Kinin B1 receptor blockade attenuates hepatic fibrosis and portal hypertension in chronic liver diseases in mice

Animal studies were approved by the Institutional Animal Care and Use Committee at Jeonbuk National University and complied with the Animal Welfare Act. In total, 96 male Balbc mice were used at 8–10 weeks and weighing 25–30 g. Mice were housed under controlled light/dark conditions and fed a standard diet with water ad libitum. All animals were observed daily for general health, and all invasive procedures were performed under aseptic conditions.

Study design for CCl₄ induced liver fibrosis model is shown in Additional file 1: Fig. S1A. Mice were randomly assigned to three study groups: (1) sham control, n = 12; (2) CCl₄ + vehicle, n = 22; (3) CCl₄ + BI 113823, n = 20. Liver fibrosis was induced by intraperitoneal administration of CCl₄ (1 ml/kg/body weight dissolved in olive oil [1:3]) twice a week for 6 weeks [19]. Sham control mice received only olive oil injection. Mice received vehicle (0.1% Natrosol, p.o.) and BI 113823 (50 mg/kg, p.o., a gift from Boehringer Ingelheim Pharma KG, Biberach, Germany) daily for 6 weeks.

Study design for the BDL induced liver fibrosis model is shown in Additional file 1: Fig. S1B. Mice were anesthetized with ketamine (100 mg kg⁻¹, i.m.) plus xylazine (10 mg kg⁻¹, i.m.). A 2 cm abdominal midline incision was made and the common bile duct was located and tightly ligated with 4–0 silk suture [19]. Sham control mice underwent identical laparotomy without BDL (n = 12). Mice received daily treatment of vehicle (0.1% Natrosol, p.o., n = 18) and BI 113823 (50 mg/kg, p.o., n = 12) for 3 weeks.

At the end of study protocol, mice were anesthetized as described above, an abdominal midline incision was made, and the portal vein was cannulated through an ileocolic vein with a 24-gauge catheter which was connected to a pressure transducer [20]. Portal vein pressure was recorded using a Powerlab data acquisition system (ADInstruments Inc., CO). At the end of the experiment, liver tissue was collected. Tissue samples were snap frozen in liquid nitrogen or fixed in buffered formalin for histopathological examination.

Liver tissue specimens were fixed in 10% formalin and stained with haematoxylin and eosin (HE) and picrosirius red. Slides were examined by light microscopy for morphological alterations and collagen accumulation in a blinded fashion. Fibrosis and inflammatory cell accumulation were depicted through immunohistochemical staining for α-SMA/B1R, CD68 and neutrophil elastase. Total collagen content was assessed by hydroxyproline assay (BioVision kit), and inflammatory responses and profibrogenic mediators assessed by western blot and reverse transcription polymerase chain reaction (RT-PCR). Detailed methods are shown in Additional file 1.

Immune cell migration assays were carried out on monocytes and neutrophils treated with tumour necrosis factor (TNF)-α with or without BI 113823. Migrating cells were counted after 12 h by haemocytometer. Human peripheral monocytes were treated with lipopolysaccharides (LPS) with or without BI 113823 and TNF-α levels were measured with immunoassay kits (PreproTech, Rocky Hill, NJ, USA). Myeloperoxidase (MPO) assay was carried out on human peripheral neutrophils treated with LPS with or without BI 113823, and MPO release measured at 460 nm. Monocyte and neutrophil activation was determined by flow cytometry measuring upregulation of CD11b and CD18 on FACSCalibur apparatus.

Des-Arg9-bradykinin (DBK), a carboxypeptidase metabolite of the kinins, preferentially activates the kinin B1 receptors [7]. In isolated pig coronary arteries exposed to LPS, DBK caused induction of kinin B1 receptor-dependent, endothelium-independent contractions [17]. The changes in HSC cell proliferation and migration were measured in the presence of selective B1R agonist (DBK) and antagonist (BI 113823) in this study.

