Prostaglandin E₂ induces contraction of liver myofibroblasts by activating EP₃ and FP prostanoid receptors

Abstract

Background and purpose:

Increased portal pressure in liver injury results from hypercontraction of perivascular non-parenchymal cells including liver myofibroblasts (MFs). Prostaglandin E₂ (PGE₂) is the major eicosanoid which is released around the venous system during liver injury, but little is known about their contractile effect on MFs.

Experimental approach:

Contraction of primary rat liver MFs was measured by a collagen gel contraction assay. Expression of E prostanoid (EP) receptor subtypes was assessed by reverse transcription-polymerase chain reaction. Fura-2 fluorescence was used to determine intracellular Ca²⁺ concentration ([Ca²⁺]_i). Phosphorylation of protein kinase C (PKC) was detected by Western blot analysis.

Key results:

Liver MFs expressed mRNAs for all four EP receptors. PGE₂ induced contraction in a dose- and time-dependent manner, and slightly increased [Ca²⁺]_i only at high concentrations (10 µmol·L⁻¹). An agonist selective for EP₃ receptors, ONO-AE-248, dose-dependently induced MF contraction but did not increase [Ca²⁺]_i. Pretreatment with rottlerin (a specific novel PKC inhibitor) and Ro 31-8425 (a general PKC inhibitor) significantly reduced 1 µmol·L⁻¹ PGE₂- or ONO-AE-248-induced contractions. Furthermore, 1 µmol·L⁻¹ PGE₂ stimulated phosphorylation of PKC isoforms PKCδ and PKCε. The F prostanoid (FP) receptor antagonist AL8810 abolished the [Ca²⁺]_i elevation and the rapid contraction induced by 10 µmol·L⁻¹ PGE₂.

Conclusions and implications:

Lower concentrations up to 1 µmol·L⁻¹ of PGE₂ induce liver MF contraction via a [Ca²⁺]_i-independent PKC-mediated pathway through the EP₃ receptor, while higher concentrations have an additional pathway leading to Ca²⁺-dependent contraction through activating the FP receptor.

Introduction

In cirrhosis, portal hypertension is caused by an increase in vascular resistance and splanchnic blood flow, which are caused primarily by structural changes such as fibrotic scar tissue and by elevated intrahepatic vascular tone (Laleman et al., 2005). Several in vitro and in vivo studies have highlighted the role of non-parenchymal cells including liver myofibroblasts (MFs) and hepatic stellate cells (HSCs). MFs are located around the central vein and the portal area, whileHSCs are in the space of Disse (Knittel et al., 1999a). In response to liver injury, these cells undergo activation, and produce increased quantities of extracellular matrix (ECM) protein (Friedman, 2000). Besides a fibrogenic response, they also acquire enhanced contractile properties upon activation, which is characterized by greater expression of α-smooth muscle actin (Knittel et al., 1999b; Desmouliere et al., 2003). It is possible that contraction of these non-parenchymal cells, in response to secreted vasoconstrictors, results in an increase in the sinusoidal pressure gradient and resistance, and consequently, in the progress of portal hypertension (Shah et al., 1998; Reynaert et al., 2002). Therefore, MFs and HSCs are now considered to be therapeutic targets to decrease portal hypertension in advanced chronic liver disease. Although most studies have focused on HSC contractility and suggested that various stimulants cause contraction both in Ca²⁺-dependent and -independent manner (Kawada et al., 1993; Pinzani et al., 1996; Melton et al., 2006; Laleman et al., 2007), few studies have yet addressed MF contractility.

Cyclooxygenase (COX) is the rate-limiting enzyme in prostaglandin biosynthesis from arachidonic acid and exists as two isoforms. COX-1 is constitutively expressed in most tissues and fulfills housekeeping functions. On the other hand, COX-2 is the inducible isoform that accounts for the increased production of PGs in response to pro-inflammatory and mitogenic stimuli (Hu, 2003). Prostaglandin E₂ (PGE₂) exerts both pro-inflammatory and anti-proliferative effects via G-protein-coupled receptors (GPCRs) (Narumiya et al., 1999; Sugimoto and Narumiya, 2007). There are four GPCRs responding to PGE₂ designated as EP₁, EP₂, EP₃ and EP₄. The E prostanoid (EP) receptors exhibit differences in signal transduction depending upon coupled G protein subtypes. EP₁ receptors couples to G_q and increases intracellular Ca²⁺. EP₂ and EP₄ receptors couples to G_s and stimulate adenylate cyclase with a subsequent increase in intracellular cyclic adenosine monophosphate (cAMP). Binding of PGE₂ to the EP₃ receptor, on the other hand, activates multiple signaling pathways, including Ca²⁺, protein kinase C (PKC) and phosphatidylinositol 3-kinase (PI3K) signaling pathways (Breyer et al., 2001). In the liver, non-parenchymal cells produce mainly PGE₂ and also other PGs, and release them around the vessels during liver injury (Enomoto et al., 2000; Neyrinck et al., 2004).

