British MBBCOI Bulled* (1992) \W 4S, No 4, pp86O-«76 O The Bntnh Council 1992
Bile secretion S Ei-linger Service d'Hepatologie and Vmti de Recherches de Physiopathologie Hipatique (INSERM U-24); Hdpital Btaujon, CUchy, France
Bile is an isotonic aqueous solution of bile acids, cholesterol, phospholipids, bile pigments and inorganic electrolytes. It is secreted by the hepatocytes into the bile canaliculi and modified in the bile ductules or ducts. The three main processes identified in the generation of bile flow are: 1. Active transport of bile acids from blood into bile canaliculi. This is responsible of the bile acid-dependent canalicular bile flow. Coupling between water flow and bile acid secretion is probably effected mainly through an osmotic mechanism. There is evidence that water flows (at least in part) through the interhepatocytic junctions. The bile acid-dependent flow accounts for 30 to 60% of spontaneous basal bile flow. 2. A canalicular, bile acid-independent secretion, probably due to transport into bile of organic solutes (glutathione) and inorganic electrolytes. This fraction of bile flow is stimulated by phenobarbital. It represents 30 to 60% of basal bile flow. Normal canalicular bile flow also depends on the integrity of intracellular cytoskeletal organelles, mostly microfilaments. 3. Reabsorption and secretion of fluid and inorganic electrolytes by the ductules and ducts. Secretion chiefly occurs in response to secretin and represents 30% of basal bile flow. Although several ion transport systems have been identified on the biliary epithelial cells (in particular a Na+/H+ exchange, a Na+:HCO3- symport and a Ch7HCO3- exchange), the cellular mechanism of secretion is not known. Abnormalities of bile duct function may account for the liver disease of cystic fibrosis, but these abnormalities have not been characterized.
The liver is involved in cystic fibrosis in 5—25% of patients. The prevalence of liver disease is even higher at autopsy.1-2 The most common reported abnormalities are 'focal' or biliary cirrhosis and strictures of intra and extrahepatic bile ducts,'>3-4 sometimes resembling primary sclerosing cholangitis (see Williams et al, this issue).5 The pathogenesis of these lesions is not known. Abnormalities of bile composition, viscosity or secretion have been suspected.2 The nature and site of these disturbances is not established. Here, I shall try to summarize the mechanisms of bile secretion, with special emphasis on secretion of electrolytes by the hepatocyte and biliary cells. It can be postulated that, in cystic fibrosis, a disturbance of electrolyte transport in the bile ductules or ducts occurs, similar to that observed in other involved epithelia. BILE COMPOSITION In general, inorganic electrolytes are present in common duct bile at concentrations closely reflecting those in plasma (Table 1). Bile concentrations of sodium, potassium, calcium and bicarbonate may, however, be appreciably higher than in plasma, while the chloride level may be lower. In spite of these variations, bile osmolality, as measured by freezing point depression, is usually approximately 300 mosmol/kg and it varies in parallel with plasma osmolality. The total osmotic activity is accounted for only by the inorganic electrolytes because it is generally assumed that bile acids, which are in micellar form, have little or no osmotic activity. The concentration of bicarbonate in bile is often higher than that in plasma. This may be due to bicarbonate transport mechanisms which have been postulated in the hepatocytes, and in the bile ductules and ducts, in response to secretin (see below). The major organic constituents of bile are the conjugated bile acids, the bile pigments, cholesterol and phospholipids. STRUCTURE-FUNCTION RELATIONSHIPS IN THE BILIARY SYSTEM Bile formed by the hepatocytes is secreted into the bile canaliculi. It is then modified during its passage in the bile ductules and ducts, and in the gallbladder, where water and inorganic electrolytes are reabsorbed, with, as a result, concentration of the organic constituents. The bile canaliculi are formed by a groove of plasma membrane formed between two hepatocytes and are about l^m in diameter. The
Table 1 Flow and electrolyte concentrations of hepatic bile Concentration (mmol/l)
Flow (jil min-'Jcg-1) Species
Man Dog Sheep Rabbit Rat Guinea pig
1 5-15.4 10 9.4 90 30-150 115.9
132-165 141-230 159.6 148-156 157-166 175
4 2-5 6 4.5-119 5.3 36-67 5.8-64 6.3
0.6-2.4 1.5-6 9 — 1.3-3 3 — —
0.7-1.5 1.1-2 7 — 0 15-0.35 — —
96-126 31-107 95 77-99 94-98 69
17-55 14-61 212 40-63 22-26 49-65
3-^5 16-187 42.5 6-24 8-25 —
Numbers indicate range or means of pubbsbed values
membrane forms numerous microvilli which increase the surface area. The bile canalicular membrane represents about 13% of the hepatocyte plasma membrane. The bile canaliculi connect to bile ducts, lined by biliary epithelial cells. As in other transporting epithelia, the canalicular lumen is sealed by intercellular junctions. The smallest bile duct, the ductule, connects the canaliculus with the portal (interlobular) bile ducts. The interlobular bile ducts drain into larger bile ducts which form the intra- and extrahepatic biliary tree (Fig. 1). With respect to bile secretion, the liver may be regarded as an epithelium transporting a variety of substrates from blood to bile. This vectonal transport is made possible by the high degree of polarization of the hepatocyte plasma membrane.
