A total of 132 Male Sprague-Dawley rats (6 weeks old) that weighed 200-220 g were purchased from Sino-British SIPPR/BK Lab Animal Ltd. (Shanghai, China). The animals were housed in standard cages and maintained under standard conditions at a constant room temperature of 20-25°C, a humidity of 40-70% and a 12 h/12 h light/dark cycle, with unrestricted access to food and water. All experiments were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Changzheng Hospital Ethics Committee [approval number, CZEC (2015)-01].

A rat model was constructed using 70% partial hepatic ischemia for 60 min as described previously (21-24). Briefly, rats were fasted for 12 h before surgery and anesthetized by intraperitoneal (i.p.) injection of 40 mg/kg pentobarbital. After midline laparotomy, the interlobular ligaments were dissected. The left hepatoduodenal ligament containing the hepatic artery, portal vein and bile duct leading to the left and median lobe was clamped in the liver hilum using a microvascular clamp for 60 min. Reperfusion was initiated by clamp removal. Sham-operated rats underwent the same surgical procedures but without vascular occlusion. The animals were sacrificed by an intraperitoneal injection of sodium pentobarbital (100 mg/kg) at 2 or 6 h after reperfusion. Death of the rats was verified by a combination of criteria, including lack of pulse, breathing, corneal reflex, response to a firm toe pinch and graying of the mucous membranes. Liver and serum samples were collected for further analysis.

To increase EP4 activity, a dose of 0.1, 0.5, 1 or 10 mg/kg of an EP4 agonist (CAY10598 [CAY]; Cayman Chemical Company) was administered subcutaneously to animals 0, 0.5 and 2.5 h prior to the onset of liver reperfusion. To increase MPTP susceptibility, a single dose of carboxyatractyloside (CATR; 5 mg/kg, Sigma-Aldrich; Merck KGaA) was dissolved in 0.9% saline and then administered intraperitoneally to animals 30 min prior to the 60 min ischemic insult. Additionally, to inhibit ERK1/2 activity, rats received an intraperitoneal injection of an ERK1/2 inhibitor [PD98059 (PD), Cell Signaling Technology, Inc.], 5 mg/kg dissolved in dimethyl sulfoxide (DMSO) (Cell Signaling Technology, Inc.), 30 min prior to the onset of liver ischemia. The sham and I/R animals received the same volume of the respective vehicles.

Liver samples were fixed in 4% formaldehyde/phosphate-buffered saline overnight at 4°C. The samples were dehydrated and embedded in paraffin. Hepatocellular necrosis was calculated in hematoxylin and eosin (H&E)-stained tissue using a semi-quantitative scale, as previously described (21). Briefly, for each H&E-stained tissue sample, a total of 30 randomly selected high-power visual fields (magnification, ×400) were analyzed in a blinded fashion to determine the percentage of necrotic cells. In the current study, only grade 3 severe injury with hepatic cord disintegration was categorized as necrosis (21).

To evaluate apoptosis, terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) staining was also performed using a commercial kit from Roche Diagnostics, according to the manufacturer's instructions. Briefly, paraffin-embedded tissue sections were dewaxed by heating at 60°C followed by washing in xylene, rehydration using a graded series of ethanol (100, 95, 90, 80 and 70%) and double-distilled water, and then pretreatment with proteinase-k. Endogenous peroxidase activity was blocked by immersing in 3% H₂O₂ in methanol for 10 min at room temperature. TUNEL reaction mixture and Converter-POD were then added, and incubated in a humidified atmosphere for 60 and 30 min at 37°C in the dark, respectively. Each slice was stained using DAB for 10 min at room temperature, and liver cell apoptosis was observed under light microscopy. In each section, areas without significant necrosis in 10 different visual fields (magnification, ×400) were observed and analyzed for TUNEL-positive cells. The TUNEL index was calculated by counting the total nuclei and cells with brown nuclei and using the following formula: [(number of stained cells)/(number of stained cells + number of unstained cells)] ×100. Six sections of each tissue sample were analyzed for each group.

