Cyclooxygenase-2 up-regulates hepatic somatostatin receptor 2 expression

Somatostatin and its analogues, which function by binding to somatostatin receptors (SSTRs) 1–5, play a protective role in liver cirrhosis. Hepatic SSTR-2 expression is up-regulated in subjects with liver cirrhosis. However, little is known about the mechanisms underlying this process. In the present study, we observed the up-regulation of hepatic SSTR-2 expression in thioacetamide (TAA)-induced cirrhotic rats and further showed that cyclooxygenase-2 (COX-2) might play a role in this process via the protein kinase C (PKC)–cAMP response element binding protein (CREB) signaling pathway. In vivo, the up-regulated SSTR-2 in liver cirrhosis was inhibited by the addition of a selective COX-2 inhibitor, such as celecoxib. In vitro, the up-regulation of COX-2 by either transfection with COX-2 plasmids or treatment with TAA increased levels of SSTR-2 and phosphorylated CREB (p-CREB) in the human hepatocyte cell line L02. Furthermore, the increase in SSTR-2 expression was inhibited by the addition of celecoxib and a PKC inhibitor. Moreover, for comparable DNA methylation levels in the region upstream of the hepatic SSTR-2 gene in normal and cirrhotic livers, DNA methylation may not contribute to the up-regulation of SSTR-2 expression in cirrhotic livers. In conclusion, the up-regulation of hepatic SSTR-2 might be induced by COX-2 via the PKC-CREB signaling pathway but is probably not induced by DNA methylation.

SSTR-2 expression has been reported to be up-regulated under inflammatory conditions, such as IgA nephropathy, adipose tissue inflammation and ileum inflammation [18][19][20] . One study conducted in conjunctival and corneal epithelial cells found that inflammatory cytokines, such as interleukin (IL)-1α, IL-1β, lipopolysaccharide, and peptidoglycan, increased SSTR-2 expression 21 . Cyclooxygenase-2 (COX-2) catalyzes the synthesis of prostanoids from arachidonic acid and plays an important role in inflammation 22 . COX-2 is induced by various stimuli under inflammatory conditions. Specifically, as chronic inflammation ultimately leads to liver cirrhosis, the anti-inflammatory treatment of liver cirrhosis has received increasing attention [23][24][25] . In our previous study, significant up-regulation of COX-2 and SSTR-2 was observed in the livers of cirrhotic patients compared with the levels in non-cirrhotic patients 26 . Additionally, the COX-2 inhibitor celecoxib and the SST analogue octreotide synergistically ameliorate portal hypertension and liver fibrosis 3 . However, the relationship between the expression of COX-2 and SSTR-2 has not yet been investigated.
In the present study, we explored the effect of DNA methylation and COX-2 on SSTR-2 expression. COX-2 may contribute to the up-regulation of hepatic SSTR-2 expression in subjects with liver cirrhosis.

Results
Establishment of a liver cirrhosis model. Compared with rats in the control group, rats in the cirrhosis group showed a lower body weight, duck, and lethargy. All rats receiving thioacetamide (TAA) developed liver cirrhosis, and none experienced toxicity-related death. Livers in the control group were generally tender and reddish, with a regular and smooth surface. However, livers in the cirrhosis group exhibited a hard texture, blunt edges, a brown color and a nodular surface (Fig. 1A). Consistently, typical liver cirrhosis was also observed in the cirrhosis group using Masson's trichrome staining (Fig. 1B) and was characterized by the accumulation of extracellular matrix, destruction of the normal architecture and the formation of pseudolobules. Liver cirrhosis was further confirmed by the presence of fibrotic areas and Ishak's fibrosis scoring. The fibrotic areas were significantly increased in the cirrhosis group compared with those in the control group (Fig. 1C). Ishak's fibrosis score for all samples in the cirrhosis group was 5.8, compared with 0 in the control group (Fig. 1D). Up-regulation of SSTR-2 and COX-2 expression in the cirrhotic liver. Levels of the SSTR-2 mRNA and protein were determined in the rat liver. As shown by immunohistochemical staining (IHC), Western blot and real-time PCR, SSTR-2 expression was up-regulated in the cirrhosis group compared with that in the control group ( Fig. 2A,C,D,F). Similarly, levels of the COX-2 mRNA and protein were also increased in the livers from the cirrhosis group (Fig. 2B,C,E,G). Interestingly, the up-regulation of SSTR-2 and COX-2 expression in the cirrhotic liver was remarkably inhibited by the administration of celecoxib (P < 0.05). A significant difference was not observed between the cirrhosis +celecoxib group and the control group (P > 0.05).
