Effect-Site Equilibration for Small-Molecule Targets

Why Add an Effect-Site Compartment?

Unbound plasma concentration is often used to drive a small-molecule target model, but the target may not experience the plasma concentration immediately. Transport into tissue, membrane passage, intracellular distribution, or local binding can delay equilibration. If the model assumes that target-site concentration changes instantly with plasma concentration, it can predict a rapid pharmacological response even when access to the target is slower.

An effect-site compartment represents this delay with a simple dynamic link between unbound plasma concentration and a notional target-site concentration. It is not necessarily a physical tissue compartment with a measurable volume. Its purpose is to describe how quickly the concentration driving target action approaches an equilibrium value.

Definition: Let $C_{u,p}(t)$ be unbound plasma concentration and $C_e(t)$ be the effect-site concentration. A first-order effect-site model is $$ \frac{dC_e}{dt}=k_{e0}\left(K_{p,uu}C_{u,p}(t)-C_e(t)\right), $$ where $k_{e0}>0$ is the equilibration rate constant, with units of inverse time, and $K_{p,uu}$ is the dimensionless equilibrium ratio of effect-site concentration to unbound plasma concentration. At constant plasma concentration, the equilibrium value is $C_e=K_{p,uu}C_{u,p}$. When no separate partition estimate is available, a common simplifying assumption is $K_{p,uu}=1$.

Assumptions and Time Course

This model assumes that the unbound plasma concentration is the relevant driving concentration, that the exchange toward equilibrium is linear and first-order, and that $k_{e0}$ and $K_{p,uu}$ remain constant over the interval considered. The plasma concentration is treated as a known input from the pharmacokinetic model. The model also assumes that one effect-site state adequately summarizes the distribution delay; it does not represent multiple tissues, saturable transport, or a concentration-dependent equilibration rate.

For a constant plasma input $C_{u,p}=C_0$ beginning at time zero, the equilibrium effect-site concentration is $K_{p,uu}C_0$. Starting from an initial value $C_e(0)=C_{e,0}$, the differential equation has the solution below. The difference from equilibrium decreases exponentially.

C_e(t)=K_{p,uu}C_0+ \left(C_{e,0}-K_{p,uu}C_0\right)e^{-k_{e0}t}.

For a step from zero effect-site concentration, $C_{e,0}=0$, so the site approaches its new equilibrium from below. The time to reach half of the change toward equilibrium is $t_{1/2,e}=\ln(2)/k_{e0}$. This is an equilibration half-time, not the elimination half-life of the drug from plasma.

Derivation: With constant input, write the model as $dC_e/dt+k_{e0}C_e=k_{e0}K_{p,uu}C_0$. Multiplication by the integrating factor $e^{k_{e0}t}$ gives $$ \frac{d}{dt}\left(e^{k_{e0}t}C_e(t)\right) =k_{e0}K_{p,uu}C_0e^{k_{e0}t}. $$ Integrating from zero to $t$, then dividing by $e^{k_{e0}t}$, yields $$ C_e(t)=C_{e,0}e^{-k_{e0}t} +K_{p,uu}C_0\left(1-e^{-k_{e0}t}\right), $$ which is the stated solution. The remaining difference from equilibrium is proportional to $e^{-k_{e0}t}$; it is halved when $e^{-k_{e0}t}=1/2$, giving $t_{1/2,e}=\ln(2)/k_{e0}$.

When plasma concentration changes continuously, the model follows the input with a lag rather than resetting to a new equilibrium at every instant. For initial concentration $C_{e,0}$, its solution is

C_e(t)=C_{e,0}e^{-k_{e0}t} +k_{e0}K_{p,uu}\int_0^t C_{u,p}(s)e^{-k_{e0}(t-s)}\,ds.

This expression weights recent plasma concentrations more strongly than distant ones. The effect-site concentration therefore reflects a smoothed history of unbound plasma exposure, with the degree of smoothing governed by $k_{e0}$.

Worked Example: A Step Increase and the Inhibition Prediction

Suppose a hypothetical drug produces a constant unbound plasma concentration of $10\ \mathrm{nM}$, beginning at time zero. Let $K_{p,uu}=1$, $k_{e0}=1\ \mathrm{h}^{-1}$, and $C_e(0)=0$. Calculate the effect-site concentration after 2.4 hours. To illustrate the downstream consequence only, suppose target inhibition is read from the stipulated relationship $I=100C_e/(C_e+4\ \mathrm{nM})$, with $I$ expressed as a percentage.

