Potential conflict of interest: Nothing to report.
Cyclooxygenase-2 (COX-2) is upregulated in many cancers, and the prostanoids synthesized increase proliferation, improve angiogenesis, and inhibit apoptosis in several tissues. To explore the function of COX-2 in liver, transgenic (Tg) mice were generated containing a fusion gene (LIVhCOX-2) consisting of human COX-2 cDNA under the control of the human ApoE promoter. Six lines were developed; all of them expressed the LIVhCOX-2 transgene selectively in hepatocytes. The Tg mice exhibited a normal phenotype, and the increased levels of PGE2 found were due to the constitutively expressed COX-2. Histological analysis of different tissues and macroscopic examination of the liver showed no differences between wild-type (Wt) and Tg animals. However, Tg animals were resistant to Fas-mediated liver injury, as demonstrated by low levels of plasmatic aminotransferases, a lesser caspase-3 activation, and Bax levels and an increase in Bcl-2, Mcl-1, and xIAP proteins, when compared with the Wt animals. Moreover, the resistance to Fas-mediated apoptosis is suppressed in the presence of COX-2–selective inhibitors, which prevented prostaglandin accumulation in the liver of Tg mice. Conclusion: These results demonstrate that expression of COX-2–dependent prostaglandins exerted a protection against liver apoptosis. (HEPATOLOGY 2007;45:631–638.)
Cyclooxygenase-1 (COX-1) and COX-2 catalyze the first step in prostanoid biosynthesis.1, 2 COX-1 is constitutively expressed in many tissues and seems to be involved in the housekeeping function of prostaglandins (PGs),3 whereas COX-2 is induced by a variety of stimuli such as growth factors, cytokines, hormones, and other cellular stresses.1, 4, 5
Adult hepatocytes fail to induce COX-2 expression regardless of the pro-inflammatory factors used; only Kupffer, stellate, and immortalized mouse liver cells retain the ability to express COX-2.6, 7 In this regard, we demonstrated that fetal hepatocytes, which exhibit a liver phenotype distinct from the adult cells, were able to express COX-2 on stimulation with lipopolysaccharide and pro-inflammatory cytokines.8 We also demonstrated that partial hepatectomy induced COX-2 in hepatocytes and contributed to the progression of cell cycle after partial hepatectomy.9 Moreover, the high levels of CCAAT/enhancer binding protein-alpha (C/EBP-α) in the adult liver, which binds to the NF-IL6 site and inhibits the activity of the COX-2 promoter, were responsible for the suppression of COX-2 inducibility in adult hepatocytes.10
In addition to inflammation, COX-2 expression has been associated with carcinogenesis, angiogenesis, and tumor development. In liver, previous work has shown that COX-2 expression and prostaglandin (PG) synthesis are key components in the secretion of gelatinases (MMP-2 and MMP-9), adhesion and migration of hepatoma cells, and, therefore, in the remodeling of extracellular matrix that occurs under pathological circumstances such as tumor invasion.11, 12 In addition to liver regeneration after partial hepatectomy,9, 13 expression of COX-2 has been detected in animal models of cirrhosis,14 after hepatitis B and C virus infection,15, 16 in human hepatoma cell lines,17, 18 in cholangiocarcinoma,19 and in HCC.20
Despite these observations, the question of whether COX-2 overexpression is sufficient to induce tumorigenesis has not been addressed. Recently, different models of transgenic (Tg) mice overexpressing COX-2 have been published. Cognitive deficits and neuronal apoptosis were reported in Tg mice that overexpressed COX-2 in neurons. Tg mice overexpressing COX-2 gene in mammary glands developed tumorigenesis,21, 22 and controversial results have been published in keratinocytes.23, 24 The last model corresponds to COX-2/microsomal prostaglandin E synthase (mPGES-1) Tg mice, which develop hyperplastic gastric tumors induced by activated macrophages; moreover, this pathway contributes to Helicobacter-associated gastric tumorigenesis.25 These Tg mice have provided valuable insight into the COX-2 contribution to tumorigenesis.
