Deriving the Operational Model of Agonism

From Receptor Occupancy to Response

Receptor occupancy and measured response are not the same quantity. An agonist can occupy a fraction of receptors without producing a proportionate fraction of the system’s maximum response: downstream signalling may amplify the stimulus, and the steps between binding and measurement may themselves saturate. The Black–Leff operational model combines an equilibrium occupancy description with a saturable stimulus-response relationship. Its purpose is to express how agonist affinity, operational efficacy, and transducer slope jointly shape a concentration-effect curve.

The derivation below concerns agonist-produced response above baseline. Let \(E_0\) be the baseline response and let \(E-E_0\) denote the response increment. The model’s \(E_m\) is the maximum response increment the transducer could produce; it need not be reached by the agonist, even at saturating concentration.

Definition: \(K_A\) is the agonist equilibrium dissociation constant in the one-site binding model; a smaller \(K_A\) corresponds to higher affinity. The dimensionless parameter \(\tau\) is operational efficacy, a composite measure of agonist efficacy and available receptor/transduction capacity relative to the stimulus needed for transducer activation. The positive, dimensionless parameter \(n\) is the transducer slope, which controls the steepness of the stimulus-response relationship. \(E_m\) is the system’s maximum possible response increment.

Deriving the Operational Model

Assume reversible binding at equilibrium to one homogeneous receptor population. Let \(A\) be the free agonist concentration at the receptor. The fractional receptor occupancy is

\theta=\frac{A}{K_A+A}.

Suppose the stimulus \(S\) generated by the agonist is proportional to occupied receptor number. If \(R_T\) denotes total receptor concentration and \(\varepsilon\) is a proportionality constant representing agonist stimulus per occupied receptor, then

S=\varepsilon R_T\theta =\varepsilon R_T\frac{A}{K_A+A}.

Now assume a saturable transducer: the fraction of its maximum response produced by stimulus \(S\) follows a Hill-type relationship with slope \(n>0\) and stimulus scale \(K_E>0\). In this relationship, \(K_E\) is the stimulus that produces half of the transducer’s maximum response.

\frac{E-E_0}{E_m} =\frac{S^n}{K_E^n+S^n} =\frac{(S/K_E)^n}{1+(S/K_E)^n}.

Define operational efficacy as \(\tau=\varepsilon R_T/K_E\). Substitution gives \(S/K_E=\tau A/(K_A+A)\). Therefore,

\frac{E-E_0}{E_m} = \frac{\left(\dfrac{\tau A}{K_A+A}\right)^n} {1+\left(\dfrac{\tau A}{K_A+A}\right)^n} = \frac{\tau^n A^n} {(K_A+A)^n+\tau^n A^n}.

This is the Black–Leff operational model. The derivation makes its assumptions explicit: occupancy is at equilibrium, the agonist acts through the stated receptor population, stimulus is proportional to occupancy, and stimulus is converted to response by the specified saturable function. It also shows why \(\tau\) is operational rather than a direct measurement of molecular efficacy: it combines \(\varepsilon\), \(R_T\), and \(K_E\).

Worked Example: Calculating Response When \(n=1\)

Suppose a hypothetical assay has \(E_m=100\) response units, \(K_A=10\) nM, \(\tau=4\), and \(n=1\). What response increment does the model predict at \(A=10\) nM? What are the limiting maximum response and the concentration producing half of that maximum?

For \(n=1\), the operational model simplifies by expanding the denominator:

E-E_0 =E_m\frac{\tau A}{K_A+(1+\tau)A}.

At \(A=10\) nM, substitution gives

E-E_0 =100\frac{4(10)}{10+(1+4)(10)} =100\frac{40}{60} =66.7\text{ response units}.