Previous in vivo studies reported that COX-2 is highly expressed in cirrhotic liver, mainly in the sinusoidal area and area around the vessels, indicating that COX-2 and the resulting product PGE₂ may modulate the function of non-parenchymal cells during liver injury (Cheng et al., 2002; Mohammed et al., 2004). However, the role of PGE₂ in liver MF bioactivity still remains unknown.

In this study, we isolated rat liver MFs and assessed the changes in their contractility and intracellular Ca²⁺ concentration ([Ca²⁺]_i) by PGE₂ treatment. Low doses of PGE₂ caused MF contraction via Ca²⁺-independent PKC activation through the EP₃ receptor. In addition to this mechanism, high doses of PGE₂ activated the F prostanoid (FP) receptor and subsequently induced Ca²⁺-dependent contraction.

Methods

Cell isolation and culture

All animal procedures were performed in accordance with the guidelines of the University of Tokyo. Rat liver MFs were isolated from male Sprague-Dawley rats (300–500 g) as described previously (Kojima et al., 2007). In brief, the liver was perfused with collagenase (Sigma, USA) and protease (Merck, Germany), following removal of hepatocytes and cell debris by low-speed centrifugation. A cell fraction containing MFs was obtained by density gradient centrifugation using 14.6% Nycodenz (Sigma, USA). The cells were cultured in Dulbecco's modified Eagle's medium (DMEM, Sigma, USA) supplemented with 10% fetal bovine serum (FBS, JRH Biosciences, USA), and passaged two to three times before use. Cells between passage 3 and 8 were used for experiments. We confirmed that isolated cells from rat liver express α-smooth muscle actin, fibulin-2 and interleukin-6 at the mRNA level prior to the experiments (data not shown). These have been suggested as specific markers for MFs (Knittel et al., 1999b; Ramadori and Saile, 2002; Tateaki et al., 2004).

Collagen gel contraction assay

Contraction of liver MFs was examined as previously described (Reynaert et al., 2001; Maruyama et al., 2008). In brief, collagen gels were prepared by mixing 70% type I collagen from porcine tendon (Nitta Gelatin, Japan), 20% 5 × DMEM and 10% 0.05 N NaOH under ice-cold conditions (final collagen concentration, 2.1 mg·mL⁻¹). The solution was added to each well of 12 well plates and incubated at 37°C for 1 h. Cells were plated on top of the collagen gels at a concentration of 1 × 10⁵ cells per well. After incubation overnight to allow cell attachment, serum-free conditions were introduced for 48 h. Gels were then detached from the plates using a microspatula. The surface area of the gels was quantified using ImageJ (National Institutes of Health, USA). Relative contraction of the gels was expressed as a percentage according to the following formula: [(gel surface area of buffer − gel surface area of test substance)/(gel surface area of buffer − gel surface area of 5% FBS stimulation for 24 h)] × 100%.

[Ca²⁺]_i measurement

[Ca²⁺]_i was measured using fura-2 AM as previously described (Kojima et al., 2007). In brief, liver MFs on glass coverslips in HEPES-buffered solution were loaded with 3 µmol·L⁻¹ fura-2 AM (Dojindo Laboratories, Japan) for 40 min in a dark room at 37°C and placed in a bath on the stage of an inverted microscope (TE-300, Nikon, Japan) equipped with a 40-fold objective lens. Using a fluorescence imaging system (Hamamatsu Photonics, Japan), images of 510 nm fluorescence were captured every 3 s using 340 nm and 380 nm wavelength light, and the fluorescence of an image at 340 nm (F340) was divided by the fluorescence at 380 nm (F380) giving a ratio (F340/F380). [Ca²⁺]_i was calculated using the following formula described previously (Takahashi et al., 1999): [Ca²⁺]_i = K_d × (S_f2/S_b2) × (R − R_min)/(R_max − R), where K_d is the effective dissociation constant of fura-2 and has a value of 224 nmol·L⁻¹, R is the fluorescence ratio, R_min and R_max are the ratios in absence and presence of Ca²⁺, respectively, and S_f2 and S_b2 are the emissions at 380 nm in the absence and presence of Ca²⁺ respectively. Calibration was accomplished after permeabilization of the cells with 3 µmol·L⁻¹ ionomycin and measurement of fluorescence at both wavelengths under Ca²⁺-free (in 0.5 mmol·L⁻¹ EGTA) or Ca²⁺-saturated (in 1.5 mmol·L⁻¹ CaCl₂) conditions to obtain R_min, R_max, S_f2 and S_b2. The area under the Δratio per time curve (area under curve, AUC) was calculated to assess the response.