.. CANALICULAR FLOW '(•/
(Erythntol or mannitol clearance)
DUCTULE -DUCT TOTAL BILE FLOW
Fig. 1 Schematic view of the biliary system Bile formed by the hepatocyte is collected in the bile canaliculus Canalicular flow is estimated indirectly by erythntol or mannitol clearance Bile is then drained into bile ductules and ducts.
Polarization of the hepatocyte plasma membrane Three domains of the hepatocyte plasma membrane may be recognized: sinusoidal (facing the blood sinusoids), lateral (or intercellular) and canalicular. They demonstrate important morphological, biochemical and enzymatic differences. Especially important for transepithelial transport is the localization of the Na+, K+-activated adenosine
triphosphatase (Na+, K+-ATPase) which is mainly in the sinusoidal and intercellular membrane, with little or no activity in the canalicular membrane.6 Alkaline phosphatase, whose role in transport and in bile secretion is unknown, is (in contrast) preferentially located on the canalicular membrane. Tight junction and the paracellular pathway A substrate in plasma can enter canalicular bile in one of two ways: the transcellular pathway (entering the hepatocyte through the sinusoidal membrane, crossing the hepatocyte and entering the canaliculus through the canalicular membrane) or the paracellular pathway. In the latter case, the solute crosses the intercellular junction. The junction includes the tight junction, which is a sealing structure between the lumen of the bile canaliculus and the intercellular space of ducts and, hence, the sinusoidal blood. In the liver, there is evidence for a paracellular ion and fluid flux into bile which could play an important role in choleresis and which may possibly be reduced in cholestasis. The existence of a negative charge barrier in the tight junction has also been postulated. Hepatocyte cytoskeleton The canaliculus is surrounded by a narrow zone of organelle-poor cytoplasm, known as pericanalicular ectoplasm, where actin microfilaments (7nm in diameter) are concentrated. They form a pericanalicular network, attached to the intercellular junction and extend into the microvilli where they appear to insert on the inner part of the membrane. They may have a key role in maintaining the shape of the cell, particularly its microvilli. Agents that interfere with the structure and function of micronlaments affect bile flow which suggests a role for these organelles in secretion.7 Microtubules, which are 24 nm in diameter, are more randomly distributed within the liver cell cytoplasm than are microfilaments. They play a role in the intracellular transport and secretion by the liver cell of proteins and lipoproteins, and possibly in the transport of bile acids and other anions into bile. Antimicrotubular agents may affect bile formation (see below). Biliary epithelium The biliary tree begins with the bile ductule (or canal of Hering) lined with small cuboidal cells which adjoin the bile canaliculi. The bile
ductule is located at the periphery of the lobule. It drains into the portal (or interlobular) bile ducts, which are accompanied by branches of the portal vein and hepatic artery. The bile ducts join progressively to form ducts of increasing diameter, ultimately forming the extrahepatic and common bile ducts, draining into the duodenum. Intrahepatic bile ducts are lined by columnar, cylindrical biliary epithelial cells, 10-15nm in diameter. They represent only 3—5% of the overall population of liver cells,8 but provide a large surface area for transport and exchange between blood and bile. They have a prominent Golgi system, abundant vesicular bodies, little endoplasmic reticulum, and numerous microvilli on their luminal surface.9-'1 These morphological characteristics strongly suggest that they are not passive conduits but play a role in the modification of canalicular bile by secretory and/or absorptive processes. Recently, biliary epithelial cells have been isolated and cultured, from both normal rat liver and livers of rats with bile ductular hyperplasia induced by bile duct ligation or by chemicals, such as anaphthylisothiocyanate.12 They strongly express y-glutamyl transpeptidase. They also express alkaline phosphatase, glutathione-S-transferases and various cytokeratins. In culture, they form tight junctions, a