Reversible Precursor-Product Exchange Models

Why Reversible Exchange Needs Two Coupled Equations

A precursor can form a product, while the product can also be converted back into the precursor. In that case, treating conversion as a one-way maturation step omits a pathway that changes both pools. The resulting error is not just a different estimate of the forward rate: reverse conversion changes the precursor balance and therefore changes the amount available to form product.

The model below describes two well-mixed pools with first-order exchange and first-order loss from each pool. It can represent a biomarker system when its assumptions are reasonable, but it is not a universal biochemical mechanism. In particular, it assumes that the exchange rates do not saturate and that the pools can be expressed in compatible units.

Assume first-order loss and exchange, a constant precursor input, and one-to-one exchange between pools measured on a compatible scale. Loss is removal from the modeled system, not conversion into the other pool. Under these assumptions, mass balance gives:

The term \(k_fP\) leaves the precursor and enters the product; \(k_rQ\) leaves the product and returns to the precursor. Each pool also loses material at its own rate. Adding the equations cancels the exchange terms:

This cancellation is a useful mass-balance check: exchange redistributes material, but does not itself create or destroy total material in the two-pool system.

Steady State and Its Conditions

Define \(a=k_P+k_f\), \(b=k_r\), \(c=k_f\), and \(d=k_Q+k_r\). These abbreviations make the coupled structure easier to see: \(dP/dt=R-aP+bQ\) and \(dQ/dt=cP-dQ\). At steady state, both derivatives equal zero. The product balance gives \(Q_{\mathrm{ss}}=(c/d)P_{\mathrm{ss}}\); substituting this into the precursor balance yields the steady-state values.

With nonnegative rates, \(D>0\) guarantees the unique finite steady-state formulas above. For example, positive \(k_f\) and \(k_Q\) ensure \(D>0\), even if \(k_P=0\). When \(D=0\), check the balances separately: if \(k_f=k_r=k_Q=0\) and \(k_P>0\), then \(P_{\mathrm{ss}}=R/k_P\) and any finite constant \(Q\) is a steady state. In all other \(D=0\) cases, positive constant input has no finite steady state; if both pools have no loss, then \(k_P=k_Q=0\), so continuous input accumulates.

Transient Profiles and the Two Decay Rates

The steady state describes a constant input maintained long enough for both pools to stop changing. To describe deviations from that state, write \(x=P-P_{\mathrm{ss}}\) and \(y=Q-Q_{\mathrm{ss}}\), assuming the input and rates remain constant. The deviations obey a homogeneous linear system. Its two decay rates determine how a perturbation fades:

The signs in the rate formula are chosen so that \(\lambda_{\mathrm{slow}}\) is the smaller rate and \(\lambda_{\mathrm{fast}}\) is the larger rate. When \(D>0\), both are positive for nonnegative rates with the exchange structure above, so the corresponding modes decay as \(e^{-\lambda t}\). Their units are inverse time. An observed pool can therefore show a combination of two exponential modes rather than a single exponential decline.

A short derivation makes the result interpretable. For constant input, deviations satisfy \(x'=-ax+by\) and \(y'=cx-dy\). Differentiating the second equation and substituting for \(x\) gives:

Thus \(y''+(a+d)y'+Dy=0\). Trying a mode \(y=e^{-\lambda t}\) gives \(\lambda^2-(a+d)\lambda+D=0\), whose roots are the two rates above. The same rates govern precursor deviations. This derivation assumes constant coefficients; a time-varying input changes the forcing term, while time-varying rates require a different solution.

What Forward and Reverse Conversion Change

At steady state, the pool ratio \(Q_{\mathrm{ss}}/P_{\mathrm{ss}}=k_f/(k_Q+k_r)\) shows that more forward conversion tends to raise product relative to precursor, while more reverse conversion lowers that ratio. The reverse rate also appears in \(d=k_Q+k_r\), so it affects both the product balance and the coupled decay rates. It is therefore not generally valid to treat reverse conversion as if it were simply an added product loss: reverse flux returns material to \(P\), where it can later be converted again.

Model Checks, Alternatives, and Limitations

Before interpreting fitted rates, check whether the model's structure and assumptions match the measurements. The equations imply nonnegative pools when the initial pools and input are nonnegative and the rates are nonnegative. They also imply that exchange terms cancel in the total balance. A numerical solution that creates a negative amount or fails this balance is a warning about implementation, parameter constraints, or units.

• Look for systematic profile mismatch. If measured product changes before the precursor could plausibly influence it, or if residuals show persistent curvature or time-dependent bias, the assumed two-pool mechanism may be inadequate. Random scatter around predictions is different from a repeated directional pattern.
• Check whether the rates can reasonably be treated as constant. Saturable conversion, time-varying drug effects, or changing pool volumes violate the simple first-order, constant-rate formulation.
• Check the meaning of “loss.” If material leaving one modeled pool enters another unmodeled pool, that process is not a true loss from the modeled system. The state equations may need another compartment.
• Check units and stoichiometry. If precursor and product are measured in different volumes or with different molar scaling, one-to-one amount exchange cannot be written with identical flux terms until the pools are put on a compatible scale.

An irreversible precursor–product model is a simpler alternative when reverse flux is chemically negligible or unsupported by the evidence. A single-pool turnover model may be adequate when a separate precursor state is not needed to describe the observed response. These alternatives are not interchangeable: the reversible model predicts feedback of product into precursor, whereas the simpler models omit that pathway. If only one pool is observed, several rate combinations may produce similar profiles, so a good-looking curve alone does not establish that reverse exchange has been identified.

A common mistake is to set the product derivative to zero and conclude that \(k_fP=k_QQ\). That omits the reverse flux. The correct steady-state product balance is \(k_fP=(k_Q+k_r)Q\); reverse conversion is part of the rate leaving the product pool. Another mistake is to infer that a product decline must begin immediately after input stops. As the worked example shows, if the pools begin at steady state, the product slope can initially be zero while the precursor has already begun to depart from steady state.