According to the preceding results, we developed Tg mice that express human COX-2 constitutively in hepatocytes under the control of the human ApoE promoter. These mice exhibit a normal phenotype but are resistant to a Fas-mediated apoptosis, suggesting an antiapoptotic effect for COX-2–dependent prostaglandins.
Antibodies were from Santa Cruz Laboratories (Santa Cruz, CA), from Becton & Dickinson (San Jose, CA), and from Assay Designs (Ann Arbor, MI). The 5,5-dimethyl-3(3-fluorophenyl)-4-(4-methylsulphonyl) phenyl-2(5H)-furanone (DFU) was obtained from Merck (Rahway, NJ). Reagents were from Roche Diagnostics (Mannheim, Germany) or Sigma Chemical Co. (St. Louis, MO). Fluorescent probes were from Molecular Probes (Eugene, OR) and Calbiochem (Merck KGaA, Darmstadt, Germany). Reagents for electrophoresis were obtained from Bio-Rad (Hercules, CA).
Generation of LIVhCOX-2 Mice.
To generate Tg mice overexpressing human COX-2 in the liver, we used a pLiv-Le6 vector that contains the constitutive human ApoE gene promoter and its hepatic control region (a gift from John Taylor, Gladstone Institute, San Francisco, CA).26 The transgenic plasmid (LIVhCOX-2) was generated by cloning a cDNA fragment encoding the hCOX-2 open reading frame into KpnI-XhoI sites of pLiv-Le6 vector. The 6,84 kb NotI-SpeI fragment of LIVhCOX-2 (Fig. 1A) was then isolated and injected into pronuclei of one-cell mouse embryos obtained from mating of hybrid (C57BL/6J X DBA) F1mice obtained from Charles River Laboratories (Wilmington, MA). Two-cell embryos were transferred into the oviducts of pseudopregnant foster CD1 mice as previously described.27 Integration of transgenic DNA was checked by Southern blot analysis and PCR of genomic tail DNA. For PCR analysis, primers specific for the hCOX-2 (forward 5′CTGCAACACCTGAGCGGTTAC3′ and reverse 5′TTGCCACTGCTTGTACAGCAA3′) were used to amplify a 303 bp fragment. The PCR conditions were 1 cycle at 94°C for 5 minutes, followed by 35 cycles of 95°C for 20 seconds, 60°C for 20 seconds, and 72°C for 30 seconds, followed by 72°C for 5 minutes.
Twelve-week-old Tg animals and corresponding wild-type (Wt) littermates were injected intraperitoneally with a single dose of purified hamster anti-mouse FAS monoclonal antibody Jo2 (0.3 μg/g body Wt) freshly dissolved in 0.9% NaCl to induce massive apoptosis. This dose was chosen on the basis of previous studies.28 Animals were killed 6 hours later, and livers were rapidly removed and freeze-clamped in liquid N2.
In some experiments, mice were treated daily for 4 days with 5, 5-Dimethyl-3(3-fluorophenyl)-4-(4-methylsulphonyl) phenyl-2(5H)-furanone (DFU) (5 mg/kg body weight) to inhibit COX-229 or with indomethacin (50 mg/kg body weight) as a nonselective COX-2 inhibitor. Animals were treated in accordance with Institutional Care Instructions.
RNA Preparation and Analysis.
Total RNA was extracted from liver with TRIzol Reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. First strand cDNA was synthesized from 1 μg total RNA using random hexamer and expand reverse transcriptase (Roche). The resulting cDNAs were amplified with the following oligonucleotide sequences: hCOX-2, 5′CTGCAACACCTGAGCGGTTAC3′ and 5′TTGCCACTGCTTGTACAGCAA3′ (fragment size 303 bp) and murine GAPDH as an endogenous control, 5′CAAGGTCATCCATGACAACTTTG3′ and 5′CTGAGTGGCAGTGATGGCAT3′ (fragment size 73 bp).
