- Email: [email protected]
Claudin 2 Deficiency Reduces Bile Flow and Increases Susceptibility to Cholesterol Gallstone Disease in Mice
Gastroenterology 2014;147:1134–1145
BASIC AND TRANSLATIONAL—BILIARY Claudin 2 Deficiency Reduces Bile Flow and Increases Susceptibility to Cholesterol Gallstone Disease in Mice Kengo Matsumoto,1,2 Mitsunobu Imasato,1 Yuji Yamazaki,1 Hiroo Tanaka,1 Mitsuhiro Watanabe,3,4 Hidetoshi Eguchi,5 Hiroaki Nagano,5 Hayato Hikita,2 Tomohide Tatsumi,2 Tetsuo Takehara,2 Atsushi Tamura,1 and Sachiko Tsukita1 1
Laboratory of Biological Science, Graduate School of Frontier Biosciences and Graduate School of Medicine, Osaka University, Osaka, Japan; 2Department of Gastroenterology and Hepatology, Graduate School of Medicine, Osaka University, Osaka, Japan; 3Department of Internal Medicine, School of Medicine, Keio University, Tokyo, Japan; 4Graduate School of Media and Governance, Faculty of Environment and Information Studies, Keio University, Kanagawa, Japan; and 5 Department of Surgery, Graduate School of Medicine, Osaka University, Osaka, Japan
See editorial on page 965.
BASIC AND TRANSLATIONAL BILIARY
BACKGROUND & AIMS: Bile formation and secretion are essential functions of the hepatobiliary system. Bile flow is generated by transepithelial transport of water and ionic/ nonionic solutes via transcellular and paracellular pathways that is mainly driven by osmotic pressure. We examined the role of tight junction–based paracellular transport in bile secretion. Claudins are cell–cell adhesion molecules in tight junctions that create the paracellular barrier. The claudin family has 27 reported members, some of which have paracellular ion- and/or water-channel–like functions. Claudin 2 is a paracellular channel-forming protein that is highly expressed in hepatocytes and cholangiocytes; we examined the hepatobiliary system of claudin 2 knockout (Cldn2/) mice. METHODS: We collected liver and biliary tissues from Cldn2/ and Cldn2þ/þ mice and performed histologic, biochemical, and electrophysiologic analyses. We measured osmotic movement of water and/ or ions in Cldn2/ and Cldn2þ/þ hepatocytes and bile ducts. Mice were placed on lithogenic diets for 4 weeks and development of gallstone disease was assessed. RESULTS: The rate of bile flow in Cldn2/ mice was half that of Cldn2þ/þ mice, resulting in significantly more concentrated bile in livers of Cldn2/ mice. Consistent with these findings, osmotic gradient-driven water flow was significantly reduced in hepatocyte bile canaliculi and bile ducts isolated from Cldn2/ mice, compared with Cldn2þ/þ mice. After 4 weeks on lithogenic diets, all Cldn2/ mice developed macroscopically visible gallstones; the main component of the gallstones was cholesterol (>98%). In contrast, none of the Cldn2þ/þ mice placed on lithogenic diets developed gallstones. CONCLUSIONS: Based on studies of Cldn2/ mice, claudin 2 regulates paracellular ion and water flow required for proper regulation of bile composition and flow. Dysregulation of this process increases susceptibility to cholesterol gallstone disease in mice.
gallstone disease.6–8 Cholesterol, phospholipids, and bile acids are the major components of bile. They are originally hydrophobic and their conjugated hydrophilic forms exist in a micellous state,2 which can aggregate and form stones in the aqueous bile in any condition.6,7 Therefore, the bile formation and bile flow are critical for hepatobiliary function. The incidence of gallstones is high in the Western world,7,8 and the pathogenesis of cholesterol gallstone disease is relatively well studied. The cholesterol saturation index (CSI), the ratio of the cholesterol concentration to that of bile acids and phospholipids, is a well-known indicator for cholesterol crystal nucleation. On the other hand, the water content of bile and bile flow are key factors in gallstone disease that have not been fully examined.8,9 In the biliary system, hepatic bile, which is 98% water, is secreted at a rate of 30–40 mL/h in humans. Bile formation is preceded by the active secretion of solutes from hepatocytes, such as bile acids by bile acid–dependent secretion, and glutathione, bilirubin, and/or HCO3 by bile acid– independent secretion. Water and electrolytes are then secreted.2,10 Water molecules transverse the epithelium through paracellular and/or transcellular routes. The identification and analysis of aquaporins (AQPs) has extended our understanding of how the transcellular water movement through epithelial cell sheets is regulated.11,12 On the other hand, it is still unsettled how and to what extent water moves paracellularly between the cells in epithelial sheets, mainly because it is difficult to separate the transcellular and paracellular pathways experimentally. The tight junction (TJ) is an adhesion apparatus that resides in the most apical part of the lateral membrane. TJs surround epithelial cells and attach them together. The TJ is also fundamentally responsible for the paracellular barrier of epithelia, including their charge- and size-dependent permeabilities; these properties enable the selective movement of solutes and water for biologic functions.13–15 On the
Keywords: Mouse Model; Claudin 2; Hepatic Microcirculation; TJ.
T
he hepatobiliary system plays critical roles in lipid and cholesterol homeostasis.1–4 Dysregulation of this system causes various diseases; impaired bile secretion leads to cholestasis5 and an imbalance in bile composition leads to
Abbreviations used in this paper: AQP, aquaporin; BW, body weight; CSI, cholesterol saturation index; IBDU, intrahepatic bile duct unit; TJ, tight junction. © 2014 by the AGA Institute 0016-5085/$36.00 http://dx.doi.org/10.1053/j.gastro.2014.07.033
November 2014
Materials and Methods Because of limitations in article length, the detailed Materials and Methods are presented in the Supplementary Material.
Results Claudin Expression Patterns in Liver and Gallbladder To determine which claudin subtypes are expressed in the TJs of the hepatobiliary system, we first examined the claudin gene expressions in liver and gallbladder by quantitative reverse transcription polymerase chain reaction. Claudins 1, 2, 3, 12, and 25 were the major claudins in the liver (Figure 1). Among these, claudin 2 is a channel-forming type, and was recently suggested to be a water-permeable claudin.19,27 Therefore, to elucidate the roles of TJs in regulating bile formation and secretion in the hepatobiliary tract, we analyzed Cldn2/ mice. We next examined the localization of claudin subtypes in the Cldn2þ/þ liver by immunofluorescence analyses. Claudin 2 was expressed in the perivenous regions, bile ducts, and gallbladder epithelium. In contrast, the claudin 1 signals were strong in the periportal zone and slightly weaker in the perivenous zone. Claudin 3 showed clear signals in both the perivenous and periportal zones. In the Cldn2þ/þ bile duct and gallbladder epithelium, claudins 2, 3, and 7 were strongly expressed (Supplementary Figures 1 and 2). In the Cldn2/ mice, the claudin 2 signal was not detected, but the other claudin expression patterns and levels were not significantly different from the Cldn2þ/þ mice in the liver, bile duct, or gallbladder epithelium. There were no differences in the expression patterns of occludin, ZO-1, or E-cadherin between the Cldn2þ/þ and Cldn2/ liver (Supplementary Figure 3).
Histologic Analysis of the Cldn2/ Liver and Gallbladder We next examined the liver and gallbladder in 8-weekold Cldn2þ/þ and Cldn2/ mice by H&E staining and freeze fracture electron microscopy and found no obvious differences between the Cldn2þ/þ and Cldn2/ samples (Figure 2A, B, and C). It was previously reported that claudin 2 knockdown prevents canalicular formation in WIF-B9 cells28 and that b-catenin knockout mice show decreased claudin 2 expression and wider bile canaliculi.29 We therefore examined the bile canaliculi more closely by scanning electron microscopy and immunofluorescence analyses. We found no obvious differences in the bile canaliculi width or shape between the Cldn2þ/þ and Cldn2/ liver, although the bile canaliculi is slightly wider in periportal zone than in perivenous zone both in Cldn2þ/þ and Cldn2/ bile canaliculi (perivenous zone: Cldn2þ/þ 0.79 ± 0.03 mm, Cldn2/ 0.78 ± 0.04 mm; periportal zone: Cldn2þ/þ 0.92 ± 0.03 mm, Cldn2/ 0.90 ± 0.02 mm) (Supplementary Figure 4A and B and Supplementary Table 1). Together, these data suggested that histologically, almost no changes were detected in the Cldn2/ liver and gallbladder.
BASIC AND TRANSLATIONAL BILIARY
freeze-fracture electron microscopy, TJ strands appear as a continuous meshwork in the apical intercellular space. These strands are thought to be derived from polymerized claudins forming the paracellular barrier, although the polymerization mechanism is unknown. The claudins are a large family of tetraspanning transmembrane proteins, with at least 27 members in human/mouse, which are thought to have various functions in claudin-based paracellular systems. Evidence suggests that claudins generally function to create the paracellular barrier function of epithelial cell sheets. However, some of them increase the paracellular ion permeability when expressed in paracellular barrier– established cell sheets; these are described as “ion-channel forming” claudins,14,16,17 although they also elicit a relatively nonspecific permeability for small nonionic solutes. Previous studies suggested that the combination of claudin species expressed in cells is important for determining the permselectivity of the TJs.14,15 Consistent with this idea, physiologic studies suggest that epithelium types can be classified as leaky or tight, based on transepithelial conductance and/or paracellular solute flux measurements.17,18 Claudin 2 was the first channel-forming claudin to be identified,13,19 with suggested roles in water and cation permselectivity.20–22 In claudin 2 knockout (Cldn2/) mice, deficiencies in paracellular permeability and cation selectiveness impair the reabsorption of Naþ, Cl, and water in renal proximal tubules.20 In the small intestine, claudin 2 loss significantly decreases the Naþ permeability.23 In addition, double-knockout mice in the intestine of claudin 2 and claudin 15, another channel-forming claudin, are deficient in sodium-driven nutrient absorption due to the insufficiency of Naþ in the intestinal lumen, which is usually supplied paracellularly from the submucosal space, which leads to infant death.24,25 Thus, the ion channel–forming claudins appear to be important for regulating biologic functions, in addition to their paracellular barrier function. In the liver, claudin 1, 2, and 3 are dominant. Among them, barrier-forming claudin 1 dysfunction is reported to lead to a rare autosomal recessive neonatal ichthyosissclerosing cholangitis syndrome, which is characterized by scalp hypotrichosis, scarring alopecia, ichthyosis, and sclerosing cholangitis, as described previously.26 In cholangitis in neonatal ichthyosis-sclerosing cholangitis syndrome liver, a bile leakage through claudin 1deficient TJs of hepatocellular and biliary cells is suggested to result in direct hepatocellular and/or biliary injuries and in cholestasis. Here we focused on the hepatobiliary system in Cldn2/ mice. The concentrations of bile acids, phospholipids, and cholesterol were significantly increased in the Cldn2/ bile. We found that Cldn2/ mice showed decreased transepithelial electrical conductance and water permeability in the hepatobiliary system, resulting in decreased bile flow. In addition, after 4 weeks on a lithogenic diet, the prevalence of cholesterol gallstones was 100% in the Cldn2/ mice vs 0% in Cldn2þ/þ mice. Our findings indicate that the channel-forming claudin 2 is critical for bile homeostasis in the hepatobiliary system. We also provide the first evidence that this claudin plays a key role in bile flow as a type of water flow in the body.
Role of Claudin 2 in Bile Flow and Gallstones 1135
1136 Matsumoto et al
Gastroenterology Vol. 147, No. 5
BASIC AND TRANSLATIONAL BILIARY
Figure 1. Localization and expression of claudins in the Cldn2þ/þ and Cldn2/ liver. (A) Immunofluorescent images of claudin 1, 2, and 3 in the liver. Claudin 1 expression was strong in the periportal zone and slightly weaker in the perivenous zone, while claudin 2 was expressed only in the perivenous zone. Claudin 3 was strongly expressed in both zones. Bars ¼ 25 mm. (B) Quantitative expression levels of representative claudins in the liver of Cldn2þ/þ and Cldn2/ mice. No differences were detected between the Cldn2þ/þ and Cldn2/ liver except for claudin 2. (C) Immunoblotting for claudins in the liver of Cldn2þ/þ and Cldn2/ mice.
