Properties of the endothelial cells and the glomerular barrier

Sammanfattning: The kidneys are vital organs that filter plasma to produce 180 liters of primary urine per day. Only 1% of this volume is excreted and the rest is reabsorbed. Naturally, such huge amounts of fluid must be well controlled to maintain the homeostasis, which is a prerequisite for life. Moreover, final urine is practically devoid of proteins, glucose and other important solutes. In fact, proteinuria is a hallmark of kidney disease. The classical view of the glomerular capillary is that the basement membrane and/or the podocytes constitute the actual barrier. In this thesis I have focused on a neglected component of the glomerular barrier namely the fenestrated endothelium. Functional studies of glomerular charge- and size-selectivity were conducted on isolated perfused rat kidneys perfused with albumin solutions at 8°C to inhibit tubular modification of urine. The fractional clearance, q, of 5 negatively charged proteins were compared with neutral solutes of similar size. The data strongly support the presence of a charge barrier amounting to 30-40 mEq/L of fixed charges. The magnitude of the charge barrier was then altered by using different ionic compositions of the perfusate. Even though this is a novel approach in this field, the principle is standard practice during ion-exchange chromatography. As expected the q for anionic proteins decreased when low ionic strength perfusates were used. However, q fell less than expected indicating that low ionic strength reversibly reduces charge density with no concomitant effects on size-selectivity. We hypothesize that the endothelial cell coat is the structure responsible for these effects. Studies on cultured endothelial cells revealed that these cells produce orosomucoid, an acute phase protein known to be important for the maintenance of capillary permselectivity. Using microphysiometry we could also demonstrate that orosomucoid has a direct cAMP-dependent effect on endothelial cells. Furthermore, we found that orosomucoid has an anti-inflammatory effect on endothelial cells. Finally, we studied the rate of proteoglycan synthesis in bovine glomerular endothelial cells in culture. The rate of synthesis was reduced by puromycin aminonucleoside (PAN), a substance known to induce nephrosis and increased by the pro-inflammatory cytokine IL-1ß. These studies strongly support the classical notion of a significant charge barrier and refute recent data suggesting a less selective glomerular barrier. Secondly, the data suggest different structures to be responsible for size- and charge-selectivity. Thirdly, the charge barrier seems to be a gel-like structure with a variable volume. Fourthly, the endothelial cell coat is a likely candidate for such a charged gel. Fifthly, the endothelial cells are affected by cytokines and certain drugs and produce e.g. orosomucoid and proteoglycans. These findings may be of clinical relevance for a number of edematous conditions. In fact, nephrotic syndromes may be due to endothelial cell dysfunction rather than being kidney-specific disorders.