For the immunohistochemical analysis of EP4, liver tissue sections (5-µm thick) were then immunostained using the EnVision method (DakoCytomation; Agilent Technologies, Inc.) according to the manufacturer's instructions. Briefly, tissue sections were pre-treated for 10 min with peroxidase-blocking reagent (Dako; Agilent Technologies, Inc.) to suppress endogenous peroxidase and pseudoperoxidase activity at room temperature, washed again in phosphate-buffered saline(PBS) and incubated for 1 h at 37°C in a humid chamber with the anti-EP4 polyclonal antibody (1:100; cat. no. SC-20677; Santa Cruz Biotechnology, Inc.). The slides were rinsed twice and then incubated for 60 min with goat anti-mouse immunoglobulins conjugated to a peroxidase-labeled polymer (1:500; cat. no. P0447; DAKO EnVision™ + System; Dako; Agilent Technologies, Inc.) at room temperature. After washing, cells were treated with 3,3′-diaminobenzidine (Sigma-Aldrich; Merck KGaA) as a chromogen substrate for 10 min at room temperature. Slides were then washed again in PBS and counterstained with hematoxylin for 2 min at room temperature. Sections stained with H&E were used for comparison and compared using a light microscope with a magnification of ×400.

The levels of EP4 (Santa Cruz Biotechnology, Inc.), ERK1/2, phosphorylated (p-) ERK1/2, janus kinase 2 (JAK2), p-JAK2, signal transducer and activator of transcription 3 (STAT3), p-STAT3, glycogen synthase kinase-3β (GSK3β) and p-GSK3β (Ser9; Cell Signaling Technology, Inc.) were evaluated in rat liver lysates using western blot analysis, as previously reported (25). Briefly, liver tissues were homogenized in RIPA lysis buffer (Wuhan Servicebio Technology Co., Ltd.), followed by centrifugation at 12,000 × g for 15 min at 4°C. The bicinchoninic acid method was used for quantitative analysis the protein samples. A total of 40 µg protein was applied to each lane and separated via 10% SDS-PAGE before being transferred onto PVDF membranes. A total of 5% bovine serum albumin (Sigma-Aldrich; Merck KGaA) and TBS with 0.1% Tween-20 (PBST) was used to block the unspecific binding of antibodies for 2 h at 4°C. Then, blots were incubated with the following primary antibodies: EP4 (1:1,000; cat. no. SC-20677; Santa Cruz Biotechnology, Inc.), ERK1/2 (1:1,000; cat. no. 4695; Cell Signaling Technology, Inc.), p-ERK1/2 (1:2,000; cat. no. 4370; Cell Signaling Technology, Inc.), JAK2 (1:1,000; cat. no. 3230; Cell Signaling Technology, Inc.), p-JAK2 (1:1,000; cat. no. 3776; Cell Signaling Technology, Inc.), STAT3 (1:1,000; cat. no. 30835; Cell Signaling Technology, Inc.), p-STAT3 (1:2,000; cat. no. 9145; Cell Signaling Technology, Inc.), GSK3β (1:1,000; cat. no. 5676; Cell Signaling Technology, Inc.), p-GSK3β (1:1,000; cat. no. 9322; Cell Signaling Technology, Inc.) and GAPDH (1:1,000; cat. no. 5174; Cell Signaling Technology, Inc.), overnight at 4°C. Blots were then incubated with a goat-anti-rabbit IgG-HRP (1:2,000; cat. no. 7074; Cell Signaling Technology, Inc.) for 1.5 h at room temperature and washed three times with TBST. An enhanced chemiluminescence (ECL) western blotting detection reagent (cat. no. 34076; Thermo Fisher Scientific, Inc.) was used for visualization. A digital gel image analysis system (model GIS-1000; Tanon) were used to evaluate the specific signals and the corresponding band intensities.

Serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels were determined after 2 and 6 h of reperfusion using an AutoAnalyzer (H-7600, Hitachi Ltd.).

Liver tissue were collected after 2 and 6 h of reperfusion. Tissue ROS levels were measured using a ROS enzyme-linked immunosorbent assay (ELISA) kit (Cell Signaling Technology, Inc.) according to the manufacturer's instructions. The results are expressed as relative fluorescence units (U/ml).

Mitochondria were isolated in liver tissue at 2 and 6 h of reperfusion using differential centrifugation at 4°C as previously described (26). Briefly, the fresh liver tissues (1 g) were homogenized with 8 ml of isolation buffer A (ISA) containing 220 mmol/l d-mannitol, 70 mmol/l sucrose, 10 mmol/l Tris-HCl, 1 mmol/l EGTA and 0.4% bovine serum albumin (pH 7.4). The homogenates were centrifuged at 850 × g for 10 min to collect supernatants, followed by centrifugation at 10,000 × g for an additional 10 min. The mitochondrial pellet was resuspended in a final washing isolation buffer B containing 220 mmol/l d-mannitol, 70 mmol/l sucrose and 10 mmol/l Tris-HCl (pH 7.4). Protein concentration was determined via the biuret method and calibrated with bovine serum albumin (BSA). A CRC assay was used to evaluate MPTP susceptibility, which was adapted from a previously described method (27). CRC is defined as the amount of Ca²⁺ required to trigger mass Ca²⁺ release from isolated liver mitochondria. The extramitochondrial Ca²⁺ concentration was recorded using a fluorescence microplate reader controlled by SoftMax Pro 4.8 software (Molecular Devices, LLC) in the presence of the Calcium Green-5N molecular probe (1 µmol/l, Invitrogen; Thermo Fisher Scientific, Inc.), with excitation and emission wavelengths of 500 and 530 nm, respectively (7). The fluorescence scan interval was set at 12 sec.