LX2 HSCs, kindly provided by Prof. Bumseok Kim, were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and antibiotics. LX2 cell proliferation was determined by BrdU Cell Proliferation Assay Kit (BioVision) after incubation with or without different concentrations of des-Arg9-bradykinin (DBK) and BI 113823. Migration of HSCs was measured after incubation with 0.5% FBS containing medium with different concentrations of DBK and BI 113823 after 24 h. Cytoselect 48-well cell concentration assay (cell biolabs) was used to evaluate BI 113823’s effect on transforming growth factor (TGF)-β mediated HSC contraction. Photographs were made with a digital camera at 0, 24, and 48 h. The size of the gels were digitally measured and normalized with their respective well size using Image J software. HSCs were treated with BI 113823 in medium containing 1% FBS and analysed by flow cytometry to detect apoptosis (FITC Annexin V Apoptosis Detection Kit with PI, Biolegend). Cell cycle analysis was carried out by incubating starved HSCs with medium containing 1% FBS with or without DBK and BI 113823, and analysed using FxcyclePI/RNase staining kit. HSC transwell migration assay was performed on LX2 cells using medium with 1% FBS with or without BI 113823. Migrating cells were counted after 12 h by haemocytometer. HSCs were treated with TGF-β and BI 113823 for 24 h before overnight incubation with α-SMA Ab and visualization using rhodamine fluorescence probe-labelled secondary Ab. A western blot was carried out according to standard protocols using primary and secondary Abs as described in Additional file 1: Table S2.

The results are presented as the mean ± SEM, unless otherwise specified. A two-group comparison was performed using unpaired Student’s t-test, while a multiple-group comparison was performed by analysis of variance for repeated measures followed by Tukey’s multiple comparison test, or Bonferroni test. Survival estimates were made by Kaplan–Meier analysis. Survival curves were analyzed by the nonparametric Mantel-Cox test.

Liver fibrosis in mice was evaluated by HE and sirius red staining, and hydroxyproline assay at 6 weeks after the initial CCl₄ challenge. HE staining of liver sections from CCl₄-treated mice showed prominent hepatic steatosis, necrosis, fibrotic septa formation, disruption of tissue architecture, and inflammatory cell infiltration (Fig. 1A). Liver sections from BI 113823-treated mice showed relatively normal architecture (Fig. 1A). Treatment with BI 113823 significantly reduced CCl₄-induced liver fibrosis, as demonstrated by reduction in sirius red staining for collagen and reduction of hydroxyproline content in liver (Fig. 1A and B). HSC activation and differentiation, as assessed by immunofluorescence staining for α-SMA, showed marked increase in CCl₄-treated mice, and this was correlated with marked induction of B1R expression in CCl₄-induced fibrotic liver (Fig. 1A). The expression of α-SMA and B1Rs in liver tissue was significantly reduced in mice treated with BI 113823 (Fig. 1A, G and H). B2R expression was not changed (Fig. 1H). Immunofluorescence double staining results showed that cells expressing the B1Rs co-localized with those expressing α-SMA (Fig. 1A). These findings indicate B1R signalling mediates CCl₄-induced liver fibrosis via activation of HSCs.

Six weeks after the initial CCl₄ challenge, bodyweight growth was slightly lower compared to the sham control (not significantly different, Fig. 1D). However, liver/body weight ratio and portal vein pressure were significantly increased in CCl₄-challenged mice, compared to the sham control (Fig. 1C). CCl₄-induced liver fibrosis and PH were associated with high mortality rate in those mice (12/22 death, Fig. 1F). In contrast, mice that received BI 113823 showed significantly lower liver/body weight ratio, lower portal vein pressure and, most importantly, improved survival (Fig. 1C, E and F).

