Distinguishing Inhibition of Input From Inhibition of Loss

Why the Direction of Change Matters

An indirect response is governed by both the rate at which it is supplied and the rate at which it is removed. An inhibitory drug effect on supply tends to lower the response; an inhibitory drug effect on removal tends to raise it. These opposite predictions make the direction of change after drug exposure a useful mechanistic clue. The size and timing of the change matter too: an observation taken before the response approaches equilibrium can differ substantially from the eventual drug-on level.

The comparison below assumes a stable baseline, a constant drug effect during the on-treatment phase, and a response that follows a first-order turnover model. Under those assumptions, the two mechanisms make explicit predictions for onset, equilibrium, and washout. These predictions are diagnostic expectations, not proof that a particular mechanism caused an observed change.

With no drug effect, \(q=1\), and either model has the baseline steady state \(R_0\). During drug exposure, the same \(q\) is applied to different processes: input in the first model and loss in the second. That distinction drives the different predictions. The fractional inhibition can be regarded as fixed here because exposure, and thus the drug effect, is assumed constant for the interval being analyzed.

Onset, Equilibrium, and the On-Treatment Trajectory

Dividing each equation by \(R_0\), and using \(k_{\text{in}}=k_{\text{out}}R_0\), gives the normalized models:

At drug onset the response is at baseline, so \(r(0)=1\). The initial slopes are therefore:

For \(I>0\), input inhibition produces an initially negative slope, while loss inhibition produces an initially positive slope. The corresponding slopes in response units per time are these normalized slopes multiplied by \(R_0\). Thus the sign predicts direction, while \(k_{\text{out}}\) and the inhibition level together determine the initial rate of change.

Setting the derivative to zero gives the drug-on equilibria. For input inhibition, \(r_{\text{ss}}=q\); for loss inhibition, \(r_{\text{ss}}=1/q\), provided \(q>0\). The first equilibrium is at or below baseline, and the second is at or above it. If \(q=0\) under loss inhibition, loss is completely blocked in this model: the response increases at the constant rate \(k_{\text{in}}\), so there is no finite drug-on equilibrium.

The trajectories follow by solving the linear differential equations with \(r(0)=1\). For input inhibition, subtract the equilibrium \(q\): the difference \(r-q\) satisfies \(d(r-q)/dt=-k_{\text{out}}(r-q)\). For loss inhibition with \(q>0\), the difference \(r-1/q\) satisfies \(d(r-1/q)/dt=-qk_{\text{out}}(r-1/q)\). Applying the initial condition yields:

The input-inhibition response approaches its equilibrium with rate constant \(k_{\text{out}}\), whereas the loss-inhibition response approaches its equilibrium with rate constant \(qk_{\text{out}}\). Their corresponding time constants, the reciprocals of these rate constants, are \(1/k_{\text{out}}\) and \(1/(qk_{\text{out}})\). Stronger loss inhibition (smaller \(q\)) both raises the predicted equilibrium and slows the approach to it. Inhibition of input lowers the equilibrium, but does not alter the model’s response-approach rate constant.

Washout and the Information It Provides

If the drug effect disappears immediately at washout, then \(q\) returns to 1 and both models reduce to \(dr/dt=k_{\text{out}}(1-r)\). Starting from any normalized response \(r_T\) at washout, the recovery trajectory is:

Thus, once the effect has fully disappeared, both mechanisms predict return toward baseline at rate constant \(k_{\text{out}}\). They differ in the level from which recovery starts: input inhibition generally leaves the response below baseline, while loss inhibition generally leaves it above baseline. In each case, the post-washout change is toward \(r=1\).

Which Observations Discriminate the Mechanisms?

The strongest basic discriminator is whether the response initially moves down or up from a stable baseline during exposure. A sustained decrease is consistent with inhibition of input in this pair of models; a sustained increase is consistent with inhibition of loss. The drug-on equilibrium provides a second signature, and washout provides context: compare the starting level and direction of recovery with the predicted trajectories rather than treating a return toward baseline as unique to either mechanism.

The following comparison assumes that the response is measured directly, the baseline is stable, the inhibitory effect is the only drug-driven change, and exposure is effectively constant during the on-treatment interval.

Assumptions, Diagnostics, and Limitations

The comparison relies on a stable baseline with \(R_0=k_{\text{in}}/k_{\text{out}}\), a single measured response governed by first-order loss, and a drug effect that acts on the specified rate without directly changing the other rate. During each on-treatment calculation, \(q\) is constant. For the simple washout formula, the drug effect must have disappeared immediately. If concentrations decline gradually, the effect generally changes over time, so the immediate-washout expression is not an exact description of that interval.

Check whether the observed pattern is compatible with these conditions before assigning a mechanism. A drifting pre-dose baseline weakens the interpretation of “up” or “down” relative to baseline. Sparse sampling can miss the onset direction or confuse a transient with an equilibrium. A changing concentration can make a constant-effect trajectory inappropriate, while feedback or a direct drug effect can produce behavior outside either simple model. The earlier material on scale confounding between production and loss rates is relevant here: baseline alone gives their ratio, not each rate separately. Time-course data add mechanistic information, but do not eliminate every source of ambiguity.

A common mistake is to infer the mechanism from the washout direction alone. For example, seeing the response fall after drug removal does not uniquely identify inhibition of loss: a response below baseline after input inhibition rises on washout, but a response above baseline after loss inhibition falls on washout. Both are returning toward baseline. Correction: consider the on-treatment direction and level, the response at washout, and the time course together. A second mistake is to treat \(1/q\) as a finite equilibrium when \(q=0\); in that limiting case the loss-inhibition model has zero loss and predicts linear accumulation, not a finite plateau.

The primary alternative within this comparison is inhibition of the other turnover process; the models predict opposite onsets and opposite sides of baseline at equilibrium. Other model structures can also explain observed changes, so matching one signature is not proof of mechanism. In particular, if the effect is not constant or the response does not follow the assumed turnover equation, the derived trajectories need not fit. The appropriate conclusion is therefore conditional: given the stated assumptions, one mechanism may be more consistent with the observed response pattern than the other.