endotoxin increases paracellular permeability of isolated rat hepatocyte couplets

Hepatology Research 20 (2001) 144–154

Endotoxin increases paracellular permeability ofisolated rat hepatocyte couplets

Takumi Kawaguchi a, Shotaro Sakisaka a, Masaru Harada a,*,Shinichiro Hanada a, Eitaro Taniguchi a, Hironori Koga a,

Kurumi Sasatomi a, Kyuichi Tanikawa b, Michio Sata a

a Second Department of Medicine, Kurume Uni6ersity School of Medicine, 67 Asahi-machi,Kurume 830-0011, Japan

b International Institute for Li6er Research, Kurume Research Center, Kurume, Japan

Received 2 June 2000; received in revised form 21 September 2000; accepted 2 October 2000

Abstract

Hyperbilirubinemia is frequently associated with endotoxemia. Regurgitation of bileconstituents including bilirubin into the sinusoidal space is prevented by tight junctionswhich maintain paracellular permeability between hepatocytes. To investigate the mechanismof endotoxin-associated hyperbilirubinemia, we have studied the changes in paracellularpermeability of primary hepatocyte couplets treated with endotoxin. In addition, we exam-ined the effects of ursodeoxycholic acid (UDCA), which has been widely used for variousliver diseases, on endotoxin-associated changes in paracellular permeability. The paracellularpermeability of hepatocyte couplets was evaluated by paracellular penetration of fluoresceinisothiocyanate (FITC)-dextran with molecular weights of 3, 10 and 70K using confocal laserscanning microscopy. Endotoxin increased the paracellular penetration of FITC-dextran 3and 10K. These changes were prevented by treatment with UDCA. There was littleparacellular penetration of FITC-dextran 70K under any conditions. These results suggestedthat endotoxin increased the paracellular permeability of hepatocyte couplets and thesechanges were prevented by treatment with UDCA. Furthermore, bile regurgitation throughthe paracellular route is involved in endotoxin-associated hyperbilirubinemia, and UDCAmight be a potential therapeutic agent for endotoxin-associated hyperbilirubinemia. © 2001Elsevier Science Ireland Ltd. All rights reserved.

www.elsevier.com/locate/ihepcom

* Corresponding author. Tel.: +81-942-317561; fax: +81-942-342623.E-mail address: [email protected] (M. Harada).

1386-6346/01/$ - see front matter © 2001 Elsevier Science Ireland Ltd. All rights reserved.

PII: S1386 -6346 (00 )00129 -7

T. Kawaguchi et al. / Hepatology Research 20 (2001) 144–154 145

Keywords: Endotoxin; Isolated rat hepatocyte couplets; Paracellular permeability; Ursodeoxycholic acid.

1. Introduction

Hyperbilirubinemia is frequently seen in patients with endotoxemia [1,2]. Variousfactors such as the impaired hepatocyte transport of organic anions [3–5] areinvolved in cholestasis. Recently, we demonstrated that endotoxin-associatedcholestasis is caused by change in structure and function of hepatocyte tightjunctions, which regulate the paracellular permeability of hepatocytes, in experi-mental rat model [5].

Multidrug resistance protein 2 (mrp2) is a canalicular multispecific organic aniontransport protein and excretes conjugated bilirubin into the bile canaliculi [6].Endotoxin-induced cholestasis involves a retrieval of mrp2 from the canalicularmembrane to subcanalicular regions and a down-regulation of mrp2 expression [7].Thus, impaired function of mrp2 is one of the causative factors of hyperbilirubine-mia [8].

The paracellular permeability of hepatocytes is regulated by tight junctions, theonly intercellular barrier between the sinusoidal and the canalicular spaces [5,9].Increased paracellular permeability of hepatocytes allows paracellular reflux of bileconstituents including conjugated bilirubin into the sinusoidal spaces [3]. In sometypes of hyperbilirubinemia, the paracellular permeability of hepatocytes is in-creased [5,9], however, little is known about the paracellular permeability ofhepatocytes in endotoxin-induced hyperbilirubinemia.

