Abstract
1. Macromolecules cross capillary walls via large vascular pores that are thought to be formed by plasmalemmal vesicles. Early hypotheses suggested that vesicles transferred plasma constituents across the endothelial wall either by a "shuttle" mechanism or by fusing to form transient patent channels for diffusion. Recent evidence shows that the transcytotic pathway involves both movement of vesicles within the cell and a series of fusions and fissions of the vesicular and cellular membranes. 2. The transfer of macromolecules across the capillary wall is highly specific and is mediated by receptors incorporated into specific membrane domains. Therefore, despite their morphological similarity, endothelial vesicles from heterogeneous populations in which the predominant receptor proteins incorporated in their membranes define the functions of individual vesicles. 3. Blood-brain barrier capillaries have very low permeabilities to most hydrophilic molecules. Their low permeability to macromolecules has been presumed to be due to an inhibition of the transcytotic mechanism, resulting in a low density of endothelial vesicles. 4. A comparison of vesicular densities and protein permeabilities in a number of vascular beds shows only a very weak correlation, therefore vesicle numbers alone cannot be used to predict permeability to macromolecules. 5. Blood-brain barrier capillaries are fully capable of transcytosing specific proteins, for example, insulin and transferrin, although the details are still somewhat controversial. 6. It has recently been shown that the albumin binding protein gp60 (also known as albondin), which facilitates the transcytosis of native albumin in other vascular beds, is virtually absent in brain capillaries. 7. It seems likely that the low blood-brain barrier permeability to macromolecules may be due to a low level of expression of specific receptors, rather than to an inhibition of the transcytosis mechanism.