Irreversible Inhibition and Target Resynthesis

Why Inactivation and Resynthesis Must Be Modeled Together

For a reversible inhibitor, target activity can return as drug dissociates from the target. A covalent inhibitor can instead leave a target molecule permanently inactive, on the timescale of the model. Recovery then depends on the production of new active target and the removal of old target. The same drug concentration can therefore produce different active-target levels depending on how long exposure lasts and how quickly the target turns over.

The model below separates active target from covalently inactivated target. It treats drug exposure as a driver of conversion from the active pool to the inactive pool, while synthesis adds new active target and degradation removes target. This distinction lets us calculate both the decline during treatment and recovery after treatment stops.

Coupled Equations and Their Assumptions

A commonly used concentration-dependent inactivation rate is \(k_{\mathrm{obs}}(t)=k_{\mathrm{inact}}C_u(t)/(K_I+C_u(t))\), where \(C_u(t)\) is unbound drug concentration at the target, \(k_{\mathrm{inact}}\) is the maximum inactivation rate, and \(K_I\) is the concentration giving half of that maximum. The concentration is understood to drive formation of a drug–target complex before covalent inactivation; this compact rate expression assumes that the relevant binding step is fast enough to be represented by the concentration-dependent rate.

The equations assume that target synthesis is constant, target degradation is first-order with the same rate for active and inactivated target, and each inactivation event removes one molecule from the active pool and adds it to the inactive pool. Inactivation is treated as irreversible over the modeled period: there is no conversion from \(I\) back to \(A\). The model also assumes that the unbound concentration at the target is known or supplied by an exposure model, and that drug does not directly alter target synthesis or degradation.

Before drug exposure, the target is at baseline steady state, with all target active. Setting \(k_{\mathrm{obs}}=0\) and \(dA/dt=0\) gives the baseline amount \(A_0=k_{\mathrm{syn}}/k_{\mathrm{deg}}\). Define the normalized active fraction \(a(t)=A(t)/A_0\). This normalization makes the results interpretable as a fraction of baseline active target.

The total target \(T=A+I\) obeys \(dT/dt=k_{\mathrm{syn}}-k_{\mathrm{deg}}T\), because inactivation moves target between pools but does not itself remove total target. Thus, if total target was at baseline before dosing, it remains at \(A_0\) under these assumptions. The fall in active target reflects a change in target state, not necessarily a fall in total target.

Constant Exposure and Recovery After Washout

For constant exposure, \(k_{\mathrm{obs}}\) is constant during treatment. The active fraction approaches a treatment steady state \(a_{\mathrm{ss}}=k_{\mathrm{deg}}/(k_{\mathrm{deg}}+k_{\mathrm{obs}})\). Starting from baseline, the solution at exposure duration \(t_d\) is:

Once drug concentration at the target falls to zero, \(k_{\mathrm{obs}}=0\). Active target then approaches baseline with rate constant \(k_{\mathrm{deg}}\). If washout occurs at time \(t_d\), then for \(t\ge t_d\):

These expressions show two separate timescales. During exposure, the approach to the inhibited level depends on \(k_{\mathrm{deg}}+k_{\mathrm{obs}}\). After exposure, recovery depends on target turnover alone. The target half-life in this first-order model is \(\ln(2)/k_{\mathrm{deg}}\).

How Duration and Turnover Shape the Result

A longer exposure allows active target to move closer to the treatment steady state. It does not necessarily drive active target to zero: under constant synthesis and degradation, the steady-state active fraction is positive whenever \(k_{\mathrm{deg}}>0\). A larger inactivation rate lowers that steady-state fraction. After washout, a smaller \(k_{\mathrm{deg}}\) means a longer target half-life and slower recovery.

Variable Exposure, Diagnostics, and Limitations

If the target-site concentration changes over time, use the coupled differential equations with the time-varying \(k_{\mathrm{obs}}(t)\), rather than treating the entire exposure as constant. For a specified rate function, the active fraction has the integrating-factor solution:

This expression follows by multiplying the active-fraction equation by the integrating factor \(e^{\int_0^t[k_{\mathrm{deg}}+k_{\mathrm{obs}}(u)]du}\) and integrating from baseline \(a(0)=1\). In practice, numerical integration is often simpler when the exposure profile has multiple peaks or when concentration is generated by a pharmacokinetic model.

Useful checks include confirming that \(k_{\mathrm{obs}}\) is nonnegative, rates and time use compatible units, and the simulated active fraction stays between zero and one for baseline initial conditions and nonnegative rates. With constant exposure, the active fraction should move toward \(a_{\mathrm{ss}}\); after washout, it should move monotonically toward one. A simulated increase above baseline or continued decline after complete washout signals a coding error or a model assumption beyond those stated here.

A reversible occupancy model is an alternative when binding is reversible and target activity returns as drug dissociates. It is not interchangeable with this model: covalent inactivation leaves a persistent inactive pool after drug concentration falls. Conversely, this irreversible model is inadequate if the covalent adduct is substantially reversed, if active and inactive target have different degradation rates, or if synthesis changes during treatment. Such cases require additional states or modified rate equations, supported by data that can inform the added parameters.

A common mistake is to assume that active target returns immediately to baseline once drug concentration reaches zero. The correction is that washout stops new inactivation but does not reactivate the existing inactive pool in this model. Active target returns through synthesis and turnover, with the recovery timescale set by \(1/k_{\mathrm{deg}}\). Another pitfall is to infer total target loss from a decline in active target alone: the model predicts that total target can remain at baseline while the active fraction falls.