Ursodeoxycholic acid (UDCA) is used to treat some types of cholestasis [10].Although effects of UDCA on hepatocyte function have been reported [10,11],those on the paracellular permeability of hepatocytes have never been evaluated.

The goals of this study are to investigate the effect of endotoxin on paracellularpermeability and to examine the effect of UDCA on endotoxin-induced changes ofparacellular permeability in isolated rat hepatocyte couplets, which possess cellpolarity and function in bile formation.

2. Materials and methods

2.1. Materials

All reagents were purchased from Wako Pure Chemical Industries (Osaka,Japan) unless otherwise indicated.

2.2. Animals

Male Wistar rats, each weighing 250 g were maintained on a standard diet andwater ad libitum. All rat experiments were conducted in accordance with the USNational Institutes of Health Guidelines for the Care and Use of Laboratory

T. Kawaguchi et al. / Hepatology Research 20 (2001) 144–154146

Animals and were approved by the Kurume University Institutional Animal Careand Use Committee. Five rats were used in each experiment.

2.3. Preparation of isolated rat hepatocyte couplets

Isolated rat primary hepatocyte couplets were obtained from rat livers by thecollagenase perfusion technique [12] with some modification. The liver was firstpre-perfused with Ca2+-free Hank’s solution supplemented with 5 mM EDTA and5 mM glucose at 30 ml/min for three different times; 5, 10 and 15 min, followed by0.0125% collagenase solution. After 7 min of collagenase perfusion at 30 ml/min,each liver was excised, dispersed in Hank’s solution, and the resulting cell suspen-sion was filtered through 300-gauge mesh. Cells (5×105) were seeded into 35-mmculture dishes (Becton Dickinson and Company, NJ), and were grown in William’smedium E (Flow Laboratories, Irvine, Scotland) with 0.1 mg/ml streptomycin(Streptomycin Sulfate, Meiji Seika Kaisya, Tokyo, Japan), 100 IU/ml penicillin(Crystalline Penicillin G Meiji, Meiji Seika Kaisya), and 10% fetal calf serum. Cellswere incubated at 37°C. Four hours after plating, cell viability examined by trypanblue exclusion exceeded 95%.

2.4. Cell culture

Isolated rat hepatocyte couplets, pre-perfused for 5 min, were used for evaluationin effects of endotoxin, endotoxin and anti-rat TNF-a antibody, or endotoxin andUDCA administrations on paracellular permeability. After 2 h incubation, endo-toxin (lipopolysaccharide from Escherichia coli 0111: B4, 100 mg/ml; Sigma, St.Louis, MO) alone, endotoxin and anti-rat TNF-a antibody (4 mg/ml; Genzyme,Cambridge, MA), or endotoxin and UDCA (50 mg/ml) were added. After 2 hincubation under each of four conditions, cell viability examined by trypan blueexclusion exceeded 95% and paracellular permeability of hepatocyte couplets wasexamined by paracellular penetration assay [13]. In addition, the medium level oflactate dehydrogenase (LDH) was measured in each condition.

2.5. Measurement of paracellular permeability of hepatocyte couplets

Isolated rat hepatocyte couplets with expanded canalicular spaces were identifieda priori. Then, each medium was replaced with medium containing fluoresceinisothiocyanate (FITC)-dextran with a molecular weight of 3, 10 or 70K (0.5 mg/ml;Molecular Probe, Eugene, OR). Paracellular penetration of FITC-dextran into thebile canalicular space was examined in 3 min using a confocal laser-scanningmicroscope (LSM-GB200; Olympus, Tokyo, Japan). When hepatocyte coupletsmaintained paracellular permeability, the canalicular space was absent with FITC-dextran (Fig. 1A). FITC-dextran was seen within the canalicular space, when theparacellular permeability of hepatocyte couplets increased (Fig. 1C). The rate ofcanalicular accumulation of FITC-dextran was expressed as the percentage ofhepatocyte couplets, having canalicular accumulation of FITC-dextran, of the total


counted hepatocyte couplets. Fifty hepatocyte couplets from five rats were analyzedin each experiment. The gain level was adjusted to an undetectable level ofautofluorescence of hepatocytes.

