Metformin Disposition Through OCT2 and MATE Transport

From Plasma to Urine Through a Renal Cell

Metformin is cleared largely by the kidneys, and its renal handling involves transport across two different membranes of the proximal tubule cell. Organic cation transporter 2 (OCT2) mediates uptake from blood across the basolateral membrane. Multidrug and toxin extrusion transporters (MATE1 and MATE2-K) mediate efflux from the cell across the apical membrane into tubular fluid. The direction of this sequence matters: changing uptake and changing efflux need not have the same effect on the amount of metformin inside the cell.

A useful model must also account for where metformin can go if it enters a cell but is not transported into urine. The examples below include passive return from the cell to blood. This makes the mass balance explicit and allows inhibition at either membrane to change net secretion. The model is a deliberately simplified teaching example, not a fitted or clinically validated metformin model.

The model assumes linear transport over the concentration range considered, a well-mixed renal cell, no metformin metabolism in the cell, and no other cellular exit pathway. It treats plasma concentration as fixed while calculating cell balance. For the systemic exposure example, filtration and nonrenal clearance are added separately. Tubular reabsorption, changes in urine flow or pH, and changes in transporter abundance are not represented.

Deriving Steady-State Secretion

The amount of metformin in the cell changes according to uptake minus both exit fluxes. If \(V_c\) is the modeled cell volume, the mass balance is:

At steady state, cellular amount is constant, so the left side is zero. Solving the resulting equation gives the intracellular concentration and the apical secretion rate:

Thus the unbound-plasma-based secretory clearance is \(CL_{\mathrm{sec},u}=J_{\mathrm{MATE}}/C_{p,u}\). If \(f_u\) is the unbound fraction in plasma and \(C_p\) is total plasma concentration, then \(C_{p,u}=f_uC_p\), and secretion expressed relative to total plasma concentration is \(f_uCL_{\mathrm{sec},u}\).

What Inhibition at Each Membrane Changes

Represent inhibition as a fractional reduction in the effective clearance of the affected transporter. If \(I_{\mathrm{OCT2}}\) and \(I_{\mathrm{MATE}}\) are inhibition fractions between 0 and 1, then:

The asterisk denotes the value in the presence of inhibitor. These expressions assume that the inhibitor changes only the specified effective clearance, with no change to \(CL_b\), plasma binding, filtration, or the other transporter. The starred clearances can be substituted directly into the steady-state equations. This simple fractional reduction is a modeling assumption; it is not a universal rule for translating inhibitor concentration into transporter inhibition.

Connecting Renal Secretion to Plasma Exposure

The renal-cell result predicts secretion at a given plasma concentration; it does not alone give the plasma concentration. To make that connection, consider a hypothetical linear systemic model at steady state during constant input. Let \(R_{\mathrm{in}}\) be the input rate, \(CL_{\mathrm{NR}}\) nonrenal clearance, \(GFR\) glomerular filtration rate, and \(f_u\) the unbound fraction. Assume filtration clearance relative to total plasma concentration is \(f_uGFR\), and secretion clearance relative to total plasma concentration is \(f_uCL_{\mathrm{sec},u}\). Then total clearance and steady-state total plasma concentration are:

This relation assumes constant input, linear clearances, constant binding, and a steady state. If inhibition reduces renal clearance while all other terms stay fixed, the model predicts higher steady-state plasma concentration. The size of that increase depends on how much of total clearance was renal secretion; it is not determined by the percentage inhibition alone.

Assumptions, Diagnostics, and Model Alternatives

The steady-state solution follows from the cellular mass balance, but its interpretation depends on the assumed pathways. A useful diagnostic is to check flux balance explicitly: at steady state, uptake must equal the sum of MATE efflux and return to blood. A calculated cell concentration that does not satisfy this equality indicates an algebra or implementation error. Units provide another check: clearance multiplied by concentration must yield amount per time, and secretory clearance must have units of volume per time.

The model also assumes linear transport. OCT2 and MATE transport can be saturable, and the linear equations are most defensible over a range where flux is approximately proportional to concentration. If concentrations approach transporter capacity, a model using saturable fluxes, such as \(J=V_{\max}C/(K_m+C)\), may be more appropriate. Then the cellular balance generally requires solving a nonlinear equation, and the effect of inhibition can depend on the concentrations and the transporter’s kinetic parameters. A constant fractional reduction in clearance may no longer describe the effect across doses.

There is an important contrast with a simpler irreversible two-step model that omits return to blood. In that model, steady state requires OCT2 uptake to equal MATE efflux. If uptake is a fixed linear function of plasma concentration, MATE clearance determines the cell concentration needed to carry the same flux, but does not change the eventual steady-state flux. MATE inhibition would then predict accumulation and a longer approach to steady state, not a lower steady-state secretion rate. That result may be unsuitable when the model needs to represent a reduction in net secretion. Adding return to blood is one explicit way to represent competing cellular exit; other models may instead include additional pathways or transport kinetics.

For metformin, transporter direction and cellular context are essential: OCT2-mediated entry is basolateral, and MATE-mediated exit is apical. MATE transport is proton-coupled, so local conditions can affect the effective apical transport. The present model does not resolve proton gradients, separate MATE1 from MATE2-K, or represent other transporters. Nor does it establish that an observed change in plasma exposure is caused exclusively by OCT2 or MATE inhibition. The parameter values and inhibition fractions must be justified for the intended application.

A practical comparison of model choices is summarized below.

The linear model is useful for learning how membrane-specific inhibition propagates through a cell balance and into a simple systemic exposure calculation. Its limitations are equally important: it cannot, by itself, establish clinical effect size, capture dose-dependent saturation, or represent changes in renal physiology.