Reverse transcription-polymerase chain reaction (RT-PCR)

Total RNA was extracted from liver MFs, and the concentration of total RNA was adjusted to 1 µg·µL⁻¹ with RNase-free distilled water. RT-PCR was performed as previously described (Kojima et al., 2007). After denaturation at 95°C for 10 min, 36 cycles of amplification at 94°C for 40 s, at 60°C for 1 min and at 72°C for 1 min were performed using a thermal cycler (Takara Bio, Japan). The PCR products were electrophoresed onto a 2% agarose gel containing ethidium bromide at 0.2 µg·mL⁻¹. The detectable fluorescence bands were visualized using an ultraviolet transilluminator. The forward and the reverse primers for EP₁ through EP₄ receptors, FP receptor, COX-1, COX-2, microsomal prostaglandin E synthase-1 (mPGES-1), mPGES-2 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were designed as shown in Table 1.

Table 1.

RT-PCR primer pairs

Target gene	5′ sequence	3′ sequence	Product (bp)
GAPDH	TCCCTCAAGATTGTCAGCAA	AGATCCACAACGGATACATT	308
EP1	ATGGTCTTCTTTGGCCTGTG	GTTCTCTCGGAAACGTCGAG	387
EP2	GAACGCTACCTCGCCATCGG	CGAAGGTGATGGTCATAATGGC	421
EP3	CTTTGCCTCCGCCTTCGCC	CTTAGCAGCAGATAAACCCAGG	363
EP4	CATCTTACTCATCGCCACCTCTC	GTTAGGTCTGGCAGGTATAGGAGG	393
FP	GACTCTTAGCTCTCGGCATCTC	CGTAGCAGAATGTAGACCCAGG	298
COX-1	CCGGATTGGTGGGGGTAG	AGGGGCAGGTCTTGGTGTTG	434
COX-2	CTGTATCCCGCCCTGCTGGTG	ACTTGCGTTGATGGTGGCTGTCTT	282
PGE synthase-1	GTTTGGTGATGGAGAACAGC	GTAGACGAAACCAAGGAAGAGG	257
PGE synthase-2	CCAGTACAAGACATGTCCCTTC	GTACACGTTGGGAGAGATGAGA	439

COX, cyclooxygenase; EP, E prostanoid; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; PGE, prostaglandin E; RT-PCR, reverse transcription-polymerase chain reaction.

Western blot analysis

To detect phosphorylated PKCδ and PKCε, Western blots were carried out in accordance with the method described previously (Vary et al., 2005), with modifications noted below. Liver MFs were transferred to ice-cold homogenization buffer [50 mmol·L⁻¹ Tris-HCl, pH 7.4, 1% NP-40, 0.25% sodium deoxycholate, 150 mmol·L⁻¹ NaCl, 1 mmol·L⁻¹ EDTA, 1 mmol·L⁻¹ Na₃VO₄, 1 mmol·L⁻¹ NaF, 1 mg·mL⁻¹ Pefabloc SC (protease inhibitor) and 1 mg·mL⁻¹ Complete (protease inhibitor cocktail)] followed by centrifugation at 20 000× g for 30 min at 4°C, and the supernatant was collected for analysis. The blots on polyvinylidene difluoride (PVDF) membranes were probed with anti-phospho-PKCδ^Thr505 and phospho-PKCε^Ser729 antibodies (1:500 each). For secondary reaction, anti-mouse IgG (Alexa Fluor 680) or anti-rabbit IgG (IRDye800) were used (1:10 000 each). Bands were detected and quantified with the Odyssey system (LI-COR Biosciences, USA). To correct for loading variations, the result was expressed as a ratio of phospho/total PKC with the control ratio set at 1.0.

Cell morphology and viability

For Giemsa staining, MFs were fixed in 4% paraformaldehyde for 5 min, incubated in Giemsa solution (2% in 10 mmol·L⁻¹ phosphate buffer, pH 7.4) for 1 h, washed in distilled water, dehydrated and covered with glass. For actin staining, cells were fixed in 4% paraformaldehyde and incubated with 0.01% Triton and 10% normal goat serum (Chemicon International, USA) for 30 min at room temperature. Cells were probed with anti-smooth muscle actin antibody (1:100) followed by anti-mouse IgG (Alexa Fluor 568, 1:100). Cells were finally incubated with DAPI (1 µg·mL⁻¹) for 5 min. For Trypan blue staining, cells were trypsinized, resuspended in 0.3% Trypan blue solution and counted using a haemocytometer. The cells with and without blue dye staining were recorded as dead and alive respectively.

