SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

The mechanisms underlying intrahepatic vasoconstriction are not fully elucidated. Here we investigated the Kupffer cell (KC)-dependent increase in portal pressure by way of actions of vasoconstrictive cysteinyl leukotrienes (Cys-LTs). Liver cirrhosis was induced in rats by bile duct ligation (BDL for 4 weeks; controls: sham-operation) and thioacetamide application (18 weeks). Infusion of leukotriene (LT) C4 or LTD4 in isolated perfused livers (20 nM, BDL and sham) demonstrated that LTC4 is a more relevant vasoconstrictor. In BDL animals the Cys-LT1 receptor inhibitor montelukast (1 μM) reduced the maximal portal perfusion pressure following LTC4 or LTD4 infusion. The infusion of LTC4 or D4in vivo (15 μg/kg b.w.) confirmed LTC4 as the more relevant vasoconstrictor. Activation of KCs with zymosan (150 μg/mL) in isolated perfused BDL livers increased the portal perfusion pressure markedly, which was attenuated by LT receptor blockade (Ly171883, 20 μM). Cys-LTs in the effluent perfusate increased with KC activation but less with additional blockade of KCs with gadolinium chloride (10 mg/kg body weight, 48 and 24 hours pretreatment). KCs were isolated from normal rat livers and activated with zymosan or lipopolysaccharide at different timepoints. This resulted in an increase in Cys-LT production that was not influenced by preincubation with montelukast (1 μM). Infusion of LTC4 (20 nM) and the thromboxane analog U46619 (0.1 μM) further enhanced portal pressure, indicating additive effects. Treatment with montelukast for 10 days resulted in an impressive reduction in the basal portal pressure and an attenuation of the KC-dependent increase in portal pressure. Conclusion: Activation of isolated KCs produced Cys-LTs. Infusion of Cys-LTs increased portal pressure and, vice versa, treatment with montelukast reduced portal pressure in rat liver cirrhosis. Therefore, montelukast may be of therapeutic benefit for patients with portal hypertension. (HEPATOLOGY 2010)

In patients with liver cirrhosis the occurrence of esophageal and gastric varices is associated with a high risk of bleeding episodes, which can be life-threatening.1 Therefore, the prevention and intervention strategies currently in place must be expanded. However, the development of new therapies is dependent on an improved understanding of the pathophysiology of portal hypertension.

Thus far, it is well accepted that portal hypertension results from a rise in intrahepatic resistance coupled with an increase in portal blood flow.2 The morphological and molecular changes that the cirrhotic liver undergoes form the basis for subsequent alterations in the generation, degradation, and the response to vasodilators and vasoconstrictors. Hepatic vascular tone rearrangements originate from an increase in the levels of vasoconstrictors, such as thromboxane (TX) A23-6 or cysteinyl leukotrienes (Cys-LTs, leukotrienes C4, D4, E4),7, 8 and a decrease in the levels of vasodilators.9-13 One mechanism in addition to the many others for the intrahepatic hyperresponsiveness to vasoconstrictors in the cirrhotic liver is the up-regulation of Rho kinase-mediated contraction.14-18

Different studies demonstrated a close association between bacterial infections and variceal bleeding.19-21 Indeed, recent studies suggested that Kupffer cell (KC) activation may lead to portal hypertension in cirrhosis by production of intrahepatic vasoconstrictors.4, 5 However, the role of KC-derived vasoconstrictors for increased intrahepatic vascular resistance of the cirrhotic liver is insufficiently understood to date. Therefore, we here investigated the role of KC-derived LTs in the elevation of portal pressure. Montelukast is an antagonist of the Cys-LT1 receptor and has been proven effective in the treatment of asthma.22 Here we investigated the role of LTs for intrahepatic vascular resistance and the effect of montelukast on portal hypertension in rats with experimentally induced cirrhosis.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Animal Models.

All animals were ethically treated according to the criteria established by the National Academy of Sciences and published by the National Institutes of Health, in addition to the legal requirements of Germany. All animal experiments were approved by the local government (Regierung von Oberbayern, Munich, Germany) and were reported to the responsible authorities every 3 months.

Induction of Liver Fibrosis by Bile Duct Ligation (BDL).

Male Sprague-Dawley rats (Charles River, Sulzfeld, Germany), weighing 180-200 g, were anesthetized by intraperitoneal injection of pentobarbital (50 mg/kg body weight [b.w.]). A midline laparotomy was performed and the common bile duct was ligated twice with 3-0 silk and cut between the two ligations.5 Sham-operated rats were subjected to laparotomy without BDL. Perfusion of the liver with Krebs-Henseleit buffer and all in vivo experiments were performed 4 weeks after BDL, as described below.