At saturating concentration, the response fraction approaches \(\tau/(1+\tau)=4/5\), so the limiting agonist-produced increment is \(100(4/5)=80.0\) response units. For \(n=1\), the concentration producing half of this limiting increment is \(K_A/(1+\tau)\), giving \(10/5=2.00\) nM. Substitution at \(A=2.00\) nM gives \(100(8)/(10+10)=40.0\) units, exactly half of 80.0.

What the Parameters Imply

The model’s limiting response follows by taking \(A\) to infinity. Dividing numerator and denominator by \(A^n\) yields

\lim_{A\to\infty}\frac{E-E_0}{E_m} = \frac{\tau^n}{1+\tau^n}.

Thus the agonist’s predicted maximum is below the system ceiling \(E_m\) for finite \(\tau\). Increasing \(\tau\) moves the asymptote closer to \(E_m\), but does not change \(E_m\) itself. The affinity parameter \(K_A\) affects the concentration scale, while \(n\) affects the shape of the transition. These parameters act jointly in the full curve; the distinctions are not a guarantee that each can be estimated precisely from any given data set.

A useful concentration summary is the model’s EC50: the concentration at which the response increment is half of its own limiting agonist response. It is not, in general, the concentration producing half occupancy. To derive it, write \(f=(E-E_0)/E_m\), \(f_{\max}=\tau^n/(1+\tau^n)\), and set \(f=f_{\max}/2\). The model gives

\frac{\tau^n A^n}{(K_A+A)^n+\tau^n A^n} = \frac{1}{2}\frac{\tau^n}{1+\tau^n}.

For positive \(\tau\) and \(A\), cancelling \(\tau^n\) and rearranging gives \((K_A+A)^n/A^n=2+\tau^n\). Taking the positive \(n\)th root and solving for \(A\) gives

EC_{50} = \frac{K_A}{(2+\tau^n)^{1/n}-1}.

For \(n=1\), this reduces to \(K_A/(1+\tau)\). The formula shows that response potency can be greater than binding affinity alone would suggest: when \(\tau\) is large, the response reaches half of its own maximum at a concentration below \(K_A\). The derivation does not imply that \(EC_{50}=K_A\), or that either quantity can be read directly from the other without assumptions about the model.

Worked Example: Inferring \(\tau\) and EC50 from a Plateau

In a hypothetical assay, the system ceiling is \(E_m=100\) units, and a concentration-response curve approaches an agonist-produced plateau of 80 units. Suppose the transducer slope is \(n=2\), and an independent binding measurement gives \(K_A=6.00\) nM. Under the operational model, calculate \(\tau\) and EC50.

The plateau is 80% of the system ceiling, so \(f_{\max}=0.80\). From \(f_{\max}=\tau^n/(1+\tau^n)\),

0.80=\frac{\tau^2}{1+\tau^2} \quad\Longrightarrow\quad 0.80+0.80\tau^2=\tau^2 \quad\Longrightarrow\quad \tau^2=4, \qquad \tau=2.

Using the EC50 formula,

EC_{50} =\frac{6.00}{(2+2^2)^{1/2}-1} =\frac{6.00}{\sqrt{6}-1} =\frac{6.00}{1.449} =4.14\text{ nM}.

The predicted response increment at this concentration is half the 80-unit plateau, or 40.0 units. This calculation depends on the assumed \(n=2\) and the operational model; a plateau by itself does not determine EC50 without a shape assumption and an affinity scale.

Worked Example: Occupancy and Response Need Not Match

Consider another hypothetical system with \(E_m=100\) units, \(K_A=5.00\) nM, \(\tau=9\), and \(n=1\). At \(A=5.00\) nM, calculate occupancy and response.

The occupancy is

\theta=\frac{5.00}{5.00+5.00}=0.500.

Using the \(n=1\) operational equation, the response increment is

E-E_0 =100\frac{9(5.00)}{5.00+(1+9)(5.00)} =100\frac{45.0}{55.0} =81.8\text{ units}.

Half the receptors are occupied, while the predicted response is 81.8% of the system ceiling. The response is not a direct readout of the occupied fraction; the transduction step and operational efficacy determine how occupancy maps to effect.