The effect-site concentration is $$ C_e(2.4)=10\left(1-e^{-1(2.4)}\right) =10(1-0.0907180)=9.09282\ \mathrm{nM}. $$

Substituting this concentration into the given illustrative inhibition relationship gives $$ I=\frac{100(9.09282)}{9.09282+4} =69.4489\%\approx69.45\%. $$

Thus the predicted inhibition at 2.4 hours is about $69.45\%$, below the eventual equilibrium prediction of $100(10)/(10+4)=71.43\%$. The difference arises because the effect-site concentration has not quite reached its equilibrium value of $10\ \mathrm{nM}$. The inhibition relationship is used here as a downstream readout; the effect-site model itself specifies the delayed concentration.

The same input concentration can yield different effect-site concentrations at different times after a change in exposure. That distinction matters when predicting target inhibition: the target model should be driven by the concentration at the site, not by an assumption that it instantly equals the current plasma value.

Worked Example: Early and Later Predictions During the Same Exposure

Use the hypothetical step exposure and inhibition relationship in the preceding example, but compare predictions at $0.3$ and $2.4$ hours. Assume the same zero initial effect-site concentration, $K_{p,uu}=1$, $k_{e0}=1\ \mathrm{h}^{-1}$, plasma concentration $10\ \mathrm{nM}$, and illustrative $4\ \mathrm{nM}$ inhibition scale. The input is constant throughout, so the step-input solution applies.

At $0.3$ hours, $$ C_e(0.3)=10(1-e^{-0.3}) =10(1-0.740818)=2.59182\ \mathrm{nM}, $$ and $$ I(0.3)=\frac{100(2.59182)}{2.59182+4} =39.3187\%\approx39.32\%. $$

At $2.4$ hours, the calculation gives $C_e(2.4)=9.09282\ \mathrm{nM}$, and therefore $I(2.4)=69.4489\%\approx69.45\%$. Although the plasma concentration is $10\ \mathrm{nM}$ at both times, the predicted inhibition differs because the effect site is still equilibrating. These calculations assume that the stated inhibition relationship applies unchanged at both times.

Changing Exposure and Washout

A delay is apparent not only as a slow rise. When plasma concentration falls, the effect-site concentration can remain elevated while the target site equilibrates back toward its new, lower equilibrium. For a step from a prior constant exposure to zero plasma concentration, the effect-site concentration decays exponentially from its value at the time of the step.

Worked Example: Effect-Site Decline After Plasma Washout

In a hypothetical washout, unbound plasma concentration falls abruptly from $8\ \mathrm{nM}$ to zero. At that moment, assume the effect site was at equilibrium, so $C_e(0)=8\ \mathrm{nM}$, with $K_{p,uu}=1$. Let $k_{e0}=0.25\ \mathrm{h}^{-1}$. Calculate the effect-site concentration four hours after the plasma decrease.

After the step to zero, the equilibrium effect-site concentration is zero. The solution is therefore $$ C_e(4)=8e^{-0.25(4)} =8e^{-1} =8(0.367879)=2.94303\ \mathrm{nM}. $$

The concentration at the site remains about $2.943\ \mathrm{nM}$ even though the plasma input is zero. The equilibration half-time is $\ln(2)/0.25=2.773$ hours, consistent with a relatively slow decline: after four hours, more than one half-life has elapsed, but the site concentration has not reached zero.

For a plasma concentration that itself decays exponentially, the effect-site trajectory can be calculated from the general input solution. This makes clear that the site may continue rising even after plasma concentration has begun to fall, if drug is still arriving at the site faster than it leaves.

Worked Example: A Decaying Plasma Input

Suppose a hypothetical input after a dose is $C_{u,p}(t)=12e^{-0.5t}\ \mathrm{nM}$, where time is in hours. Set $K_{p,uu}=1$, $k_{e0}=1\ \mathrm{h}^{-1}$, and $C_e(0)=0$. Find the effect-site concentration at two hours and compare it with plasma concentration then.

Substitution into the integral solution gives $$ C_e(t)=\int_0^t 12e^{-0.5s}e^{-(t-s)}\,ds =24\left(e^{-0.5t}-e^{-t}\right)\ \mathrm{nM}. $$

At two hours, $$ C_e(2)=24(e^{-1}-e^{-2}) =24(0.367879-0.135335) =5.58106\ \mathrm{nM}. $$ The plasma concentration at the same time is $$ C_{u,p}(2)=12e^{-1}=4.41455\ \mathrm{nM}. $$

Thus the effect-site concentration is higher than the contemporaneous plasma concentration at two hours. This is not a contradiction: the effect site is responding to the earlier, higher plasma input as well as the current value. The calculation assumes the specified exponential plasma profile continues to drive a linear effect compartment.