Induction of Liver Cirrhosis by Thioacetamide (TAA).

Rats were treated with TAA for 18 weeks.23 Briefly, 200-250 g rats were weighed once per week. A starting dose of 0.3% TAA was added to their drinking water. If body weight varied by 20% or more from one week to the next, the dose of TAA was adapted as described.23

In Situ Rat Liver Perfusion Studies.

Rats were anesthetized by intraperitoneal injection of sodium pentobarbital (50 mg/kg b.w.). After incision of the abdominal wall, ≈3-4 mL of blood were taken from the inferior vena cava in order to determine serum parameters. The portal vein was then cannulated with a 14G Teflon intravenous catheter and the liver was perfused at a constant flow rate as described.5, 24 The details regarding the isolated liver perfusion system and the detailed protocols for the infusion of LTC4, LTD4, montelukast, the thromboxane analog 9,11-dideoxy-9,11-methanoepoxy-prostaglandin F2 (U46619), zymosan for KC activation, gadoliniumchloride for KC blockade, and the Rho kinase inhibitor (+)-(R)-trans-4-(1-aminoethyl)-N-(4-pyridyl)cyclohexanecarboxamide dihydrochloride (Y27632) are provided in the Supporting data.

In Vivo Measurement of Portal Pressure.

Body weight was determined directly prior to performing all experiments. Surgery (4 weeks after BDL or 18 weeks after TAA) was begun following intraperitoneal injection of sodium pentobarbital (50 mg/kg b.w.). The arterial blood pressure was continuously monitored by way of a carotid catheter. A laparotomy was performed and a PE-tube (23G, 0.6 × 30 mm) was inserted over the ileocolic vein and advanced toward the confluence of the portal and splenic veins.5, 25 This cannula was used for simultaneous infusion of LTC4, LTD4, zymosan, or lipopolysaccharide (LPS) and to monitor portal pressure, as described,5, 25 by way of a transducer system (Sirecust 404 from Siemens, Germany, and Transducer Gabarith from Becton Dickinson, Singapore). LTC4 (15 μg/kg b.w., minute 40 to 46, n = 5) or LTD4 (15 μg/kg b.w., minute 40 to 46, n = 5) were infused into the livers in order to analyze the in vivo effects on portal pressure.

Montelukast Treatment and KC Activation In Vivo.

In each of the following experimental groups three different conditions were investigated: (1) infusion of 0.9% NaCl solvent (100 μL/min, over 6 minutes); (2) KC activation with Zymosan A (intraportal application, 3.2 mg/min, from minute 0 to 6)5; and (3) activation of KCs by administration of intraportal LPS (LPS from E. coli 026:B6, 4 mg/kg b.w., 100 μL/min over 6 minutes). Animals in group 1 (solvent n = 5, zymosan n = 5, LPS n = 5) underwent BDL for 28 days. Rats were fed with a control diet (C1000, Altromin, Lage) for 28 days, after which 0.9% NaCl solvent, zymosan, or LPS were administered. Animals in group 2 (solvent n = 5, zymosan n = 7, LPS n = 7) underwent BDL, were fed with control C1000 for 18 days, and were subsequently fed C1000 containing montelukast for the next 10 days (Altromin, average of 0.125 mg/kg b.w. per day, from days 18 to 28). Animals in group 3 (solvent n = 5, zymosan n = 5, LPS n = 5) were administered TAA for 126 days in the drinking water and were fed with control C1000. Animals in group 4 (solvent n = 5, zymosan n = 7, LPS n = 7) were administered TAA in the drinking water for 126 days. Beginning on day 116, rats in group 4 were fed with C1000 containing montelukast (average of 0.125 mg/kg b.w. per day, from days 116 to 126) instead of the control diet.

Isolation of KC and Activation with Zymosan and LPS.

Isolation of KCs was performed according to a recently published protocol26; for details, see Supporting data. For activation, KCs were suspended in RPMI 1640 supplemented with penicillin/streptomycin, without fetal calf serum, for 24 hours. Afterwards, KCs were activated with 0.5 mg/mL zymosan in RPMI 1640 for 6 minutes, 1, 3, and 12 hours or with 10 μg/mL LPS for 6 minutes, 1, 3, and 12 hours. At these timepoints, media were collected and stored at −80°C for analysis of Cys-LT levels by enzyme-linked immunosorbent assay (ELISA). In addition, KCs were scraped off the culture vessels and their protein content was measured using the bicinchoninic acid protein assay reagents (BCA method).

Measurement of Serum Parameters.