Materials

Drug and molecular target nomenclature conforms to the BJP's Guide to Receptors and Channels (Alexander et al., 2008). The chemicals obtained were: PGE₂ and AL8810 (Cayman Chemical, USA), ATP disodium salt, ionomycin calcium salt and DAPI (Sigma, USA), Ro 31-8425 (Calbiochem, Germany), phorbol 12,13-dibutyrate (PDBu) and rottlerin (Biomol International, USA), Pefabloc SC and Complete (Roche Applied Science, USA), anti-PKCδ and anti-PKCε antibodies (BD Biosciences, USA), anti-phospho-PKCδ^Thr505 antibody (Cell Signaling Technology, USA), anti-phospho-PKCε^Ser729 antibody (Upstate Biotechnology, USA), anti-rabbit IgG conjugated to Alexa Fluor 680 and anti-mouse IgG conjugated to Alexa Fluor 568 (Molecular Probes, USA), anti-mouse IgG conjugated to IRDye800 (Rockland, USA) and anti-smooth muscle actin antibody (DAKO, Denmark). ONO-DI-004 (EP₁ receptor agonist), ONO-AE1-259-01 (EP₂ receptor agonist), ONO-AE-248 (EP₃ receptor agonist) and ONO-AE1-329 (EP₄ receptor agonist) were gifts kindly provided by Ono Pharmaceutical Company Ltd. (Osaka, Japan).

Statistical analysis

The results of the experiments are expressed as mean ± SEM. Statistical evaluation of the data was performed by paired or unpaired Student's t-test for comparison between two groups, and by one-way analysis of variance followed by Dunnett's test for comparisons between more than two groups. Statistical significance was established at P-values lower than 0.05.

Results

EP receptor subtypes

We performed RT-PCR in liver MFs to investigate the expression of the four subtypes of EP receptor. As shown in Figure 1A, MFs expressed mRNAs for all four EP receptor subtypes: EP₁, EP₂, EP₃ and EP₄. They also expressed COX-1, COX-2, mPGES-1 and mPGES-2, which are the essential enzymes in PGE₂ biosynthesis from arachidonic acid (Figure 1B).

Figure 1.

Expression of the four subtypes of E prostanoid (EP) receptor EP₁ through EP₄ (A) and cyclooxygenase-1 (COX-1), COX-2, microsomal prostaglandin E synthase-1 (mPGES-1) and mPGES-2 (B) in liver myofibroblasts (MFs). Figures show representative results from RT-PCR. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; RT-PCR, reverse transcription-polymerase chain reaction.

Effect of PGE₂ on liver MF contraction

We next performed the collagen gel contraction assay in order to investigate the contractile effect of PGE₂. Hydrated collagen gels provide a potent model for examining in vitro the reciprocal mechanical interactions that occur between cells and the ECM (Tomasek et al., 2002). Liver MFs cultured on collagen gels were stimulated with PGE₂ after cell attachment and serum starvation. As shown in Figure 2A, PGE₂ induced cell contraction in a dose-dependent manner (0.1–10 µmol·L⁻¹, n = 5). Contractile effects of PGE₂ were observed at each time point during the experiment. PGE₂ (1 µmol·L⁻¹) induced cell contraction of 44.8 ± 3.0% at 8 h. Data are shown as the percentage of maximal contraction induced by 5% FBS.

Figure 2.

Prostaglandin E₂ (PGE₂) induced collagen gel contraction in liver myofibroblasts. (A) Collagen gels were treated in DMEM containing a range of concentrations of PGE₂: 0.1, 0.3, 1 and 10 µmol·L⁻¹. (B) Effect of E prostanoid receptor subtype-selective agonists on collagen gel contraction: EP₁ (ONO-DI-004), EP₂ (ONO-AE1-259-01), EP₃ (ONO-AE-248) and EP₄ (ONO-AE1-329). All the agonists were used at a concentration of 1 µmol·L⁻¹. (C) Effect of EP₃ receptor agonist ONO-AE-248 over a range of concentrations: 0.1, 1 and 10 µmol·L⁻¹. Gels were photographed to quantify the surface area. Relative contraction was calculated as shown in Methods. Data are presented as mean ± SEM. *P < 0.05, **P < 0.01 compared with unstimulated condition. DMEM, Dulbecco's modified Eagle's medium.