well developed basement membrane and are positive for laminin and fibronectin. CANALICULAR BILE FLOW The maximal bile secretory pressure (about 25-30cm H2O) exceeds the sinusoidal perfusion pressure (about 5—10 cm H2O) which excludes hydrostatic filtration as an important mechanism of canalicular bile secretion. Canalicular bile flow is regarded mainly as an osmotic water flow in response to active solute transport. Bile acids, which are potent choleretics, are most probably one of these solutes but the transport of other organic solutes and probably inorganic ions also plays a role. Estimation of canalicular bile flow Canalicular bile flow may be estimated by measuring the biliary clearance of non-metabolized solutes that enter canalicular bile by passive processes and are neither secreted nor reabsorbed by the biliary epithelium. The most widely used of such solutes are erythritol and mannitol (labelled with 14Q. The original assumption that erythritol and mannitol do not cross the biliary epithelium was based on the observation that their clearance was not modified by secretin which acts presumably on the ductules or
ducts. However, small increases in erythritol and mannitol biliary clearance in response to secretin have been observed in dogs because either canalicular bile flow is stimulated by this hormone, or there is some permeability of the bile ductules or ducts to erythritol and mannitol. Bile acid dependent-flow Canalicular bile flow includes two major components, designated as bile acid-dependent and bile acid-independent flow. Bile acid-dependent flow is related to bile acid transport and secretion into the canalicular lumen. Bile acid transport by the hepatocyte involves three steps: uptake by the sinusoidal membrane, intracellular transport and canalicular secretion. Uptake of bile acids The uptake of conjugated bile acids by the hepatocyte is an active process utilizing the transmembrane sodium gradient as its source of energy. The transmembrane sodium gradient is continuously maintained by the Na+, K+ -ATPase of the plasma membrane (see above). The enzyme uses ATP as its energy source and exchanges three sodium ions for two potassium ions. Sodium is pumped out of the cell and the intracellular sodium concentration is maintained at a low level. The sodium concentration difference (high outside, low inside) can be used to drive substrates into the cell. This is the case for conjugated bile acids. The protein responsible for the transport has been identified by photoaffinity labelling techniques3 as a 48 kD protein. The system is called a symport, or cotransport. It has been reconstituted in artificial liposomes and in Xenopus oocytes.14 Intracellular transport Cytosolic protein binding After uptake by the sinusoidal membrane, bile acids are, in part, bound to cytosolic proteins. Glutathione S-transferases of the Ya family have affinity binding sites for monohydroxylated and dihydroxylated bile acids.15 Binders I and II, also designated as Y', which are proteins with a molecular weight of approximately 33 000, have high affinity binding sites for several bile acids with overlapping, but distinct specificities.15 They exhibit 3a-hydroxysteroid dehydrogenase activity and may reduce 3-oxo-bile acids into 3a-hydroxy bile acids. After injection of a labelled bile acid, 30-60% of the intracellular bile acid is found in the cytosolic
fraction. The exact role of protein binding in the transport of bile acids is not known. It is assumed that protein binding plays a role in facilitating bile acid entry into the cell and/or in reducing bile acid efflux out of die cell after entry. Evidence for vesicular transport Existing evidence supports the view diat intracellular organelles play a role in the vectorial transport of bile acids by the hepatocyte toward the canalicular membrane. High affinity binding sites have been identified in the Golgi apparatus: the binding affinity was similar to that of the plasma membrane. The existence of a vesicular transport for bile acids is supported by studies with microtubule inhibitors, morphological observations, and experiments using autoradiography and immunoperoxidase techniques. Treatment of rats with microtubule inhibitors, such as colchicine and vinblastine, inhibits bile acid secretion. Studies