Bile Flow Retardation and Bile Lipid Condensation in the Hepatic and Gallbladder Bile of Cldn2/ Mice Next, we examined the functional effects of claudin 2 deficiency on the hepatobiliary system in vivo. Because claudin 2 resides in the TJs of bile canaliculi, we first focused on the bile of Cldn2þ/þ and Cldn2/ mice. The mice were fasted for 4 hours with drinking water ad libitum, then the hepatic and gallbladder bile were collected separately. The bile of Cldn2/ mice was a deeper yellow in color and
appeared more concentrated than the Cldn2þ/þ bile (Figure 3A). There was no significant difference in gallbladder volume between the 2 types of mice (Figure 3B). The bile flow rate in the Cldn2þ/þ mice was 5.8 ± 0.32 mL/min/ 100 g body weight (BW), and was about half that rate in the Cldn2/ mice (3.0 ± 0.40 mL/min/100 g BW) (Figure 3C). Consistent with these findings, the hepatic bile in the Cldn2/ mice was also a deeper yellow than that in the Cldn2þ/þ mice (Figure 3A). Spectroscopy over a broad range (200–800 nm) showed that the absorbance at 420 nm, which
November 2014
Role of Claudin 2 in Bile Flow and Gallstones 1137
corresponds to yellow, was significantly greater in the Cldn2/ than in the Cldn2þ/þ gallbladder bile. We next examined the composition of the hepatic bile. The concentrations of total cholesterol (27 ± 4.1 mg/dL vs 62 ± 5.4 mg/dL in Cldn2þ/þ and Cldn2/, respectively), phospholipids (5 ± 0.6 mg/dL vs 12.3 ± 1.5 mg/dL), total bile acid (18 ± 2.9 mmol/L vs 35 ± 4.8 mmol/L), and total bilirubin (6 ± 0.67 mg/dL vs 12.3 ± 0.83 mg/dL) were significantly increased in the Cldn2/ mice (Figure 3D). Similar results were obtained for the gallbladder bile: total
BASIC AND TRANSLATIONAL BILIARY
Figure 2. Histologic analysis of the liver. (A) H&Estained images of the Cldn2þ/þ and Cldn2/ liver. Bar ¼ 50 mm. (B) Thinsection electron microscopic images of the Cldn2þ/þ and Cldn2/ liver. Bar ¼ 500 nm. (C) Electron microscopic images of freeze-fracture replicas of the Cldn2þ/þ and Cldn2/ bile duct. TJ strands were observed in the bile ducts of both mice. No differences were detected. Bar ¼ 500 nm. Table shows statistical information about the numbers and width of TJ strands in the Cldn2þ/þ and Cldn2/ bile ducts.
cholesterol (110 ± 14.4 mg/dL vs 160 ± 7.0 mg/dL in Cldn2þ/þ and Cldn2/, respectively), phospholipids (12.2 ± 1.5 mg/dL vs 24.8 ± 0.8 mg/dL), total bile acid (83 ± 9.5 mmol/L vs 115 ± 6.7 mmol/L), and total bilirubin (5.5 ± 1.2 mg/dL vs 7.8 ± 1.3 mg/dL) (Figure 3E). The concentration of nonesterified fatty acids in the Cldn2/ hepatic bile (42.6 ± 9.7 mEq/L) was about 16 times that in the Cldn2þ/þ hepatic bile (2.6 ± 0.8 mEq/L). Because bile lipids and bile acids are synthesized and secreted by hepatocytes, we examined whether these
1138 Matsumoto et al
Gastroenterology Vol. 147, No. 5
BASIC AND TRANSLATIONAL BILIARY
Figure 3. Bile flow rate and bile composition. (A) Images of hepatic bile and gallbladder bile. The bile of the Cldn2/ mice was a deeper yellow color than that of the Cldn2þ/þ mice. (B) Gallbladder bile volume. The volume of the Cldn2/ mice was as same as that of the Cldn2þ/þ mice (n ¼ 15). (C) Bile flow rates of hepatic bile. The bile flow rate of the Cldn2/ mice was about half that of the Cldn2þ/þ mice (n ¼ 6). Hepatic (D) and gallbladder (E) biliary lipid (total cholesterol [T-Cho], phospholipid (PL), nonesterified fatty acids [NEFA]), bile acid, and total bilirubin (T-Bil) concentrations. These concentrations were higher in the Cldn2/ mice than in the Cldn2þ/þ mice (n ¼ 6). *P < .05.
differences between the Cldn2þ/þ and Cldn2/ mice could be explained by changes in the intrahepatic or serum lipid concentrations. There were no significant differences in either the intrahepatic or the serum lipids between the Cldn2þ/þ and Cldn2/ mice (Supplementary Figure 5A and B). The calculated daily secretion rate of the total bile acid in the Cldn2þ/þ and Cldn2/ mice was almost the same: 158 ± 20 vs 148 ± 18 mmol/100 g BW/day, respectively. Together, these findings indicated that the
decreased bile flow and high bile concentration in the Cldn2/ mice were probably caused by the decreased water volume in the bile. Expressions of water-permeable AQPs and bile transporters in the hepatocytes were not affected by the claudin 2 deletion (Supplementary Figure 5C and 6), suggesting that the reduced bile flow might be caused, at least in part, by a decrease in water supplied to the bile through the paracellular TJ route.
Role of Claudin 2 in Bile Flow and Gallstones 1139
Figure 4. Transepithelial water permeability in IBDUs. (A) Schematic drawing of IBDU preparation from mouse liver. (B) Immunofluorescent images of claudin 2 and ZO-1 in IBDUs of Cldn2þ/þ mice. Claudin 2 was uniformly expressed in the IBDUs. (C) Differential interference contrast (DIC) micrographs of mouse IBDUs in (i) hypotonic or (ii) hypertonic buffer at 0 seconds or 90 seconds. Representative results are shown. Dotted white line surrounds the luminal area of IBDU. (D) Changes in the luminal area of mouse IBDUs over time in (i) hypotonic or (ii) hypertonic buffer (n ¼ 8). The change rates of the Cldn2/ IBDUs were about half those of the Cldn2þ/þ IBDUs.
Decreased Transepithelial Water Permeability in the Cldn2/ Bile Duct and Bile Canaliculi One of the main forces driving water into the bile tract is believed to be a small osmotic gradient created by biliary transport processes. We next investigated the role of paracellular water permeability in ductal bile formation using enclosed isolated mouse intrahepatic bile duct units (IBDUs)30 (Figure 4A). When the IBDUs of Cldn2þ/þ mice were placed in a hypotonic buffer, the luminal space immediately increased and continued to grow for 90 seconds (Figure 4B and C). In contrast, when IBDUs were placed in a hypertonic buffer, the luminal space of the Cldn2þ/þ IBDUs
decreased rapidly, and continued to shrink for 90 seconds (Figure 4C and D). When the IBDUs of Cldn2/ mice were exposed to hypotonic or hypertonic buffer, the luminal area similarly increased or decreased, but the rates of change were about half those of the Cldn2þ/þ IBDUs. The Cldn2þ/þ and Cldn2/ IBDUs in hypotonic buffer grew 25.3% ± 1.9% and 7.3% ± 1.1%, and in hypertonic buffer was reduced by 17.5% ± 1.5% and 6.6% ± 1.0% in 90 seconds, respectively (Figure 4D). We next examined the role of paracellular water permeability in canalicular bile formation using primary cultured hepatocytes.31 After 48 hours of culture, bile
BASIC AND TRANSLATIONAL BILIARY
November 2014
1140 Matsumoto et al
canaliculi-like structures had formed between 2 cells, the interior of which was enclosed from the extracellular buffer (Figure 5A). There was no significant difference in the size of the canaliculi-like structures between the Cldn2þ/þ and Cldn2/ hepatocytes (121.8 ± 6.7 mm2 vs 121.1 ± 6.5 mm2, respectively). The Cldn2þ/þ hepatocytes were positive for claudin 2 protein, and the other claudin expressions were the same as those in hepatocytes in vivo (Figure 5C). When these Cldn2þ/þ cells were exposed to hypertonic buffer, the luminal space of the bile canalicular like structures decreased immediately, shrinking by 16.1% ± 0.7% in 60 seconds. In contrast, the structures between the Cldn2/ hepatocytes shrank by only 7.8% ± 0.7% in 60 seconds (Figure 5B and D, Supplementary videos). Therefore, it was likely that the transepithelial water permeability was decreased consistently in the Cldn2/ bile duct and bile canaliculi and in isolated hepatocyte in culture.
Gastroenterology Vol. 147, No. 5
Claudin 2 Deficiency Markedly Decreases Transepithelial Conductance in the Common Bile Duct Because ion movement is thought to be associated with water movement in the hepatobiliary system, we examined ion concentrations in the bile and serum. [Naþ] and [Kþ] in the Cldn2þ/þ hepatic bile were 140.2 ± 2.3 mEq/L and 6.1 ± 0.4 mEq/L, respectively, and in the gallbladder bile were 131.5 ± 1.6 mEq/L and 6.4 ± 0.7 mEq/L. The [Naþ] and [Kþ] of the Cldn2/ hepatic bile were 132.6 ± 2.2 mEq/L and 8.3 ± 0.5 mEq/L, respectively, and in the gallbladder bile were 123.9 ± 2.9 mEq/L and 13.0 ± 1.6 mEq/L, respectively (Figure 6A). There was a small significant decrease in sodium concentration and increase in potassium concentration in the bile of Cldn2/ mice compared with Cldn2þ/þ mice. These results are consistent with a decrease in
BASIC AND TRANSLATIONAL BILIARY
Figure 5. Transepithelial water permeability in the bile canaliculi of primary cultured hepatocytes. (A) Differential interference contrast (DIC) micrograph of mouse primary hepatocytes after 24 hours of culture. Asterisk indicates canalicular structure. Representative data. Bar ¼ 10 mm. (B) DIC micrographs of mouse primary hepatocytes in hypertonic buffer at 0 or 60 seconds. Representative results are shown. (C) Major claudin gene expression in the Cldn2þ/þ and Cldn2/ primary hepatocytes demonstrated by reverse transcription polymerase chain reaction. (D) Reduction of the canalicular area in hypertonic buffer at 60 seconds (n ¼ 70). The change rate of the Cldn2/ canalicular area was about half that of Cldn2þ/þ.
November 2014
Role of Claudin 2 in Bile Flow and Gallstones 1141
paracellular sodium permeability in Cldn2/ mice. Unlike other bile components, the Naþ was not concentrated in the Cldn2/ mice compared with the Cldn2þ/þ mice. To examine the claudin 2based paracellular ion permeabilities, we measured the electrical conductance across the bile duct of adult Cldn2þ/þ and Cldn2/ mice (Figure 6B). The transepithelial conductance of the Cldn2/ bile duct was 8.1 ± 0.7 mS/cm2, approximately 45% of that of the Cldn2þ/þ bile duct (18.1 ± 1.0 mS/cm2) (Figure 6C), indicating that the ion permeability was decreased in the Cldn2/ bile duct.
Claudin 2 Deficiency Markedly Decreases the Transepithelial Conductance and Selective Paracellular Permeability for Monovalent Cations in the Gallbladder Epithelium A large amount of water is reabsorbed by the gallbladder epithelium in an isosmotic manner. As the method for measuring gallbladder epithelial transport is well established, we examined the transepithelial conductance and selective paracellular permeability for monovalent cations of the gallbladder epithelium by Ussing chamber assays. The transepithelial conductance in the Cldn2/ gallbladder was 28.7 ± 8.3 mS/cm2, which was approximately half that of
BASIC AND TRANSLATIONAL BILIARY
Figure 6. Transepithelial conductance and selective paracellular permeability for sodium. (A) Concentrations of sodium and potassium ions in serum, hepatic bile, and gallbladder bile (n ¼ 6). The sodium ion concentration was significantly lower, but the potassium ion concentration was significantly higher in the Cldn2/ hepatic and gallbladder bile than in those of Cldn2þ/þ. (B) Method for measuring the transepithelial conductance of bile ducts. (C, D) Conductance across the Cldn2þ/þ and Cldn2/ bile duct (C) or gallbladder epithelium (D) (n ¼ 8). The transepithelial conductance of the Cldn2/ bile duct and gallbladder was about half that of the Cldn2þ/þ bile duct and gallbladder. *P < .05 (E) Permeability of the Cldn2þ/þ and Cldn2/ gallbladder epithelium for Naþ and Cl. *P < .05
1142 Matsumoto et al
the Cldn2þ/þ gallbladder (60.5 ± 7.9 mS/cm2) (Figure 6D). Because claudin 2 is reported to be involved in paracellular cation permeability, we examined the NaCl dilution potential and estimated the Naþ and Cl permeabilities (Figure 6E). The results suggested that the Cldn2/ gallbladder epithelium was less permeable to Naþ than the Cldn2þ/þ gallbladder epithelium.
Gastroenterology Vol. 147, No. 5
Increased Susceptibility to Cholesterol Gallstone Disease in Cldn2/ Mice To examine the effects of claudin 2 on gallstone formation, we fed Cldn2þ/þ and Cldn2/ mice a lithogenic diet, which contained high fat, high cholesterol, and an additional 0.5% cholic acid. Strikingly, 4-week administration of this diet induced the formation of macroscopically visible
BASIC AND TRANSLATIONAL BILIARY
Figure 7. Gallstone prevalence rate after 4 weeks on a lithocholic diet. (A) Macroscopic images of the gallbladders of Cldn2þ/þ and Cldn2/ mice. (B) Interference microscopy of hepatic bile from Cldn2þ/þ and Cldn2/ mice. (C) (i) Gallstone prevalence in the Cldn2þ/þ and Cldn2/ gallbladders. (n ¼ 11 [Cldn2þ/þ], 12 [Cldn2/]). (ii) Hepatic bile flow rate after 4 weeks on a lithocholic diet (n ¼ 7). *P < .05 (D) Cholesterol saturation index (CSI) of hepatic bile under a normal and lithogenic diet (n ¼ 6). (E) Quantitative expressions of claudin 1, 2, 3 in human liver. (F) Schematic drawing of the proposed water microcirculation system in the biliary tract. BD, bile duct; CV, central vein; HA, hepatic artery; PV, portal vein.
gallstones in all the Cldn2/ mice (n ¼ 12, Figure 7AC). In contrast no gallstones were formed in the wild-type mice after 4 weeks on a lithogenic diet (n ¼ 11). The main component of the gallstones was cholesterol (>98%). Eightweek administration of lithogenic diet did not induce clear additional changes in biliary system compared with 4-week administration, which did in the Cldn2/ mice, and gallstones were found in about half the wild-type mice (n ¼ 5 of 9) after 8 weeks on the diet. We next observed the cholesterol crystallization in the hepatic bile by interference microscopy. Many microcrystals were detected in the Cldn2/ bile, and no microcrystals were detected in the Cldn2þ/þ bile. The bile flow rate of both the Cldn2þ/þ and Cldn2/ mice was slightly increased compared with nonchallenged mice, but that of the Cldn2/ mice (4.4 ± 1.1 mL/min/100 g BW) was still significantly lower than that of the challenged Cldn2þ/þ mice (Figure 7D). There was no significant difference in CSI between the Cldn2þ/þ and Cldn2/ hepatic bile either before or after 4 weeks of the lithogenic diet administration (Figure 7E). As for claudin expression, there is no significant difference in the expression of major claudins in the gallbladder of Cldn2þ/þ and Cldn2/ mice with control diet or lithogenic diet for 4 weeks (Supplementary Figure 7). A small but significant increase in claudin 2 expression was detected in the liver and bile duct in Cldn2þ/þ mice with lithogenic diet. This might induce the increase of bile flow rate observed in lithogenic-administered mice. Clear histologic changes were not found after lithogenic diet administration for 4 weeks (Supplementary Figure 8). These results suggest that bile concentration and changes of bile component in Cldn2/ mice lead to increase the susceptibility to gallstone formation regardless of the CSI.