Effect of EP receptor-selective agonists on cell contraction

Liver MFs cultured on collagen gels were stimulated with a selective agonist of each EP receptor to examine which subtype is responsible for the contractile effect of PGE₂. The EP₃ receptor agonist ONO-AE-248 (1 µmol·L⁻¹) produced contraction of 31.7 ± 1.7% at 8 h after stimulation (P < 0.01), whereas neither ONO-DI-004 (EP₁ receptor agonist), ONO-AE1-259-01 (EP₂ receptor agonist), nor ONO-AE1-329 (EP₄ receptor agonist) induced contraction. The EP₄ receptor agonist ONO-AE1-329 showed a slight suppressive effect at 1 h (Figure 2B, P < 0.05, n = 6). As shown in Figure 2C, the contractile effects of ONO-AE-248 were dose-dependent (0.1–10 µmol·L⁻¹, n = 7). These results suggest that PGE₂ induces contraction in MFs via the EP₃ subtype of the PGE receptor.

[Ca²⁺]_i response to PGE₂ and a selective agonist of the EP₃ receptor

To explore the downstream pathways of EP₃ receptor, we examined the involvement of intracellular Ca²⁺ in liver MFs treated with PGE₂ or an EP₃ selective agonist. PGE₂ (1 µmol·L⁻¹) and ONO-AE-248 (1–10 µmol·L⁻¹) did not increase [Ca²⁺]_i, while PGE₂ induced a slight but apparent increase in [Ca²⁺]_i at a concentration of 10 µmol·L⁻¹ (Figure 3A). We previously reported that ATP induces sufficient [Ca²⁺]_i increase in MFs (Kojima et al., 2007). Compared with [Ca²⁺]_i increase by 10 µmol·L⁻¹ ATP, the amount of change in AUC induced by 10 µmol·L⁻¹ PGE₂ was only 23% (AUC, 0–2 min). Because of their similarity of chemical structure, prostanoids are known to react with different prostanoid receptors, especially at higher concentration. We thus examined the effect of antagonism of the FP receptor on [Ca²⁺]_i elevation induced by10 µmol·L⁻¹ PGE₂. As shown in Figure 3B, the FP receptor antagonist AL8810 (3 µmol·L⁻¹) completely inhibited this [Ca²⁺]_i increase (P < 0.01, n = 4).

Figure 3.

Effect of prostaglandin E₂ (PGE₂) and ONO-AE-248 on [Ca²⁺]_i. (A) Liver myofibroblasts cultured on glass coverslips were stimulated with PGE₂ and ONO-AE-248 (1 or 10 µmol·L⁻¹), or ATP (10 µmol·L⁻¹). Changes in [Ca²⁺]_i were measured using Ca²⁺ fluorescent dye fura-2. (B) The area under [Ca²⁺]_i-time curve after stimulation (AUC: 0–2 min). Data are presented as mean ± SEM of 52–110 cells from four separate experiments. **P < 0.01 compared with 10 µmol·L⁻¹ PGE₂-treated cells. AUC, area under curve.

Effect of PKC inhibitors on cell contraction

We next examined the effect of PKC inhibitors on cell contraction induced by PGE₂ or the EP₃ receptor agonist. Liver MFs on collagen gels were pretreated with compounds that differentially inhibit different PKC isozymes. Ro 31-8425 is a broad range PKC inhibitor for both Ca²⁺-dependent conventional PKC (cPKC) and Ca²⁺-independent novel PKC (nPKC), showing threefold to fourfold selectivity for cPKCs versus PKCε. Rottlerin specifically inhibits nPKC isozyme PKCδ. In the present study, cells were pretreated with these inhibitors for 30 min prior to the treatment with 1 µmol·L⁻¹ of PGE₂ or ONO-AE-248. As shown in Figure 4A,B, Ro 31-8425 (30–300 nmol·L⁻¹) significantly inhibited PGE₂- and ONO-AE-248-induced MF contractions (P < 0.05, n = 5). Rottlerin (0.3–3 µmol·L⁻¹) also inhibited cell contraction in a dose-dependent manner (Figure 4C,D, P < 0.01, n = 5).

Figure 4.

Ro 31-8425 and rottlerin inhibited liver myofibroblast (MF) contractions induced by PGE₂- or ONO-AE-248. Cells were stimulated with 1 µmol·L⁻¹ of PGE₂ (A, C) or ONO-AE-248 (B, D) after pretreatment for 30 min with either protein kinase C (PKC) inhibitor: Ro 31-8425 (general PKC inhibitor: 30, 100 and 300 nmol·L⁻¹) (A, B) or rottlerin (specific nPKC inhibitor: 0.3, 1 and 3 µmol·L⁻¹) (C, D). Data are presented as mean ± SEM. *P < 0.05, **P < 0.01 compared with 1 µmol·L⁻¹ PGE₂- or ONO-AE-248-treated cells.

