Blood‐Brain Barrier Ultrastructure: Beyond Tight Junctions

Abstract

Listen

Translate article

It is well known that the structural basis of the blood-brain barrier (BBB) is the continuous interendothelial tight junctions; however, the significance of clefts in the junctional zone and of other structural features to BBB function is not as well understood. Because descriptive studies are insensitive and vulnerable to both subjectivity and perceptual errors, we have quantified BBB ultrastructure in a variety of species (FIGURE 1). BBB capillary walls are 30% thinner than walls in other continuous capillaries in the mouse.’ Since important nutrients cross the endothelial membrane via membranebound transport molecules and then diffuse across the endothelial cytoplasm, a thinner capillary wall would provide for a shorter diffusion path and therefore may be an advantage in vessels with restricted permeability. In the human, gray matter capillary walls are thinner than those in white matter. This may reflect the higher metabolic demands of gray matter. In the developing mouse, BBB capillary walls are thick in the fetus and become thin in parallel with the maturation of the BBB.2 In the aging human, capillary walls in both gray and white matter become thinner. This is due to both thinning of the endothelial layer and a decline in both size and number of pericytes.” If pericytes form a “second line of defense” in the BBB by phagocytosing molecules that escape the endothelial barrier, then their gradual loss during aging suggests that the BBB in the elderly may not be as able to compensate for transient leaks as in younger brains. We have serially reconstructed pericytes from normal human brain and have shown that at least 95% of them are granular (p<O.O2).‘ In non-BBB capillaries serial reconstruction of interendothelial junctions has shown that the junctional clefts form channels, some of which span the junction from lumen to ablumen.’ Our studies have shown that in the normal BBB, junctional clefts constitute 26% of the junctional length in both mouse and human (FIGURE 2). In the developing mouse this “cleft index” is high in the fetus and declines during the maturation of the BBB in parallel with the declining permeability? Furthermore, the clefts enlarge and can be filled with vascular tracer under some pathological conditions. A possible explanation for these observations is that in the normal BBB the channels formed by the clefts are not continuous across the entire junction, but that when the BBB is permeable, some of the clefts anastornose and form complete channels. Endothelial vesicles are much more numerous in permeable endothelium than in blood-brain barrier endothelium and so have been implicated in vascular permeability. Barrier vessels have densities of vesicles ranging from 3 to 11 per pm2 cytoplasm.’.’