Celecoxib inhibited SSTR-2 expression in L02 hepatocytes. To investigate the role of COX-2 in the regulation of SSTR-2 expression, we induced COX-2 expression in L02 hepatocytes by transfecting cells with a COX-2 overexpression plasmid or by incubating cells with TAA. The COX-2 gene was successfully cloned into the pcDNA3.1 expression plasmid, which was confirmed by DNA sequencing. COX-2-pcDNA3.1 was then transfected into L02 cells and was confirmed by Western blotting for COX-2. A significantly higher level of the SSTR-2 protein was observed in L02 cells transfected with COX-2-pcDNA than in cells transfected with empty-pcDNA3.1 (P < 0.01). However, this up-regulation was dramatically suppressed by the addition of celecoxib at concentrations of 20 μM and 40 μM (P < 0.01; Fig. 3A,B). Consistent with these findings, levels of the SSTR-2 and COX-2 proteins were both significantly increased in L02 cells incubated with TAA in a dose-dependent manner (P < 0.05). Similarly, this TAA-induced (80 mg/L) up-regulation of SSTR-2 was also inhibited by the addition of celecoxib (P < 0.01; Fig. 3C,D). COX-2 up-regulated SSTR-2 via the PKC-CREB signaling pathway. The cAMP response element binding protein (CREB) is an important transcription factors that binds to the promoter region of the SSTR-2 gene. To investigate the possible mechanisms by which SSTR-2 expression was up-regulated in the cirrhotic liver, we detected CREB levels in cirrhotic livers and L02 cells. In vivo, levels of phosphorylated CREB (p-CREB) were markedly increased in the cirrhosis group compared with those in the control group. However, the increased p-CREB level was significantly reduced in the celecoxib group (Fig. 4A,B). In vitro, p-CREB levels were significantly increased in L02 cells transfected with the COX-2 overexpression plasmid compared with the levels in the control group, and this change was inhibited by the addition of celecoxib in a dose-dependent manner (Fig. 4C,D). P-CREB is an important downstream transcription factor of the PKC and p38 signaling pathways. The PKC and p38 pathways were then inhibited by selective inhibitors in L02 cells transfected with COX-2. The PKC inhibitor, but not the p38 inhibitor, significantly decreased the up-regulation of SSTR-2 expression induced by COX-2 overexpression (Fig. 5). Based on these results, COX-2 might up-regulate SSTR-2 expression via the PKC-CREB pathway.
DNA methylation of the SSTR-2 gene remained unchanged in rats with liver cirrhosis. Sequences between −2000 and +1000 relative to the transcription start site were analyzed using MethPrimer. Two CpG islands located between −939 to −545 and −337 to +590, respectively, were identified. Methylation levels of 49 CpG sites between −2000 and +1000 were analyzed using the Sequenom MassARRAY EpiTYPER. The methylation levels of the distal region (−2000 to −1200) were relatively high, while the methylation levels of the proximal region (−1200 to +1000) tended to be low. Overall, no significant difference in methylation levels was observed between the control group and the cirrhosis group (Fig. 6). A DNA fragment of 499 bp containing 32 CpGs located within the regulatory area (−413 to +85) was further analyzed using bisulfite sequencing. Consistently, low methylation levels were detected in the control group (3.7%) and the cirrhosis group (3.3%; P > 0.05, Fig. 7).