Claudin Expressions in Human Liver With or Without Gallstones To examine the effects of claudin 2 on gallstone formation in human, the expression levels of claudins were estimated in the human liver with or without gallstones (n ¼ 15, respectively) (Figure 7F and Supplementary Table 2). As a result, there was a tendency for the expression level of claudin 2 to be decreased in human liver with gallstones, as compared with that without gallstones. Although many factors supposedly regulate the formation of gallstones, our findings suggest the possibility that claudin 2 contributes to creating the microenvironmental setting for the formation of gallstones in human biliary system.
Discussion In this study, we provide the first in vivo evidence that a change in the paracellular pathway (ie, claudin 2 deficiency) of the hepatobiliary tract altered the bile content and volume, and induced a susceptibility to gallstone formation. These findings suggest that a biologic system can be regulated by paracellular permeation, the disruption of which can induce disease in a pathologic microenvironment.
Role of Claudin 2 in Bile Flow and Gallstones 1143
It is generally accepted that bile is initially formed by the secretion of bile acids from hepatocytes into bile canaliculi and is modified when it passes through the bile duct and gallbladder.2,23 Bile acids and other organic anions (such as glutathione or bilirubin), which are taken up from the bloodstream or produced in hepatocytes, are transported into the canalicular lumen across the apical cell membrane of the hepatocytes. These solutes are concentrated in the canalicular lumen and create osmotic pressure, which induces transepithelial water transport into the bile canaliculi to generate the bile flow.1,10 Cholangiocytes, as well as hepatocytes, contribute to the water output of the bile.32 Cholangiocytes secrete HCO3 and Cl and reabsorb bile acids, glutamate, glucose, and other molecules. The water volume is regulated by both secretion and reabsorption during its passage along intrahepatic bile ducts.33,34 Evidence has been accumulated that there are 2 pathways for transepithelial water transport: a transcellular route via transporters and a paracellular route through the TJs between cells.11,12 Water permeation through the transcellular pathway in the hepatobiliary system occurs via AQPs.3,35 Rodent hepatocytes express AQP0, 8, and 9. AQP8 is localized to intracellular vesicles and the canalicular plasma membrane, and is proposed to be involved in canalicular water secretion, and AQP9 is localized to the basolateral plasma membrane.35 This distribution of AQPs facilitates water transport into the canalicular lumen via the transcellular route in vitro. However, AQP8 knockout mice do not show any bile flow phenotypes.36 Similarly, cholangiocytes express AQP1 and 4, and AQP1 deletion reduces the osmotic water permeability in vitro, but bile flow is not affected in AQP1 knockout mice.37 These findings might be explained by redundant characteristics of the AQPs or some biologic compensation. On the other hand, it is plausible that the paracellular route is another critical regulator of water transport into the canalicular lumen in vivo. In this study, Cldn2/ mice showed an obvious bile flow phenotype; the bile flow rate was about 50% that of Cldn2þ/þ mice. Although bile acids and organic anions are a major driving force of the bile flow, it was unlikely that they substantially contributed to the bile flow retardation of the Cldn2/ mice because their daily generation was nearly the same as that in the Cldn2þ/þ mice. We also examined the expression of transporters in hepatocytes that are involved in the export of canalicular bile acids and lipids and found no differences between the Cldn2þ/þ and Cldn2/ mice (Supplementary Figure 5C). There was also no significant difference in the expression of AQPs examined between Cldn2þ/þ and Cldn2/ hepatocytes analyzed by reverse transcription polymerase chain reaction, immunofluorescence, and immunoblotting (Supplementary Figure 6). These findings support the idea that the paracellular transport contributes to the bile flow, along with the transcellular transport. Collectively, our findings suggest that claudin 2 is directly involved in the transepithelial water transport in the hepatobiliary system. Claudin 2 was one of the first claudins to be described as a TJ claudin, and it was the first-described and most-studied
BASIC AND TRANSLATIONAL BILIARY
November 2014
1144 Matsumoto et al
BASIC AND TRANSLATIONAL BILIARY
channel-forming type of claudin.13 Claudin 2 forms a cationselective channel and its presence increases paracellular cation permeability and decreases transepithelial resistance in cell lines.19,38 In the kidney, the cation selectivity in mouse proximal tubules is lost by claudin 2 deletion, and the increased paracellular resistance and loss of cation selectivity result in extreme urinary NaCl loss in response to NaCl challenge.20 In our study, the [Naþ] in the Cldn2/ bile was slightly lower than in the Cldn2þ/þ bile, even though the bile acids and lipids in the Cldn2/ bile were more concentrated. We also noted that the transepithelial conductance of the Cldn2/ bile duct was decreased significantly. It is likely that the deletion of claudin 2 reduces the cation permeability, resulting in attenuated bile flow. Claudin 2 was recently reported to form a paracellular water channel that mediates water transport across the TJs in MDCK cells.21 Simulated by a simple model using organic cations and alkali metal ions, the pore of the claudin 2 channel was determined to be about 6.5 Å.27 Because the molecular diameter of a water molecule is about 2.8 Å and of hydrated Naþ is about 3.7 Å, a water or hydrated Naþ molecule could permeate the claudin 2 channel. Although it is difficult to separate the transcellular and paracellular routes of transepithelial water transport experimentally, our results suggest that water can travel through any number of pathways, and that claudin 2 deletion in the biliary system attenuates the paracellular water transport. The current findings provide a new microcirculation model in which the amount of transepithelial water transport into bile canaliculi varies in different zones of the liver (Figure 7G). Because claudin 2 is expressed preferentially in the perivenous area, the water transport into bile canaliculi might take place mainly in this area. The strong expression of AQP9 in the perivenous area of the basolateral plasma membrane supports the water flow31,35 (Supplementary Figure 6). The bile canaliculi have blind ends in the perivenous area. The localized water transport, in which more water is transported at the beginning of the bile canaliculi, also promotes smooth bile flow in the bile canaliculi. Therefore, the uneven distribution of claudin 2 contributes to smooth and effective exchanges or transport between blood in the sinusoid and bile in the bile canaliculi in the liver. In this study, we also demonstrated that claudin 2 deletion in the biliary system promotes cholesterol gallstone formation in mice. The pathogenesis of cholesterol gallstones is known to be multifactorial, including genetic and environmental factors.6,7,9 Among these factors, cholesterol supersaturation appears to be the critical factor for cholesterol crystal nucleation, and the CSI is considered to be a reliable parameter for the biliary cholesterol saturation.39,40 Although there are only a few reports with human samples,8 the bile concentration is also thought to contribute to cholesterol gallstone formation. In the current study, the Cldn2/ bile was concentrated with decreased water input and promoted cholesterol gallstone formation, by administration of lithogenic diet even though the CSI was not significantly different from the Cldn2þ/þ bile. We
Gastroenterology Vol. 147, No. 5
experimentally suggest that the bile concentration is a factor for cholesterol crystal nucleation. This study provides the first evidence that the bile content and volume are regulated at least in part by paracellular water permeability due to claudin 2, which suggests a new therapeutic target for preventing gallstone disease. In addition, the regulation of claudin 2based paracellular permeability may be a potential strategy for treating cholestasis due to various etiologies.
Supplementary Material Note: To access the supplementary material accompanying this article, visit the online version of Gastroenterology at www.gastrojournal.org, and at http://dx.doi.org/10.1053/ j.gastro.2014.07.033.
References 1. Nicolaou M, Andress EJ, Zolnerciks JK, et al. Canalicular ABC transporters and liver disease. J Pathol 2012; 226:300–315. 2. Esteller A. Physiology of bile secretion. World J Gastroenterol 2008;14:5641. 3. Portincasa P, Calamita G. Water channel proteins in bile formation and flow in health and disease: when immiscible becomes miscible. Mol Aspects Med 2012;33: 651–664. 4. Matsuo M. ATP-binding cassette proteins involved in glucose and lipid homeostasis. Biosci Biotechnol Biochem 2010;74:899–907. 5. Hirschfield GM, Heathcote EJ, Gershwin ME. Pathogenesis of cholestatic liver disease and therapeutic approaches. Gastroenterology 2010;139:1481–1496. 6. Maurer KJ, Carey MC, Fox JG. Roles of infection, inflammation, and the immune system in cholesterol gallstone formation. Gastroenterology 2009;136:425–440. 7. Stokes CS, Krawczyk M, Lammert F. Gallstones: environment, lifestyle and genes. Dig Dis 2011;29:191–201. 8. van Erpecum KJ, van Berge Henegouwen GP, Stoelwinder B, et al. Bile concentration is a key factor for nucleation of cholesterol crystals and cholesterol saturation index in gallbladder bile of gallstone patients. Hepatology 1990;11:1–6. 9. van Erpecum KJ. Biliary lipids, water and cholesterol gallstones. Biol Cell 2005;97:815–822. 10. Ballatori N, Truong AT. Glutathione as a primary osmotic driving force in hepatic bile formation. Am J Physiol 1992; 263:G617–G624. 11. Jessner W, Zsembery A, Graf J. Transcellular water transport in hepatobiliary secretion and role of aquaporins in liver. Wien Med Wochenschr 2008;158:565–569. 12. Huebert RC, Splinter PL, Garcia F, et al. Expression and localization of aquaporin water channels in rat hepatocytes. Evidence for a role in canalicular bile secretion. J Biol Chem 2002;277:22710–22717. 13. Tsukita S, Furuse M, Itoh M. Multifunctional strands in tight junctions. Nat Rev Mol Cell Biol 2001;2:285–293. 14. Gunzel D, Yu AS. Claudins and the modulation of tight junction permeability. Physiol Rev 2013;93:525–569.
15. Steed E, Balda MS, Matter K. Dynamics and functions of tight junctions. Trends Cell Biol 2010;20:142–149. 16. Mineta K, Yamamoto Y, Yamazaki Y, et al. Predicted expansion of the claudin multigene family. FEBS Lett 2011;585:606–612. 17. Krug SM, Gunzel D, Conrad MP, et al. Charge-selective claudin channels. Ann N Y Acad Sci 2012;1257:20–28. 18. Shen L, Weber CR, Raleigh DR, et al. Tight junction pore and leak pathways: a dynamic duo. Annu Rev Physiol 2011;73:283–309. 19. Furuse M, Furuse K, Sasaki H, et al. Conversion of zonulae occludentes from tight to leaky strand type by introducing claudin-2 into Madin-Darby canine kidney I cells. J Cell Biol 2001;153:263–272. 20. Muto S, Hata M, Taniguchi J, et al. Claudin-2-deficient mice are defective in the leaky and cation-selective paracellular permeability properties of renal proximal tubules. Proc Natl Acad Sci U S A 2010;107:801–8016. 21. Rosenthal R, Milatz S, Krug SM, et al. Claudin-2, a component of the tight junction, forms a paracellular water channel. J Cell Sci 2010;123:1913–1921. 22. Amasheh S, Meiri N, Gitter AH, et al. Claudin-2 expression induces cation-selective channels in tight junctions of epithelial cells. J Cell Sci 2002;115:4969–4976. 23. Tamura A, Hayashi H, Imasato M, et al. Loss of claudin-15, but not claudin-2, causes Naþ deficiency and glucose malabsorption in mouse small intestine. Gastroenterology 2011;140:913–923. 24. Tamura A, Kitano Y, Hata M, et al. Megaintestine in claudin-15-deficient mice. Gastroenterology 2008;134: 523–534. 25. Wada M, Tamura A, Takahashi N, et al. Loss of claudins 2 and 15 from mice causes defects in paracellular Naþ flow and nutrient transport in gut and leads to death from malnutrition. Gastroenterology 2013;144:369–380. 26. Hadj-Rabia S, Baala L, Vabres P, et al. Claudin-1 gene mutations in neonatal sclerosing cholangitis associated with ichthyosis: a tight junction disease. Gastroenterology 2004;127:1386–1390. 27. Yu AS, Cheng MH, Angelow S, et al. Molecular basis for cation selectivity in claudin-2-based paracellular pores: identification of an electrostatic interaction site. J Gen Physiol 2009;133:111–127. 28. Son S, Kojima T, Decaens C, et al. Knockdown of tight junction protein claudin-2 prevents bile canalicular formation in WIF-B9 cells. Histochem Cell Biol 2009; 131:411–424. 29. Yeh TH, Krauland L, Singh V, et al. Liver-specific betacatenin knockout mice have bile canalicular abnormalities, bile secretory defect, and intrahepatic cholestasis. Hepatology 2010;52:1410–1419. 30. Cova E, Gong A, Marinelli RA, et al. Water movement across rat bile duct units is transcellular and channelmediated. Hepatology 2001;34:456–463.