Serum parameters were measured with an Olympus AU2700 analyzer (Olympus Germany, Hamburg) according to standard tests established by the International Federation of Clinical Chemistry (IFCC).

ELISA.

The levels of LTC4/D4/E4 (Cys-LTs) were measured in the effluent perfusate and in the KC media, as described above, in duplicate using an enzyme immunoassay (Cayman Chemical, Ann Arbor, MI).

Western Blot Analyses for the Cys-LT1 Receptor.

To assess the expression of the Cys-LT1 receptor, western blot analyses were performed according to standard procedures. Briefly, after sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and blotting, proteins on the membranes were detected by incubation with primary antibodies (anti-Cys-LT1 receptor, Santa Cruz Biotechnology, Santa Cruz, CA), followed by incubation with a horseradish peroxidase (HRP)-conjugated secondary antibody (Santa Cruz Biotechnology). Bands were visualized using a chemiluminescent detection kit (ECL Plus, Amersham Pharmacia, Uppsala, Sweden). β-actin was used as an internal control and detected with an appropriate anti-β-actin antibody (Millipore, Billerica, MA).

Histological Evaluation.

Liver tissues were fixed in 4% buffered formalin, dehydrated in graded ethanol, and embedded in paraffin using standard procedures. Four-μm-thick longitudinal sections were stained with H&E (hematoxylin/eosin) and Elastica van Gieson (EvG) using standard procedures.

Drugs and Reagents.

Zymosan and E. coli (026:B06) LPS were obtained from Sigma Chemical (St. Louis, MO). Ly171883 was purchased from Calbiochem (Darmstadt, Germany). U46619, LTC4, LTD4, and montelukast were purchased from Cayman Chemical (Ann Arbor, MI). LTC4, LTD4, and montelukast were dissolved in nitrogen-bubbled dimethyl sulfoxide (DMSO).

Statistical Analysis.

All data are expressed as the mean ± standard deviation (SD). Statistical analysis of the data was performed using analysis of variance (ANOVA) and Student's t test, where appropriate. P-values <0.05 were considered statistically significant. Statistical preparation and analysis was performed in collaboration with the Institute for Biometrics and Epidemiology of the University of Munich (IBE, Munich, Germany).

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Increases in Portal Pressure by LTC4 and LTD4.

The first set of experiments compared the increase in portal perfusion pressure by infusion of LTC4 and LTD4 into the livers of sham-operated rats and fibrotic rats 4 weeks after BDL. All BDL rats developed ascites and portal hypertension as determined by an increase in portal pressure measured in vivo immediately prior to the start of liver perfusion. Furthermore, the weight of the spleen also increased in the rats at this time (data not shown). Characteristic histological findings for biliary fibrosis were confirmed in all livers used for experiments (data not shown).

Infusion of LTC4 increased the portal perfusion pressure to a higher degree as infusion of LTD4 in sham-operated rats (P < 0.002) and BDL livers (P < 0.026) (Fig. 1A). The basal and the maximal portal perfusion pressure in BDL rats was higher than in sham-operated rats (Fig. 1A), but the absolute degree in portal perfusion pressure increase was not significantly different between BDL rats and sham-operated rats (Fig. 1A). According to the elevated basal portal perfusion pressure, the basal efflux (minute 20) of Cys-LTs measured by ELISA was significantly higher in BDL animals than in sham-operated animals (68 ± 17 pg/min × g liver versus 21 ± 14 pg/min × g liver). Compatible to the hypothesis of Cys-LT-mediated portal perfusion pressure increase, the coinfusion of the Cys-LT1 receptor inhibitor montelukast reduced the maximal portal perfusion pressure after LTC4 and LTD4 infusion in BDL animals (Fig. 1B).

thumbnail image

Figure 1. Vasoconstriction by LTC4 and LTD4. Data are expressed as the mean ± SD. (A) Basal portal perfusion pressure was higher (+P < 0.05) in BDL animals than in sham-operated animals. Infusion of LTC4 (20 nM, minute 40 to 46, n = 5) increased the portal perfusion pressure in sham-operated animals and in biliary fibrosis to a higher extent (#P < 0.0001, ##P < 0.0005) than did infusion of LTD4 (20 nM, minute 40 to 46, n = 5, *P < 0.0002, **P < 0.0004). This transient increase eventually returned to basal levels. (B) The additional infusion of montelukast (1 μM, minute 30 to 60) reduced the maximal portal perfusion pressure following infusion of LTC4 (20 nM, minute 40 to 46, n = 5) or LTD4 (20 nM, minute 40 to 46, n = 5) in BDL animals (*P < 0.0004, #P < 0.0005, **P < 0.001, ##P < 0.0004) (C) After a stabilization period of 10 minutes, in vivo infusion of LTC4 (15 μg/kg b.w., minute 0 to 6, n = 5) or LTD4 (15 μg/kg b. w., minute 0 to 6) resulted in a significant increase in portal pressure (*P < 0.05) compared to the controls. Again, portal pressure increased to a higher extent with the infusion of LTC4 than with infusion of LTD4 (**P < 0.05).

