Glomerular permeability: a never-ending saga.

THE PAPER BY TANNER ET AL. (27) revisits the critical question of how the intricate glomerular capillary membrane provides such a highly restrictive barrier to macromolecules while still allowing the movement of vast volumes of protein-free fluid (1, 7, 13, 14, 20). In recent years, this topic has emerged from dormancy because of exciting new technology and because several recent studies presenting provocative data have challenged well-established concepts regarding glomerular permselectivity (5, 6, 20, 24). In particular, it has been proposed that vast quantities of protein are normally filtered and that tubular reabsorption of intact proteins is primarily responsible for preventing loss of protein from the body (6, 24). The complex nature of the glomerular capillary wall with three distinct components has given rise to many questions regarding the specific roles of each of the major barriers, with arguments going from one extreme to the other (2, 6–9, 16, 20). Furthermore, the role of charge on the glomerular endothelium and basement membrane has been challenged anew (21, 26), even though it was once thought to be comfortably settled. In essence, almost every aspect of glomerular permselectivity is under new scrutiny, and we can only hope that the new studies will bring an enlightened understanding rather than further confusion. The utilization of fluorescently labeled probes of varying molecular size and charge coupled with applications of modern two-photon confocal microscopy has provided a new approach to answer these unresolved questions. Although numerous studies have examined the permeability properties of the glomerular membrane, the authors emphasize that the high resolution of the new technology coupled with the extraordinary width of the glomerular basement membrane (GBM) in kidneys from Necturus maculosus (salamander or commonly known as the “mud puppy”) allows visualization and quantitation of the fluorescent molecules in the GBM as well as in Bowman’s space. This allows for a more detailed consideration of the specific barrier functions of the individual components of the capillary wall. Volumes have been written about barrier functions of capillary systems throughout the body (17, 20, 28), but it is recognized that the glomerular capillaries are among the most complex and challenging to understand. Tanner and colleagues (27) make several important points. Certainly, one of the most important conclusions is that the overall glomerular permeability to large macromolecules, in particular albumin, is extremely low with a sieving coefficient in the range of 0.005. These data support many of the “classic findings” (1, 7, 12, 14) and are at variance with suggestions by Comper and associates (5, 6, 23, 24) that much greater amounts of protein are filtered by even normal glomeruli. Another conclusion is that charge selectivity could not be detected in the glomerular membrane of Necturus kidneys (4, 15, 18, 23, 26, 29). However, the data indicating that anionic charges do not seem to have an important role is a difficult one to generalize (7, 15, 16, 20). This issue is physiologically relevant as charge has often been used to explain the fact that serum albumin has a much lower glomerular sieving coefficient than would be expected from just its molecular size and/or effective molecular radius. So, if the charge on albumin does not restrict passage, then what does? Until this is answered, the statement that “charge has no detectable effect on filterability of these macromolecules” must remain in question. We should not forget the numerous studies that contributed to the conceptual framework regarding the importance of the electrostatic barrier (1, 7, 13, 15, 16, 18). For the same equivalent radius, the fractional clearances of albumin (3.6 nm) and negatively charged dextran sulfate are considerably lower than the clearances of uncharged molecules (4, 18). Thus differences in the transport of electrically charged macromolecules have been thought to be due to the membrane-bound polyanionic glycoproteins that are rich in sialic acid and heparan sulfate residues, which set up a negative electrostatic field that repels polyanions (9, 16). These are associated with the glycoprotein coat that covers the endothelial fenestrations, the basement membrane, and the epithelial cells. Partial loss of these anionic sites can lead to albuminuria in the absence of any gross structural abnormalities and in cases of mild glomerulonephritis (11, 22). Such a loss has been induced experimentally by neutralization of the electrostatic barrier with the polycation protamine. In more severe glomerular injury-associated proteinuria, a larger fraction of the filtrate appears to pass through a population of large-diameter, nonselective pores (7, 8, 11, 12). Importantly, the authors point out that the Necturus kidney has a much lower glycosaminoglycan concentration in the GBM than exists in the rat. Nevertheless, the sieving coefficient for albumin was still very low. In addition to size and charge, molecular configuration influences the sieving coefficient (3, 19). Rigid or globular molecules such as horseradish peroxidase or Ficoll have lower glomerular sieving coefficients for any given molecular size than neutral dextran polymers with highly deformable linear structures (29). Because shape, flexibility, and deformability contribute to the quantitative relationship between molecular size and transglomerular solute flux, it has been challenging to establish the true dimensions of the extracellular channels. Data currently available indicate that the effective radius of the channels in the glomerular membrane is in the range of 4.5 to 6 nm (7, 8, 12, 14, 29). The data provided by Tanner et al. (27) provide fresh support to the conclusions based on older more conventional techniques. Perhaps the most challenging issue addressed by Tanner et al. regards the barrier function of the main components of the glomerular capillary wall. As mentioned, the greater width of the GBM and the capabilities of the new technology allowed Address for reprint requests and other correspondence: L. G. Navar, Dept. of Physiology, SL39, Tulane Univ. Health Sciences Center, 1430 Tulane Ave., New Orleans, LA 70112 (e-mail: [email protected]). Am J Physiol Renal Physiol 296: F1266–F1268, 2009; doi:10.1152/ajprenal.00152.2009.