Role of Claudin 2 in Bile Flow and Gallstones 1145 31. Larocca MC, Soria LR, Espelt MV, et al. Knockdown of hepatocyte aquaporin-8 by RNA interference induces defective bile canalicular water transport. Am J Physiol Gastrointest Liver Physiol 2009;296:G93–G100. 32. Tabibian JH, Masyuk AI, Masyuk TV, et al. Physiology of cholangiocytes. Compr Physiol 2013;3:541–565. 33. Masyuk A, Marinelli R, Larusso N. Water transport by epithelia of the digestive tract. Gastroenterology 2002; 122:545–562. 34. Boyer JL. Bile duct epithelium: frontiers in transport physiology. Am J Physiol 1996;270:G1–G5. 35. Matsuzaki T, Tajika Y, Ablimit A, et al. Aquaporins in the digestive system. Med Electron Microsc 2004;37:71–80. 36. Yang B, Song Y, Zhao D, et al. Phenotype analysis of aquaporin-8 null mice. Am J Physiol Cell Physiol 2005; 288:C1161–C1170. 37. Mennone A, Verkman AS, Boyer JL. Unimpaired osmotic water permeability and fluid secretion in bile duct epithelia of AQP1 null mice. Am J Physiol Gastrointest Liver Physiol 2002;283:G739–G746. 38. Schnermann J, Huang Y, Mizel D. Fluid reabsorption in proximal convoluted tubules of mice with gene deletions of claudin-2 and/or aquaporin1. Am J Physiol Renal Physiol 2013;305:F1352–F1364. 39. Carey MC. Critical tables for calculating the cholesterol saturation of native bile. J Lipid Res 1978;19:945–955. 40. Paumgartner G. Biliary physiology and disease: reflections of a physician-scientist. Hepatology 2010;51: 1095–106. 41. Mitaka T. The current status of primary hepatocyte culture. Int J Exp Pathol 1998;79:393–409. 42. Beyenbach KW, Frömter E. Electrophysiological evidence for Cl secretion in shark renal proximal tubules. Am J Physiol 1985;248:F282–F295. Author names in bold designate shared co-first authorship. Received November 5, 2013. Accepted July 19, 2014. Reprint requests Address requests for reprints to: Atsushi Tamura, PhD, or Sachiko Tsukita, PhD, Laboratory of Biological Science, Graduate School of Frontier Biosciences, Osaka University, 2-2 Yamadaoka, Suita, Osaka 565-0871, Japan. e-mail: [email protected] or [email protected]; fax: 81-6-6879-3329. Acknowledgment The authors thank Professor K. W. Beyenbach (Cornell University) and all the members of our laboratories for helpful discussion. We also thank Drs Grace Gray and Leslie Meglietta for proofreading the manuscript. Conflicts of interest The authors disclose no conflicts. Funding This work was supported in part by Grants-in-Aid for Scientific Research (A) and Creative Scientific Research, and by CREST (Core Research for Evolutional Science and Technology) of Japan Science and Technology Agency to Sachiko Tsukita, and by Grant-in-Aid for Exploratory Research to Yuji Yamazaki, and by Scientific Research (C) to Atsushi Tamura from the Ministry of Education, Culture, and Sports, Science and Technology of Japan.
BASIC AND TRANSLATIONAL BILIARY
November 2014
1145.e1 Matsumoto et al
Supplementary Materials Animals Generation of claudin 2deficient mice was described previously20,23; these mice were backcrossed to C57BL/6J for more than 6 generations. All animal experiments were performed in accordance with protocols approved by the Animal Studies Committee of Osaka University School of Medicine.
Antibodies Rat anti–E-cadherin monoclonal antibody was generously provided by Dr M. Takeichi (Riken CDB, Kobe, Japan). Rat anti-occludin and rat anti-ZO-1 monoclonal antibodies were generated in our laboratory.25 Rabbit anti-claudin 1, 2, 3 (Invitrogen, Life Technologies, Carlsbad, CA), rabbit antiAQP1 and 8 (Merck Millipore, Billerica, MA), rabbit antiAQP9 (Abcam, Bristol, England) polyclonal antibodies, Alexa-488labeled anti-rabbit IgG, and Alexa-546labeled anti-rat IgG (Molecular Probes, Eugene, OR) were purchased.
H&E Staining and Light Microscopy Mouse liver and gallbladder were fixed in 10% formalin/phosphate-buffered saline, embedded in paraffin, cut into 8-mm–thick sections, and stained with H&E.24
Bile Flow Rate Measurements Mice were fasted for 4 hours with free access to water. After ligation of the cystic duct, the common bile duct was cannulated with a 30-gauge needle, and hepatic bile was collected for 1 hour to determine the flow rate and analyze the bile composition. During surgery and hepatic bile collection, the mice were under anesthesia and maintained at 37 C on a hot plate.
Biochemical Measurements of Bile
Gastroenterology Vol. 147, No. 5
Measurement of Water Movement Across Intrahepatic Bile Duct Units in Response to Osmotic Gradient The enclosed IBDUs were preincubated for 15 minutes in HEPES-based buffer (35 mM NaCl or 120 mM NaCl, osmolality 300 mOsm) at room temperature, then placed in hypotonic HEPES buffer (35 mM NaCl, osmolality 100 mOsm), or hypertonic HEPES buffer (120 mM NaCl, osmolality 500 mOsm). Water movement into or out of the enclosed IBDUs was estimated by measuring the change in luminal area using Image-J software on a 2-dimensional micrograph, using the method described for rat bile duct units.30
Canalicular Water Movement in Response to Osmotic Gradient Hepatocytes with bile canaliculus-like structures were preincubated using the method described for IBDUs.31 After a 15-minute incubation, the hepatocytes were placed in hypertonic HEPES buffer (120 mM NaCl, osmolality 500 mOsm). Water movement out of the canalicular space was estimated by the same method used for IBDUs.
Electrophysiological Measurements of Gallbladder Epithelium The gallbladder was cut, flattened, and mounted in an Ussing chamber with an exposed area of 0.03 cm2 to examine the transepithelial conductance and NaCl dilution potential, as described previously.25
Thin-Section Electron Microscopy and Scanning Electron Microscopy Samples were processed for and subjected to ultra-thin section electron microscopy and scanning electron microscopy as described previously.24
Freeze-Fracture Replica Electron Microscopy
Hepatic and gallbladder bile samples were collected. Ionized sodium and potassium were measured by a portable ion meter (Horiba, Kyoto, Japan). Total cholesterol, phospholipids, nonesterified fatty acids, total bile acids, and total bilirubin levels in the bile samples were also determined (Nagahama Life Science Laboratory, Shiga, Japan).
Cldn2þ/þ and Cldn2/ bile ducts were fixed in 2% paraformaldehyde and 2.5% glutaraldehyde in 0.1 M phosphate buffer, and processed for and subjected to freeze-fracture electron microscopy, as described previously.24
Intrahepatic Bile Duct Unit Isolation
Immunofluorescence Analysis
Portal veins were cannulated, and the livers were perfused with ethylene glycol-bis(b-aminoethyl ether)N,N,N0 ,N0 -tetraacetic acid for 10 minutes followed by 0.02% type II collagenase for 7 minutes at 37 C. The hepatic capsule was removed, and the portal tissue was mechanically dissociated from the hepatic parenchymal tissue in cold Hank’s balanced salt solution as described previously.30 The portal tissue was then cut into 2- to 3-mm segments and cultured in Dulbecco’s modified Eagle medium for 12 hours. After incubation, both ends of the IBDU had sealed, enclosing the lumen.
Mouse liver and gallbladder were dissected and frozen in liquid N2. Frozen sections were cut and processed for, and subjected to, indirect immunofluorescence analysis, as described previously.25
Reverse Transcription Polymerase Chain Reaction and Quantitative Real-Time Reverse Transcription Polymerase Chain Reaction Reverse transcription polymerase chain reaction and quantitative real-time reverse-transcription polymerase chain reaction were performed as described previously.24
November 2014
Serum Biochemical Analysis Blood was collected from the inferior vena cava of mice after a 4-hour fast, and analyzed (Nagahama Life Science Laboratory, Shiga, Japan).
Role of Claudin 2 in Bile Flow and Gallstones 1145.e2
pipette coated with Sylgard 184. Through one barrel of the pipette, a 500-nA current was applied to the lumen, the transepithelial voltage was measured through the other barrel of the pipette, and an electrode placed near the proximal side of the duct in the bath.
Primary Culture of Hepatocytes The primary culture of hepatocytes was performed using the method described for rat hepatocytes.41 Briefly, mice were anesthetized and the liver was perfused with 5 mL/min HEPES-based buffer containing 0.5 mM ethylene glycol-bis(b-aminoethyl ether)-N,N,N0 ,N0 tetraacetic acid for 10 minutes at 37 C, and then with 5 mL/ min collagenase A buffer (HEPES-based buffer with 0.5 mg/ mL collagenase A) for 7 minutes at 37 C. The perfused liver was cut into pieces and passed through 70-mm nylon mesh. The filtrate was centrifuged at 50g for 5 minutes at 4 C. The hepatocytes in the pellet were then mixed with Percoll solution and centrifuged at 50g for 20 minutes to remove dead cells. These hepatocytes were seeded on 35-mm glassbottom dishes coated with type IA collagen and cultured for 24 hours at 37 C, as described. After this incubation, bile canaliculi-like structures had formed between 2 cells, the interior of which was enclosed from the extracellular buffer.
Electrophysiological Measurements of the Bile Duct The electrophysiological parameters of the common bile duct were measured using the method described for shark renal tubules as a reference.42 The diagram of the circuit is shown in Figure 6B. Briefly, the common bile duct was gently isolated, then the proximal side of the duct was cannulated with a double-barreled perfusion pipette and tied gently. The distal end of the duct was sucked into a
Lithogenic Diet
Cldn2þ/þ and Cldn2/ mice (n ¼ 14 each), at 8–12 weeks of age, were fed a lithogenic diet (containing 19% butter fat, 1.25% cholesterol, and 0.5% cholic acid) for 4 weeks. Mice that lost body weight were excluded (3 Cldn2þ/þ mice and 2 Cldn2/ mice). Although no gallstones formed in the wildtype mice after 4 weeks on a lithogenic diet, gallstones were found in about half the wild-type mice after 8 weeks on the diet. Experiments were performed after the mice were fasted for 4 hours.
Human Liver Samples All the liver samples were obtained from noncancerous part of the tissue resected for hepatocellular carcinoma. The presence of gallstones was examined by computed tomography images acquired before surgical resection. There were not significant difference in age, sex, etiology, and histology between the control group (n ¼ 15) and gallstones group (n ¼ 15) (Supplementary Table 1). The use of liver samples for this study was approved by the ethical committee of Osaka University Hospital and Osaka University Graduate School of Medicine.
Statistical Analysis Results are expressed as mean ± SEM. P values were calculated using an independent t-test. P < .05 was considered statistically significant.
1145.e3 Matsumoto et al
Gastroenterology Vol. 147, No. 5
Supplementary Figure 1. Localization and expression of claudins in the Cldn2þ/þ and Cldn2/ gallbladder. (A) Immunofluorescent images of claudin 1, 2, and 3 in the gallbladder. Bar ¼ 25 mm. (B) Quantitative expression levels of representative claudins in the Cldn2þ/þ and Cldn2/ (i) gallbladder and (ii) common bile duct. E-cad, E-cadherin.
November 2014
Supplementary Figure 2. Localization and expression of claudins in the Cldn2þ/þ and Cldn2/ intrahepatic and common bile ducts. Immunofluorescent images of claudin 2, 3, and 7 in the (A) intrahepatic bile ducts and (B) common bile ducts. Bar ¼ 20 mm.
Role of Claudin 2 in Bile Flow and Gallstones 1145.e4
1145.e5 Matsumoto et al
Gastroenterology Vol. 147, No. 5
Supplementary Figure 3. Immunofluorescent staining in the liver. Immunofluorescent images of occludin, ZO-1, and E-cadherin (E-cad) in the liver. No differences were detected between Cldn2þ/þ and Cldn2/. Bars ¼ 50 mm.
November 2014
Supplementary Figure 4. Histologic analysis of the liver and gallbladder. (A) Scanning electron microscopic images (upper panels for low magnifications; bars ¼ 20 mm. Middle panels for high magnifications; bars ¼ 10 mm) and immunofluorescent images (lower panels; bars ¼ 10 mm) of bile canaliculi in the liver. The canalicular diameters in the perivenous region were smaller than in the periportal region, but there were no differences between Cldn2þ/þ and Cldn2/. (B) H&E-stained images of the Cldn2þ/þ and Cldn2/ gallbladder. Bars ¼ 50 mm.
Role of Claudin 2 in Bile Flow and Gallstones 1145.e6
1145.e7 Matsumoto et al
Gastroenterology Vol. 147, No. 5
Supplementary Figure 5. Biochemical analyses of serum and hepatic lipids in Cldn2þ/þ and Cldn2/ mice. (A) Serum concentrations of aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), and lipids (total cholesterol [TCHO], triglyceride [TG], phospholipid [PL]) (n ¼ 6). (B) Concentrations of intrahepatic lipids (T-CHO, TG, nonesterified fatty acids [NEFA], total bile acids [TBA]) (n ¼ 10). (C) Quantitative expression levels of representative lipids and bile acid transporters in the Cldn2þ/þ and Cldn2/ liver.
November 2014
Supplementary Figure 6. Localization and expression of AQPs in the Cldn2þ/þ and Cldn2/ liver. (A) Immunofluorescent images of AQP 1, 8, and 9 in the liver. Bar ¼ 50 mm. (B) (i) Expression of major Aqp genes in the Cldn2þ/þ and Cldn2/ liver, examined by reverse transcription polymerase chain reaction (RT-PCR). (ii) Immunoblotting for AQP8 and 9 in the liver of Cldn2þ/þ and Cldn2/ mice.
Role of Claudin 2 in Bile Flow and Gallstones 1145.e8
1145.e9 Matsumoto et al
Gastroenterology Vol. 147, No. 5
Supplementary Figure 7. Gene expressions in the liver, gallbladder, and bile duct. Quantitative expression levels of representative claudins in the (A) liver, (B) gallbladder, and (C) common bile duct of Cldn2þ/þ mice under a control diet or lithogenic diet for 4 weeks. *P < .05
November 2014
Role of Claudin 2 in Bile Flow and Gallstones 1145.e10
Supplementary Figure 8. H&E-stained images of the gallbladder and liver of Cldn2þ/þ and Cldn2/ mice after 4 weeks on a lithogenic diet or control diet. (A) H&Estained images of the gallbladder of Cldn2þ/þ and Cldn2/ mice under a lithogenic diet. The gallbladder wall of the Cldn2/ mice was thickened compared with that of the Cldn2þ/þ mice. Bars ¼ 50 mm. (B) H&E-stained images of the under the lithogenic diet. No differences were detected. Bar ¼ 50 mm.
Supplementary Table 1.The Bile Canalicular Diameter in Perivenous and Periportal Zone in Mouse Liver
Perivenous (mm) Periportal (mm)
Cldn2þ/þ
Cldn2/
P value
0.79 ± 0.03 0.92 ± 0.03
0.78 ± 0.04 0.90 ± 0.02
.84 .66
NOTE. The width of bile canaliculi of perivenous zone and periportal zone was similar between in Cldn2þ/þ and Cldn2/ mice, respectively.
Supplementary Table 2.Sex, Age, Etiologic and Histologic Profiles of Human Liver
Sex (men/women) Age Etiology (HBV/HCV) Histology (CH/LC)
Control
Gallstone
3/12 61.2 ± 2.5 4/11 7/8
1/14 63 ± 2.8 4/10 4/8
NOTE. There were not significant difference in age, sex, etiology, and histology between controls group and gallstone group.
