Solute recirculation

The compartmental model of Larsen et al. (2000, 2002) for solute-coupled water transport across toad intestine simulates a leaky epithelium involved in isotonic fluid absorption, allowing the osmolality of the transported fluid to either be fixed at isotonic or to vary depending on parameter selection. When the measured toad intestine water and solute permeabilities and active transport rates are incorporated into the model, a markedly hypertonic absorbate is predicted to exit from the lateral intercellular spaces. Most previous compartmental models of isotonic fluid absorption encountered the hypertonic absorbate problem and solved it either by increasing the water flux into the lateral intercellular spaces or by decreasing the solute diffusion coefficient within those spaces. In the present model, even when the water flux is increased by postulating that 75 % of the water flows across a highly water-permeable tight junction, formation of a hypertonic absorbate is predicted unless solute efflux is greatly retarded. The solution proposed by Larsen et al. (2000, 2002) is that 60-80 % of the solute that has been actively transported out of the cell re-enters it in a reversed flow across the basal membrane. This phenomenon has been denoted ‘solute recirculation’ and was first suggested by Ussing & Nedergaard (1993) as a way of explaining the non-unity caesium flux ratios in the toad intestine. The concept underwent further refinement and development in their investigations of toad intestine (Nedergaard et al. 1999) and the exocrine glands of frog skin (Ussing et al. 1996).

In the Larsen et al. model, the proposed recirculating solute flux is accompanied by a water flow into the cell across the basolateral membrane equal to 50 % of the absorbate. When a water-impermeable tight junction is assumed, the model predicts that the fraction of recycled solute would rise to 87 %, implying that the rate of isotonic fluid absorption would diminish to 13 % in the absence of a basal bathing solution or that the solute concentration of the transported fluid would rise to over 500 mm. Such a central role for solute recirculation in the formation of an isotonic absorbate has not been envisioned by other investigators or tested experimentally except by the Copenhagen group.

The conclusion from the Larsen et al. model that isotonic absorption requires a water-permeable tight junction and substantial solute recirculation depends critically on the caesium flux ratio. A previous attempt to separate the solute fluxes across an epithelium into transcellular and paracellular components used the caesium flux ratio as a putative measure of paracellular permeation and found values about double that expected for passive diffusion (Nedergaard et al. 1999). Ussing & Nedergaard (1993) were the first to utilize caesium as a paracellular tracer. I question whether it is an appropriate tracer for such purposes because caesium is far from inert. It readily enters cells and accumulates in place of potassium, blocking some potassium channels. In the Larsen et al. model, the observed sodium tracer fluxes are partitioned into transcellular and paracellular components on the basis of the caesium flux ratio. The calculated paracellular sodium flux ratio then amounts to 3.66, much greater than the unity value expected for a passive pathway. The authors conclude from the high flux ratio that significant solute-solvent interaction, i.e. solvent drag, must occur in the paracellular pathway.

According to the Larsen et al. (2002) model calculations, pseudo solvent drag, arising in unstirred layers on either side or within the epithelium, cannot yield the observed flux ratio. Pseudo solvent drag occurs because fluid flow in the lateral intercellular spaces sweeps solutes away from the tight junction and results in large local solute concentration differences. The resultant diffusional solute movement may then be mistaken for solvent drag as its magnitude varies with water flow. Larsen et al. (2002) note that more slowly diffusing paracellular tracers, such as sucrose, inulin or Dextran, can produce flux ratios substantially greater than unity as a result of pseudo solvent drag, but conclude that the caesium diffusion coefficient in free solution is sufficiently high to rule this possibility out. However, the relevant parameter – the caesium diffusion coefficient within the lateral intercellular spaces – has never been measured.

The first direct measurements of water flux across the tight junctions of a leaky epithelium (MDCK cells) did not demonstrate any significant contribution of transjunctional water movement to the transepithelial fluid flux (Kovbasnjuk et al. 1998). Water permeability of the tight junctions was not ruled out by this investigation; it was merely shown to be insignificant in comparison with the transepithelial flow. Thus, the tracer caesium flux ratio could be influenced by solvent drag within the pores of the tight junction without transjunctional water flow significantly contributing to net fluid absorption. Since the model by Larsen et al. (2000, 2002) depends critically on the caesium flux ratio value, there is a clear need for experiments using other accepted paracellular tracers showing that their flux ratio is similarly asymmetric and that this asymmetry is due to true solvent drag within the tight junctions.

Finally, although Larsen et al. (2000, 2002) make a number of arguments in their papers that negating half or more of the work done to actively transport solutes out of the cell is both energetically feasible and serves to explain several previously unexplained phenomena, I am not persuaded. Evidence for such large solute fluxes would have appeared in tracer flux studies decades ago if they were any more than a mathematical convenience or the consequence of relying on a poorly chosen paracellular tracer species.