Effect of PKC inhibitors on cell morphology and viability

We confirmed whether both PKC inhibitors influence MF morphology and viability under the present conditions. As shown in Figure 5A, we could not observe any changes in MF cell shape and expression of α-smooth muscle actin after 8 h treatment with Ro 31-8425 (300 nmol·L⁻¹) or rottlerin (3 µmol·L⁻¹). We also examined cell viability using Trypan blue staining. MFs were stimulated for 8 h with Ro 31-8425 (300 nmol·L⁻¹) or rottlerin (3 µmol·L⁻¹) and the percentage of viable cells were not different from that under unstimulated condition (Figure 5B, n = 5).

Figure 5.

Ro 31-8425 and rottlerin did not influence myofibroblast (MF) morphology and viability. MFs were treated with Ro 31-8425 (300 nmol·L⁻¹) or rottlerin (3 µmol·L⁻¹) for 8 h. (A) MF were fixed in paraformaldehyde and stained in Giemsa solution (upper row) or probed with anti-smooth muscle actin antibody (lower row). (B) Cells after 8 h stimulation were resuspended in 0.3% Trypan blue solution, and viable and dead cells were counted. Data are presented as mean ± SEM. No significant difference was observed between control and 8 h stimulated conditions.

PGE₂ promotes PKCδ and PKCε phosphorylation

To directly examine whether PGE₂ activates nPKC, we next evaluated the phosphorylation of nPKC isozymes, PKCδ (Thr505, Figure 6A) and PKCε (Ser729, Figure 6B). While 5 min stimulation with 1 µmol·L⁻¹ PGE₂ did not phosphorylate both PKC isozymes, 15 min stimulation significantly increased phosphorylation of both PKCδ^Thr505 (P < 0.01, n = 4) and PKCε^Ser729 (P < 0.05, n = 4). For both PKC isozymes, we confirmed that a known PKC activator PDBu (1 µmol·L⁻¹, 5 min) increased significantly their phosphorylation (Figure 6A,B, far right lane). We next examined whether PKC inhibitors can block these phosphorylations. 30 min pretreatment with Ro 31-8425 (300 nmol·L⁻¹) and rottlerin (3 µmol·L⁻¹) influenced neither PKCδ (Figure 6C) nor PKCε (Figure 6D) phosphorylation induced by PGE₂ (1 µmol·L⁻¹, 15 min).

Figure 6.

Effect of prostaglandin E₂ (PGE₂) on protein kinase C (PKC) phosphorylations. Cell homogenates were analysed by Western blotting after stimulation with: (A, B) 1 µmol·L⁻¹ PGE₂ (5 or 15 min) or 1 µmol·L⁻¹ PDBu (5 min), (C, D) 30 min pretreatment with Ro 31-8425 or rottlerin followed by 1 µmol·L⁻¹ PGE₂ (15 min). Anti-phospho-PKCδ^Thr505 antibody (A, C) or anti-phospho-PKCε^Ser729 (B, D) antibody was used. The phosphorylation was measured by densitometry and normalized as described in Methods. Representative pictures are also shown. Data are presented as mean ± SEM. *P < 0.05, **P < 0.01 compared with basal control.

FP receptor antagonist partially blocked 10 µmol·L⁻¹ PGE₂-induced cell contraction

We examined the effect of the FP receptor antagonist AL8810 on cell contraction induced by PGE₂. As shown in Figure 7A, 30 min pretreatment with AL8810 (3 µmol·L⁻¹) partially blocked the contraction induced by 10 µmol·L⁻¹ PGE₂. ONO-AE-248 (10 µmol·L⁻¹) induced only 59.7% contraction as compared with PGE₂ at 1 h, while the difference was abolished by AL8810 treatment (P < 0.05, n = 4). In contrast, AL8810 did not affect cell contraction induced by ONO-AE-248 or 1 µmol·L⁻¹ PGE₂ (n = 4, data not shown). When assessed by RT-PCR, MFs expressed FP receptor as well as EP receptors (Figure 7B).

Figure 7.

The F prostanoid (FP) receptor selective antagonist AL8810 partially inhibited 10 µmol·L⁻¹ prostaglandin E₂ (PGE₂)-induced liver myofibroblast (MF) contraction. (A) Cells were stimulated with 10 µmol·L⁻¹ PGE₂ after pretreatment for 30 min with AL8810 at a concentration of 3 µmol·L⁻¹. Data are presented as mean ± SEM. **P < 0.01 compared with 10 µmol·L⁻¹ PGE₂-treated cells. (B) Representative results from RT-PCR showing the expression of the FP receptor. RT-PCR, reverse transcription-polymerase chain reaction.

Discussion and conclusions

In the present study, we examined the effect of PGE₂ on isolated liver MF contractility. We demonstrated that PGE₂ up to 1 µmol·L⁻¹ induced cell contraction via the EP₃-PKC pathway which was not accompanied by a detectable [Ca²⁺]_i increase and that at higher concentrations, PGE₂ activated the FP receptor resulting in a [Ca²⁺]_i increase and additional cell contraction.