from our laboratory and by others have indicated that these agents do not inhibit basal bile acid secretion or bile acid secretion after a tracer dose of a labelled bile salt, but strongly inhibit secretion after injection of a substantial load of bile acid.16 More direct evidence for such vesicular transport is derived from morphological observations. After a 48-h taurocholate loading, significant increases in the amounts of Golgi-rich area, Golgi membranes and pericanalicular vesicles have been observed. Autoradiographic studies using two labelled bile acid analogues (cholyglycylhistamine and cholylglycyltyrosine) have shown preferential distribution of grains over the Golgi complex and the endoplasmic reticulum. Immunoperoxidase experiments with specific anti-bile acid polyclonal antibodies have shown preferential staining of the Golgi apparatus, and, to a lesser extent, of the endoplasmic reticulum.17 These experiments also suggest that, at some stage of their intracellular transport, bile acids interact with these organelles. It is noteworthy that a taurocholate transport system has been identified in isolated purified Golgi fractions from rat liver. This system appears to be sodium-independent and, therefore, distinct from the sinusoidal transporter. Collectively, these observations support the view that after uptake by die sinusoidal membrane and cytosolic protein binding, bile acids may be transported by the hepatocyte by two distinct padiways (Fig. 2). One is a non-vesicular pathway by simple diffusion to the canalicular carrier and secretion into bile by carrier-mediated transport, facilitated by die membrane potential. The other is a vesicular pathway involving
the Golgi apparatus: bile acids are transported by a specific earner of the Golgi membrane into the Golgi lumen, and then into Golgi-derived vesicles. These vesicles move toward the canaliculus by a microtubuledependent process and could be secreted into bile by exocytosis. The latter pathway becomes increasingly important when the transcellular flux of bile acid increases.
Blood (space of Disse) Liver cell Canaliculus Tight junction (paracellular pathway)
Fig. 2 Pathways of bile acid transport by the hepatocyte After uptake (1), bile acids may diffuse toward the canalicular membrane and be secreted by one or several canalicular carriers (2), or, at higher loads, be transported by a vesicular pathway involving the Golgi apparatus (3, 4 5)
Canalicular secretion From the preceding discussion, two mechanisms can be proposed for canalicular secretion. One is via the canalicular carrier protein. This protein has been identified by photoaffinity labelling and reconstituted in artificial liposomes.18 It is a 100 kD transmembrane protein. This protein does not use energy but utilizes the membrane potential (-35 mV inside the cell, -5 mV in the canalicular lumen) to drive bile acids out of the cell into the canalicular lumen: it is a typical example of facilitated diffusion. Recently, an ATP-dependent carrier has been identified on canalicular membrane vesicles.19 It is not clear whether this carrier is identical to or different from the 100 kD protein. The second mechanism could be exocytosis of Golgi-derived vesicles after membrane fusion. Direct proof of such a mechanism, as indicated above, is currently lacking.
Effect of bile acid secretion on canalicular flow An apparently linear relationship between bile acid secretion rate and bile flow has been demonstrated in many animal species, including man (Fig. 3). Bile acid induced choleresis is presumably of canalicular origin, since bile acids are secreted into the canaliculi, and since there is a linear relationship between erythritol clearance and bile acid secretion (Fig. 3). The slope of the regression line between bile flow and bile acid secretion (usually 7-12 uJ/u.mol) is generally termed apparent choleretic activity.
Ductal/ductular flow Erythritol clearance Slope (ul/umol) -Kbile acid-dependent flow) Estimation of canalicular bile acidindependent flow
Bile acid output umol.min"1 100g -1
Fig. 3 Relationship between total bile flow, canalicular bile flow (erythritol clearance) and bile acid output in bile. The slope of the regression line between flow and bile acid output is the apparent choleretic activity of the bile acids.