Determination of Metabolites.
Endogenous PGE2 levels in liver tissue were determined by specific immunoassay (GE Healthcare Bio-Sciences AB, Uppsala, Sweden). Briefly, liver samples were homogenized in 0.1 M phosphate buffer pH 7.5 containing 0.9% bovine serum albumin (BSA) and 0.5% Kathon (assay buffer). After centrifugation, the lipids were then analyzed according to manufacturer's instructions.
Amino transferases, ALT and ASP, total bilirubin, and alkaline phosphatase were assayed spectrophotometrically in plasma.8, 30 Protein levels were determined with the Bradford reagent.
Histochemistry and Immunofluorescence.
For hematoxylin-eosin staining, liver was fixed in 10% paraformaldehyde before embedding into paraffin. Histopathology analysis was carried out by the Department of Medicine and Animal Surgery from the Faculty of Veterinary Medicine, Madrid, Spain.
For immunohistochemical detection of hCOX-2, liver samples were snap-frozen in 2-methylbutane immersed in liquid N2, and serial 8-μm-thick sections were cut onto gelatinized glass with a Microm HM550 sledge cryostat (Microm International GmbH, Walldorf, Germany). The preparations were fixed with 4% paraformaldehyde pH 7 for 30 minutes at room temperature (RT), washed with phosphate-buffered saline, and permeabilized with methanol for 15 minutes at RT. After blocking with 3% BSA for 1 hour at RT, the sections were incubated for 1 hour with COX-2 antiserum (diluted 1:10; Assay Design) in 1% BSA at RT. After that, the sections were incubated with the appropriate secondary antibody coupled to Alexa Fluor 488 for 45 minutes. TO-PRO-3 (Molecular Probes) was used for DNA staining. Confocal images were obtained with a Leica TCS SP2 Spectral microscope and a 63×/1.40 NA oil objective (Leica Microsystems, Wetzlar, Germany). For the detection and quantification of apoptosis, the terminal deoxynucleotidyl transferase-mediated nick-end labeling (TUNEL) commercial kit for cell death detection (Roche) was used, following the instructions of the manufacturer. The numbers of positive staining with apoptotic morphology were counted in five random fields per sample, and averaged numbers were calculated.
Tissue samples (100 mg) were homogenized in a lysis buffer containing 10 mM Tris-HCl (pH 7.5), 1 mM MgCl2, 1 mM EGTA, 10% glycerol, 0.5% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate, 1 mM β-mercaptoethanol, and 0.1 mM phenylmethyl sulfonyl fluoride. Extracts were vortexed for 30 minutes at 4°C and, after centrifuging for 20 minutes at 15,000 g, the supernatants were stored at −20°C. For Western blot analysis, extracts were boiled for 5 minutes in Laemmli sample buffer, and equal amounts of protein (20-30 μg) were separated by 10% to 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). The relative amounts of each protein were determined in total extracts with the following polyclonal or monoclonal antibodies: COX-2 (Assay Design and Santa Cruz), COX-1 and mPGES (Cayman), inhibitors of apoptosis (xIAP) (R & D Systems), and Bcl-2 family proteins (Santa Cruz). After incubation with the corresponding anti-rabbit or anti-mouse horseradish peroxidase conjugated secondary antibody, blots were developed by the ECL protocol. Target protein band densities were normalized by calculating the ratio to the corresponding densities of αp85 protein. Different exposure times were performed on each blot to ensure linearity of the band intensities. Densitometric analysis was expressed in arbitrary units.
