The dimensions and numbers of small vesicles in cells, endothelial and mesothelial and the significance of these for endothelial permeability

Measurements were made of the diameters, numbers and other parameters of small, smooth-surfaced vesicles in the endothelium of blood capillaries and lymphatics, and in the mesothelial cells of the diaphragms of mice. Some measurements were also made on the aortic endothelium. With a few exceptions, there were no morphological differences between the various sites. It was found that between 25% and 35% of the non-nuclear cell volume was composed of vesicles, whose membranes accounted for about 55% of their volumes. Their internal volumes were similar to 70,000 nm(3), totalling similar to 0.04 mu m(3)/mu m(2) of luminal surface area. For each 1 mu m(2) there were similar to 135 vesicles attached to each surface membrane of the cell, and between similar to 200 and similar to 350 vesicles lying free in the cytoplasm. There was probably a slight amount of shrinkage during the preparation of the material, and the true linear dimensions were probably similar to 105% of those actually observed. Thus the values for the internal volumes were probably similar to 85,000 nm(3) and similar to 0.05 mu m(3) respectively; the vesicular numbers were probably similar to 125 attached to each surface and between similar to 175 and similar to 300 free. The vesicles attached to the plasma membranes often had quite long stalks; these were estimated to be similar to 30 nm at the moment of rupture. Thus the vesicles must be released an appreciable distance away from the membrane. This modifies the conclusions of Shea & Karnovsky (1966), since it can now be shown that Brownian movement alone is capable of accounting for the release of the vesicles, their movements within the cells and their transportation of material. By combining these results with others estimating the endothelial permeability coefficients, it can be calculated that the average free lifetime of a vesicle is similar to 1 1/2 sec, from union with one plasma membrane to the next. It can also be shown that the average time of such an attachment is similar to 2 1/2 sec. There are many possible sources of error relating to these measurements; they must only be regarded as tentative. It appears likely, however, that they are of about the correct order of magnitude as they accord well with other data.