2.6. Statistical Analysis

All data were expressed as the mean9S.E. Comparisons among multiple groupswere analyzed using the Kruskal–Wallis analysis of variance. Differences betweentwo groups were analyzed by the Mann–Whitney U test. A P value B0.05 wasconsidered statistically significant.

3. Results

3.1. Effect of pre-perfusion time on paracellular permeability of isolated rathepatocyte couplets

Canalicular accumulation rates of FITC-dextran 3 (Fig. 2A) and 10K (Fig. 2B)

Fig. 1. Representative photograph of maintained and increased paracellular permeability in isolated rathepatocyte couplets incubated with FITC-dextran for 3 min. (A) FITC-dextran was absent in the bilecanaliculus, when paracellular permeability was maintained. (B) Transmission image of the couplet seenin (A). (C) FITC-dextran was present in the bile canaliculus (arrow), when paracellular permeabilityincreased. (D) Transmission image of the couplet seen in (C).


Fig. 2. Effect of pre-perfusion time on paracellular permeability of hepatocyte couplets. (A) Paracellularpermeability to FITC-dextran 3K. (B) Paracellular permeability to FITC-dextran 10K. (C) Paracellularpermeability to FITC-dextran 70K. Values are expressed as mean9S.E. (n=50 hepatocyte coupletsfrom five rats for each group). 5 min, 5 min pre-perfused hepatocyte couplets; 10 min, 10 minpre-perfused hepatocyte couplets; 15 min, 15 min pre-perfused hepatocyte couplets. * PB0.01.


were significantly increased in 10 and 15 min compared with 5 min pre-perfusedhepatocyte couplets. The canalicular accumulation rate of FITC-dextran 70K wasless than 10% in any pre-perfusion condition (Fig. 2C).

3.2. Effect of endotoxin on LDH le6els in medium

There were no significant changes in medium levels of LDH between control andendotoxin-administered hepatocytes (Fig. 3).

3.3. Effect of endotoxin on paracellular permeability of hepatocyte couplets

Endotoxin administration significantly increased the canalicular accumulationrate of FITC-dextran 3 and 10K compared with those in control (Fig. 4).There were no significant changes in the canalicular accumulation rate ofFITC-dextran 70K between control and endotoxin-administered hepatocyte cou-plets (Fig. 4).

3.4. Effects of anti-rat TNF-a antibody or UDCA on endotoxin-induced change inparacellular permeability of hepatocyte couplets

Administration of anti-rat TNF-a antibody with endotoxin produced no signifi-cant changes in the canalicular accumulation rates of FITC-dextran 3 and 10Kcompared with those in endotoxin-administered hepatocyte couplets (Fig. 5).Administration of UDCA with endotoxin significantly decreased canalicular accu-

Fig. 3. Effect of endotoxin on LDH levels in medium; Values are expressed as mean9S.E. (n=6 foreach group). Control, hepatocyte couplets incubated with William’s medium E. Endotoxin, hepatocytecouplets incubated with William’s medium E containing endotoxin (100 mg/ml).


Fig. 4. Effect of endotoxin on paracellular permeability of hepatocyte couplets. Values are expressed asmean9S.E. (n=50 hepatocyte couplets from five rats for each group). Control, hepatocyte coupletsincubated with William’s medium E. Endotoxin, hepatocyte couplets incubated with William’s mediumE containing endotoxin (100 mg/ml). * PB0.01.

mulation rates of FITC-dextran 3 and 10K compared with those in endotoxin-ad-ministered hepatocyte couplets (Fig. 5).