Mouse liver was homogenized in 10 mM HEPES pH 7.9; 1 mM EGTA, 1 mM EDTA, 120 mM NaCl, 1 mM DTT, 0.5 mM phenylmethyl sulfonyl fluoride, 2 μg/ml aprotinin, 10 μg/ml leupeptin, 2 μg/ml TLCK, 5 mM NaF, 1 mM NaVO4, 10 mM Na2MoO4, and 0.5% Nonidet P-40. The extracts were centrifuged at 8,000g for 15 minutes at 4°C, and the supernatants were stored at −80°C. Protein content was assayed with the Bradford reagent. Caspase activities were measured by cleavage of specific fluorogenic substrates in accordance with the supplier's instructions (Calbiochem). Substrates were N-acetyl-DEVD-7-amino-4-trifluoromethylcoumarin for caspase 3, N-acetyl-IETD-7-amino-4-trifluoromethylcoumarin for caspase 8, and N-acetyl-LEHD-7-amino-4-trifluoromethylcoumarin for caspase 9, respectively. The linearity of caspase assays was determined over a 30-minute reaction.31
Data are expressed as means ± SD. Statistical significance of differences between the control and transgenic groups was evaluated by the Mann-Whitney test. All tests have been calculated two-tailed, and the significance has been considered at P < 0.05.
Generation and Characterization of LIVhCOX-2-Tg Mice.
To achieve selective expression of COX-2 in hepatocytes, we generated Tg mice carrying the human isoform of COX-2 (hCOX-2) under the control of the human ApoE promoter and its endogenous hepatic control region. Founder mice were generated by injecting one-cell embryos with the construct (Fig. 1A) and used to establish six lines of LIVhCOX-2 Tg mice. Screening of mice was done by PCR analysis (Fig. 1B) and confirmed by Southern blot hybridization of tail DNA (data not shown). The heterozygous LIVhCOX-2-4 Tg line expressing 55 copies of the Tg was used.
The Tg expression was examined by reverse transcription PCR analysis in five different tissues and was restricted to the liver and kidney (Fig. 1C), consistently with the known pattern of expression of the ApoE promoter.26 Despite the expression of the mRNA in the kidney, the human protein was undetectable by immunoblotting in this organ. In contrast, the liver expressed the highest amount of hCOX-2 (Fig. 1D-H). The Tg animals were phenotypically similar to their normal littermates and did not exhibit a detectable histological change in the liver at 12 weeks of age (Fig. 1E-F). We confirmed the functional activity of Tg COX-2 by measuring the levels of prostaglandin E2 (PGE2), which showed a fivefold increase. Selective COX-2 inhibitors reduced the levels of PG to that of Wt animals (Fig. 1I).
Plasma Levels of AST and ALT Are Decreased in Tg Mice After Jo2 Treatment.
To determine whether expression of COX-2 could influence the apoptotic response in liver, activation of the Fas pathway was used. Challenge of mice with Jo2, an activator of the Fas pathway and inducer of apoptosis, showed significant differences between Wt and Tg mice. The plasma levels of AST,31 ALT, and total bilirubin were determined, and they were much lower in COX-2-Tg mice than in Wt counterparts. Serum alkaline phosphatase remained unchanged under these conditions (Table 1).
Table 1. Plasma Levels of Liver Injury Markers
NOTE. Plasma levels of aminotransferases (AST, ALT), total bilirubin (TBIL), and alkaline phosphatase (ALP) were assayed spectrophotometrically from the different groups of animals. Results are mean ± SD of 5 animals per condition.
P < 0.01 versus the corresponding Wt condition, respectively.
P < 0.005 versus the corresponding Wt condition, respectively.
The levels of hCOX-2, COX-1 and microsomal PGE synthase, which catalyzes the isomerization of PGH2 to PGE2 and is regulated in a coordinated manner with COX-2, were analyzed in liver extracts from control and Tg mice after Jo2 administration. We found hCOX-2 was significantly expressed in Tg animals but was undetectable in the Wt (Fig. 2A). Total COX-2 (human and endogenous) was also increased in Tg mice. The levels of mPGES increased in Tg animals and COX-1 was present at low levels in Wt and Tg mice. Treatment with Jo2 did not change COX-2, COX-1, and mPGES protein levels (Fig. 2A). The contribution of COX-2 to this process was confirmed by inhibition (>70%) of PGE2 synthesis after treatment of the animals with the COX-2 selective inhibitor DFU (Fig. 2B).