Electron microscopy was used to examine the ultrastructure of the hepatocytes in liver tissue samples after 6 h of reperfusion. Briefly, the liver samples were fixed in a 1% osmic acid fixative solution (Wuhan Servicebio Technology Co., Ltd.) at 4°C for 3 h and then washed with phosphate-buffered saline (PBS) three times at 4°C for 30 min. Dehydration was carried out in graded ethanol solutions (50, 70 and 90%), in a 90% solution (ethanol:acetone; 1:1; v/v) and finally in a 90% acetone solution at 4°C for 20 min. Then, the samples were washed three times with acetone at room temperature for 30 min. The tissue samples were embedded in a solution [acetone: Optimal cutting temperature compound (Wuhan Servicebio Technology Co., Ltd.); 2:1; v/v] for 4 h and then stored overnight at room temperature. Subsequently, the tissues were embedded in 100% embedding medium for 3 h at 37°C. Next, the tissues were cured in a drying oven at 37°C overnight, at 45°C for 12 h, and at 60°C for 24 h. Finally, the tissues were cut to 50-nm thick sections using an ultramicrotome and stained with 2% uranyl acetate and lead citrate (Wuhan Servicebio Technology Co., Ltd.) with pH 12 at room temperature for 10 min. The ultrathin sections were then examined and imaged using a Hitachi H-7650 transmission electron microscope (TEM; Hitachi, Ltd.), with magnification ×3,500.

Liver tissue were collected at 6 h of reperfusion. A cAMP complete ELISA kit (cat. no. ab133051; Abcam) was used to determine the cAMP levels in liver tissue samples, according to the manufacturer's instructions.

SPSS 11.0 (SPSS, Inc.) was used for the statistical analysis. The data are expressed as the mean ± standard deviation (SD). Statistical analyses were performed using unpaired two-tailed Student's t-test or one-way ANOVA followed by Tukey's post-hoc test, with P<0.05 considered to indicate statistical significance.

In the present study, EP4 protein expression during rat hepatic I/R injury was examines using immunochemistry and western blot analyses. As indicated in Fig. 1, EP4 expression was downregulated in hepatocytes compared with the sham liver samples. After 2 h of reperfusion following 60 min of ischemia, the EP4 protein was prominently expressed in the membranes as well as in the cytoplasm of hepatocytes, indicating that I/R induced EP4 protein expression during early reperfusion. The expression of EP4 was lower in I/R-injured livers after 6 h compared with after 2 h of reperfusion (Fig. 1B), indicating that decreased EP4 expression may be associated with liver injury during reperfusion. Thus, in the following experiments, the effects of activated EP4 signaling on hepatic I/R injury were examined by subcutaneously injecting rats with the EP4-specific agonist CAY prior to reperfusion.

To examine the dose-response relationship between CAY and hepatic I/R, rats received CAY three times at a dose of 0.1, 0.5, 1 or 10 mg/kg. Liver sections obtained from the ischemic lobe after 6 h of reperfusion were subjected to histopathological analysis (Fig. 2A). A semiquantitative scale was used to further determine the level of hepatocellular necrosis. The necrosis rates after 6 h of reperfusion following 60 min of ischemia for the 0.1, 0.5, 1 and 10 mg/kg treatments were 33.5±9.4, 31.7±8.3, 22.5±6.0 and 20.0±5.1, respectively (Fig. 2C). The current data indicate that this EP4 agonist inhibits hepatic I/R injury in a dose-dependent manner. Therefore, the minimal effective dose of CAY (1 mg/kg) was used to further investigate the protective mechanisms of the EP4 agonist against hepatic I/R.