4. Discussion

In the present study, we demonstrated that endotoxin increased the paracellularpermeability of isolated rat hepatocyte couplets and UDCA prevented the endo-toxin-induced increase in paracellular permeability.

Extracellular Ca2+ is required for tight junctional integrity [14]. Low Ca2+

medium affects the paracellular permeability of hepatocyte couplets [15]. In apreliminary study, we examined the effect of pre-perfusion time on the paracellularpermeability of hepatocyte couplets, since Ca2+-free EDTA containing Hank’ssolution was used for pre-perfusion process in preparation of hepatocyte couplets.Canalicular accumulation rates of FITC-dextran 3 and 10K in 5 min pre-perfusedhepatocyte couplets were significantly lower than those in 10 and 15 min pre-per-fused hepatocyte couplets, indicating that paracellular permeability was mostmaintained in 5 min pre-perfused hepatocyte couplets in these conditions. Althoughsome canalicular accumulation of FITC-dextran were caused by disruption of TJdue to dynamic canalicular contraction [16], here we revealed that pre-perfusionwith Ca2+-free EDTA containing Hank’s solution affected the paracellular perme-ability of hepatocyte couplets. Therefore, we applied 5 min pre-perfused hepatocytecouplets in the following studies.


Isolated rat hepatocyte couplets are primary bile secretory units and have theadvantage that they are avoid of the confounding influences of vascular perfusionand bile duct epithelial transport [17]. The paracellular permeability of hepatocytecouplets has been evaluated by measuring electrical resistance [18]. However, it ispossible that reduction of electrical resistance could be caused by increased plasmamembrane permeability to ions. Therefore, paracellular permeability was evaluatedby canalicular accumulation rate of FITC-dextran in this study [13] and weperformed experiments in 3 min, to exclude the involvement of the very earlytranscytotic vesicle pathway (4–6 min) [19]. In addition, this assay also evaluatesthe extent of increased paracellular permeability by changing of dextran MW.

Impaired function of hepatic organic anion transporters is part of the mecha-nisms of endotoxin-associated hyperbilirubinemia [3–5]. We have demonstratedthat endotoxin increased the paracellular permeability of hepatocyte couplets toFITC-dextran 3 and 10K. Because the MW of bilirubin is less than 3K, our findingssuggested that endotoxin affected hepatocyte tight junction with a subsequent bilereflux into the sinusoidal space from the canalicular space and induced hyperbiliru-binemia. The increased paracellular permeability was unlikely as a result of severecell injury, because paracellular permeability of hepatocyte couplets to FITC-dex-tran 70K was not affected by treatment with endotoxin, and there was no differencein medium LDH levels between control and endotoxin-treated groups. Recently wedemonstrated that increased paracellular permeability of hepatocytes and alter-ations in hepatocyte tight junctions in a rat model of endotoxin-associated cholesta-sis [5]. Collectively, endotoxin-associated hyperbilirubinemia may result from notonly impairment of mrp2 [3–5] but also increased paracellular permeability ofhepatocytes.

Fig. 5. Effects of anti-rat TNF-a antibody or UDCA on the endotoxin-induced increase in paracellularpermeability of hepatocyte couplets. Values are expressed as mean9S.E. (n=50 hepatocyte coupletsfrom five rats for each group). Endotoxin, hepatocyte couplets incubated with William’s medium Econtaining endotoxin (100 mg/ml); anti TNF-a antibody, hepatocyte couplets incubated with William’smedium E containing a mixture of endotoxin and anti-rat TNF-a antibody (4 mg/ml); UDCA,hepatocyte couplets incubated with William’s medium E containing a mixture of endotoxin and UDCA(50 mg/ml). * PB0.01.