BASIC AND TRANSLATIONAL—BILIARY Claudin 2 Deficiency Reduces Bile Flow and Increases Susceptibility to Cholesterol Gallstone Disease in Mice Kengo Matsumoto,1,2 Mitsunobu Imasato,1 Yuji Yamazaki,1 Hiroo Tanaka,1 Mitsuhiro Watanabe,3,4 Hidetoshi Eguchi,5 Hiroaki Nagano,5 Hayato Hikita,2 Tomohide Tatsumi,2 Tetsuo Takehara,2 Atsushi Tamura,1 and Sachiko Tsukita1 1
Laboratory of Biological Science, Graduate School of Frontier Biosciences and Graduate School of Medicine, Osaka University, Osaka, Japan; 2Department of Gastroenterology and Hepatology, Graduate School of Medicine, Osaka University, Osaka, Japan; 3Department of Internal Medicine, School of Medicine, Keio University, Tokyo, Japan; 4Graduate School of Media and Governance, Faculty of Environment and Information Studies, Keio University, Kanagawa, Japan; and 5 Department of Surgery, Graduate School of Medicine, Osaka University, Osaka, Japan
See editorial on page 965.
BASIC AND TRANSLATIONAL BILIARY
BACKGROUND & AIMS: Bile formation and secretion are essential functions of the hepatobiliary system. Bile flow is generated by transepithelial transport of water and ionic/ nonionic solutes via transcellular and paracellular pathways that is mainly driven by osmotic pressure. We examined the role of tight junction–based paracellular transport in bile secretion. Claudins are cell–cell adhesion molecules in tight junctions that create the paracellular barrier. The claudin family has 27 reported members, some of which have paracellular ion- and/or water-channel–like functions. Claudin 2 is a paracellular channel-forming protein that is highly expressed in hepatocytes and cholangiocytes; we examined the hepatobiliary system of claudin 2 knockout (Cldn2/) mice. METHODS: We collected liver and biliary tissues from Cldn2/ and Cldn2þ/þ mice and performed histologic, biochemical, and electrophysiologic analyses. We measured osmotic movement of water and/ or ions in Cldn2/ and Cldn2þ/þ hepatocytes and bile ducts. Mice were placed on lithogenic diets for 4 weeks and development of gallstone disease was assessed. RESULTS: The rate of bile flow in Cldn2/ mice was half that of Cldn2þ/þ mice, resulting in significantly more concentrated bile in livers of Cldn2/ mice. Consistent with these findings, osmotic gradient-driven water flow was significantly reduced in hepatocyte bile canaliculi and bile ducts isolated from Cldn2/ mice, compared with Cldn2þ/þ mice. After 4 weeks on lithogenic diets, all Cldn2/ mice developed macroscopically visible gallstones; the main component of the gallstones was cholesterol (>98%). In contrast, none of the Cldn2þ/þ mice placed on lithogenic diets developed gallstones. CONCLUSIONS: Based on studies of Cldn2/ mice, claudin 2 regulates paracellular ion and water flow required for proper regulation of bile composition and flow. Dysregulation of this process increases susceptibility to cholesterol gallstone disease in mice.
gallstone disease.6–8 Cholesterol, phospholipids, and bile acids are the major components of bile. They are originally hydrophobic and their conjugated hydrophilic forms exist in a micellous state,2 which can aggregate and form stones in the aqueous bile in any condition.6,7 Therefore, the bile formation and bile flow are critical for hepatobiliary function. The incidence of gallstones is high in the Western world,7,8 and the pathogenesis of cholesterol gallstone disease is relatively well studied. The cholesterol saturation index (CSI), the ratio of the cholesterol concentration to that of bile acids and phospholipids, is a well-known indicator for cholesterol crystal nucleation. On the other hand, the water content of bile and bile flow are key factors in gallstone disease that have not been fully examined.8,9 In the biliary system, hepatic bile, which is 98% water, is secreted at a rate of 30–40 mL/h in humans. Bile formation is preceded by the active secretion of solutes from hepatocytes, such as bile acids by bile acid–dependent secretion, and glutathione, bilirubin, and/or HCO3 by bile acid– independent secretion. Water and electrolytes are then secreted.2,10 Water molecules transverse the epithelium through paracellular and/or transcellular routes. The identification and analysis of aquaporins (AQPs) has extended our understanding of how the transcellular water movement through epithelial cell sheets is regulated.11,12 On the other hand, it is still unsettled how and to what extent water moves paracellularly between the cells in epithelial sheets, mainly because it is difficult to separate the transcellular and paracellular pathways experimentally. The tight junction (TJ) is an adhesion apparatus that resides in the most apical part of the lateral membrane. TJs surround epithelial cells and attach them together. The TJ is also fundamentally responsible for the paracellular barrier of epithelia, including their charge- and size-dependent permeabilities; these properties enable the selective movement of solutes and water for biologic functions.13–15 On the
Keywords: Mouse Model; Claudin 2; Hepatic Microcirculation; TJ.
T
he hepatobiliary system plays critical roles in lipid and cholesterol homeostasis.1–4 Dysregulation of this system causes various diseases; impaired bile secretion leads to cholestasis5 and an imbalance in bile composition leads to
Abbreviations used in this paper: AQP, aquaporin; BW, body weight; CSI, cholesterol saturation index; IBDU, intrahepatic bile duct unit; TJ, tight junction. © 2014 by the AGA Institute 0016-5085/$36.00 http://dx.doi.org/10.1053/j.gastro.2014.07.033
November 2014
Materials and Methods Because of limitations in article length, the detailed Materials and Methods are presented in the Supplementary Material.
Results Claudin Expression Patterns in Liver and Gallbladder To determine which claudin subtypes are expressed in the TJs of the hepatobiliary system, we first examined the claudin gene expressions in liver and gallbladder by quantitative reverse transcription polymerase chain reaction. Claudins 1, 2, 3, 12, and 25 were the major claudins in the liver (Figure 1). Among these, claudin 2 is a channel-forming type, and was recently suggested to be a water-permeable claudin.19,27 Therefore, to elucidate the roles of TJs in regulating bile formation and secretion in the hepatobiliary tract, we analyzed Cldn2/ mice. We next examined the localization of claudin subtypes in the Cldn2þ/þ liver by immunofluorescence analyses. Claudin 2 was expressed in the perivenous regions, bile ducts, and gallbladder epithelium. In contrast, the claudin 1 signals were strong in the periportal zone and slightly weaker in the perivenous zone. Claudin 3 showed clear signals in both the perivenous and periportal zones. In the Cldn2þ/þ bile duct and gallbladder epithelium, claudins 2, 3, and 7 were strongly expressed (Supplementary Figures 1 and 2). In the Cldn2/ mice, the claudin 2 signal was not detected, but the other claudin expression patterns and levels were not significantly different from the Cldn2þ/þ mice in the liver, bile duct, or gallbladder epithelium. There were no differences in the expression patterns of occludin, ZO-1, or E-cadherin between the Cldn2þ/þ and Cldn2/ liver (Supplementary Figure 3).
Histologic Analysis of the Cldn2/ Liver and Gallbladder We next examined the liver and gallbladder in 8-weekold Cldn2þ/þ and Cldn2/ mice by H&E staining and freeze fracture electron microscopy and found no obvious differences between the Cldn2þ/þ and Cldn2/ samples (Figure 2A, B, and C). It was previously reported that claudin 2 knockdown prevents canalicular formation in WIF-B9 cells28 and that b-catenin knockout mice show decreased claudin 2 expression and wider bile canaliculi.29 We therefore examined the bile canaliculi more closely by scanning electron microscopy and immunofluorescence analyses. We found no obvious differences in the bile canaliculi width or shape between the Cldn2þ/þ and Cldn2/ liver, although the bile canaliculi is slightly wider in periportal zone than in perivenous zone both in Cldn2þ/þ and Cldn2/ bile canaliculi (perivenous zone: Cldn2þ/þ 0.79 ± 0.03 mm, Cldn2/ 0.78 ± 0.04 mm; periportal zone: Cldn2þ/þ 0.92 ± 0.03 mm, Cldn2/ 0.90 ± 0.02 mm) (Supplementary Figure 4A and B and Supplementary Table 1). Together, these data suggested that histologically, almost no changes were detected in the Cldn2/ liver and gallbladder.
BASIC AND TRANSLATIONAL BILIARY
freeze-fracture electron microscopy, TJ strands appear as a continuous meshwork in the apical intercellular space. These strands are thought to be derived from polymerized claudins forming the paracellular barrier, although the polymerization mechanism is unknown. The claudins are a large family of tetraspanning transmembrane proteins, with at least 27 members in human/mouse, which are thought to have various functions in claudin-based paracellular systems. Evidence suggests that claudins generally function to create the paracellular barrier function of epithelial cell sheets. However, some of them increase the paracellular ion permeability when expressed in paracellular barrier– established cell sheets; these are described as “ion-channel forming” claudins,14,16,17 although they also elicit a relatively nonspecific permeability for small nonionic solutes. Previous studies suggested that the combination of claudin species expressed in cells is important for determining the permselectivity of the TJs.14,15 Consistent with this idea, physiologic studies suggest that epithelium types can be classified as leaky or tight, based on transepithelial conductance and/or paracellular solute flux measurements.17,18 Claudin 2 was the first channel-forming claudin to be identified,13,19 with suggested roles in water and cation permselectivity.20–22 In claudin 2 knockout (Cldn2/) mice, deficiencies in paracellular permeability and cation selectiveness impair the reabsorption of Naþ, Cl, and water in renal proximal tubules.20 In the small intestine, claudin 2 loss significantly decreases the Naþ permeability.23 In addition, double-knockout mice in the intestine of claudin 2 and claudin 15, another channel-forming claudin, are deficient in sodium-driven nutrient absorption due to the insufficiency of Naþ in the intestinal lumen, which is usually supplied paracellularly from the submucosal space, which leads to infant death.24,25 Thus, the ion channel–forming claudins appear to be important for regulating biologic functions, in addition to their paracellular barrier function. In the liver, claudin 1, 2, and 3 are dominant. Among them, barrier-forming claudin 1 dysfunction is reported to lead to a rare autosomal recessive neonatal ichthyosissclerosing cholangitis syndrome, which is characterized by scalp hypotrichosis, scarring alopecia, ichthyosis, and sclerosing cholangitis, as described previously.26 In cholangitis in neonatal ichthyosis-sclerosing cholangitis syndrome liver, a bile leakage through claudin 1deficient TJs of hepatocellular and biliary cells is suggested to result in direct hepatocellular and/or biliary injuries and in cholestasis. Here we focused on the hepatobiliary system in Cldn2/ mice. The concentrations of bile acids, phospholipids, and cholesterol were significantly increased in the Cldn2/ bile. We found that Cldn2/ mice showed decreased transepithelial electrical conductance and water permeability in the hepatobiliary system, resulting in decreased bile flow. In addition, after 4 weeks on a lithogenic diet, the prevalence of cholesterol gallstones was 100% in the Cldn2/ mice vs 0% in Cldn2þ/þ mice. Our findings indicate that the channel-forming claudin 2 is critical for bile homeostasis in the hepatobiliary system. We also provide the first evidence that this claudin plays a key role in bile flow as a type of water flow in the body.
Role of Claudin 2 in Bile Flow and Gallstones 1135
1136 Matsumoto et al
Gastroenterology Vol. 147, No. 5
BASIC AND TRANSLATIONAL BILIARY
Figure 1. Localization and expression of claudins in the Cldn2þ/þ and Cldn2/ liver. (A) Immunofluorescent images of claudin 1, 2, and 3 in the liver. Claudin 1 expression was strong in the periportal zone and slightly weaker in the perivenous zone, while claudin 2 was expressed only in the perivenous zone. Claudin 3 was strongly expressed in both zones. Bars ¼ 25 mm. (B) Quantitative expression levels of representative claudins in the liver of Cldn2þ/þ and Cldn2/ mice. No differences were detected between the Cldn2þ/þ and Cldn2/ liver except for claudin 2. (C) Immunoblotting for claudins in the liver of Cldn2þ/þ and Cldn2/ mice.
