Why Map Emax Effects to Rates?
In an indirect response model, a drug changes the rate at which a response is produced, the rate at which it is lost, or both. An Emax function describes how the drug effect changes with concentration. Mapping that function to a rate makes the mechanism explicit: the drug changes the rate multiplier, while the response itself continues to evolve according to production and loss.
This distinction matters because a rate effect can alter both the eventual response level and the speed at which the response approaches it. The steady-state response depends on the balance between modified production and modified loss; the approach rate depends on the effective loss rate. A multiplier applied to production therefore has a different role from the same multiplier applied to loss.
For the standard inhibitory Emax form, the fractional drug effect is \(I_{\max}C/(IC_{50}+C)\). The corresponding rate multiplier is one minus that effect. For a stimulatory Emax form, the fractional increase is \(S_{\max}C/(SC_{50}+C)\), so the rate multiplier is one plus that increase. The symbols \(IC_{50}\) and \(SC_{50}\) denote the concentrations at which the relevant fractional effect reaches half its maximum; \(I_{\max}\) and \(S_{\max}\) set the maximum effect magnitude.
Equilibrium and Its Stability
At a fixed concentration, equilibrium means that the response is no longer changing, so \(dR/dt=0\). Provided \(k_{\mathrm{out}}f_{\mathrm{out}}(C)>0\), solving the rate equation gives
This expression shows why the effect on equilibrium is a ratio: production sets the input, while loss determines how quickly response is removed. A production multiplier acts in the numerator; a loss multiplier acts in the denominator. If both rates are modified, both multipliers appear in the ratio.
The equilibrium is stable when the effective loss coefficient is positive. To see this, write \(b=k_{\mathrm{out}}f_{\mathrm{out}}(C)\) and \(a=k_{\mathrm{in}}f_{\mathrm{in}}(C)\). With \(b>0\), the equilibrium is \(a/b\), and the deviation \(D(t)=R(t)-a/b\) satisfies
Thus \(D(t)=D(0)e^{-bt}\), which tends to zero as time increases. The effective time constant is \(1/b=1/[k_{\mathrm{out}}f_{\mathrm{out}}(C)]\). A production multiplier changes the equilibrium but not this time constant in the canonical model; a loss multiplier changes both. If the loss multiplier is zero, this stability argument fails: with positive production the response continues to increase rather than approaching a finite equilibrium.
Worked Example: Inhibiting Response Production
Suppose a hypothetical response has baseline production \(k_{\mathrm{in}}=12\) response units per hour and baseline loss \(k_{\mathrm{out}}=0.30\) per hour. At a fixed concentration \(C=4\) mg/L, production is inhibited with \(I_{\max}=0.80\) and \(IC_{50}=4\) mg/L. Find the baseline response, the production multiplier, and the drug-on equilibrium.
First, the baseline equilibrium is
The inhibition fraction at the stated concentration is \(0.80(4)/(4+4)=0.40\). Therefore the remaining production multiplier is
Loss is unmodified, so \(f_{\mathrm{out}}=1\). The equilibrium is
At this concentration the production rate is \(12(0.60)=7.2\) response units/h, and the loss coefficient remains \(0.30\ \text{h}^{-1}\). Their ratio is \(7.2/0.30=24\) response units, confirming the result. The equilibrium is stable, and its time constant remains \(1/0.30=3.33\) h because the drug modifies production, not loss.
Worked Example: Inhibiting Response Loss
Consider a hypothetical response with \(k_{\mathrm{in}}=10\) response units/h and \(k_{\mathrm{out}}=0.50\ \text{h}^{-1}\). Loss is inhibited with \(I_{\max}=0.75\) and \(IC_{50}=3\) mg/L. At \(C=6\) mg/L, calculate the loss multiplier, the equilibrium, and the time constant.
The baseline equilibrium is \(10/0.50=20\) response units. The fractional inhibition is
Consequently, the remaining loss multiplier is \(f_{\mathrm{out}}=1-0.50=0.50\). Production remains unmodified, giving
The effective loss coefficient is \(0.50(0.50)=0.25\ \text{h}^{-1}\), so the time constant is \(1/0.25=4.00\) h. The equilibrium doubles because the same production is balanced by half the baseline loss coefficient. The positive effective loss coefficient ensures that the equilibrium is stable.
Worked Example: Stimulation of Response Loss
For a hypothetical response, let \(k_{\mathrm{in}}=8\) response units/h and \(k_{\mathrm{out}}=0.40\ \text{h}^{-1}\). Suppose loss is stimulated with \(S_{\max}=1.5\) and \(SC_{50}=3\) mg/L. At \(C=3\) mg/L, calculate the response equilibrium and compare its time constant with baseline.
The baseline equilibrium is \(8/0.40=20\) response units. At \(C=SC_{50}\), the multiplier is
Production is unchanged, so
The effective loss coefficient is \(0.40(1.75)=0.70\ \text{h}^{-1}\), making the time constant \(1/0.70=1.429\) h. Baseline time constant is \(1/0.40=2.50\) h. The increased loss therefore lowers the equilibrium and shortens the time constant; the calculations use the same multiplier in the denominator of the equilibrium and in the effective loss coefficient.