Endotoxin stimulates Kupffer cells to release TNF-a [20], which increases para-cellular permeability [21]. Kupffer cells were hardly seen in isolated hepatocytesuspension by microscopic observation (data not shown). With only this finding,however, it cannot be excluded that the possibility of involvement of TNF-a fromcontaminated Kupffer cells in endotoxin-induced increase in paracellular permeabil-ity. To exclude this possibility, we used anti-rat TNF-a antibody. Treatment withanti-rat TNF-a antibody did not prevent the increase in paracellular permeability.Since endotoxin is trapped by both Kupffer cells and hepatocytes [22], thesefindings suggest that endotoxin directly affects the function of hepatocyte tightjunctions. Although it is still unknown how endotoxin increases the paracellularpermeability of hepatocyte couplets, one would assume that endotoxin disturbsCa2+ homeostasis in hepatocytes [23]. Intracellular Ca2+ has been reported to beimportant in the regulation of paracellular permeability [14,15].

UDCA is widely used for treatment with cholestasis [10,11]. However, it hasremained uncertain whether UDCA affects the paracellular permeability of hepato-cytes. Here we showed UDCA prevented the increase in paracellular permeabilityassociated with endotoxin. The mechanisms that underlie the effects of UDCA onparacellular permeability are uncertain. Two possibilities, however, exist. UDCA isknown to activate extracellular signal-related (ERK) mitogen-activated proteinkinases signal transduction pathway, which is associated for formation of tightjunctions [24]. Activation of ERK signal transduction pathway leads increase inboth transepithelial electrical resistance [24] and bile acid secretion [25]. Alterna-tively, UDCA enhances the transport of endotoxin across the hepatocytes fromblood to bile [26]. Thus, UDCA may prevent hepatocyte damage by activatingERK signaling pathway and/or enhancing the excretion of endotoxin into bile. Thedetailed intracellular mechanisms of UDCA action on the impaired paracellularpermeability should be further elucidated.

In conclusion, we have demonstrated that endotoxin increased the paracellularpermeability of hepatocyte couplets and UDCA prevented the increase in paracellu-lar permeability associated with endotoxin. These findings suggest that increasedparacellular permeability may contribute to endotoxin-associatedhyperbilirubinemia.

Acknowledgements

We thank Kaori Maeda for technical assistance.

References

[1] Franson TR, Hierholzer WJ, Jr, LaBrecque DR. Frequency and characteristics of hyperbilirubine-mia associated with bacteremia. Rev Infect Dis 1985;7:1–9.

[2] Pirovino M, Meister F, Rubli E, Karlaganis G. Preserved cytosolic and synthetic liver function injaundice of severe extrahepatic infection. Gastroenterology 1989;96:1589–95.


[3] Trauner M, Meier PJ, Boyer JL. Molecular pathogenesis of cholestasis. N Engl J Med1998;339:1217–27.

[4] Trauner M, Arrese M, Soroka CJ, et al. The rat canalicular conjugate export pump (Mrp2) isdown-regulated in intrahepatic and obstructive cholestasis. Gastroenterology 1997;113:255–64.

[5] Kawaguchi T, Sakisaka S, Mitsuyama K, et al. Cholestasis with altered structure and function ofhepatocyte tight junction and decreased expression of canalicular multispecific organic aniontransporter in a rat model of colitis. Hepatology 2000;31:1285–95.

[6] Keppler D, Cui Y, Konig J, Leier I, Nies A. Export pumps for anionic conjugates encoded by MRPgenes. Adv Enzyme Regul 1999;39:237–46.

[7] Kubitz R, Wettstein M, Warskulat U, Haussinger D. Regulation of the multidrug resistance protein2 in the rat liver by lipopolysaccharide and dexamethasone. Gastroenterology 1999;116:401–10.

[8] Toh S, Wada M, Uchiumi T, et al. Genomic structure of the canalicular multispecific organicanion-transporter gene (MRP2/cMOAT) and mutations in the ATP-binding-cassette region inDubin–Johnson syndrome. Am J Hum Genet 1999;64:739–46.

[9] Kawaguchi T, Sakisaka S, Sata M, Mori M, Tanikawa K. Different lobular distributions of alteredhepatocyte tight junctions in rat models of intrahepatic and extrahepatic cholestasis. Hepatology1999;29:205–16.