Bile Flow Retardation and Bile Lipid Condensation in the Hepatic and Gallbladder Bile of Cldn2/ Mice Next, we examined the functional effects of claudin 2 deficiency on the hepatobiliary system in vivo. Because claudin 2 resides in the TJs of bile canaliculi, we first focused on the bile of Cldn2þ/þ and Cldn2/ mice. The mice were fasted for 4 hours with drinking water ad libitum, then the hepatic and gallbladder bile were collected separately. The bile of Cldn2/ mice was a deeper yellow in color and
appeared more concentrated than the Cldn2þ/þ bile (Figure 3A). There was no significant difference in gallbladder volume between the 2 types of mice (Figure 3B). The bile flow rate in the Cldn2þ/þ mice was 5.8 ± 0.32 mL/min/ 100 g body weight (BW), and was about half that rate in the Cldn2/ mice (3.0 ± 0.40 mL/min/100 g BW) (Figure 3C). Consistent with these findings, the hepatic bile in the Cldn2/ mice was also a deeper yellow than that in the Cldn2þ/þ mice (Figure 3A). Spectroscopy over a broad range (200–800 nm) showed that the absorbance at 420 nm, which
November 2014
Role of Claudin 2 in Bile Flow and Gallstones 1137
corresponds to yellow, was significantly greater in the Cldn2/ than in the Cldn2þ/þ gallbladder bile. We next examined the composition of the hepatic bile. The concentrations of total cholesterol (27 ± 4.1 mg/dL vs 62 ± 5.4 mg/dL in Cldn2þ/þ and Cldn2/, respectively), phospholipids (5 ± 0.6 mg/dL vs 12.3 ± 1.5 mg/dL), total bile acid (18 ± 2.9 mmol/L vs 35 ± 4.8 mmol/L), and total bilirubin (6 ± 0.67 mg/dL vs 12.3 ± 0.83 mg/dL) were significantly increased in the Cldn2/ mice (Figure 3D). Similar results were obtained for the gallbladder bile: total
BASIC AND TRANSLATIONAL BILIARY
Figure 2. Histologic analysis of the liver. (A) H&Estained images of the Cldn2þ/þ and Cldn2/ liver. Bar ¼ 50 mm. (B) Thinsection electron microscopic images of the Cldn2þ/þ and Cldn2/ liver. Bar ¼ 500 nm. (C) Electron microscopic images of freeze-fracture replicas of the Cldn2þ/þ and Cldn2/ bile duct. TJ strands were observed in the bile ducts of both mice. No differences were detected. Bar ¼ 500 nm. Table shows statistical information about the numbers and width of TJ strands in the Cldn2þ/þ and Cldn2/ bile ducts.
cholesterol (110 ± 14.4 mg/dL vs 160 ± 7.0 mg/dL in Cldn2þ/þ and Cldn2/, respectively), phospholipids (12.2 ± 1.5 mg/dL vs 24.8 ± 0.8 mg/dL), total bile acid (83 ± 9.5 mmol/L vs 115 ± 6.7 mmol/L), and total bilirubin (5.5 ± 1.2 mg/dL vs 7.8 ± 1.3 mg/dL) (Figure 3E). The concentration of nonesterified fatty acids in the Cldn2/ hepatic bile (42.6 ± 9.7 mEq/L) was about 16 times that in the Cldn2þ/þ hepatic bile (2.6 ± 0.8 mEq/L). Because bile lipids and bile acids are synthesized and secreted by hepatocytes, we examined whether these
1138 Matsumoto et al
Gastroenterology Vol. 147, No. 5
BASIC AND TRANSLATIONAL BILIARY
Figure 3. Bile flow rate and bile composition. (A) Images of hepatic bile and gallbladder bile. The bile of the Cldn2/ mice was a deeper yellow color than that of the Cldn2þ/þ mice. (B) Gallbladder bile volume. The volume of the Cldn2/ mice was as same as that of the Cldn2þ/þ mice (n ¼ 15). (C) Bile flow rates of hepatic bile. The bile flow rate of the Cldn2/ mice was about half that of the Cldn2þ/þ mice (n ¼ 6). Hepatic (D) and gallbladder (E) biliary lipid (total cholesterol [T-Cho], phospholipid (PL), nonesterified fatty acids [NEFA]), bile acid, and total bilirubin (T-Bil) concentrations. These concentrations were higher in the Cldn2/ mice than in the Cldn2þ/þ mice (n ¼ 6). *P < .05.
differences between the Cldn2þ/þ and Cldn2/ mice could be explained by changes in the intrahepatic or serum lipid concentrations. There were no significant differences in either the intrahepatic or the serum lipids between the Cldn2þ/þ and Cldn2/ mice (Supplementary Figure 5A and B). The calculated daily secretion rate of the total bile acid in the Cldn2þ/þ and Cldn2/ mice was almost the same: 158 ± 20 vs 148 ± 18 mmol/100 g BW/day, respectively. Together, these findings indicated that the
decreased bile flow and high bile concentration in the Cldn2/ mice were probably caused by the decreased water volume in the bile. Expressions of water-permeable AQPs and bile transporters in the hepatocytes were not affected by the claudin 2 deletion (Supplementary Figure 5C and 6), suggesting that the reduced bile flow might be caused, at least in part, by a decrease in water supplied to the bile through the paracellular TJ route.
Role of Claudin 2 in Bile Flow and Gallstones 1139
Figure 4. Transepithelial water permeability in IBDUs. (A) Schematic drawing of IBDU preparation from mouse liver. (B) Immunofluorescent images of claudin 2 and ZO-1 in IBDUs of Cldn2þ/þ mice. Claudin 2 was uniformly expressed in the IBDUs. (C) Differential interference contrast (DIC) micrographs of mouse IBDUs in (i) hypotonic or (ii) hypertonic buffer at 0 seconds or 90 seconds. Representative results are shown. Dotted white line surrounds the luminal area of IBDU. (D) Changes in the luminal area of mouse IBDUs over time in (i) hypotonic or (ii) hypertonic buffer (n ¼ 8). The change rates of the Cldn2/ IBDUs were about half those of the Cldn2þ/þ IBDUs.
Decreased Transepithelial Water Permeability in the Cldn2/ Bile Duct and Bile Canaliculi One of the main forces driving water into the bile tract is believed to be a small osmotic gradient created by biliary transport processes. We next investigated the role of paracellular water permeability in ductal bile formation using enclosed isolated mouse intrahepatic bile duct units (IBDUs)30 (Figure 4A). When the IBDUs of Cldn2þ/þ mice were placed in a hypotonic buffer, the luminal space immediately increased and continued to grow for 90 seconds (Figure 4B and C). In contrast, when IBDUs were placed in a hypertonic buffer, the luminal space of the Cldn2þ/þ IBDUs
decreased rapidly, and continued to shrink for 90 seconds (Figure 4C and D). When the IBDUs of Cldn2/ mice were exposed to hypotonic or hypertonic buffer, the luminal area similarly increased or decreased, but the rates of change were about half those of the Cldn2þ/þ IBDUs. The Cldn2þ/þ and Cldn2/ IBDUs in hypotonic buffer grew 25.3% ± 1.9% and 7.3% ± 1.1%, and in hypertonic buffer was reduced by 17.5% ± 1.5% and 6.6% ± 1.0% in 90 seconds, respectively (Figure 4D). We next examined the role of paracellular water permeability in canalicular bile formation using primary cultured hepatocytes.31 After 48 hours of culture, bile
BASIC AND TRANSLATIONAL BILIARY
November 2014
1140 Matsumoto et al
canaliculi-like structures had formed between 2 cells, the interior of which was enclosed from the extracellular buffer (Figure 5A). There was no significant difference in the size of the canaliculi-like structures between the Cldn2þ/þ and Cldn2/ hepatocytes (121.8 ± 6.7 mm2 vs 121.1 ± 6.5 mm2, respectively). The Cldn2þ/þ hepatocytes were positive for claudin 2 protein, and the other claudin expressions were the same as those in hepatocytes in vivo (Figure 5C). When these Cldn2þ/þ cells were exposed to hypertonic buffer, the luminal space of the bile canalicular like structures decreased immediately, shrinking by 16.1% ± 0.7% in 60 seconds. In contrast, the structures between the Cldn2/ hepatocytes shrank by only 7.8% ± 0.7% in 60 seconds (Figure 5B and D, Supplementary videos). Therefore, it was likely that the transepithelial water permeability was decreased consistently in the Cldn2/ bile duct and bile canaliculi and in isolated hepatocyte in culture.
Gastroenterology Vol. 147, No. 5
Claudin 2 Deficiency Markedly Decreases Transepithelial Conductance in the Common Bile Duct Because ion movement is thought to be associated with water movement in the hepatobiliary system, we examined ion concentrations in the bile and serum. [Naþ] and [Kþ] in the Cldn2þ/þ hepatic bile were 140.2 ± 2.3 mEq/L and 6.1 ± 0.4 mEq/L, respectively, and in the gallbladder bile were 131.5 ± 1.6 mEq/L and 6.4 ± 0.7 mEq/L. The [Naþ] and [Kþ] of the Cldn2/ hepatic bile were 132.6 ± 2.2 mEq/L and 8.3 ± 0.5 mEq/L, respectively, and in the gallbladder bile were 123.9 ± 2.9 mEq/L and 13.0 ± 1.6 mEq/L, respectively (Figure 6A). There was a small significant decrease in sodium concentration and increase in potassium concentration in the bile of Cldn2/ mice compared with Cldn2þ/þ mice. These results are consistent with a decrease in
BASIC AND TRANSLATIONAL BILIARY
Figure 5. Transepithelial water permeability in the bile canaliculi of primary cultured hepatocytes. (A) Differential interference contrast (DIC) micrograph of mouse primary hepatocytes after 24 hours of culture. Asterisk indicates canalicular structure. Representative data. Bar ¼ 10 mm. (B) DIC micrographs of mouse primary hepatocytes in hypertonic buffer at 0 or 60 seconds. Representative results are shown. (C) Major claudin gene expression in the Cldn2þ/þ and Cldn2/ primary hepatocytes demonstrated by reverse transcription polymerase chain reaction. (D) Reduction of the canalicular area in hypertonic buffer at 60 seconds (n ¼ 70). The change rate of the Cldn2/ canalicular area was about half that of Cldn2þ/þ.
November 2014
Role of Claudin 2 in Bile Flow and Gallstones 1141
paracellular sodium permeability in Cldn2/ mice. Unlike other bile components, the Naþ was not concentrated in the Cldn2/ mice compared with the Cldn2þ/þ mice. To examine the claudin 2based paracellular ion permeabilities, we measured the electrical conductance across the bile duct of adult Cldn2þ/þ and Cldn2/ mice (Figure 6B). The transepithelial conductance of the Cldn2/ bile duct was 8.1 ± 0.7 mS/cm2, approximately 45% of that of the Cldn2þ/þ bile duct (18.1 ± 1.0 mS/cm2) (Figure 6C), indicating that the ion permeability was decreased in the Cldn2/ bile duct.
Claudin 2 Deficiency Markedly Decreases the Transepithelial Conductance and Selective Paracellular Permeability for Monovalent Cations in the Gallbladder Epithelium A large amount of water is reabsorbed by the gallbladder epithelium in an isosmotic manner. As the method for measuring gallbladder epithelial transport is well established, we examined the transepithelial conductance and selective paracellular permeability for monovalent cations of the gallbladder epithelium by Ussing chamber assays. The transepithelial conductance in the Cldn2/ gallbladder was 28.7 ± 8.3 mS/cm2, which was approximately half that of
BASIC AND TRANSLATIONAL BILIARY
Figure 6. Transepithelial conductance and selective paracellular permeability for sodium. (A) Concentrations of sodium and potassium ions in serum, hepatic bile, and gallbladder bile (n ¼ 6). The sodium ion concentration was significantly lower, but the potassium ion concentration was significantly higher in the Cldn2/ hepatic and gallbladder bile than in those of Cldn2þ/þ. (B) Method for measuring the transepithelial conductance of bile ducts. (C, D) Conductance across the Cldn2þ/þ and Cldn2/ bile duct (C) or gallbladder epithelium (D) (n ¼ 8). The transepithelial conductance of the Cldn2/ bile duct and gallbladder was about half that of the Cldn2þ/þ bile duct and gallbladder. *P < .05 (E) Permeability of the Cldn2þ/þ and Cldn2/ gallbladder epithelium for Naþ and Cl. *P < .05
1142 Matsumoto et al
the Cldn2þ/þ gallbladder (60.5 ± 7.9 mS/cm2) (Figure 6D). Because claudin 2 is reported to be involved in paracellular cation permeability, we examined the NaCl dilution potential and estimated the Naþ and Cl permeabilities (Figure 6E). The results suggested that the Cldn2/ gallbladder epithelium was less permeable to Naþ than the Cldn2þ/þ gallbladder epithelium.
Gastroenterology Vol. 147, No. 5
Increased Susceptibility to Cholesterol Gallstone Disease in Cldn2/ Mice To examine the effects of claudin 2 on gallstone formation, we fed Cldn2þ/þ and Cldn2/ mice a lithogenic diet, which contained high fat, high cholesterol, and an additional 0.5% cholic acid. Strikingly, 4-week administration of this diet induced the formation of macroscopically visible
BASIC AND TRANSLATIONAL BILIARY
Figure 7. Gallstone prevalence rate after 4 weeks on a lithocholic diet. (A) Macroscopic images of the gallbladders of Cldn2þ/þ and Cldn2/ mice. (B) Interference microscopy of hepatic bile from Cldn2þ/þ and Cldn2/ mice. (C) (i) Gallstone prevalence in the Cldn2þ/þ and Cldn2/ gallbladders. (n ¼ 11 [Cldn2þ/þ], 12 [Cldn2/]). (ii) Hepatic bile flow rate after 4 weeks on a lithocholic diet (n ¼ 7). *P < .05 (D) Cholesterol saturation index (CSI) of hepatic bile under a normal and lithogenic diet (n ¼ 6). (E) Quantitative expressions of claudin 1, 2, 3 in human liver. (F) Schematic drawing of the proposed water microcirculation system in the biliary tract. BD, bile duct; CV, central vein; HA, hepatic artery; PV, portal vein.
gallstones in all the Cldn2/ mice (n ¼ 12, Figure 7AC). In contrast no gallstones were formed in the wild-type mice after 4 weeks on a lithogenic diet (n ¼ 11). The main component of the gallstones was cholesterol (>98%). Eightweek administration of lithogenic diet did not induce clear additional changes in biliary system compared with 4-week administration, which did in the Cldn2/ mice, and gallstones were found in about half the wild-type mice (n ¼ 5 of 9) after 8 weeks on the diet. We next observed the cholesterol crystallization in the hepatic bile by interference microscopy. Many microcrystals were detected in the Cldn2/ bile, and no microcrystals were detected in the Cldn2þ/þ bile. The bile flow rate of both the Cldn2þ/þ and Cldn2/ mice was slightly increased compared with nonchallenged mice, but that of the Cldn2/ mice (4.4 ± 1.1 mL/min/100 g BW) was still significantly lower than that of the challenged Cldn2þ/þ mice (Figure 7D). There was no significant difference in CSI between the Cldn2þ/þ and Cldn2/ hepatic bile either before or after 4 weeks of the lithogenic diet administration (Figure 7E). As for claudin expression, there is no significant difference in the expression of major claudins in the gallbladder of Cldn2þ/þ and Cldn2/ mice with control diet or lithogenic diet for 4 weeks (Supplementary Figure 7). A small but significant increase in claudin 2 expression was detected in the liver and bile duct in Cldn2þ/þ mice with lithogenic diet. This might induce the increase of bile flow rate observed in lithogenic-administered mice. Clear histologic changes were not found after lithogenic diet administration for 4 weeks (Supplementary Figure 8). These results suggest that bile concentration and changes of bile component in Cldn2/ mice lead to increase the susceptibility to gallstone formation regardless of the CSI.