Assumptions, Diagnostics, and Alternatives

The operational model is most interpretable when its mechanistic assumptions are plausible. The concentration \(A\) should represent the free concentration at the receptor and the binding-response system should be sufficiently close to equilibrium for an equilibrium occupancy equation to apply. The one-site occupancy expression presumes a homogeneous receptor population with reversible binding. The stimulus must be adequately represented as proportional to occupancy, and a single saturable transducer with one slope \(n\) must describe the response mapping. The response baseline must be separated from the agonist-produced increment.

In practice, examine whether the data cover enough of the concentration range to inform the lower response region, transition, and plateau. If concentrations are only low relative to \(K_A\), the data may constrain a combination of parameters without distinguishing affinity from operational efficacy. If the plateau is not observed, the fitted maximum and \(\tau\) may be weakly determined and strongly correlated with other curve parameters. Sparse measurements near the transition also make the slope difficult to estimate. These are limitations of information in the data, not problems fixed simply by reporting more decimal places.

Systematic residual patterns can indicate that the assumed curve shape or mechanism is inadequate. For example, consistent underprediction on one side of the transition and overprediction on the other can suggest a slope mismatch; a poor fit near the plateau can indicate an incorrectly specified maximum or an unobserved process. A single fitted curve cannot by itself establish that equilibrium, one-site binding, or proportional stimulus generation is biologically correct. Binding data, concentration coverage, and mechanistic knowledge help assess those assumptions.

An empirical Hill concentration-response curve is a useful alternative when the goal is to summarize the observed curve without assigning a receptor mechanism. It can describe a maximum, midpoint, and slope, but its fitted midpoint is not automatically \(K_A\), and its parameters do not provide the occupancy-to-stimulus interpretation supplied by the operational model. The simpler occupancy-only approach predicts occupancy, not response, unless a response mapping is added. The operational model is therefore useful when its added mechanistic structure is scientifically relevant and supported by the available information; that structure also brings assumptions that must be checked.

Common pitfall: Treating \(E_m\) as the response that every agonist must reach, or interpreting the observed plateau as a direct measure of affinity, confuses system capacity with agonist-specific response. The model predicts an agonist plateau of \(E_m\tau^n/(1+\tau^n)\); affinity enters the concentration scale, while \(\tau\) helps determine how much of the system ceiling is reached.

Check Your Understanding

Answer each question, then compare with the concise solutions below.

For \(n=1\), \(K_A=12\) nM, and \(\tau=3\), what is the predicted response fraction at \(A=6\) nM?
For \(n=1\) and \(\tau=1\), what fraction of \(E_m\) is reached at saturating agonist concentration?
With \(K_A=8\) nM, \(\tau=3\), and \(n=1\), calculate the EC50 and state why it differs from \(K_A\).
What does \(K_A\) describe, and why is it not generally equal to the operational model EC50?
Name two assumptions whose failure could make the operational model an inadequate description of an observed curve.

Solutions. (1) For \(n=1\), \(f=\tau A/[K_A+(1+\tau)A]\), so \(f=3(6)/[12+4(6)]=18/36=0.500\). (2) The limiting fraction is \(\tau/(1+\tau)=1/2=0.500\). (3) \(EC_{50}=K_A/(1+\tau)=8/4=2.00\) nM; it differs from \(K_A\) because stimulus-response amplification makes half of the agonist’s own maximum occur below the binding dissociation constant. (4) \(K_A\) is the equilibrium dissociation constant in the occupancy model; EC50 is the concentration producing half of the response plateau and depends on both \(K_A\) and operational efficacy (and, generally, \(n\)). (5) Examples include nonequilibrium conditions, multiple receptor populations, non-proportional stimulus generation, or a transducer relationship that is not adequately represented by the assumed saturable function.

Deriving the Operational Model of Agonism

What You'll Learn