[10] Poupon R, Poupon RE. Ursodeoxycholic acid therapy of chronic cholestatic conditions in adultsand children. Pharmacol Ther 1995;66:1–15.

[11] Cirillo NW, Zwas FR. Ursodeoxycholic acid in the treatment of chronic liver disease. Am JGastroenterol 1994;89:1447–52.

[12] Harada M, Sakisaka S, Yoshitake M, et al. Role of cytoskeleton and acidification of endocyticcompartment in asialoglycoprotein metabolism in isolated rat hepatocyte couplets. Hepatology1995;21:1413–21.

[13] Burgstahler AD, Nathanson MH. NO modulates the apicolateral cytoskeleton of isolatedhepatocytes by a PKC-dependent, cGMP-independent mechanism. Am J Physiol 1995;269:G789–99.

[14] Roma MG, Ahmed-Choudhury J, Coleman R. The protein kinase inhibitor 1-(5-isoquinolinylsul-fonyl)-2-methyl-piperzine (H-7) prevents and reverses Ca(2+ )-mediated injury in isolated rathepatocyte couplets. Toxicol Appl Pharmacol 1999;161:192–201.

[15] Anwer MS, Clayton LM. Role of extracellular Ca2+ in hepatic bile formation and taurocholatetransport. Am J Physiol 1985;249:G711–8.

[16] Watanabe S, Tomono M, Takeuchi M, et al. Bile canalicular contraction in the isolated hepatocytedoublet is related to an increase in cytosolic free calcium ion concentration. Liver 1988;8:178–83.

[17] Boyer JL. Isolated rat hepatocyte couplets. In: Tavoloni N, Berk PD, editors. Hepatic Transportand Bile Secretion. New York: Raven Press, 1993:597–606.

[18] Graf J, Henderson RM, Krumpholz B, Boyer JL. Cell membrane and transepithelial voltages andresistances in isolated rat hepatocyte couplets. J Membr Biol 1987;95:241–54.

[19] Hayakawa T, Ng OC, Ma A, Boyer JL. Taurocholate stimulates transcytotic vesicular pathwayslabeled by horseradish peroxidase in the isolated perfused rat liver. Gastroenterology 1990;99:216–28.

[20] Whiting JF, Green RM, Rosenbluth AB, Gollan JL. Tumor necrosis factor-alpha decreaseshepatocyte bile salt uptake and mediates endotoxin-induced cholestasis. Hepatology 1995;22:1273–8.

[21] Marano CW, Lewis SA, Garulacan LA, Soler AP, Mullin JM. Tumor necrosis factor-alphaincreases sodium and chloride conductance across the tight junction of CACO-2 BBE, a humanintestinal epithelial cell line. J Membr Biol 1998;161:263–74.

[22] Mimura Y, Sakisaka S, Harada M, Sata M, Tanikawa K. Role of hepatocytes in direct clearanceof lipopolysaccharide in rats. Gastroenterology 1995;109:1969–76.

[23] Spitzer JA, Deaciuc IV. IP3-dependent Ca2+ release in permeabilized hepatocytes of endotoxemicand septic rats. Am J Physiol 1987;253:E130–4.


[24] Kinugasa T, Sakaguchi T, Gu X, et al. Claudins regulate the intestinal barrier in response toimmune mediators. Gastroenterology 2000;118:1001–11.

[25] Schliess F, Kurz AK, vom Dahl S, et al. Mitogen-activated protein kinases mediate the stimulationof bile acid secretion by tauroursodeoxycholate in rat liver. Gastroenterology 1997;113:1306–14.

[26] Hori Y, Ohyanagi H. Protective effect of the intravenous administration of ursodeoxycholic acidagainst endotoxemia in rats with obstructive jaundice. Surg Today 1997;27:140–4.

.

endotoxin increases paracellular permeability of isolated rat hepatocyte couplets

Documents