Recent studies have clearly shown that MFs possess a different origin from HSCs (Knittel et al., 1999a; Kinnman et al., 2003). Although both cells exhibit a similar phenotype regarding contractility and ECM secretion, some reports have provided evidence that these cells differentially contribute to liver cirrhosis. Proliferation of MFs observed in a cirrhotic liver suggests a pivotal role of this cell type, while HSCs also undergo activation during liver injury but fail to proliferate in vivo and also in vitro (Ramadori and Saile, 2002; Saile et al., 2002). In this study, we sought to investigate the contractility of MFs which still remained largely unknown.

Many reports have shown that non-parenchymal cells in sinusoidal spaces were positively stained with COX-2 in various models of liver disease (Yamamoto et al., 2003; Planaguma et al., 2005). These results allow us to hypothesize that liver MFs are one of the major sources of PGE₂ upon liver injury. Indeed, MFs expressed COX-2 as well as mPGES-1 and mPGES-2, indicating that secreted PGE₂ may control the function of adjacent non-parenchymal cells in an autocrine and/or paracrine manner. Previous reports show that liver injury or inflammation increases the local PGE₂ concentration in sinusoids up to several hundred or thousand nanomolar (Neuschafer-Rube et al., 1993; Devaux et al., 2001). These reports support that PGs in our study are within a physiologically relevant concentration in the liver. In the present study, MFs expressed a noticeable level of COX-2 expression even under unstimulated conditions (Figure 1B). We previously reported that scaffold materials influence MF bioactivity (Kojima et al., 2007). Therefore incubation on plastic dishes may have provided some stimulation of COX-2 expression in MFs. Further investigation is needed to clarify this point.

Prostaglandin E₂ is thought to regulate important liver functions such as glucose homeostasis, delivery of blood substances to hepatocytes and lipid oxidation (Bradford et al., 1999; Enomoto et al., 2000; Pestel et al., 2002). In this study, we focused on its contractile effects on liver MFs and demonstrated that exogenous administration of PGE₂ promoted cell contraction. PGE₂ is known to produce a broad range of biological effects through its binding to specific receptors namely EP₁, EP₂, EP₃ and EP₄. In liver MFs, all four EP receptor subtypes were observed at mRNA level. We then applied four EP subtype-selective agonists to see their contractile effects individually and found that only ONO-AE-248, a selective agonist for EP₃, receptors, induced contraction. Our results are consistent with the previous reports showing that EP₃ receptors mediate PGE₂-induced contraction in human pulmonary artery (Qian et al., 1994). Another group also suggested that both EP₁ and EP₃ receptors are involved in constriction of adult porcine large cerebral arteries (Jadhav et al., 2004). However, the selective agonist for EP₁ receptors did not induce contraction in rat liver MFs, even at a concentration of 10 µmol·L⁻¹ (data not shown).

Some other studies clarified that PGE₂ exhibits vasodilator effects through EP₂ receptor in human dermal fibroblasts, or through EP₄ receptor in mice aorta and human middle cerebral arteries, which couple efficiently to G_s or G_i (Davis et al., 2004; Sandulache et al., 2006). Consistent with these observations, we showed that the EP₄ selective agonist ONO-AE1-329 slightly inhibited rat liver MF contraction at an early period of PGE₂ treatment. This dilation may be mediated through intracellular cAMP generation resulting from EP₄ receptor-coupled G_s activation. In this study, we could not observe any contractile effect after EP₁ and EP₂ receptor stimulation. These receptors may be involved in other pathophysiological roles in MFs, such as cell proliferation, migration, or ECM secretion.

Although a previous study indicated that EP₃ receptor stimulation caused smooth muscle contraction through Ca²⁺-signaling pathways in guinea-pig aorta (Jones et al., 1998), our study demonstrated that PGE₂ or a EP₃ selective agonist caused MF contraction without changing [Ca²⁺]_i at a concentration of 1 µmol·L⁻¹. We therefore sought to reveal the Ca²⁺-independent pathway in MF contraction induced by 1 µmol·L⁻¹ of PGE₂ or ONO-AE-248.