In our next model, liver fibrosis in mice was evaluated 3 weeks after BDL. Liver sections from BDL mice showed marked hepatic steatosis, necrosis, inflammatory cell infiltration and fibrotic septa formation, as well as extensive collagen deposition as accessed by HE and sirius red staining, and hydroxyproline assay (Fig. 2A and B). Similarly, liver fibrosis in BDL mice was significantly reduced in mice treated with BI 113823, compared to vehicle controls (Fig. 2A). Liver fibrosis in BDL mice was also associated with marked increase of α-SMA expression, as well as co-localized induction of B1Rs. BI 113823 reduced the expression of α-SMA and B1Rs in BDL mice (Fig. 2A, G and H). These findings further support the hypothesis that B1R signalling mediates chronic liver fibrosis via activation of HSCs.

Three weeks after BDL, liver/body weight ratio was not significantly different among groups (Fig. 2C) and body weight was significantly decreased in vehicle-treated BDL mice, compared to sham control (Fig. 2D). However, the decrease of body weight after BDL was less in BI 113823-treated mice (Fig. 2D). BDL resulted in severe PH, and BI 113823 significantly decreased portal vein pressure (Fig. 2E) and improved survival (Fig. 2F).

Altogether, these data demonstrate that, in experimental models of CLD, B1R inhibition with BI 113823 attenuates hepatic fibrosis formation, PH and, most importantly improves survival.

We next determined the expression of profibrotic mediators in CCl₄- and BDL-induced liver fibrosis. Western blot analysis showed an extensive increase in the expression of profibrotic mediators α-SMA, collagen I, VEGF and proliferating cell nuclear antigen (PCNA) in CCl₄- and BDL-induced liver fibrosis (Fig. 3A and B, Additional file 1: Fig. S2A and B). Compared to vehicle-treated mice, expression of α-SMA, collagen I, VEGF and PCNA were all significantly reduced in mice treated with BI 113823 (Fig. 3A and B). Furthermore, hepatic expression of phosphorylated Akt protein was significantly increased in CCl₄- and BDL-induced liver fibrosis (Fig. 3A and B), and BI 113823 reduced this expression (Fig. 3A and B, Additional file 1: Fig. S2A and S2B).

RT-PCR showed that hepatic mRNA expressions of growth factors PDGF, TGF and CTGF, as well as ECM molecules Col-1, Col-3 and Col-4 increased significantly in CCl₄- and BDL-induced liver fibrosis, and all were reduced in mice treated with BI 113823 (Fig. 3C and D). Together, these findings further demonstrate that BI 113823 attenuates hepatic fibrosis via inhibition of HSC activation and proliferation, downregulates the expression of fibrogenic mediators via inhibition of the Akt signalling pathway.

Chronic CCl₄ challenge and BDL resulted in marked hepatic inflammatory cell infiltration in mice (Fig. 4A and B). Immunohistochemical staining of liver sections showed a marked increase of CD68 positive macrophages and neutrophils (positive for neutrophil elastase) in both CCl₄- and BDL-induced liver fibrosis (Fig. 4A and B). BI 113823 reduced hepatic inflammatory cell infiltration of macrophages and neutrophils, and hepatocyte apoptosis (Fig. 4A and B).

To provide further evidence that B1Rs mediate inflammatory cell infiltration and inflammatory mediator production in CCl₄- and BDL-induced liver fibrosis, we assessed protein expression of inflammatory cell markers CD68 and neutrophil elastase, and chemoattractant markers MCP-1 and COX-2 in liver tissues. Western blot of liver lysates showed that inflammatory molecules, COX-2, MCP-1, CD68, and neutrophil elastase were strongly increased in CCl₄- and BDL-induced liver fibrosis, compared to sham control mice (Fig. 5A, B, Additional file 1: Fig. S3A and B). BI 113823 significantly reduced hepatic expression of these inflammatory mediators (Fig. 5A, B, Additional file 1: Fig. S3A and B).

RT-PCR analyses supported these observations by showing that mRNAs of inflammatory mediators, IL-1β, IL-6, MCP-1, MCP-3 and TIMP-1 increased in CCl₄- and BDL-induced liver fibrosis, and decreased in BI 113823-treated mice (Fig. 5C and D). These findings indicate that B1R-mediated inflammatory responses are highly implicated in the pathogenesis of liver fibrosis.