Discussion
Previous studies have demonstrated that the expression of SSTR-2 can be differentially regulated by hormones (estrogen, testosterone and glucocorticoids), inflammatory cytokines (IL-1α and IL-1β), epigenetics (DNA methylation and histone acetylation), lipopolysaccharides and peptidoglycans 15,21,27 . In the present study, hepatic SSTR-2 expression was up-regulated in the TAA-induced cirrhotic rat model. Furthermore, COX-2 might play a role in this process via the PKC-CREB signaling pathway in vivo and in vitro. Since comparable DNA methylation levels were observed upstream of the hepatic SSTR-2 promoter region in the normal and cirrhotic livers, DNA methylation may not contribute to the up-regulation of SSTR-2 expression in the cirrhotic rat liver. A cirrhotic rat model was established by an intraperitoneal injection of TAA, a hepatotoxic chemical, to investigate the mechanism regulating SSTR-2 expression in the cirrhotic liver. Consistent with a previous study, hepatic SSTR-2 expression was up-regulated in the liver cirrhosis group 10 . The DNA methylation level in the promoter region of the SSTR-2 gene was reported to be inversely correlated with its expression level in several cancer cell lines, and DNA methylation silences the SSTR-2 promoter in vitro 15 . Based on accumulating evidence, aberrant global DNA methylation contributes to the progression of liver cirrhosis [28][29][30] . Specific loci, including collagen 1A1, secreted phosphoprotein 1, SPRR3 and TNFSF15, have been confirmed to be hypomethylated, resulting in the up-regulation of these genes in patients with liver cirrhosis of various etiologies [29][30][31] . However, researchers have not determined whether DNA methylation causes SSTR-2 up-regulation in the liver of a cirrhotic rat. Accordingly, the Sequenom MassARRAY assay and bisulfite sequencing were then used to detect the methylation level upstream of the SSTR-2 gene. However, no significant difference in methylation level of the SSTR-2 gene was observed within −2000 to +1000 relative to the transcriptional start site using the Sequenom MassARRAY assay. Since the Sequenom MassARRAY assay was not able to examine every CpG site in the detected region, the bisulfite sequencing assay was further applied to detect the region from −413 to +85, which contains most of the regulatory elements in the SSTR-2 promoter 32 . Nevertheless, similar to the results from the Sequenom MassARRAY, no significant difference was observed between the cirrhosis group and the control group. Although the reason is not quite clear, the difference in the cell proliferation state between the cirrhotic liver and cultured human hepatocytes might explain this discrepancy. Notably, the sample size for the methylation analysis (3 per group) in the present study was sufficient to detect a 1% difference between the two groups. Torrisani J identified two CpG islands and a new transcription start site located 3.8 kb upstream (−3939 to −4205) 15 . Hence, NCBI has updated the transcription start site and defined Torrisani J. 's upstream transcription start site as +1. By performing a CpG analysis using the Methyl primer express software, we found that most CpG islands were located within −2000 to +1000 in rats (Fig. 6A). Importantly, Torrisani J also verified that the new transcription start site is located in close proximity to a region in which the transcription start site is located in rat 15 . Thus, the promoter region analyzed in the present study covered the upstream promoter region reported by Torrisani J. Accordingly, the DNA methylation at the promoter region may not regulate SSTR-2 expression in the liver of a cirrhotic rat. However, further study is needed to determine whether DNA methylation at other sites regulates SSTR-2 expression.