Download figure to PowerPoint

In vivo portal pressure measurements confirmed that the infusion of LTC4 increased the portal pressure to a higher degree as LTD4 in BDL rats (Fig. 1C). The systemic mean arterial pressure was affected neither by the infusion of LTC4 nor by the infusion of LTD4 (Table 1). The heart rate was stable in both experimental settings (data not shown).

Table 1. Systemic Mean Arterial Pressure
Mean Arterial Pressure (mm Hg)Minute 5Minute 6Minute 40
  1. In vivo portal pressure measurements in bile duct ligated animals (BDL, 4 weeks) or TAA administerd animals (TAA, 18 weeks) fed with the control diet C1000 with or without addition of montelukast to block the CysLT1 receptor. The systemic mean arterial pressure decreased only in the rats treated additionally with zymosan (*P < O.05).

BDL
 LTC4114 ± 12112 ± 8115 ± 11
 LTD4110 ± 11115 ± 9111 ± 15
BDL
 C1000 + Zymosan116 ± 14107 ± 1282 ± 9*
 C1000/Montelukast + Zymosan112 ± 12108 ± 1097 ± 8*
 C1000 + LPS110 ± 12112 ± 7115 ± 17
 C1000/Montelukast + LPS114 ± 11115 ± 12117 ± 15
TAA
 C1000 + Zymosan110 ± 1493 ± 1275 ± 9*
 C1000/Montelukast + Zymosan112 ± 1290 ± 983 ± 11*
 C1000 + LPS109 ± 14110 ± 12112 ± 12
 C1000/Montelukast + LPS112 ± 15114 ± 8111 ± 14

KC-Dependent Increase of Portal Pressure by Way of Cys-LTs.

Compared to vehicle administration, KC activation with zymosan increased the portal perfusion pressure in rat fibrotic livers (Fig. 2A). This zymosan-induced increase in portal perfusion pressure was attenuated by infusion of the LT receptor antagonist Ly171883 (Fig. 2A). The increase in portal perfusion pressure following KC activation was paralleled by an increase in Cys-LT efflux into the perfusate (Fig. 2B). This LT efflux was reduced by pretreatment with the KC blocker GdCl3 (Fig. 2B). Isolated KCs (Fig. 2C) produced vasoconstrictive Cys-LTs upon activation with zymosan or LPS at different timepoints (Fig. 2D). These results indicate that KCs have a high potential to produce LTs, which may have a functional role.

thumbnail image

Figure 2. Cys-LTs produced by KCs increase portal pressure. (A) KC activation with Zymosan A (•, Zy, 150 μg/mL, minute 40 to 46, n = 5) increased the portal perfusion pressure (#P < 0.05) when compared to control experiments (○, 100 minutes, n = 4) in bile duct-ligated animals. This increase was attenuated by additional treatment with the LT receptor antagonist Ly171883 (▪, 20 μM, minute 25 to 55, n = 5, *P < 0.05). (B) The increase in portal perfusion pressure with Zymosan A was paralleled by the efflux of Cys-LTs (○, Zy, 150 μg/mL, minute 40 to 46, n = 5), as determined by ELISA. This efflux was reduced by pretreatment with GdCl3 (10 mg/kg b.w., intraperitoneal, 48 and 24 hours before experiments, n = 5, *P < 0.05). (C) KCs were isolated from normal livers. Confocal microscopic analysis of CD163 expression revealed an 85% purity of KCs. (D) Activation of KCs with Zymosan A (0.5 mg/mL, *P < 0.05) or LPS (10 μg/mL, **P < 0.05, for 6 minutes, 1, 3, or 12 hours, n = 4) resulted in a significant increase in Cys-LT efflux into the media compared to unstimulated cells. The additional incubation with montelukast (1 μM) did not alter the LT efflux into the media following activation of KCs with Zymosan A (0.5 mg/mL) or LPS (10 μg/mL, *P < 0.05, for 6 minutes, 1, 3, or 12 hours, n = 4).

Download figure to PowerPoint

Increases in Portal Perfusion Pressure Mediated by TXA2 and LTC4, and the Role of Rho Kinase.