Claudin Expressions in Human Liver With or Without Gallstones To examine the effects of claudin 2 on gallstone formation in human, the expression levels of claudins were estimated in the human liver with or without gallstones (n ¼ 15, respectively) (Figure 7F and Supplementary Table 2). As a result, there was a tendency for the expression level of claudin 2 to be decreased in human liver with gallstones, as compared with that without gallstones. Although many factors supposedly regulate the formation of gallstones, our findings suggest the possibility that claudin 2 contributes to creating the microenvironmental setting for the formation of gallstones in human biliary system.
Discussion In this study, we provide the first in vivo evidence that a change in the paracellular pathway (ie, claudin 2 deficiency) of the hepatobiliary tract altered the bile content and volume, and induced a susceptibility to gallstone formation. These findings suggest that a biologic system can be regulated by paracellular permeation, the disruption of which can induce disease in a pathologic microenvironment.
Role of Claudin 2 in Bile Flow and Gallstones 1143
It is generally accepted that bile is initially formed by the secretion of bile acids from hepatocytes into bile canaliculi and is modified when it passes through the bile duct and gallbladder.2,23 Bile acids and other organic anions (such as glutathione or bilirubin), which are taken up from the bloodstream or produced in hepatocytes, are transported into the canalicular lumen across the apical cell membrane of the hepatocytes. These solutes are concentrated in the canalicular lumen and create osmotic pressure, which induces transepithelial water transport into the bile canaliculi to generate the bile flow.1,10 Cholangiocytes, as well as hepatocytes, contribute to the water output of the bile.32 Cholangiocytes secrete HCO3 and Cl and reabsorb bile acids, glutamate, glucose, and other molecules. The water volume is regulated by both secretion and reabsorption during its passage along intrahepatic bile ducts.33,34 Evidence has been accumulated that there are 2 pathways for transepithelial water transport: a transcellular route via transporters and a paracellular route through the TJs between cells.11,12 Water permeation through the transcellular pathway in the hepatobiliary system occurs via AQPs.3,35 Rodent hepatocytes express AQP0, 8, and 9. AQP8 is localized to intracellular vesicles and the canalicular plasma membrane, and is proposed to be involved in canalicular water secretion, and AQP9 is localized to the basolateral plasma membrane.35 This distribution of AQPs facilitates water transport into the canalicular lumen via the transcellular route in vitro. However, AQP8 knockout mice do not show any bile flow phenotypes.36 Similarly, cholangiocytes express AQP1 and 4, and AQP1 deletion reduces the osmotic water permeability in vitro, but bile flow is not affected in AQP1 knockout mice.37 These findings might be explained by redundant characteristics of the AQPs or some biologic compensation. On the other hand, it is plausible that the paracellular route is another critical regulator of water transport into the canalicular lumen in vivo. In this study, Cldn2/ mice showed an obvious bile flow phenotype; the bile flow rate was about 50% that of Cldn2þ/þ mice. Although bile acids and organic anions are a major driving force of the bile flow, it was unlikely that they substantially contributed to the bile flow retardation of the Cldn2/ mice because their daily generation was nearly the same as that in the Cldn2þ/þ mice. We also examined the expression of transporters in hepatocytes that are involved in the export of canalicular bile acids and lipids and found no differences between the Cldn2þ/þ and Cldn2/ mice (Supplementary Figure 5C). There was also no significant difference in the expression of AQPs examined between Cldn2þ/þ and Cldn2/ hepatocytes analyzed by reverse transcription polymerase chain reaction, immunofluorescence, and immunoblotting (Supplementary Figure 6). These findings support the idea that the paracellular transport contributes to the bile flow, along with the transcellular transport. Collectively, our findings suggest that claudin 2 is directly involved in the transepithelial water transport in the hepatobiliary system. Claudin 2 was one of the first claudins to be described as a TJ claudin, and it was the first-described and most-studied
BASIC AND TRANSLATIONAL BILIARY
November 2014
1144 Matsumoto et al
BASIC AND TRANSLATIONAL BILIARY
channel-forming type of claudin.13 Claudin 2 forms a cationselective channel and its presence increases paracellular cation permeability and decreases transepithelial resistance in cell lines.19,38 In the kidney, the cation selectivity in mouse proximal tubules is lost by claudin 2 deletion, and the increased paracellular resistance and loss of cation selectivity result in extreme urinary NaCl loss in response to NaCl challenge.20 In our study, the [Naþ] in the Cldn2/ bile was slightly lower than in the Cldn2þ/þ bile, even though the bile acids and lipids in the Cldn2/ bile were more concentrated. We also noted that the transepithelial conductance of the Cldn2/ bile duct was decreased significantly. It is likely that the deletion of claudin 2 reduces the cation permeability, resulting in attenuated bile flow. Claudin 2 was recently reported to form a paracellular water channel that mediates water transport across the TJs in MDCK cells.21 Simulated by a simple model using organic cations and alkali metal ions, the pore of the claudin 2 channel was determined to be about 6.5 Å.27 Because the molecular diameter of a water molecule is about 2.8 Å and of hydrated Naþ is about 3.7 Å, a water or hydrated Naþ molecule could permeate the claudin 2 channel. Although it is difficult to separate the transcellular and paracellular routes of transepithelial water transport experimentally, our results suggest that water can travel through any number of pathways, and that claudin 2 deletion in the biliary system attenuates the paracellular water transport. The current findings provide a new microcirculation model in which the amount of transepithelial water transport into bile canaliculi varies in different zones of the liver (Figure 7G). Because claudin 2 is expressed preferentially in the perivenous area, the water transport into bile canaliculi might take place mainly in this area. The strong expression of AQP9 in the perivenous area of the basolateral plasma membrane supports the water flow31,35 (Supplementary Figure 6). The bile canaliculi have blind ends in the perivenous area. The localized water transport, in which more water is transported at the beginning of the bile canaliculi, also promotes smooth bile flow in the bile canaliculi. Therefore, the uneven distribution of claudin 2 contributes to smooth and effective exchanges or transport between blood in the sinusoid and bile in the bile canaliculi in the liver. In this study, we also demonstrated that claudin 2 deletion in the biliary system promotes cholesterol gallstone formation in mice. The pathogenesis of cholesterol gallstones is known to be multifactorial, including genetic and environmental factors.6,7,9 Among these factors, cholesterol supersaturation appears to be the critical factor for cholesterol crystal nucleation, and the CSI is considered to be a reliable parameter for the biliary cholesterol saturation.39,40 Although there are only a few reports with human samples,8 the bile concentration is also thought to contribute to cholesterol gallstone formation. In the current study, the Cldn2/ bile was concentrated with decreased water input and promoted cholesterol gallstone formation, by administration of lithogenic diet even though the CSI was not significantly different from the Cldn2þ/þ bile. We
Gastroenterology Vol. 147, No. 5
experimentally suggest that the bile concentration is a factor for cholesterol crystal nucleation. This study provides the first evidence that the bile content and volume are regulated at least in part by paracellular water permeability due to claudin 2, which suggests a new therapeutic target for preventing gallstone disease. In addition, the regulation of claudin 2based paracellular permeability may be a potential strategy for treating cholestasis due to various etiologies.
Supplementary Material Note: To access the supplementary material accompanying this article, visit the online version of Gastroenterology at www.gastrojournal.org, and at http://dx.doi.org/10.1053/ j.gastro.2014.07.033.
References 1. Nicolaou M, Andress EJ, Zolnerciks JK, et al. Canalicular ABC transporters and liver disease. J Pathol 2012; 226:300–315. 2. Esteller A. Physiology of bile secretion. World J Gastroenterol 2008;14:5641. 3. Portincasa P, Calamita G. Water channel proteins in bile formation and flow in health and disease: when immiscible becomes miscible. Mol Aspects Med 2012;33: 651–664. 4. Matsuo M. ATP-binding cassette proteins involved in glucose and lipid homeostasis. Biosci Biotechnol Biochem 2010;74:899–907. 5. Hirschfield GM, Heathcote EJ, Gershwin ME. Pathogenesis of cholestatic liver disease and therapeutic approaches. Gastroenterology 2010;139:1481–1496. 6. Maurer KJ, Carey MC, Fox JG. Roles of infection, inflammation, and the immune system in cholesterol gallstone formation. Gastroenterology 2009;136:425–440. 7. Stokes CS, Krawczyk M, Lammert F. Gallstones: environment, lifestyle and genes. Dig Dis 2011;29:191–201. 8. van Erpecum KJ, van Berge Henegouwen GP, Stoelwinder B, et al. Bile concentration is a key factor for nucleation of cholesterol crystals and cholesterol saturation index in gallbladder bile of gallstone patients. Hepatology 1990;11:1–6. 9. van Erpecum KJ. Biliary lipids, water and cholesterol gallstones. Biol Cell 2005;97:815–822. 10. Ballatori N, Truong AT. Glutathione as a primary osmotic driving force in hepatic bile formation. Am J Physiol 1992; 263:G617–G624. 11. Jessner W, Zsembery A, Graf J. Transcellular water transport in hepatobiliary secretion and role of aquaporins in liver. Wien Med Wochenschr 2008;158:565–569. 12. Huebert RC, Splinter PL, Garcia F, et al. Expression and localization of aquaporin water channels in rat hepatocytes. Evidence for a role in canalicular bile secretion. J Biol Chem 2002;277:22710–22717. 13. Tsukita S, Furuse M, Itoh M. Multifunctional strands in tight junctions. Nat Rev Mol Cell Biol 2001;2:285–293. 14. Gunzel D, Yu AS. Claudins and the modulation of tight junction permeability. Physiol Rev 2013;93:525–569.
15. Steed E, Balda MS, Matter K. Dynamics and functions of tight junctions. Trends Cell Biol 2010;20:142–149. 16. Mineta K, Yamamoto Y, Yamazaki Y, et al. Predicted expansion of the claudin multigene family. FEBS Lett 2011;585:606–612. 17. Krug SM, Gunzel D, Conrad MP, et al. Charge-selective claudin channels. Ann N Y Acad Sci 2012;1257:20–28. 18. Shen L, Weber CR, Raleigh DR, et al. Tight junction pore and leak pathways: a dynamic duo. Annu Rev Physiol 2011;73:283–309. 19. Furuse M, Furuse K, Sasaki H, et al. Conversion of zonulae occludentes from tight to leaky strand type by introducing claudin-2 into Madin-Darby canine kidney I cells. J Cell Biol 2001;153:263–272. 20. Muto S, Hata M, Taniguchi J, et al. Claudin-2-deficient mice are defective in the leaky and cation-selective paracellular permeability properties of renal proximal tubules. Proc Natl Acad Sci U S A 2010;107:801–8016. 21. Rosenthal R, Milatz S, Krug SM, et al. Claudin-2, a component of the tight junction, forms a paracellular water channel. J Cell Sci 2010;123:1913–1921. 22. Amasheh S, Meiri N, Gitter AH, et al. Claudin-2 expression induces cation-selective channels in tight junctions of epithelial cells. J Cell Sci 2002;115:4969–4976. 23. Tamura A, Hayashi H, Imasato M, et al. Loss of claudin-15, but not claudin-2, causes Naþ deficiency and glucose malabsorption in mouse small intestine. Gastroenterology 2011;140:913–923. 24. Tamura A, Kitano Y, Hata M, et al. Megaintestine in claudin-15-deficient mice. Gastroenterology 2008;134: 523–534. 25. Wada M, Tamura A, Takahashi N, et al. Loss of claudins 2 and 15 from mice causes defects in paracellular Naþ flow and nutrient transport in gut and leads to death from malnutrition. Gastroenterology 2013;144:369–380. 26. Hadj-Rabia S, Baala L, Vabres P, et al. Claudin-1 gene mutations in neonatal sclerosing cholangitis associated with ichthyosis: a tight junction disease. Gastroenterology 2004;127:1386–1390. 27. Yu AS, Cheng MH, Angelow S, et al. Molecular basis for cation selectivity in claudin-2-based paracellular pores: identification of an electrostatic interaction site. J Gen Physiol 2009;133:111–127. 28. Son S, Kojima T, Decaens C, et al. Knockdown of tight junction protein claudin-2 prevents bile canalicular formation in WIF-B9 cells. Histochem Cell Biol 2009; 131:411–424. 29. Yeh TH, Krauland L, Singh V, et al. Liver-specific betacatenin knockout mice have bile canalicular abnormalities, bile secretory defect, and intrahepatic cholestasis. Hepatology 2010;52:1410–1419. 30. Cova E, Gong A, Marinelli RA, et al. Water movement across rat bile duct units is transcellular and channelmediated. Hepatology 2001;34:456–463.
Role of Claudin 2 in Bile Flow and Gallstones 1145 31. Larocca MC, Soria LR, Espelt MV, et al. Knockdown of hepatocyte aquaporin-8 by RNA interference induces defective bile canalicular water transport. Am J Physiol Gastrointest Liver Physiol 2009;296:G93–G100. 32. Tabibian JH, Masyuk AI, Masyuk TV, et al. Physiology of cholangiocytes. Compr Physiol 2013;3:541–565. 33. Masyuk A, Marinelli R, Larusso N. Water transport by epithelia of the digestive tract. Gastroenterology 2002; 122:545–562. 34. Boyer JL. Bile duct epithelium: frontiers in transport physiology. Am J Physiol 1996;270:G1–G5. 35. Matsuzaki T, Tajika Y, Ablimit A, et al. Aquaporins in the digestive system. Med Electron Microsc 2004;37:71–80. 36. Yang B, Song Y, Zhao D, et al. Phenotype analysis of aquaporin-8 null mice. Am J Physiol Cell Physiol 2005; 288:C1161–C1170. 37. Mennone A, Verkman AS, Boyer JL. Unimpaired osmotic water permeability and fluid secretion in bile duct epithelia of AQP1 null mice. Am J Physiol Gastrointest Liver Physiol 2002;283:G739–G746. 38. Schnermann J, Huang Y, Mizel D. Fluid reabsorption in proximal convoluted tubules of mice with gene deletions of claudin-2 and/or aquaporin1. Am J Physiol Renal Physiol 2013;305:F1352–F1364. 39. Carey MC. Critical tables for calculating the cholesterol saturation of native bile. J Lipid Res 1978;19:945–955. 40. Paumgartner G. Biliary physiology and disease: reflections of a physician-scientist. Hepatology 2010;51: 1095–106. 41. Mitaka T. The current status of primary hepatocyte culture. Int J Exp Pathol 1998;79:393–409. 42. Beyenbach KW, Frömter E. Electrophysiological evidence for Cl secretion in shark renal proximal tubules. Am J Physiol 1985;248:F282–F295. Author names in bold designate shared co-first authorship. Received November 5, 2013. Accepted July 19, 2014. Reprint requests Address requests for reprints to: Atsushi Tamura, PhD, or Sachiko Tsukita, PhD, Laboratory of Biological Science, Graduate School of Frontier Biosciences, Osaka University, 2-2 Yamadaoka, Suita, Osaka 565-0871, Japan. e-mail: [email protected] or [email protected]; fax: 81-6-6879-3329. Acknowledgment The authors thank Professor K. W. Beyenbach (Cornell University) and all the members of our laboratories for helpful discussion. We also thank Drs Grace Gray and Leslie Meglietta for proofreading the manuscript. Conflicts of interest The authors disclose no conflicts. Funding This work was supported in part by Grants-in-Aid for Scientific Research (A) and Creative Scientific Research, and by CREST (Core Research for Evolutional Science and Technology) of Japan Science and Technology Agency to Sachiko Tsukita, and by Grant-in-Aid for Exploratory Research to Yuji Yamazaki, and by Scientific Research (C) to Atsushi Tamura from the Ministry of Education, Culture, and Sports, Science and Technology of Japan.