An MPTP opener, CATR (28), was used to determine the effect of the EP4 agonist on MPTP modulation during I/R. As indicated in Fig. 3A, I/R insult significantly induced hepatic injury by increasing serum ALT and AST levels in animals over 2 and 6 h of reperfusion, and treatment with 1 mg/kg CAY markedly decreased the serum ALT and AST levels in I/R animals. Next, hepatocyte necrosis and apoptosis were evaluated in each group after 6 h of reperfusion using H&E and TUNEL staining. As indicated in Fig. 3B and C, the hepatocyte necrosis rates and TUNEL index values were significantly lower in the I/R + CAY group compared with the I/R group. Notably, the protective effect produced by CAY was partially reversed when the MPTP opener CATR was administered 30 min prior to I/R insult. The current results indicate that EP4 activation with CAY significantly protects the liver, and that MPTP opening may serves an important role in this effect.

To further clarify the effect of EP4 signaling on MPTP modulation during I/R, ROS levels were examined in hepatic tissues and isolated hepatocyte mitochondria, in order to determine the CRC, which is an index used to evaluate MPTP susceptibility, after 2 and 6 h of reperfusion. As shown in Fig. 4, compared with the I/R group, the I/R + CAY group exhibited significantly reduced liver ROS levels (Fig. 4A) and increased mitochondrial CRC values (Fig. 4B), indicating that MPTP opening is inhibited by EP4 activation. However, when the MPTP opener CATR was administered prior to ischemia, the protective effects induced by the EP4 agonist were significantly reversed; the ROS levels in liver tissues were significantly increased (Fig. 4A) and the CRC levels significantly reduced (Fig. 4B) in the animals treated with both CATR and CAY compared with those treated with CAY alone, over 2 and 6 h of reperfusion. To further confirm this result, hepatocyte ultrastructure was observed after 6 h of reperfusion using transmission electron microscopy. As revealed in Fig. 4C, hepatocyte nuclei remained intact in both the sham and I/R + CAY groups, with most of the mitochondrial morphology remaining intact. However, nuclear chromatin condensation was found in the I/R group, and most of the mitochondria were swollen or destroyed in both the I/R and I/R + CAY + CATR groups. Given that MPTP opening serves a key role in I/R-induced cell death, the current data further confirm that the protective effect of EP4 activation on hepatic I/R is achieved by inhibition of MPTP opening.

To further investigate the signaling cascades of EP4-assicated MPTP modulation during I/R, the changes in cAMP levels and ERK1/2 signaling were measured in hepatocytes. As expected, EP4 activation significantly increased cAMP concentrations and subsequent ERK1/2 signaling activation by enhancing p-ERK1/2 levels in the liver tissue after 6 h of reperfusion (Fig. 5A and B). ERK1/2 is an intra-cellular signaling molecule that influences various functions (including cell metabolism and development) and interacts with a variety of intracellular signaling cascades. Within these cascades, GSK3β and STAT3 have been confirmed to be key molecules regulating MPTP opening (2,6,29). Thus, changes in expression of GSK3β and JAK-STAT3 signaling proteins were evaluated in hepatocytes after 6 h of reperfusion. As indicated in Fig. 5C, the I/R group exhibited significantly downregulated p-GSK3β expression, compared with the sham group. However, compared with I/R alone, the EP4 agonist significantly increased p-GSK3β expression in the liver tissue, suggesting that EP4 activation may increase GSK3β phosphorylation during reperfusion. In addition, as revealed in Fig. 5D and E, I/R stimulated increases in JAK2 phosphorylation and subsequent STAT3 phosphorylation, compared with sham treatment. However, compared with I/R alone, the EP4 agonist failed to further stimulate the phosphorylation of JAK2 and STAT3. The current results indicate that EP4 activation-associated MPTP inhibition, and subsequent hepatoprotection, may be associated with increased cAMP concentration and subsequent ERK1/2-GSK3β signaling activation.

Finally, to further confirm the role of ERK1/2-GSK3β signaling in EP4-associated MPTP modulation, PD, an ERK1/2 inhibitor, was used to block the liver protection mediated by CAY. As depicted in Fig. 6, the necrotic hepatocyte percentages and ROS levels were significantly higher in the I/R + CAY + PD group than in the I/R + CAY group (Fig. 6A and D), while the CRC level was significantly lower (Fig. 6C). Notably, p-GSK3β expression was inhibited in the animals pretreated with PD after 6 h of reperfusion (Fig. 6B). Thus, the intraperitoneal injection of the ERK1/2 inhibitor prior to ischemia significantly reversed the protective effect of CAY during hepatic I/R, further confirming that EP4 regulates the signaling cascade of MPTPs, via the ERK1/2-GSK3β pathway.

Taken together, the present results indicate that EP4 activation protects the liver against I/R by activating ERK1/2-GSK3β signaling and further inhibiting MPTP opening.