In addition to LTC4 and LTD4, TXA2 is also a well-known vasoconstrictor of the intrahepatic vasculature. We tested whether these vasoconstrictors have additive effects. Infusion of the TXA2 analog U46619 increased the portal perfusion pressure in fibrotic livers to approximately the same extent as did LTC4 (Fig. 3A). Coinfusion of both vasoconstrictors increased the portal perfusion pressure to a higher degree than did infusion of each substance alone (Fig. 3A). The increase in portal perfusion pressure observed with coinfusion was nearly additive. In order to delineate the possible mechanisms of action of these vasoconstrictors, the Rho kinase inhibitor Y27632 was coinfused with zymosan (for KC activation), U46619 (a TX analog), LTC4, or both U46619 and LTC4. The increase in portal perfusion pressure was attenuated in all four experimental settings by coinfusion with Y27632 (Fig. 3B). In contrast to BDL rats, following KC activation, Y27632 did not reduce the maximal portal perfusion pressure in sham-operated animals (data not shown).

thumbnail image

Figure 3. The additive effects of coinfusion of LTC4 and the TXA2 agonist U46619 in increasing portal perfusion pressure and their influence on Rho kinase-containing elements. Data are expressed as the mean ± SD. (A) The increase in portal perfusion pressure in bile duct-ligated animals following LTC4 infusion (20 nM, minute 40 to 46, n = 5) was augmented by the infusion of the TXA2 analog U46619 (0.1 μM, minute 40 to 46, n = 5, coinfusion n = 5). (B) Infusion of the Rho kinase inhibitor Y27632 (10 μM, minute 25 to 55) attenuated the increase in portal perfusion pressure following zymosan infusion (150 μg/mL, minute 40 to 46, n = 5), U46619 infusion (0.1 μM, minute 40 to 46, n = 5), LTC4 infusion (20 nM, minute 40 to 46, n = 5), and coinfusion of LTC4 (20 nM, minute 40 to 46) and U46619 (0.1 μM, minute 40 to 46, n = 5).

Download figure to PowerPoint

Effects of Treatment with Montelukast on Portal Hypertension of TAA and BDL Rats In Vivo.

Both TXA2 and LTC4 appear to be of major importance for the regulation of intrahepatic vasculature. The possible treatment of patients suffering from liver cirrhosis with a TXA2 receptor antagonist is limited due to its deleterious effects on blood coagulation. In contrast, the administration of montelukast, which has been approved for the treatment of asthma, may represent a conceivable treatment option for patients with liver cirrhosis. Therefore, the effectiveness of montelukast in lowering portal pressure was tested in two different models of liver cirrhosis, the BDL and the TAA models.

Effects of Montelukast on Histology and Serum Parameters.

EvG staining of rat livers revealed typical biliary fibrosis following BDL and severe cirrhosis following TAA administration (Fig. 4A,B) and all rats developed ascites and had an increased spleen weight (data not shown). Animals were fed with the control diet (C1000) or with C1000 mixed with montelukast. Liver and kidney function in all treatment groups was determined by analysis of the serum at the end of the experiments. Sodium, potassium, creatinine, bilirubin, alanine aminotransferase (ALT), aspartate aminotransferase (AST), aP, and γGT levels (Table 2) and the fibrosis stage (data not shown) were not influenced by montelukast treatment (BDL-C1000 versus BDL-C1000/Montelukast, TAA-C1000 versus TAA-C1000/Montelukast). The expression of the Cys-LT1 receptor was enhanced in BDL and TAA animals (Fig. 4C).

thumbnail image

Figure 4. Histological analysis of rat livers 4 weeks after BDL and 18 weeks after administration of TAA. (A) Animals underwent BDL for 4 weeks. EvG staining showed a typical picture for biliary fibrosis. (B) Administration of TAA in the drinking water for 18 weeks, with a weekly adaptation of the dose according to body weight, resulted in a histological pattern characteristic of severe cirrhosis. (C) The expression of the Cys-LT1 receptor was investigated by western blot analyses, which showed an increase in BDL animals and in TAA animals compared to sham-operated animals.

Download figure to PowerPoint

Table 2. Serum Parameters
 BDL + C1000BDL + C1000/MontelukastTAA + C1000TAA + C1000/Montelukast
  1. Bile duct ligated animals (BDL, 4 weeks) and thioacetamide administered animals (TAA, 18 weeks) recieving the control diet C1000 or the control diet C1000 mixed with montelukast to block the leukotriene receptors. The bilirubin was elevated in the animals following BDL (*), but not in the TAA groups. The AST, ALT, and aP values increased in all cirrhotic animals (**). The montelukast treatment did not alter the serum parameters in one of the cirrhosis groups.