BASIC AND TRANSLATIONAL BILIARY
November 2014
1145.e1 Matsumoto et al
Supplementary Materials Animals Generation of claudin 2deficient mice was described previously20,23; these mice were backcrossed to C57BL/6J for more than 6 generations. All animal experiments were performed in accordance with protocols approved by the Animal Studies Committee of Osaka University School of Medicine.
Antibodies Rat anti–E-cadherin monoclonal antibody was generously provided by Dr M. Takeichi (Riken CDB, Kobe, Japan). Rat anti-occludin and rat anti-ZO-1 monoclonal antibodies were generated in our laboratory.25 Rabbit anti-claudin 1, 2, 3 (Invitrogen, Life Technologies, Carlsbad, CA), rabbit antiAQP1 and 8 (Merck Millipore, Billerica, MA), rabbit antiAQP9 (Abcam, Bristol, England) polyclonal antibodies, Alexa-488labeled anti-rabbit IgG, and Alexa-546labeled anti-rat IgG (Molecular Probes, Eugene, OR) were purchased.
H&E Staining and Light Microscopy Mouse liver and gallbladder were fixed in 10% formalin/phosphate-buffered saline, embedded in paraffin, cut into 8-mm–thick sections, and stained with H&E.24
Bile Flow Rate Measurements Mice were fasted for 4 hours with free access to water. After ligation of the cystic duct, the common bile duct was cannulated with a 30-gauge needle, and hepatic bile was collected for 1 hour to determine the flow rate and analyze the bile composition. During surgery and hepatic bile collection, the mice were under anesthesia and maintained at 37 C on a hot plate.
Biochemical Measurements of Bile
Gastroenterology Vol. 147, No. 5
Measurement of Water Movement Across Intrahepatic Bile Duct Units in Response to Osmotic Gradient The enclosed IBDUs were preincubated for 15 minutes in HEPES-based buffer (35 mM NaCl or 120 mM NaCl, osmolality 300 mOsm) at room temperature, then placed in hypotonic HEPES buffer (35 mM NaCl, osmolality 100 mOsm), or hypertonic HEPES buffer (120 mM NaCl, osmolality 500 mOsm). Water movement into or out of the enclosed IBDUs was estimated by measuring the change in luminal area using Image-J software on a 2-dimensional micrograph, using the method described for rat bile duct units.30
Canalicular Water Movement in Response to Osmotic Gradient Hepatocytes with bile canaliculus-like structures were preincubated using the method described for IBDUs.31 After a 15-minute incubation, the hepatocytes were placed in hypertonic HEPES buffer (120 mM NaCl, osmolality 500 mOsm). Water movement out of the canalicular space was estimated by the same method used for IBDUs.
Electrophysiological Measurements of Gallbladder Epithelium The gallbladder was cut, flattened, and mounted in an Ussing chamber with an exposed area of 0.03 cm2 to examine the transepithelial conductance and NaCl dilution potential, as described previously.25
Thin-Section Electron Microscopy and Scanning Electron Microscopy Samples were processed for and subjected to ultra-thin section electron microscopy and scanning electron microscopy as described previously.24
Freeze-Fracture Replica Electron Microscopy
Hepatic and gallbladder bile samples were collected. Ionized sodium and potassium were measured by a portable ion meter (Horiba, Kyoto, Japan). Total cholesterol, phospholipids, nonesterified fatty acids, total bile acids, and total bilirubin levels in the bile samples were also determined (Nagahama Life Science Laboratory, Shiga, Japan).
Cldn2þ/þ and Cldn2/ bile ducts were fixed in 2% paraformaldehyde and 2.5% glutaraldehyde in 0.1 M phosphate buffer, and processed for and subjected to freeze-fracture electron microscopy, as described previously.24
Intrahepatic Bile Duct Unit Isolation
Immunofluorescence Analysis
Portal veins were cannulated, and the livers were perfused with ethylene glycol-bis(b-aminoethyl ether)N,N,N0 ,N0 -tetraacetic acid for 10 minutes followed by 0.02% type II collagenase for 7 minutes at 37 C. The hepatic capsule was removed, and the portal tissue was mechanically dissociated from the hepatic parenchymal tissue in cold Hank’s balanced salt solution as described previously.30 The portal tissue was then cut into 2- to 3-mm segments and cultured in Dulbecco’s modified Eagle medium for 12 hours. After incubation, both ends of the IBDU had sealed, enclosing the lumen.
Mouse liver and gallbladder were dissected and frozen in liquid N2. Frozen sections were cut and processed for, and subjected to, indirect immunofluorescence analysis, as described previously.25
Reverse Transcription Polymerase Chain Reaction and Quantitative Real-Time Reverse Transcription Polymerase Chain Reaction Reverse transcription polymerase chain reaction and quantitative real-time reverse-transcription polymerase chain reaction were performed as described previously.24
November 2014
Serum Biochemical Analysis Blood was collected from the inferior vena cava of mice after a 4-hour fast, and analyzed (Nagahama Life Science Laboratory, Shiga, Japan).
Role of Claudin 2 in Bile Flow and Gallstones 1145.e2
pipette coated with Sylgard 184. Through one barrel of the pipette, a 500-nA current was applied to the lumen, the transepithelial voltage was measured through the other barrel of the pipette, and an electrode placed near the proximal side of the duct in the bath.
Primary Culture of Hepatocytes The primary culture of hepatocytes was performed using the method described for rat hepatocytes.41 Briefly, mice were anesthetized and the liver was perfused with 5 mL/min HEPES-based buffer containing 0.5 mM ethylene glycol-bis(b-aminoethyl ether)-N,N,N0 ,N0 tetraacetic acid for 10 minutes at 37 C, and then with 5 mL/ min collagenase A buffer (HEPES-based buffer with 0.5 mg/ mL collagenase A) for 7 minutes at 37 C. The perfused liver was cut into pieces and passed through 70-mm nylon mesh. The filtrate was centrifuged at 50g for 5 minutes at 4 C. The hepatocytes in the pellet were then mixed with Percoll solution and centrifuged at 50g for 20 minutes to remove dead cells. These hepatocytes were seeded on 35-mm glassbottom dishes coated with type IA collagen and cultured for 24 hours at 37 C, as described. After this incubation, bile canaliculi-like structures had formed between 2 cells, the interior of which was enclosed from the extracellular buffer.
Electrophysiological Measurements of the Bile Duct The electrophysiological parameters of the common bile duct were measured using the method described for shark renal tubules as a reference.42 The diagram of the circuit is shown in Figure 6B. Briefly, the common bile duct was gently isolated, then the proximal side of the duct was cannulated with a double-barreled perfusion pipette and tied gently. The distal end of the duct was sucked into a
Lithogenic Diet
Cldn2þ/þ and Cldn2/ mice (n ¼ 14 each), at 8–12 weeks of age, were fed a lithogenic diet (containing 19% butter fat, 1.25% cholesterol, and 0.5% cholic acid) for 4 weeks. Mice that lost body weight were excluded (3 Cldn2þ/þ mice and 2 Cldn2/ mice). Although no gallstones formed in the wildtype mice after 4 weeks on a lithogenic diet, gallstones were found in about half the wild-type mice after 8 weeks on the diet. Experiments were performed after the mice were fasted for 4 hours.
Human Liver Samples All the liver samples were obtained from noncancerous part of the tissue resected for hepatocellular carcinoma. The presence of gallstones was examined by computed tomography images acquired before surgical resection. There were not significant difference in age, sex, etiology, and histology between the control group (n ¼ 15) and gallstones group (n ¼ 15) (Supplementary Table 1). The use of liver samples for this study was approved by the ethical committee of Osaka University Hospital and Osaka University Graduate School of Medicine.
Statistical Analysis Results are expressed as mean ± SEM. P values were calculated using an independent t-test. P < .05 was considered statistically significant.
1145.e3 Matsumoto et al
Gastroenterology Vol. 147, No. 5
Supplementary Figure 1. Localization and expression of claudins in the Cldn2þ/þ and Cldn2/ gallbladder. (A) Immunofluorescent images of claudin 1, 2, and 3 in the gallbladder. Bar ¼ 25 mm. (B) Quantitative expression levels of representative claudins in the Cldn2þ/þ and Cldn2/ (i) gallbladder and (ii) common bile duct. E-cad, E-cadherin.
November 2014
Supplementary Figure 2. Localization and expression of claudins in the Cldn2þ/þ and Cldn2/ intrahepatic and common bile ducts. Immunofluorescent images of claudin 2, 3, and 7 in the (A) intrahepatic bile ducts and (B) common bile ducts. Bar ¼ 20 mm.
Role of Claudin 2 in Bile Flow and Gallstones 1145.e4
1145.e5 Matsumoto et al
Gastroenterology Vol. 147, No. 5
Supplementary Figure 3. Immunofluorescent staining in the liver. Immunofluorescent images of occludin, ZO-1, and E-cadherin (E-cad) in the liver. No differences were detected between Cldn2þ/þ and Cldn2/. Bars ¼ 50 mm.
November 2014
Supplementary Figure 4. Histologic analysis of the liver and gallbladder. (A) Scanning electron microscopic images (upper panels for low magnifications; bars ¼ 20 mm. Middle panels for high magnifications; bars ¼ 10 mm) and immunofluorescent images (lower panels; bars ¼ 10 mm) of bile canaliculi in the liver. The canalicular diameters in the perivenous region were smaller than in the periportal region, but there were no differences between Cldn2þ/þ and Cldn2/. (B) H&E-stained images of the Cldn2þ/þ and Cldn2/ gallbladder. Bars ¼ 50 mm.
Role of Claudin 2 in Bile Flow and Gallstones 1145.e6
1145.e7 Matsumoto et al
Gastroenterology Vol. 147, No. 5
Supplementary Figure 5. Biochemical analyses of serum and hepatic lipids in Cldn2þ/þ and Cldn2/ mice. (A) Serum concentrations of aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), and lipids (total cholesterol [TCHO], triglyceride [TG], phospholipid [PL]) (n ¼ 6). (B) Concentrations of intrahepatic lipids (T-CHO, TG, nonesterified fatty acids [NEFA], total bile acids [TBA]) (n ¼ 10). (C) Quantitative expression levels of representative lipids and bile acid transporters in the Cldn2þ/þ and Cldn2/ liver.
November 2014
Supplementary Figure 6. Localization and expression of AQPs in the Cldn2þ/þ and Cldn2/ liver. (A) Immunofluorescent images of AQP 1, 8, and 9 in the liver. Bar ¼ 50 mm. (B) (i) Expression of major Aqp genes in the Cldn2þ/þ and Cldn2/ liver, examined by reverse transcription polymerase chain reaction (RT-PCR). (ii) Immunoblotting for AQP8 and 9 in the liver of Cldn2þ/þ and Cldn2/ mice.
Role of Claudin 2 in Bile Flow and Gallstones 1145.e8
1145.e9 Matsumoto et al
Gastroenterology Vol. 147, No. 5
Supplementary Figure 7. Gene expressions in the liver, gallbladder, and bile duct. Quantitative expression levels of representative claudins in the (A) liver, (B) gallbladder, and (C) common bile duct of Cldn2þ/þ mice under a control diet or lithogenic diet for 4 weeks. *P < .05
November 2014
Role of Claudin 2 in Bile Flow and Gallstones 1145.e10
Supplementary Figure 8. H&E-stained images of the gallbladder and liver of Cldn2þ/þ and Cldn2/ mice after 4 weeks on a lithogenic diet or control diet. (A) H&Estained images of the gallbladder of Cldn2þ/þ and Cldn2/ mice under a lithogenic diet. The gallbladder wall of the Cldn2/ mice was thickened compared with that of the Cldn2þ/þ mice. Bars ¼ 50 mm. (B) H&E-stained images of the under the lithogenic diet. No differences were detected. Bar ¼ 50 mm.
Supplementary Table 1.The Bile Canalicular Diameter in Perivenous and Periportal Zone in Mouse Liver
Perivenous (mm) Periportal (mm)
Cldn2þ/þ
Cldn2/
P value
0.79 ± 0.03 0.92 ± 0.03
0.78 ± 0.04 0.90 ± 0.02
.84 .66
NOTE. The width of bile canaliculi of perivenous zone and periportal zone was similar between in Cldn2þ/þ and Cldn2/ mice, respectively.
Supplementary Table 2.Sex, Age, Etiologic and Histologic Profiles of Human Liver
Sex (men/women) Age Etiology (HBV/HCV) Histology (CH/LC)
Control
Gallstone
3/12 61.2 ± 2.5 4/11 7/8
1/14 63 ± 2.8 4/10 4/8
NOTE. There were not significant difference in age, sex, etiology, and histology between controls group and gallstone group.