Reading the Four Mappings Together
The four models can be compared by tracking the direction of the modified rate and its location in the equilibrium ratio. The comparison assumes a fixed concentration, positive \(k_{\mathrm{in}}\), and a positive effective loss coefficient. Each Emax multiplier equals 1 at \(C=0\), so the model returns to baseline when there is no drug effect.
| Drug effect | Modified rate | Equilibrium relative to \(R_0\) | Effect on time constant |
|---|---|---|---|
| Inhibit production | Production decreases | Decreases by \(f_{\mathrm{in}}\) | Unchanged |
| Stimulate production | Production increases | Increases by \(f_{\mathrm{in}}\) | Unchanged |
| Inhibit loss | Loss coefficient decreases | Increases by \(1/f_{\mathrm{out}}\) | Lengthens |
| Stimulate loss | Loss coefficient increases | Decreases by \(1/f_{\mathrm{out}}\) | Shortens |
For example, multiplying production by 0.5 halves the equilibrium if loss is unchanged. Multiplying loss by 0.5 doubles the equilibrium, rather than halving it, because loss is in the denominator. This reciprocal effect is a frequent source of sign and direction errors.
Assumptions, Diagnostics, and Limitations
The mappings above are appropriate when the intended model is a first-order turnover model and the Emax function describes a concentration-dependent change in a rate. The rate multipliers are assumed to be nonnegative, and a finite stable equilibrium requires \(f_{\mathrm{out}}(C)>0\). In the usual interpretation, the baseline rates remain constant and concentration is treated as fixed while calculating the drug-on equilibrium. If concentration changes over time, the instantaneous equilibrium calculated from this formula need not equal the response at that time; the response can lag behind changing rates.
Several checks help identify a mistaken mapping or an implausible parameterization:
- Check the zero-concentration limit. Substituting \(C=0\) should give both multipliers equal to 1 and \(R_{\mathrm{ss}}=R_0\).
- Check the high-concentration limit. For inhibited production, \(f_{\mathrm{in}}\) tends to \(1-I_{\max}\); for stimulated production it tends to \(1+S_{\max}\). The corresponding limits apply to loss multipliers. These limits should agree with the maximum effects intended by the model.
- Check the loss coefficient. If \(k_{\mathrm{out}}f_{\mathrm{out}}\le 0\), a positive finite stable equilibrium is not supported by this model. Do not interpret the ratio as a valid stable response in that case.
- Check units and rate placement. Multipliers are dimensionless. A production rate has response-per-time units; a loss coefficient has inverse-time units and multiplies \(R\).
- Check predicted direction. Inhibition of production should not raise equilibrium when loss is unchanged; inhibition of loss should not lower it. If the result goes the other way, inspect whether the multiplier has been placed in the correct part of the ratio.
A direct Emax model applied to the response itself is a different modeling choice. It specifies a concentration-response relationship for the response level, whereas an indirect-response model specifies a concentration effect on production or loss and lets the response follow its turnover dynamics. A direct response curve may describe steady-state observations compactly, but by itself it does not encode the production-loss mechanism or the rate-dependent time constant. Conversely, a rate-based model requires assumptions about turnover and may be difficult to distinguish from alternatives using only a few steady-state observations.
These formulas describe the equilibrium implied by a chosen mechanism; they do not establish that the mechanism is uniquely identified by observed response data. A similar equilibrium change can arise from different combinations of production and loss effects. The Emax mapping is therefore most useful when its mechanistic assumptions are scientifically defensible and when response-time information is also considered.
Check Your Understanding
Answer each question, then compare your reasoning with the solutions below.
- A drug stimulates production by a factor of 1.4 and does not change loss. What is the drug-on equilibrium in terms of \(R_0\), and what happens to the time constant?
- A loss-inhibition model has \(I_{\max}=0.60\), \(C=IC_{50}\), and baseline \(R_0=30\) response units. Find the loss multiplier and equilibrium.
- Why does stimulation of loss lower equilibrium but shorten the time constant?
- What condition on the loss multiplier is required for a finite stable equilibrium at fixed concentration?
Solutions. (1) The production multiplier is 1.4, so \(R_{\mathrm{ss}}=1.4R_0\); the time constant is unchanged because loss is unmodified. (2) At \(C=IC_{50}\), the inhibition fraction is \(0.60/2=0.30\), so \(f_{\mathrm{out}}=0.70\). Thus \(R_{\mathrm{ss}}=30/0.70=42.86\) response units. (3) The loss multiplier appears in the denominator of \(R_{\mathrm{ss}}=R_0/f_{\mathrm{out}}\), lowering the equilibrium when it increases; it also raises the effective loss coefficient \(k_{\mathrm{out}}f_{\mathrm{out}}\), whose reciprocal is the shorter time constant. (4) The effective loss coefficient must be positive, which requires \(f_{\mathrm{out}}(C)>0\) when \(k_{\mathrm{out}}>0\).