LAUSR

To the editor, I read with interest the review published by Vartak et al.[1] that questions the site of water transport associated with the bile excretion of choleretic bile acids and organic anions. The authors deny a water efflux associated with bile salt export pump (BSEP)and multidrug resistance– associated protein 2 (MRP2)dependent transport at the canaliculi membrane because no bile flow was observed with intravital microscopy. In contrast, bile flow is well evidenced in ductules and ducts where water efflux into ductular lumen might originate from cholangiocytes in association with HCO− 3 excretion. It is understandable that the primary bile is stagnant in canaliculi constrained by surrounding liver tissue. Progression of this stagnant flow is improved by closely related actin and myosin that contract canaliculi. It is possible that water transport across the canalicular membrane might remain unnoticed within such constrained canaliculi. Water transport would increase canaliculi volume and decrease concentrations of choleretic substrates. Besides the paracellular transport of water through claudins in tight junctions, a transmembrane pathway exists through aquaporin 8 (AQP8). BSEP, MRP2, and AQP8 are located in cholesterolenriched membrane microdomains that favor their functional interactions. In normal rat livers, AQP8 located in intracellular and subapical vesicles migrate to the canalicular membrane when concentrations of choleretic substrates increase in hepatocytes. To compensate for cellular water loss, AQP9 must transport water when these substrates cross the basolateral membrane. In cholestatic rat livers, the downregulation of BSEP and MRP2 is associated with a reduced expression of AQP8.[2] Delivery of an adenovector encoding for human AQP1 redistributes BSEP and MRP2 in the canalicular cholesterolrich microdomains and improves bile secretion. AQP1 is expressed in cholangiocytes, and water excretion can be triggered by secretin acting at the basolateral membrane. However, the choleretic effect of the organic anion gadobenate dimeglumine (BOPTA) is observed in perfused rat livers when the hepatic artery is not perfused.[3] Secretin can only reach cholangiocytes by capillaries originating from the hepatic artery. Interestingly, numerous cilia extend from the apical membrane of cholangiocytes and are ideally positioned to detect changes in bile flow, bile composition, and bile osmolality.[4] These cilia might play a role in the regulation of ductal bile formation by acting as osmosensors. Whether cilia detect bile concentrations to trigger water efflux from cholangiocytes is unknown. It is clear that the choleresis associated with BOPTA transport into bile canaliculi through MRP2 is correlated to hepatocyte concentrations.[3] Otherwise, the signal and mechanism that efflux water from cholangiocytes in the absence of capillary perfusion should be defined.