Sodium (mmol/l)142 ± 4137 ± 6141 ± 5139 ± 3
Potassium (mmol/L)4.2 ± 0.34.3 ± 0.84.0 ± 0.94.0 ± 0.5
Creatinine (mg/dL)0.3 ± 0.10.3 ± 0.10.4 ± 0.10.3 ± 0.1
Bilirubin (mg/dL)8.2 ± 1.3*7.6 ± 1.7*0.5 ± 0.20.4 ± 0.3
AST (U/L)545 ± 36**498 ± 39**157 ± 13**143 ± 17**
ALT (U/L)134 ± 29**138 ± 24**142 ± 21**145 ± 14**
γGT (U/L)31 ± 5.232 ± 4.318 ± 3.517 ± 2.7
aP (U/L)650 ± 23**672 ± 40**712 ± 40**634 ± 40**

Effect of Montelukast on Portal Hypertension.

Montelukast treatment for 10 days lowered the basal portal pressure in both animal models of cirrhosis (Fig. 5A,B). Portal pressure increased acutely with administration of either β-glycan-rich zymosan or bacterial LPS in both cirrhosis models (Fig. 5A,B). The absolute increase in portal pressure (BDL-Zy: 9.2 ± 1.7 versus 5.3 ± 1.4* cm H2O, BDL-LPS: 5.4 ± 1.2 versus 3.1 ± 1.1* cm H2O; TAA-Zy: 10.3 ± 1.8 versus 7.3 ± 1.4* cm H2O, TAA-LPS: 7.1 ± 1.7 versus 4.1 ± 1.9 cm H2O; *P < 0.05), and hence the maximal portal pressure, were attenuated by montelukast treatment in both the BDL and TAA groups (Fig. 5A,B). However, the mean arterial pressure was not influenced by montelukast treatment (Table 1). The activation of KCs with zymosan reduced the mean arterial pressure in both cirrhosis models and in both the C1000 control-fed and the C1000/montelukast-fed groups (Table 1). The heart rate was not influenced by any of the treatments in any of the groups (data not shown).

thumbnail image

Figure 5. Montelukast treatment lowers the basal portal pressure and attenuates the KC-mediated increase in portal pressure in two different rat models of liver cirrhosis. Data are expressed as the mean ± SD. (A) Animals underwent BDL for 28 days. The first group of rats was fed with the control diet (C1000) followed by infusion of 0.9% NaCl solvent (n = 5, as indicated), zymosan (n = 7, as indicated), or LPS (n = 7, as indicated). The second group of rats was fed with C1000 containing montelukast from days 18 to 28 (on average 0.125 mg/kg b.w. per day). This orally substituted treatment lowered the basal portal pressure, the absolute increase in portal pressure, and the maximal portal pressure following KC activation with Zymosan A (intraportal, 3.2 mg/min, minute 0 to 6, maximal portal pressure at minute 6, n = 5) or LPS (LPS from E. coli 026:B6, 4 mg/kg b.w., 100 μL/min over 6 minutes, maximal portal pressure at minute 6, n = 5). (B) TAA was administered in the drinking water to rats weighing between 200 g and 250 g for 126 days. The control group was fed with C1000. During the experimental procedure, 0.9% NaCl solvent (n = 5, as indicated), zymosan (n = 5, as indicated), or LPS (n = 5, as indicated) were infused. Again, C1000 containing montelukast lowered the basal portal pressure, the absolute increase in portal pressure, and the maximal portal pressure following zymosan (n = 7, as indicated) or LPS infusion (n = 7, as indicated).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

This study investigated the KC-dependent production of LTs and their effects on intrahepatic vasoconstriction. We also investigated the use of the LT receptor antagonist montelukast as a possible new treatment option for portal hypertension in patients with liver cirrhosis. The principal findings of our study are: (1) activated KCs produce significant amounts of Cys-LTs; (2) LTC4 is a more potent vasoconstrictor than LTD4; (3) the vasoconstrictors thromboxane A2 and LTC4 additively increase the portal perfusion pressure and each effect is reduced by administration of a Rho kinase inhibitor; and (4) treatment with montelukast reduces portal hypertension in two different rat models of liver cirrhosis and prevents the KC-dependent increases of portal pressure (for summary, see Fig. 6). These results were obtained in vitro using isolated KCs, in situ by performing isolated liver perfusion, and in vivo using two different rat models of liver cirrhosis.

LTs from KCs Increased Portal Pressure.