We next determined the effect of B1R-mediated inflammatory responses and activation in human monocytes and neutrophils. BI 113823 inhibited TNF-α-induced monocyte and neutrophil transmigration (Fig. 6A and D), reduced LPS-induced TNF-α production in monocytes and reduced LPS-stimulated MPO activity in neutrophils (Fig. 6B and E). LPS treatment resulted in a significant increase in activation of human monocytes and neutrophils, as evidenced by an increase in cell surface molecule CD11 and CD18 expression (Fig. 6C and F, Additional file 1: Fig. S4A and S4B). LPS-induced monocyte and neutrophil activation were inhibited by BI 113823 treatment (Fig. 6C and F, Additional file 1: Fig. S4A and B).

TGF-β is a major profibrogenic cytokine responsible for initiation and perpetuation of HSC activation. In this study, TGF-β stimulated strong intracellular expression of α-SMA in LX2 hHSCs, indicating HSC activation switching from a quiescent state to a myofibroblast phenotype, which was inhibited by BI 113823 (Fig. 7A). In collagen gel contraction assays, BI 113823 reduced TGF-β-stimulated hHSC contractility measured at 48 h by final gel area (Fig. 7C and D). Next, we determined the downstream effect of myofibroblast transdifferentiation, specifically fibrosis molecules production. Western blot analysis showed that TGF-β stimulated a significant increase in the expression of profibrogenic proteins α-SMA, Col-1, VEGF and MCP-1, as well as the expression of B1Rs and B2Rs in LX-2 hHSCs. Treatment of hHSCs with BI 113823 reduced the expression of profibrotic proteins and B1Rs (Fig. 7B). Furthermore, BI 113823 treatment inhibited TGF-β stimulated phosphorylation of Akt in LX-2 hHSCs (Fig. 7B).

HSCs can migrate to the sites of tissue injury during fibrogenesis and differentiate into contractile myofibroblasts that promote liver stiffness and PH. We further examined the effects of BI 113823 using migration (Fig. 7E) and wound-healing assays (Fig. 7F and G), where the selective B1R agonist [21] DBK stimulated hHSCs migration in a dose-dependent manner (Fig. 7F). BI 113823 treatment significantly inhibited FBS- and DBK-stimulated migration in LX-2 hHSCs (Fig. 7E–G).

In this study, we found that the selective B1R agonist DBK enhances the proliferation of hHSCs (Fig. 8A). BI 113823 dose-dependently inhibited DBK-stimulated hHSC proliferation (Fig. 8B) but did not affect non-DBK-stimulated hHSC growth or apoptosis (Fig. 8C and D). Flow cytometry analysis showed that DBK (2 µM) stimulated the entry of hHSCs into S phase of the cell cycle. At 1 µM, BI 113823 reduced DBK-stimulated G1 to S phase cell cycle transition by 64% (Fig. 8E, Additional file 1: Fig. S5A). Taken together, the present study provides direct evidence that DBK induces hHSC proliferation and migration via activation of B1Rs.

Western blot analysis showed that B1Rs were weakly expressed in serum-free quiescent or growing hHSCs, but were strongly upregulated by DBK (Fig. 8F, Additional file 1: Fig. S5B). DBK also stimulated phosphorylation of P70S6 kinase, PI3K, GSK3, ERK and increased the expression of PCNA in HSCs (Fig. 8F, Additional file 1: Fig. S5B). Interestingly, phosphorylation of P70S6 kinase was not detected in serum-free quiescent HSCs with or without DBK, but DBK stimulated strong phosphorylation of P70S6 kinase in HSCs with 1% FBS. BI 113823 treatment inhibited phosphorylation of P70S6 kinase, PI3K, GSK3 and ERK, blocked B1R induction, and reduced expression of PCNA (Fig. 8F, Additional file 1: Fig. S5B). Together, these findings demonstrate that DBK-stimulated B1R signalling mediates HSC proliferation and migration via activation of the PI3K/Akt pathway.

