5 nm, the endothelial vesicular system has been the best structural candidate for the large pore system. As large pores are far fewer HSP inhibitor in number than small pores and are expected to undergo a dynamic fluctuation between open and closed states, the occurrence of large pores in an EM section should be infrequent. The dynamics and interactions of endothelial vesicles are unknown. Palade [13,14] first described endothelial vesicles and postulated a discontinuous mechanism of transport whereby vesicles shuttled solutes between luminal and abluminal surface.

Simionescu et al. [19] described transendothelial channels of fused vesicular compartments that were true pores through which solutes could move. However, Bundgaard et al. [1] detected very few if any free vesicles in serial section reconstructions

of the capillary wall, which showed that the standard configuration of vesicular compartments was fused clusters of vesicles connected to either surface but not both. These Selleck BAY 80-6946 studies were based on reconstructions of ultrathin (25 nm) sections through randomly chosen regions. As large pores need only to occur at a frequency 1/μm2 of capillary wall [17], it is possible that free vesicles and open channels may have been missed in these studies. In contrast, Wagner and Robinson [26] examined stereopairs of high-voltage electron images of thick (0.5–1.0 μm) sections and detected free vesicles not connected to either surface. Distribution of perfused tracer through serial sections of the capillary wall has also provided evidence that the vesicular system is involved in transport [25]. These previous 3D studies have limitations that leave uncertainty regarding the structure of the vesicular system and have sometimes produced conflicting results. Another uncertainty lies in whether

or not conventional methods of chemical fixation produce artifactual vesicular configurations. Comparing cryofixation with chemical fixation, Frøkjaer-Jensen et al. [5] showed that interconnection of vesicular structures persisted regardless of the type of fixation. Wagner and Andrews [22] demonstrated that chemically fixed capillaries had significantly more vesicular profiles per unit area than isothipendyl cryofixed capillaries, which suggested that vesicle formation may be stimulated by aldehyde fixation. However, comparisons between this study using aldehyde fixation and those of Lebbink et al. [9] on cryofixed endothelial cells indicate that free vesicles and transendothelial channels persist regardless of fixation method and are most likely bona fide biological structures. This study constitutes a new approach, marrying a previous technique of perfusing tracers through capillaries with TEM tomography. As perfused agents that increase permeability in capillaries may also affect the conformation of vesicular structures [2], it could be reasonably argued that terbium might induce the formation of transendothelial channels and/or free vesicles.