Our initial experiments demonstrated that infusion of LTC4 or LTD4 increased the portal perfusion pressure and in vivo the portal pressure in the livers of rats that underwent BDL. The differences observed in the in vivo and in the in situ settings may be explained by many factors. In vivo, the effect of LTD4 and LTC4 may be limited by serum dipeptidases that degrade LTs. These are lacking in the in situ perfusion models. The increase of the portal perfusion pressure by way of LTD4 and LTC4 was in accordance with previous studies that delineated the importance of the enhanced vasoconstrictory potential of LTs in the cirrhotic liver.7, 8 In the present study the increase in portal pressure caused by LTC4 was greater than the caused by LTD4 in BDL animals and in sham-operated animals. The portal perfusion pressure increase in sham-operated animals was comparable to that of previous studies with normal livers.27 The doses of LTC4 and LTD4 used in the present study were chosen on the basis of previous dose-finding studies using normal livers.28 Furthermore, in previous studies it has been shown that the portal perfusion pressure increase by LTC4 or LTD4 in normal livers can be counteracted almost completely by the LT receptor inhibitor Ly171883. In the present study the Cys-LT1 receptor blocker montelukast massively reduced the portal perfusion pressure increase by LTC4 and LTD4. Together, these earlier and the present findings support the hypothesis of LT receptor-mediated intrahepatic vasoconstriction.

thumbnail image

Figure 6. Summary. Activation of KCs, e.g., by bacterial products results in the production of Cys-LTs, which increase the portal pressure by way of activation of the Cys-LT1 receptor on contractile cells. The Cys-LT1 receptor mediates vasocontraction by activation of Rho kinase in cells with contractile elements like hepatic stellate cells and myofibroblasts. Montelukast blocks the Cys-LT1 receptor and attenuates portal hypertension.

Download figure to PowerPoint

KCs isolated from normal livers produced significant amounts of Cys-LTs after activation with zymosan or LPS. The amount of LTs generated was maximal after 1 hour of zymosan treatment. This was in accordance with a previous report demonstrating that, among the different types of liver cells, only KCs express 5-lipoxygenase messenger RNA (mRNA).8 However, it was previously shown that hepatocytes exhibit the highest potential to generate Cys-LTs from the unstable intermediate LTA4.8 We did not further investigate this here, but we found that KCs produced significant amounts of Cys-LTs upon activation with bacterial products. Patients with liver cirrhosis have reduced bowel motility and increased bowel permeability, which results in an enhanced load of bacterial products in the portal vein.29-31 Furthermore, variceal bleeding in patients with liver cirrhosis is closely associated with bacterial infection.19-21 Therefore, activation of KCs by bacterial products is important in the context of liver cirrhosis and portal hypertension. Accordingly, the functional relevance of the data we obtained using isolated KCs was further confirmed both in situ and in vivo. Activation of KCs with β-glycan-rich zymosan increased the portal perfusion pressure markedly. However, this increase in portal perfusion pressure was dampened by the infusion of the LT receptor antagonist Ly171883. In addition, this increase in portal perfusion pressure was paralleled by an efflux of Cys-LTs, which in turn was dampened by blockade of KCs with GdCl3 pretreatment. GdCl3 has been used in different studies for the blockade of KCs.24, 32, 33 In summary, our results using isolated KCs are in line with our results using isolated livers and the measurements obtained for portal pressure in vivo.

Portal Pressure Is Further Increased by Thromboxanes and LTs.

The vasoconstrictor thromboxane A2 has also been shown to increase portal pressure.3-5 In addition, significant amounts of TXA2 are produced by KCs.4, 5, 34, 35 In additional experiments, we investigated whether LTC4 and TXA2 have additive effects in the intrahepatic microvasculature. Four weeks after the rats underwent BDL the increase in portal perfusion pressure caused by LTC4 was comparable to the increase caused by the TXA2 analog U46619. Interestingly, coinfusion of both vasoconstrictors increased the portal perfusion pressure in a manner that was almost equivalent to the sum of the increase in portal perfusion pressure caused by the individual vasoconstrictors.

Role of Rho Kinase in LT-Mediated Vasocontraction.

The postulated targets of these vasoconstrictors are liver cells with contractile potential, such as hepatic stellate cells and myofibroblasts. Contractile mechanisms can be differentiated into Ca2+-dependent and -independent pathways. Both pathways have been shown to be up-regulated in the cirrhotic liver, but Ca2+-independent pathways, such as the Rho kinase pathway, have been demonstrated to be more relevant.14, 15, 17, 36 Rho kinase mediates contraction of hepatic stellate cells and hepatic myofibroblasts. Consequently, intrahepatic vascular resistance can be substantially reduced by Rho kinase inhibitors.18 In the present study, infusion of the Rho kinase inhibitor Y27632 attenuated the increase in portal perfusion pressure observed following infusion of zymosan for KC activation in BDL animals, but not in sham-operated animals. This confirmed the particular importance of the Rho kinase pathway in fibrotic livers, as described.14, 15, 18 Accordingly in the present study, the maximal portal perfusion pressure increase by infusion of U46619 or LTC4 or coinfusion of both U46619 and LTC4 was attenuated by the infusion of Y27632. This result suggests that the effects of these vasoconstrictors are dependent on the function of Rho kinase, which is supposed to act in hepatic stellate cells and myofibroblasts.14, 15, 17 The interaction of Cys-LTs and hepatic stellate cells for intrahepatic contraction has been described.8

Effects of Treatment with the LT Receptor Blocker Montelukast.

