Lack of selectivity to small ions in paracellular pathways in cerebral and muscle capillaries of the frog.

Selectivity to passive permeation of small ions through capillary walls was studied by measurements of diffusion potentials in response to ionic gradients established across capillary walls in frog brain and muscle in superfusion and perfusion experiments. Average dilution potentials in response to 2:1 or 10:1 gradients of NaCl across brain capillaries were 4.2 and 10.2 mV, respectively. The 'diluted' side was negative with respect to the 'undiluted' side, reflecting higher mobility of Cl- than of Na+ ions. Bi-ionic potentials in response to isosmotic KCl:NaCl gradients averaged 5.0 mV in brain capillaries, negative on the KCl side, reflecting higher mobility of K+ than of Na+ ions. The potential variations were symmetrical across the capillary wall. From Planck-Henderson formalism, the relative permeabilities in brain capillaries of Na+, K+ and Cl- were PCl/PNa:1.54 and PK/PNa:1.56, rather close to mobility ratios in free solution. Experiments on muscle capillaries showed similar results to those in brain. Streaming potentials created with excess (200 mosmol) mannitol or sucrose were congruent to 1 mV in brain capillaries and zero in muscle capillaries. It is concluded that the transcapillary permeation pathway in muscle and brain is neutral or weakly charged. The dominant ion permeation is paracellular in all 'continuous' capillaries. The large range of ion permeability of 'continuous' capillaries may be explained by variation in the length of the effective open portion of interendothelial junctions. The very low passive permeability of the blood-brain barrier may be due to an almost closed endothelial junction, leaving only about 0.1% of the length open. The open fraction is functionally similar to that in muscle and mesentery. This interpretation is in accordance with a finite, but very low, permeability to hydrophilic non-electrolytes. Such a brain capillary would still display strong preference for lipid-soluble solutes, but its behaviour cannot be satisfactorily understood from a simple analogy to a cellular plasma membrane.