The hypothesis that bile acids increase bile flow by providing an osmotic driving force for water and electrolytes was proposed by Sperber.20 However, because bile acids are in micellar form in bile, most osmotic activity must be accounted for by their counter-ions (or cations accompanying the bile acid anions to maintain electroneutrality). Hypercholeretic bile acids and the chole-hepatic shunt hypothesis In 1980, we demonstrated that ursodeoxycholic acid (3a, 7p"-dihydroxycholanoic acid, the 7|J epimer of chenodeoxycholic acid), a bile acid which is widely used in the medical dissolution of cholesterol gallstones, had a much more pronounced choleretic effect that
taurocholate21 or even than its own conjugate, tauroursodeoxycholate (Fig. 4). Subsequently, several other bile acids were shown to have a similar 'hypercholeretic effect': for example, 23-norursodeoxycholate or 23-nor-chenodeoxycholate. In all cases, hypercholeresis is associated with a marked stimulation of bicarbonate concentration and secretion into bile.
Bile Flow pi mm-' 100g-'
•--r—'*-' • URSODEOXYCHOLATE — TAUROURSODEOXYCHOLATE
200 400 600 800 Bile Acid Output nmol. mirrMOOg- 1
Fig. 4 Effect of ursodeoxycholate on bile flow Bile flow induced by ursodeoxycholate is much higher than that induced by tauroursodeoxycholate. This is designated as hypercholeresis
The mechanism of this curious phenomenon is not yet established. At present, the best explanation is probably the so-called chole-hepatic shunt hypothesis, proposed by Hofmann and colleagues.22 According to this hypothesis, ursodeoxycholate is transported into bile in part in the unconjugated form. It becomes protonated in the biliary lumen into ursodeoxycholic acid. The proton comes from H2CO3 and the process generates one bicarbonate which is secreted into bile. The protonated ursodeoxycholic acid is lipid soluble and readily absorbed by the biliary epithelial cells into the peribih'ary vascular plexus and returns to the hepatocyte which re-secretes it into bile. The bile acid does not appear in final bile, but, at each cycle, one bicarbonate is secreted and stimulates bile flow (Fig. 5). Other explanations have been proposed, involving hepatocytic mechanisms, including stimulation of the
Na + /H + exchange in the plasma membrane. This hypothesis rests on the observation that amiloride, an inhibitor of the Na+/H+ exchanger, inhibits urdoseoxycholate-induced choleresis. Direct proof, however, is not available.
Fig. 5 The chole-hepatic shunt hypothesis See text for explanation.
Bile acid-independent flow When the linear relationship between bile flow and bile acid secretion is extrapolated for a zero bile acid secretion, a positive intercept is obtained (Fig. 3). This bile flow is generally designated as canalicular bile acid-independent flow. Although mis procedure of estimation is subject to criticism, the value obtained is similar to that measured directly in the isolated perfused liver in the absence of bile acids in die system. Bile acid-independent flow differs markedly between animal species. It is increased by phenobarbital, other inducers of the microsomal drugmetabolizing system, and thyroid hormones. It is inhibited by chronic estrogen administration. The mechanism is not fully established. Organic ion transport has been postulated. Glutathione has been implicated, because there is a good correlation between canalicular bile salt independent flow and glutathione secretion.23 Inorganic ion transport is also possible. Several ion pumps have been identified on the hepatocyte plasma membrane. In addition to die Na+, K+-ATPase discussed earlier, a Na + /H + exchanger is present on the basolateral plasma membrane.24 It plays a role in intracellular pH
regulation. An anion transporter and a Na+/HCO-f symport have also been characterized on the basolateral membrane.24 A bicarbonate/chloride exchanger (antiporter) has been characterized on the canalicular membrane.24 These ion pumps are represented schematically in Figure 6. Their role in bile flow has not been conclusively established.