Hepatocyte damage results in the recruitment of inflammatory cells, and upregulation of inflammatory and fibrogenetic mediators, including various cytokines, chemokines that activate HSCs and drive liver fibrosis, and activation of Kupffer cells [2, 21]. Kinins mediate liver inflammatory injury and fibrosis through constitutively expressed B2 receptors and through induction of B1 receptors [6, 7]. Inflammatory cells are also a major source of tissue kallikrein, which increases kinin production [22]. Increased blood level of tissue kallikrein and reduced hepatic clearance of plasma kallikrein was associated with liver cirrhosis patients and cirrhotic rats [23, 24]. B1Rs are induced by stimulation of lipopolysaccharides, IL-1β and TNF-α, by tissue injury and exposure to kinins (the receptor agonists) via activation of NF-ĸB signaling pathway [6, 7, 15‐18, 25]. B1 receptor induction elicits persistent inflammatory responses and its up-regulation is further boosted by ligand binding [17, 26]. Therefore, kinin B1 receptors may be important in maintaining chronic inflammatory processes, such as chronic liver injury [8].

The present study discovered that DBK stimulates LX2 human HSC proliferation. In addition to induction of B1Rs by inflammatory mediator TGF-β, we also show that in HSCs, B1Rs are strongly induced by the receptor agonist DBK. This is consistent with a previous report that B1Rs can be induced by exposure to inflammatory mediators and its receptor agonists [25]. Using human peripheral blood inflammatory cells, we showed that B1Rs mediate human monocytes and neutrophil activation, migration and cytokine production. These findings suggest that blocking B1Rs to control inflammatory responses would be protective for liver fibrosis. In this study, CCl₄- and BDL-induced liver fibrosis was associated with a marked increase of macrophage and neutrophil accumulation and strong upregulation of B1Rs in liver tissue. B1R blockade with BI 113823 inhibited the accumulation of macrophages and neutrophils in both CCl₄ and BDL fibrosis livers, and reduced the expression of inflammatory mediators, B1Rs, COX-2, IL-1β, IL-6, MCP-1, MCP-3 and TIMP-1 in liver tissue. Taken together, our results demonstrated that blocking inflammatory responses with BI 113823 contributed to reduced chronic liver fibrosis in mice.

Kinin B1 receptors is strongly expressed in regenerating and cirrhotic livers [8, 27‐29]. In human lung myofibroblasts, B1R activation by DBK triggers synthesis and secretion of matrix metalloproteases, type I collagen synthesis and the expression of connective tissue growth factor [29, 30]. B1R blockade inhibits vascular cell proliferation and vascular remodelling, and attenuates cardiac fibrosis and hypertrophy [13, 31]. Mice lacking B1Rs display an improvement on leptin and insulin sensitivity and are protected from non-alcoholic fatty liver disease (NAFLD) after a high-fat diet treatment [32]. Collectively, these observations suggest that B1Rs may play an important role in hepatic fibrosis.

In the present study, B1R mRNA and protein were weakly expressed in normal mice liver, but were strongly increased in CCl₄ and BDL fibrosis mice. The induction of B1Rs was correlated with severe hepatic fibrosis in CCl₄ and BDL mice, with excessive collagen deposition and fibrotic molecules expression. B1R blockade with BI 113823 inhibited HSC activation, fibrotic mediator expression and collagen production, and reduced mouse liver fibrosis induced by treatment with CCl₄ and BDL. These findings support the hypothesis that B1 receptors may be important in the pathogenesis of liver fibrosis.