In vascular smooth muscle, a Ca²⁺-independent pathway, known as Ca²⁺ sensitization, is proposed to be an important component of the constrictor response to many receptor agonists. Several key intracellular protein kinases including PKC have been thought to be involved in pathways leading to Ca²⁺-sensitization (Somlyo and Somlyo, 2003). PKC can be subdivided into three classes based on primary structure and biological properties: conventional PKC isozymes (cPKC), novel PKC isozymes (nPKC) and atypical PKC isozymes (aPKC) (Ward et al., 2004). Our results showing no [Ca²⁺]_i increase in the presence of 1 µmol·L⁻¹ of PGE₂ or ONO-AE-248 rule out the involvement of cPKC, which requires [Ca²⁺]_i increase on activation. In our study, we revealed that rottlerin (a specific nPKC inhibitor) as well as Ro 31-8425 (a general PKC inhibitor) strongly inhibited 1 µmol·L⁻¹ PGE₂- or ONO-AE-248-induced contraction. The general PKC inhibitor Ro 31-8425 is known to inhibit both cPKCs- and PKCε-activity while rottlerin inhibits PKCδ more potently than other nPKC isozymes. The effects of these inhibitors were specific to the contractile response against PGE₂, because the administration of each PKC inhibitor alone did not induce toxicity in terms of cell morphology and viability. We also observed that PKCδ and PKCε undergo rapid phosphorylation in response to MF stimulation with PGE₂. These data show the possible involvement of PKCδ and PKCε in the primary component of EP₃ selective agonist action.

Ro 31-8425 and rottlerin are known to bind to the ATP binding site at the catalytic domain of PKC (Muid et al., 1991; Gschwendt et al., 1994). Although it is well-known that cPKCs are phosphorylated by a Ro 31-8425- and rottlerin-insensitive kinase, PDK-1 (phosphoinositide-dependent kinase-1), the detailed relationship of nPKC phosphorylation and the inhibitory effect of PKC inhibitors is still unclear (Way et al., 2000; Steinberg, 2004). As shown in Figure 6C,D, pretreatment with Ro 31-8425 and rottlerin did not influence nPKC phosphorylations. Our results suggest that nPKCs may be phosphorylated by a kinase insensitive to Ro 31-8425 and rottlerin in liver MFs. Further investigation should be made to reveal the detailed interactions between PKC inhibition and the inhibition of contractile activity by these agents.

As well as PKC signaling pathways, it has been suggested that the other pathways also involve in Ca²⁺ sensitization. Recently, the small GTPase, Rho, and Rho-kinase has been investigated as an important regulator of HSC proliferation, migration and contraction (Tangkijvanich et al., 2001; Ramm et al., 2003; Laleman et al., 2007) but the involvement of Rho and other related factors in MF contraction still remains to be tested.

A higher concentration of PGE₂ (10 µmol·L⁻¹) caused rapid and greater MF contraction compared with that with the low concentration (1 µmol·L⁻¹). A notable finding in the [Ca²⁺]_i measurement is that this PGE₂-induced contraction at 10 µmol·L⁻¹ is accompanied by a small increase of [Ca²⁺]_i, suggesting that different signal pathways are involved. PGE₂ and prostaglandin F_2α (PGF_2α) are structurally identical, except at the C-9 position in the cyclopentane ring where PGE₂ has a keto substituent and PGF_2α has a hydroxyl. Also, PGE₂ binds to the FP receptor in CHO cells and HEK-293 cells (Kiriyama et al., 1997; Fujino et al., 2004). Therefore, it is possible that the [Ca²⁺]_i increase by 10 µmol·L⁻¹ PGE₂ in rat liver MFs was due to FP receptor activation, whose downstream pathway has been classically characterized by Ca²⁺ signaling (Breyer et al., 2001). We demonstrated here that the selective antagonist for FP receptor, AL8810, significantly inhibited [Ca²⁺]_i increase by 10 µmol·L⁻¹ PGE₂. AL8810 also inhibited the rapid contraction by 10 µmol·L⁻¹ PGE₂, which indicates the involvement of the FP receptor in PGE₂-induced MF contraction, especially at higher concentrations.

In summary, we showed that low concentrations of PGE₂ induced contraction in liver MFs via EP receptor subtype EP₃. This effect appears to involve an enhancement of the Ca²⁺ sensitization through a Ca²⁺-independent, novel PKCδ and/or PKCε-mediated pathway. We also demonstrated that a high concentration of PGE₂ activated the PGF_2α receptor FP and increased [Ca²⁺]_i, leading to stronger MF contraction than through the EP₃-PKC pathway. A better understanding of the intracellular signal transduction mechanisms leading to MF contraction might lead to the identification of novel potential targets for the treatment of portal hypertension and cirrhosis. The present results may provide new insights into the development of a pharmacological therapeutic strategy targeting PGE₂ and its receptor subtypes.

Glossary

Abbreviations:

ECM

extracellular matrix

GPCR

G-protein-coupled receptor

HSC

hepatic stellate cell

MF

myofibroblast

mPGES

microsomal prostaglandin E synthase

PGE₂

prostaglandin E₂

PGF_2α

prostaglandin F_2α

Conflicts of interest

None.