We previously explored the therapeutic efficacy of celecoxib for liver cirrhosis and found that celecoxib attenuated TAA-induced liver cirrhosis by inhibiting fibrosis, angiogenesis, inflammation and the epithelial-to-mesenchymal transition of hepatocytes in animal models [23][24][25] . In the present study, SSTR-2 expression was inhibited by the addition of celecoxib, a COX-2 inhibitor that functions by directly binding to the active site of COX-2. In one previous study, which investigated the correlation between the expression of SSTR2 and COX-2, the author applied an immunochemistry method to examine the expression of COX-2 and SSTR2 in gastroenteropancreatic neuroendocrine tumors. Interestingly, an inverse correlation was identified between the expression of COX-2 and SSTR2 in the foregut, but not the hindgut, in subjects with neuroendocrine tumors 33 . However, the limitations of this study were the exclusive use of the immunochemistry method and the correlation coefficient was set at a low level of 0.3. In the present study, COX-2 induced SSTR-2 expression in vitro. We noticed an up-regulation of SSTR-2 expression when COX-2 expression was induced, and this up-regulation was diminished when the COX-2 signaling pathway was inhibited by celecoxib. In our study and a previous study, SSTR-2 was mainly expressed in hepatocytes of cirrhotic livers. COX-2 was also mainly expressed in hepatocytes and macrophages 10,34 . Accordingly, strategies targeting the COX-2/SSTR-2 signaling pathway are postulated to either promote liver fibrosis by reducing hepatic inflammation or improve liver fibrosis by inhibiting the epithelial-to-mesenchymal transition of hepatocytes.
Activating transcription factor-2 (ATF-2), CREB and c-Jun are important transcriptional factors that directly bind to the ATF/CRE site within the SSTR-2 promoter region, thereby activating SSTR-2 transcription 32,35 . Levels of p-CREB and c-Jun are both increased in cirrhotic livers or in activated HSCs 36,37 . Meanwhile, COX-2 increases the phosphorylation of CREB and c-Jun by activating several signaling pathways 38 . Consistent with previous studies, we also observed increased p-CREB levels in cirrhotic livers, a change that was inhibited by the addition of celecoxib. Furthermore, the COX-2-PKC-CREB signaling pathway might participate in the mechanism regulating SSTR2 expression in vitro. Based on these data, COX-2 might up-regulate hepatic SSTR2 expression in the cirrhotic liver via the PKC-CREB signaling pathway. Notably, our study only concentrated on the role of one signaling pathway in the mechanism regulating SSTR-2 expression, and further studies are required to determine whether other pathways are involved in this process.
Based on results of the present study, COX-2 inhibitors may influence the therapeutic efficacy of SST or its analogues, particularly when their effects are mediated by SSTR-2. A combination of a COX-2 inhibitor and an SST or its analogues is frequently employed as a treatment for pancreatitis and various tumors 39,40 . In a mouse model of pancreatitis, the administration of octreotide in combination with diclofenac sodium (a COX inhibitor) did not show superior benefits compared to the administration of each agent separately 41 . The author suggested that this lack of change may be due to the interaction between the two agents. Diclofenac sodium may have  inhibited the anti-secretory effect of octreotide on the exocrine pancreas in this model. In contrast to this finding, one of our previous studies found that nude mice with HepG2 xenografts that were treated with SOM230 and celecoxib together exhibited better survival than mice given SOM230 or celecoxib alone 42 . As shown in our previous study, celecoxib and the SST analogue octreotide synergistically ameliorate portal hypertension and liver fibrosis in the cirrhotic liver by inhibiting angiogenesis 3 . The probable explanation may be that the two agents exert their anti-tumor and anti-fibrosis effects through similar mechanisms, including inhibiting proliferation and angiogenesis and promoting apoptosis 3,27,43 . Moreover, the inhibition of SSTR-2 expression by celecoxib might lead to the synergistic therapeutic effects of celecoxib and octreotide on HepG2 xenografts and cirrhotic rats.