HSCs are recognised as the primary cellular source of matrix components in patients with CLD [33]. When activated, HSCs proliferate and transdifferentiate into myofibroblasts, and these cells produce fibrogenic molecules in the hepatic ECM [1‐5]. Pro-inflammatory and pro-fibrotic cytokines such as TGF-β and PDGF are recognised as common mediators of HSC activation and proliferation [34]. In the present study, we discovered that the selective B1 agonist DBK is a strong mediator of HSC activation and proliferation by stimulation of the B1R pathway. B1Rs were weakly expressed in quiescent hHSCs, but were strongly induced in response to TGF-β and DBK. DBK enhanced HSC proliferation. Interestingly, B1Rs blockade with BI 113823 inhibited DBK-stimulated HSC proliferation, but did not affect the proliferation and apoptosis in the normal culture of HSCs. BI 113823 blocked DBK-stimulated HSC proliferation by preventing G1 to S phase cell cycle transition. The activation of HSCs can induce α-SMA-positive myofibroblast transition, strengthen their proliferation and increase collagen synthesis [33]. In this study, B1Rs blockade with BI 113823 inhibited TGF-β and DBK stimulated HSC activation, migration and reduced the expression of fibrotic molecules such as α-SMA, collagen-1, VEGF and MCP-1. Our findings demonstrate that DBK is a strong activator of HSC and the DBK-stimulated B1R signalling pathway, which regulates HSC activation, proliferation and migration, and contributes to fibrogenesis in chronic liver injury.

The PI3K signalling pathway promotes cell proliferation and collagen synthesis in HSCs, and inhibition of PI3K signalling in activated HSCs and fibrosis models inhibit the expression of collagen and attenuates the development of liver fibrosis [33, 35]. In the present study, the increase in HSC activation, cell proliferation and collagen production in CCl₄- and BDL-induced fibrotic liver was correlated with strong phosphorylation of Akt, and B1R blockade with BI 113823, inhibited the phosphorylation of Akt, and reduced liver fibrosis. We also discovered that DBK induced a marked increase of expression of B1Rs in quiescent and growing hHSCs. The induction of B1Rs in hHSCs was correlated with an increase of PCNA and the phosphorylation of P70S6 kinase, PI3K, GSK3 and ERK. Treatment with BI 113823 inhibited DBK-stimulated phosphorylation of P70S6 kinase, PI3K, GSK3 and ERK, and inhibited DBK-induced HSC proliferation and migration. Taken together, our results demonstrated that the inhibition of HSC activation, proliferation and fibrosis formation by BI 113823 involves blocking PI3K/Akt signalling pathways both in vitro and in vivo.

In portal hypertension, damaged parenchymal and non- parenchymal cells contribute to the increase of intrahepatic vascular resistance (IHVR) in cirrhosis through two major mechanisms: a profound alteration in liver architecture and a pathological increase in the hepatic vascular tone [36‐38]. Hepatic fibrosis and sinusoidal capillarization are major contributors to these structural changes and IHVR in chronic liver diseases [36]. IHVR also attributed to deregulated contractile elements, such as endothelin 1 (ET1) and thromboxane A2 (TXA2), and are hyporesponsive to vasodilators, such as nitric oxide (NO) and prostacyclin [36, 37, 39‐41]. B1Rs are induced in rat portal veins and pig coronary arteries and stimulation triggers vasoconstriction via the cyclooxygenase-2 (COX-2) and TXA2 pathway [14, 17]. Elevated VEGF levels exacerbate portal hypertension by increasing nitric oxide production and angiogenesis.

In this study, B1R expression was strongly increased in hHSCs in response to TGF-β and DBK, and is also increased in the livers of CCl₄ and BDL fibrosis mice. B1R blockade with BI 113823 inhibited hHSC contractility in collagen gel contraction assays, attenuated PH in CCl₄ and BDL fibrosis mice. This reduction in portal vein pressure was associated with downregulation of VEGF and COX-2 expression in both mice models, as well as reduction of VEGF protein expression in LX-2 hHSC cells. These results provide novel insight into the role of B1R expression and the regulation of hepatic vascular tone in diseases.