Based on these pathophysiological considerations that KCs upon activation by bacterial constituents produce Cys-LTs, which induce intrahepatic vasoconstriction, we treated two rat models of liver cirrhosis with montelukast. The daily dose of montelukast was chosen according to the standard dose used to treat asthma in humans (10 mg/day, accordingly for a man of ≈80 kg: 0.125 mg/kg b.w. per day). Montelukast acts as a competitive inhibitor by binding to the CysLT1 receptor.22, 37 Both Cys-LTs, the LTC4 and the LTD4 bind to the CysLT1 receptor.22, 37 Our data show that the expression of the Cys-LT1 receptor is up-regulated in BDL livers and in TAA livers. Accordingly, treatment of two rat models of liver cirrhosis with montelukast for 10 days lowered the basal portal pressure, the absolute zymosan-induced increase in portal pressure, and the maximal portal pressure following activation of KCs with two different activators. Therefore, the effects of montelukast treatment were independent of the type of rat liver cirrhosis model chosen, although the effects of montelukast were more pronounced in the TAA model. A significant amount of LTs are eliminated by secretion into the bile, and the relative amounts of LTD4 secreted in the bile are increased in the TAA model.38 Therefore, competitive inhibition of the CysLT1 receptor is easy to achieve with the less potent LTD4 in the TAA model with an intact bile duct. However, montelukast treatment was still effective even in the presence of an underlying liver disease affecting the bile duct.

However, the effect of the Cys-LT1 receptor inhibitor MK-571 observed in earlier studies7 was mild and the effect was less pronounced than with AA-861 (inhibition of 5-lipoxygenase for LT synthesis). The discrepancy may be explained by the use of another perfusion model (recirculating) and another animal model (CCl4). Thus, liver-derived LTs that accumulate in the recirculating perfusion system may have limited the effect of MK-571 in that study. However, as the 5-lipoxygenase-inhibitor AA861 was effective, this may support the concept of the LT system as a new therapeutic target in portal hypertension.

Systemic blood pressure was not influenced by montelukast treatment (Table 1) in the rat models of liver cirrhosis. However, infusion of β-glycan-rich zymosan resulted in a decrease in systemic blood pressure, as reported.5 This effect was not observed by infusion of LPS. A possible explanation for the observed decrease in blood pressure is the systemic effect of cytokines produced by immunological cells, such as KCs or leukocytes, following their activation by bacterial products. The observed increase in portal pressure was higher after zymosan infusion than after LPS infusion. Higher doses of LPS would perhaps increase the portal pressure to a degree comparable to zymosan, and hence decrease systemic blood pressure more dominantly, as is observed in sepsis.

Injury of the cirrhotic livers following montelukast treatment was excluded by analyzing unaltered serum parameters (Table 2). Recent studies in children with asthma have shown that the clinical and laboratory safety profile of montelukast was similar to that observed in the placebo group.39 The big advantage of montelukast treatment, in comparison to treatment with a thromboxane inhibitor, is that blood coagulation is not affected by montelukast. This is a significant advantage for the use of montelukast to treat patients with liver cirrhosis.

In conclusion, LTs increase portal pressure in rat liver cirrhosis. Our results suggest that KCs play an important role in the generation of LTs following activation by bacterial products. In addition, the vasoconstrictors TXA2 and LTC4 increase the portal perfusion pressure additively. This effect, in turn, is attenuated by a Rho kinase inhibitor. These pathophysiological findings were tested as a treatment option for liver cirrhosis. The LT receptor blocker montelukast reduced the portal pressure in two different rat models of liver cirrhosis and reduced the acute increase in portal pressure following the infusion of two different microbial constituents. Therefore, montelukast may be of therapeutic benefit for patients with liver cirrhosis-related bacterial infections and portal hypertension.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

The authors thank Ingrid Liss and Christoph v. Hesler for excellent technical assistance. We thank Peter Dirschedl for support in statistical preparation and analysis of this work and Frigga Beitinger for the microscopic evaluation of the liver histology.

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
HEP_23596_sm_suppinfo.doc31KSupporting Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.