Adding Endosomal Exchange To A Tissue Model

Why Add an Endosomal Compartment?

A tissue model with extracellular and intracellular spaces can represent entry into cells, but it cannot distinguish drug in the cell from drug that has moved into an endosomal space. That distinction matters whenever transfer across the cell–endosome membrane takes time: the endosomal amount can change gradually rather than appearing immediately in a fixed proportion to the intracellular amount. The model therefore needs a separate endosomal state and a finite flux connecting it to the intracellular state.

The earlier material on choosing tissue compartments and deriving membrane fluxes supplies the basic ideas: compartments have amounts and concentrations, and a permeability–area product converts a concentration difference into an amount rate. Here those ideas are assembled into one three-compartment model. The purpose is to write the balances and make each flux direction explicit, not to assume that the compartments are instantaneously equilibrated.

For clarity, the model below uses passive exchange on a common concentration basis: each membrane flux is conductance multiplied by the concentration difference across that membrane. This corresponds to equal partitioning on the chosen concentration basis. If a model requires a membrane partition factor or a more complex transport law, the driving-force expression must be modified consistently; it should not be added as an unexplained multiplier. Conductances are nonnegative and finite. Volumes are fixed, and the balances initially omit metabolism, binding reactions, and elimination within these three spaces.

Define Fluxes Before Writing Balances

Choose a positive direction for each flux once and use it consistently. Let \(J_{ec}\) be positive from extracellular space into the cell, and let \(J_{ie}\) be positive from the cell into the endosome. With conductances \(G_{ec}\) and \(G_{ie}\), the passive fluxes are:

Each flux has units of amount/time. For example, a positive \(J_{ie}\) means that the intracellular concentration exceeds the endosomal concentration, so the net flux is into the endosome. If the concentration difference reverses, the flux is negative and the net direction reverses. A negative value is not a negative amount in a compartment; it is a signed rate under the direction convention just defined.

Let \(R_e(t)\) denote any net non-membrane input to the extracellular compartment, in amount/time. It may represent a prescribed delivery rate or the net result of a more detailed vascular exchange model. It is not itself a membrane flux in the three-compartment system. Setting \(R_e(t)=0\) gives a closed model with no external delivery or removal.

Derive the Coupled Amount Balances

For each compartment, the rate of change of amount is the sum of its inflows minus the sum of its outflows. The extracellular compartment receives \(R_e(t)\) and loses \(J_{ec}\). The intracellular compartment gains \(J_{ec}\) and loses \(J_{ie}\). The endosomal compartment gains \(J_{ie}\). Applying this bookkeeping gives:

Substituting the flux expressions produces the coupled model directly:

These are amount balances, so their left-hand sides and right-hand sides have units of amount/time. Because \(C_j=A_j/V_j\), fixed compartment volumes also allow the model to be written as concentration-rate equations:

The volume division is essential: an amount rate is not a concentration rate. The same flux can produce a larger concentration change in a smaller compartment. If volumes change with time, the simple conversion \(dC_j/dt=(dA_j/dt)/V_j\) no longer suffices; differentiating \(A_j=V_jC_j\) also introduces the changing-volume term.

Direction Reversal and Membrane Scaling

A finite flux is signed, so neither membrane should be treated as a one-way transfer unless the biology specifically requires a one-way process. In a passive model, the concentration difference determines the net direction at each instant. This makes the same equations usable during uptake and return without switching formulas.

If permeability \(P_{ie}\) and membrane area \(S_{ie}\) are supplied separately, set \(G_{ie}=P_{ie}S_{ie}\) after checking their units. This gives a useful scale check before integrating the model.

Assumptions, Checks, and Alternatives

The equations rely on a well-mixed concentration in each compartment, fixed volumes, passive linear fluxes, and the stated common concentration basis. A compartment concentration is therefore treated as representative of its entire volume. Conductance is assumed not to change over the calculation interval. If transport is saturable, active, or directionally asymmetric, the flux law must reflect that biology; replacing it with a passive concentration-difference term would be a model assumption, not a neutral simplification.

Several checks help identify a model that is being implemented incorrectly. At equal concentrations across a membrane, its passive flux must be zero. If the source-side concentration exceeds the destination-side concentration, the flux should be positive under the chosen direction. Every membrane term must appear as an outflow in one balance and an inflow in the adjacent balance with the same magnitude. Conductance times concentration must have units of amount/time, and concentration rates require division by volume. Finally, amounts should not become negative in a physically valid simulation; negative values often indicate an inconsistent sign, unsuitable numerical step size, or an invalid input rather than a meaningful prediction.

An explicit endosomal compartment is more informative than a single intracellular compartment when the amount in the endosome is a desired state or when finite transfer is important to the question. A one-cell-compartment model is simpler, but it cannot report a distinct endosomal amount. Conversely, an instantaneous partitioning approximation replaces dynamic exchange with an algebraic constraint and removes the endosomal transfer rate as a dynamic process. That approximation is only appropriate when its assumptions are justified for the application; it is not the same model as finite-rate exchange.

The three-compartment structure is still a simplification. It does not specify vascular delivery, binding, pH-dependent ionization, endosomal processing, or intracellular metabolism unless those processes are added explicitly. Nor does the presence of an endosomal state establish that available data can estimate its conductance. The model provides a clear bookkeeping structure; deciding whether its additional detail is needed and supported remains a separate modeling judgment.