Blood (space of Disse) Hepatocyte Canaliculus Tight junctions (paracellular pathway)
Fig. 6 Major transport systems and pumps for inorganic ions on the hepatocyte. On the basolateral membrane: the Na+, K+-ATPase, a Na+/H+ antiport and a Na+ HCO3- symport. On the canalicular membrane: a Q-/HCO3- antiport and probably a sulfate/hydroxyl ant]port In addition, there are Cr~ and K+ conductance pathways
ROLE OF DUCTULES AND DUCTS Secretion (ductular/ductal bile acid-independent flow) Secretion occurs in the ductules and ducts in many species including man, mostly in response to secretin administration.25 Secretin choleresis is generally accompanied by changes in bile composition, chiefly a rise in bicarbonate and pH, and a fall in bile acids. The inlraduodenal infusion of HC1 in dogs induces the same response as endogenous secretin. The evidence for a ductular/ductul site of action of secretin is: (1) secretin choleresis does not enhance the maximal biliary secretory capacity of BSP whereas bile acids, which act on the bile canaliculi, do; (2) the biliary 'wash-out' volume during constant rate BSP infusions is less with secretin choleresis than with bile acid choleresis, suggesting that secretin acts distal to the canaliculi; (3) biliary clearances of ery-
thritol and mannitol are increased during bile acid choleresis and not (or minimally) during secretin choleresis. Secretion in the bile ductules or ducts is inhibited by somatostatin and prostaglandins. There is evidence of a marked secretory activity by hyperplastic ductular cells after bile duct ligation of a-naphtylisothiocyanate treatment.26 Choleresis is proportional to the extent of bile ductular proliferation27, stimulated by secretin and rich in bicarbonate. The mechanism of this choleresis is not known. In the pig, secretin choleresis was not inhibited by amiloride.28 Dicyclohexylcarbodiimide (DCCD), a proton-ATPase inhibitor, inhibited bile flow and HCO 3 " secretion, without affecting bile salt secretion.2^ These findings suggest that secretin-stimulated bile flow may involve a proton pump. Morphological studies on bile duct cells showed a disappearance of 'tubulovesicles' from these cells after secretin administration.30 This is similar to observations on pancreatic duct cells and is consistent with discharge of cytoplasmic vesicles containing proton pumps at the basolateral membrane. Apical secretion of HCO3- would be via a CWHCO3- exchanger after opening of apical Cl~ channels.31 Direct proof of the existence of such transport systems should be obtained on isolated biliary epithelial cells. Studies on isolated rat bile duct epithelial cells have shown results consistent with Na+/H+ exchange, Na+: HCO3~ symport and Q-/HCO3- exchange.32 These transport systems play a role in intracellular pH regulation, but the results suggest that bile duct epithelial cells may be capable of transepithelial U+/HCOy- transport. The secretion of biliary cells is inhibited in vivo by somatostatin and prostaglandins. No data exist on Cl~ channels or transporters on these cells. Results from in vivo studies have suggested that certain macromolecules (eg ceruloplasmin and carcinoembryonic antigen) may be secreted into bile by bile duct epithelial cells. Isolated bile duct epithelial cells participate in receptor-mediated endocytosis of epidermal growth factor.33 Therefore, in addition to ion transport, bile duct epithelial cells could also play a role in the transport of macromolecules. The secretory activity of the bile ductules and ducts explains the choleresis mat occurs in certain diseases. Elevated bile flow has been recorded in patients with cirrhosis, other chronic liver diseases associated with ductular proliferation, and in congenital dilatation of the intrahepatic biliary tree (Caroli's syndrome). An augmentated surface of the biliary epithelium is common to these conditions.