In conclusion, as depicted in Fig. 8, COX-2 up-regulates hepatic SSTR-2 expression in a rat liver cirrhosis model via the PKC-CREB signaling pathway, while DNA methylation is not likely responsible for this process. Histological studies. Livers were fixed with 4% neutral buffered paraformaldehyde, embedded in paraffin, sectioned consecutively at a thickness of 3 μm and stained with Masson's trichrome (MT). The degree of fibrosis was assessed using Ishak's scoring system. IHC staining. Sections were deparaffinized in xylene and serial ethanol dilutions. Antigen retrieval was performed by heating the sections in 10 mM sodium citrate buffer. Sections were blocked with 10% goat serum and incubated with the primary antibody overnight at 4 °C. SSTR-2 (1:100, Santa Cruz Biotechnology, Santa Cruz, CA, USA) and COX-2 (1:200, Abcam, Cambridge, UK) antibodies were used. After an incubation with biotinylated secondary antibodies and the streptavidin-biotin-complex, sections were stained with a solution of 3,3-diaminobenzidine tetrahydrochloride and counterstained with hematoxylin.

Methods
Protein extraction and Western blotting. Proteins were extracted from liver tissues or cultured cells using a protein extraction kit (Keygen Biotec, Nanjing, China), according to the manufacturer's protocol. Protein concentrations were determined using a BCA protein assay kit (Biotime, Shanghai, China). Next, each sample was mixed with 6× SDS sample buffer and boiled for 3 min. Total proteins (20 μg) were separated on 8% SDS-PAGE gels and electrotransferred to polyvinylidene fluoride membranes (Millipore, Billerica, MA, USA). Membranes were incubated with appropriate primary antibodies overnight at 4 °C and then incubated with horseradish peroxidase-conjugated secondary antibody (ZSGB-bio, Beijing, China) at room temperature for two hours. Enhanced chemiluminescence (Biotime) was used to visualize the protein bands. Protein expression was normalized to GAPDH. The SSTR-2 antibody was obtained from Santa Cruz Biotechnology, Santa Cruz, CA, USA; the COX-2 antibody was obtained from Biworld Technology, Co., Nanjing, China; CREB and p-CREB antibodies were obtained from ImmunoWay Biotechnology Co., Newark, DE, USA; and the GAPDH antibody was obtained from Goodhere, Hangzhou, China. Detection of DNA methylation levels using the Sequenom MassARRAY. A 3 kb DNA sequence from −2000 to +1000 of the SSTR-2 (GeneBank NC_005109.4, 102136283-102143449) transcriptional start site was subjected to an analysis of CpG islands using MethPrimer (http://www.urogene.org/methprimer) with the following parameters: island size > 200 bp, GC Percent > 50.0, Obs/Exp > 0.6. Genomic DNA was extracted using the Blood and Tissue DNA purification kit (Qiagen) and subjected to bisulfite modification using EpiTect Bisulfite Kits (Qiagen) according to the manufacturer's protocol. Primers were designed using MethPrimer (http://www. urogene.org/methprimer/) and Primer Premier 5.0 (Premier, Canada), as shown in Table 1. Sequences of interest were then amplified using HotStar Taq DNA Polymerase (Qiagen). The Sequenom MassARRAY was followed by in vitro RNA transcription and base-specific cleavage (MassCCLEAVE Kit, San Diego, USA, SEQUENOM), according to the manufacturer's protocol. Mass spectra were analyzed using the MassARRAY Compact System (SEQUENOM). Methylation ratios were then generated by the EpiTYPER software (SEQUENOM). The Sequenom methylation analysis was performed at CapitalBio, Beijing, China.
Detection of DNA methylation levels by bisulfite sequencing. For bisulfite sequencing, the PCR product was cloned into PMD 19-T vector (Takara, Dalian, China), transformed into E. coli and grown on LB agar plates, containing 100 μg/mL ampicillin. Positive colonies were selected, and the plasmid was extracted using a plasmid mini kit (Omega Bio-tek, Norcross, USA). Eight clones were randomly selected and both strands were sequenced (Invitrogen).  Statistical analysis. Quantitative data are presented as means ± standard deviations and analyzed using SPSS 13.0. Analysis of variance (ANOVA) and the Student-Newman-Keuls (SNK) analysis were used for multiple comparisons. Student's t test was used for comparisons between two groups. Cell-based experiments were performed in triplicate. The difference was considered statistically significant when P < 0.05.