dose result.
Checks, Alternatives, and Limits
A useful calculation check is to follow the units and physical direction at every step. Both binding fractions and inhibition are dimensionless; multiplying the plasma concentration by them leaves concentration units unchanged. \(B_{\mathrm{eq}}\) has the same units as \(B\), while \(k_{\mathrm{out}}\Delta t\) is dimensionless. With positive inhibition, the equilibrium biomarker is at or below baseline. If inhibition weakens while \(B\) is below its new equilibrium, the model predicts recovery rather than an instantaneous return to baseline.
The midpoint calculation is convenient but not exact for a continuously changing concentration. If the exposure profile changes substantially within a four-hour interval, shorter intervals or numerical integration of the time-varying equation are preferable. A direct-response approximation would set \(B(t)=B_{\mathrm{eq}}(t)\) at every instant; that is appropriate only when biomarker adjustment is effectively much faster than the exposure changes, and it discards the turnover lag. If target-site equilibration is not rapid, the instantaneous conversion from unbound plasma to target-site concentration must be replaced by a dynamic effect-site description. That changes the input to the inhibition model as well as the predicted timing.
The model also assumes that the chosen \(IC_{50}\) applies at the relevant target site and conditions, and that one target’s inhibition controls biomarker production in the stated proportion. Substrate dependence, mixed or irreversible inhibition, target turnover, active metabolites, feedback, or a nonzero residual production fraction can require a different target or pathway model. These assumptions should be checked against the pharmacology and biomarker data rather than inferred from a visually plausible curve.
Common warning signs include predicted inhibition above one or below zero under a model intended to constrain it to the range zero to one, a biomarker that changes instantaneously despite a finite turnover rate, and a calculated target concentration that mistakenly uses total plasma concentration as if it were unbound. Another frequent error is applying the unbound fraction twice or confusing \(K_{p,uu}\), defined here as a ratio of unbound concentrations, with a ratio involving total concentrations. Verify the definition before multiplying by any binding fraction.
Check Your Understanding
Answer each question, then compare with the solutions below.
- At a time when total plasma concentration is \(2.00\ \mu\mathrm{M}\), what is \(C_{T,u}\) under the stated binding assumptions?
- Using that target-site concentration and \(IC_{50}=0.0500\ \mu\mathrm{M}\), calculate fractional inhibition and the corresponding equilibrium biomarker value.
- Why is the biomarker not predicted to equal its equilibrium value immediately when inhibition changes?
- What does the four-hour update factor \(e^{-0.4}\) represent, and what would a shorter interval change?
- Name one situation in which the direct conversion from unbound plasma concentration to target-site concentration would be inadequate.
Solutions. (1) \(C_{T,u}=0.0500(2.00)=0.100\ \mu\mathrm{M}\). (2) \(I=0.100/(0.0500+0.100)=0.6667\), so \(B_{\mathrm{eq}}=100(1-0.6667)=33.33\ \mathrm{pg/mL}\). (3) Biomarker production and first-order removal take time to change its concentration; the turnover equation makes \(B\) approach, rather than jump to, the current equilibrium. (4) It is the fraction of the difference \(B-B_{\mathrm{eq}}\) remaining after four hours under constant inhibition; shorter intervals improve the approximation when inhibition varies over time. (5) A material delay in target-site equilibration would require a dynamic effect-site model rather than the instantaneous concentration ratio.