Reabsorption The bile ductules or ducts may also have a reabsorptive function. Thus in cholecystectomized dogs, the composition of bile stored in the common bile duct was similar to typical gallbladder bile. Bile-to-plasma concentration ratios above unity in the steady state have been found for mannitol and erythritol in various species which suggest there is water reabsorption distal to the canaliculi because neither solute is thought to be transported by concentrative processes. No studies in man are available. As indicated above, prostaglandins appear to enhance reabsorption by the bile duct epithelium. The main site of reabsorption in the biliary tree is the gallbladder. Reabsorption by the gallbladder epithelium involves transepithelial NaCl and fluid transport. A detailed discussion of the processes involved is beyond the scope of this review and may be found elsewhere.34 BILE SECRETION IN CYSTIC FIBROSIS Little is known on bile secretion in cystic fibrosis. An electron microscopy study of the liver of children with cystic fibrosis without evidence of liver disease did not show signs of cholestasis: the bile canaliculi were not dilated and did not contain bile plugs. There was no bile pigment in hepatocytes. There was some hypertrophy of the smooth endoplasmic reticulum and Golgi apparatus, and large lysosomes containing lipofuscin were seen.35 The bile ducts were not dilated. These ultrastructural findings do not provide a basis for liver cell damage. There is no similar study of the liver in patients with liver disease. Radiolucent gallstones are common in young adults with cystic fibrosis. Although it was originally thought that these stones were made predominantly of cholesterol, recent studies have challenged this view and shown that stones contained large amounts of pigment material (calcium bilirubinate) and protein.36 Ursodeoxycholic acid has been shown to improve liver function tests and the nutritional state in patients with cystic fibrosis and liver disease.37 The bile acid pool of these patients became enriched with ursodeoxycholic acid.38 Whether the improvement is due to an increase in bile secretion or to the modification of the bile acid pool is not known (see Williams et al, this issue). REFERENCES 1 Nagel RA, Westaby D, Javaid A et al Liver disease and bile duct abnormalities in adults with cystic fibrosis. Lancet 1989, 2 1422-1425
2 Isenberg JI. Cystic fibrosis its influence on the liver, biliary tree, and bile salt metabolism. Semin Liver Dis 1982, 4 302-313 3 Bass S, Connon JJ, Ho CS Biliary tree in cystic fibrosis. Biliary tract abnormalities in cystic fibrosis demonstrated by endoscopic retograde cholangiography. Gastroenterology 1983; 84: 1592-1596 4 Gaskin KJ, Waters DLM, Howman-Giles R et al Liver disease and common-bile-duct stenosis in cystic fibrosis. N Engl J Med 1988, 318: 340-346 5 Strandvik B, Huelte L, Gabnelsson N, Glaumann H. Sclerosing cholangitis in cystic fibrosis Scand J Gastroenterol 1988, 23 (Supp 143): 121-124 6 Bhtzer BL, Boyer JL Cytochemical localization of Na+, K+-ATPase in the rat hepatocyte. J Chn Invest 1978; 62: 1104-1108 7 Erhnger S. Role of intracellular organelles in the hepatic transport of bile acids. Biomed Pharmacother 1990; 44: 409-416 8 Tavoloni N. The intrahepatic biliary epithelium an area of growing interest in hepatology. Semin Liver Dis 1987, 7 280-292 9 Sterner JW, Carruthers JS. Studies on the fine structure of the terminal branches of the biliary tree. Am J Pathol 1961; 38 639-661 10 Schaffner F, Popper H Electron microscopic studies of normal and proliferated bile ductules. Am J Pathol 1961, 38: 393-410 11 Stemlieb I, Quintana N. Biliary proteins and ductular ultrastructure Hepatology 1985, 5 139-143 12 Sinca AE, Mathis GA, Sano N, Elmore LW Isolation, culture, and transplantation of intrahepatic biliary epithelial cells and oval cells Pathobiology 1990; 58 44—64 13 Kramer W, Bickel U, Buscher HP, Gerok W, Kurz G Bile-salt-binding polypepndes in plasma membranes of hepatocytes revealed by photoaffinity labelling. Eur J Biochem 1982, 129: 13-24 14 Hagenbuch B, LUbbert H, Stieger B, Meier PJ Expression of the hepatocyte Na+/bile acid cotransporter