Choosing and Checking the Model

An effect-site compartment is useful when an instantaneous link from plasma concentration to target action is too restrictive, but the data do not support a detailed distribution model. Its rate constant controls the speed of equilibration; $K_{p,uu}$ controls the equilibrium ratio. These parameters have different roles and should not be treated as interchangeable. A smaller $k_{e0}$ produces a longer delay, whereas a different $K_{p,uu}$ changes the long-run concentration ratio.

Useful checks include confirming that the effect-site concentration approaches $K_{p,uu}C_{u,p}$ under a sustained constant input, and that it decays toward zero after a zero input. Compare predicted timing with observed response timing when such data exist. Persistent timing mismatches, systematic prediction errors during rising versus falling exposure, or a fitted $k_{e0}$ near a boundary can indicate that the single first-order link is inadequate. Sparse observations around concentration changes may also make the equilibration rate difficult to estimate.

The main alternatives answer different modeling needs. An instantaneous equilibrium link sets $C_e=K_{p,uu}C_{u,p}$ at every time; it is simpler but cannot represent a delay. A physical tissue compartment can represent distribution volumes and mass transfer, but requires additional structural assumptions and often more data. A chain of transit compartments can represent a broader delay pattern than one exponential process, at the cost of additional parameters. The first-order effect site is a compact approximation, not proof of a particular anatomical mechanism.

Common Pitfall and Limitations

Using total plasma concentration as the driver without justification. The model here is driven by unbound plasma concentration, consistent with the preceding material on unbound drug at the intracellular target. If another concentration is used, its relationship to the target-site driving concentration must be justified.
Confusing $k_{e0}$ with a plasma elimination rate. It describes approach to effect-site equilibrium, not drug removal from plasma. Their numerical values can differ.
Assuming the effect site is a measured, mass-conserving tissue compartment. In its usual form, it is a phenomenological state variable and need not have a physical volume or explicitly account for drug mass.
Assuming a delay changes the equilibrium concentration. With constant input, $k_{e0}$ changes the time to equilibrium; $K_{p,uu}$ determines the equilibrium ratio.

The model can fail when transport is saturable, the target-site environment changes over time, the equilibrium ratio is not constant, or multiple processes create a non-single-exponential delay. Also, concentration and response data alone may not uniquely distinguish slow equilibration from a slow downstream biological response. Additional measurements or carefully designed exposure time courses may be needed to identify the source of the lag.

Key takeaway: The first-order effect-site model makes target-site concentration a dynamic response to unbound plasma concentration. Its rate constant determines how quickly the site tracks exposure, while its equilibrium ratio determines the eventual concentration relationship. Delayed concentration can change the predicted timing and magnitude of target inhibition.

Check Your Understanding

Answer each question, then compare with the solutions below.

With $k_{e0}=0.5\ \mathrm{h}^{-1}$, what is the equilibration half-time?
If constant unbound plasma concentration is $6\ \mathrm{nM}$, $K_{p,uu}=0.8$, and the system reaches equilibrium, what is $C_e$?
After an abrupt change to zero plasma concentration, $C_e(0)=5\ \mathrm{nM}$ and $k_{e0}=0.4\ \mathrm{h}^{-1}$. What is $C_e$ after two hours?
For the step input $C_{u,p}=10\ \mathrm{nM}$, $K_{p,uu}=1$, $k_{e0}=1\ \mathrm{h}^{-1}$, and $C_e(0)=0$, is $C_e$ at one hour above or below $10\ \mathrm{nM}$? Calculate it.

Solutions. (1) $t_{1/2,e}=\ln(2)/0.5=1.386\ \mathrm{h}$. (2) At equilibrium, $C_e=K_{p,uu}C_{u,p}=0.8(6)=4.8\ \mathrm{nM}$. (3) $C_e(2)=5e^{-0.4(2)}=5e^{-0.8}=2.247\ \mathrm{nM}$, approaching the new zero equilibrium. (4) $C_e(1)=10(1-e^{-1})=6.321\ \mathrm{nM}$, below the equilibrium value because the effect site has not fully equilibrated.