in Xenopus laevis oocytes J Biol Chem 1990, 265: 5357-5360 15 Stolz A, Taiikawa H, Ookhtens M, Kaplowitz N The role of cytoplasmic proteins in hepatic bile acid transport Annu Rev Physiol 1989; 51 161-176 16 Crawford JM, Berken CA, Gollan JL. Role of the hepatocyte microtubular system in the excretion of bile salts and biliary lipid: implications for intracellular vesicular transport. J Lipid Res 1988, 29: 144-156 17 Lamn Y, Roda A, Dumont M, Feldmann G, Erlinger S. Immunoperoxidase locahzation of bile salts in rat liver cells Evidence for a role of the Golgi apparatus in bile salt transport J Clin Invest 1988, 82 1173-1182 18 Ruetz S, Hugentobler G, Meier PJ Functional reconstitution of the canalicular bile salt transport system of rat liver. Proc Nat] Acad Sci USA 1988, 85: 6147-6151 19 MQUer M, Ishikawa T, Berger U et al ATP-dependent transport of taurocholate across the hepatocyte canalicular membrane mediated by a 110-kDa glycoprotein binding ATP and bile salt J Biol Chem 1991; 266 18920-18926 20 Sperber I Secretion of organic anions in the formation of unne and bile. Pharmacol Rev 1959, 11 109-134 21 Dumont M, Uchman S, Erlinger S Hypercholeresis induced by ursodeoxycholic acid and 7-ketolithochohc acid in the rat Possible role of bicarbonate transport Gastroenterology 1980; 79 82-89 22 Yoon Y, Hagey LR, Hofmann AF, Gurantz D, Michelotti EL, Steinbach JH. Effect of side-chain shortening on the physiologic properties of bile acids, hepatic transport and effect on biliary secretion of 23-nor-ursodeoxycholate in rodents. Gastroenterology 1986, 90: 837-852 23 Ballatori N, Truong T. Relation between biliary glutathione excretion and bile acidlndependent flow. Am J Physiol 1989, 256 G22-G30 24 Nathanson MH, Boyer JL. Mechanisms and regulation of bile secretion. Hepatology 1991; 14: 551-566 25 Preisig R, Cooper HL, Wheeler HO. The relationship between taurocholate secretion rate and bile production in the unanesthetized dog during cholinergic blockade and during secretin administration J Clin Invest 1962, 41: 1152-1162
26 Alpini G, Lenzi R, Sarkozi L, Tavoloni N. Biliary physiology in rats with bile ductular cell hyperplasia. Evidence for a secretory function of proliferated bile ductules. J Chn Invest 1988; 81 569-578 27 Alpini G, Lenzi R, Zhai WR et al Bile secretory function of intrahepatic biliary epithelium in the raL Am J Physiol 1989, 257 G124-G133 28 Grotmol T, Buanes T, Raeder MG The effect of amilonde on biliary HCO3- secretion in the anaesthetized pig Acta Physiol Scand 1987 130. 447^*55 29 Grotmol T, Buanes T, Raeder MG DCCD (N, N'-Dicyclohexylcarbodiimide) inhibits biliary secretion of HCO3-. Scand J Gastroenterol 1987; 22: 207-213 30 Buanes T, Grotmol T, Landsverk T, Raeder M. Secrenn empties bile duct cell cytoplasm of vesicles when it inmates ductular HCO3 secretion in the pig. Gastroenterology 1988, 95- 417^t24 31 Grotmol T, Buanes T, Veel T, Raeder M Secretin-dependent HCO3- secretion from pancreas and liver. J Intern Med 1990, 228, Suppl 1. 47-51 32. Strazzabosco M, Mennone A, Boyer JL Intracellular pH regulation in isolated rat bile duct epithelial cells. J Clin Invest 1991, 87- 1503-1512 33 Ishii M, Vroman B, LaRusso NF Morphologic demonstration of receptor-mediated endocytosis of epidermal growth factor by isolated bile duct epithelial cells Gastroenterology 1990; 98- 1284-1291 34 Wood JR, Svanvik J Gallbladder water and electrolyte transport and its regulation. Gut 1983; 24. 579-593 35 Hultcrantz R, Mengarelli S, Strandvik B. Morphological findings in the liver of children with cystic fibrosis a light and electron microscopical study. Hepatology 1986; 6: 881-889 36 Angehco M, Gandin C, Canuzzi P et al Gallstones in cystic fibrosis' a critical reappraisal. Hepatology 1991; 14: 768-775 37 Cottmg J, Lentze MJ, Reichen J. Effects of ursodeoxycholic acid treatment on nutrition and liver function in patients with cystic fibrosis and longstanding cholestasis. Gut 1990; 31: 918-921 38 Nakagawa M, Colombo C, Setchell KDR Comprehensive study of the biliary bile acid composition of patients with cystic fibrosis and associated liver disease before and after UDCA administration. Hepatology 1990, 12: 322-334