TUNEL Staining and Caspase-like Activities Are Decreased in Transgenic Liver.
Next, we analyzed the effect of COX-2 expression on Jo2-mediated liver cell apoptosis. Figure 3A shows representative macroscopic images of Wt and Tg livers treated with Jo2. Six hours after Jo2 intraperitoneal injection, the livers from Wt mice turned dark red because of massive hepatic hemorrhage, although the mice were still alive. In contrast, livers from Tg mice were architecturally preserved with sporadic reddish changes. When Tg animals were treated with the COX-2 inhibitor DFU, the liver aspect was similar to the Wt condition. These data provide evidence that PGE2 synthesized by COX-2 suppresses apoptosis in Tg liver. Next we analyzed apoptosis by TUNEL of histological liver sections and by measurement of the caspase 3, 8, and 9 activities. A significant decrease in TUNEL staining was observed in Tg animals compared with Wt (Fig. 3B,C). Moreover when the Tg animals were treated with DFU and challenged with Jo2, TUNEL increased, resembling the Wt situation. The activities of caspases 3 and 9 were markedly lower in Tg liver after Jo2 treatment than in Wt counterparts (Fig. 4). Caspase 8 activity was also inhibited in Tg-Jo2 liver, but to a lesser extent with respect to caspases 3 and 9.
Antiapoptotic Markers Are Induced in Liver Expressing COX-2.
To address the antiapoptotic mechanism of COX-2 acting on Fas-mediated liver injury, we investigated the expression of apoptosis-related proteins (Fig. 5). As shown in Fig. 5A and 5B, Bax expression was detected in Wt liver treated with Jo2 but was lower in Tg mice under similar conditions. The protein levels of Bcl-2, xIAP, and Mcl-1 were increased significantly in Tg mice compared with the Wt littermates. Moreover, an increase in the Bcl-2/Bax ratio was demonstrated (Fig. 5B). However, equivalent levels of Fas were detected in both types of mice.
In this study, we developed Tg mice expressing hCOX-2 in hepatocytes under the control of human ApoE promoter. The Tg mice exhibited a normal phenotype, and the increased levels of PGE2 found are mainly attributable to the constitutively expressed COX-2. This indicates a functional coupling in Tg liver between COX-2 and other enzymes involved in PG biosynthesis, such as phospholipase A2 and prostanoid synthases. Histological analysis of different tissues and macroscopic examination of the liver showed no differences between Wt and Tg animals. However, Tg animals were resistant to Fas-mediated apoptosis, suggesting an antiapoptotic role for COX-2–dependent PGs. Moreover, the resistance to Fas-mediated apoptosis is suppressed in the presence of the COX-2 selective inhibitor DFU.
Although several reports have demonstrated a role for COX-2 in tumorigenesis and inhibition of apoptosis, many questions remain unsolved regarding the sufficiency of this enzyme.32, 33 Furthermore, nonsteroidal antiinflammatory drugs protected in various animal models of tumorigenesis; however, issues of non-specificity and the high doses used pointed to a revision of these studies.34 In liver, increased COX-2 expression in hepatocytes has been demonstrated in patients with HCC, especially in cirrhotic tissue and in well-differentiated HCC tumors.20, 32 In HCC, COX-2 is highly expressed at early stages and down-regulated in advanced stages, raising the question of why COX-2 is not expressed in advanced HCC.35 These results are controversial because advanced HCC produces vascular endothelial growth factor (VEGF), exhibits high invasiveness, and possesses antiapoptotic properties; all these processes can be potentially regulated by COX-2. However, in vitro studies showed that NS-398, celecoxib, and sulindac effectively inhibited growth of human hepatoma cell lines.17, 18 In mice implanted with hepatoma cells, nimesulide inhibited tumor growth by inducing apoptosis and overexpression of Bax over Bcl-2.36 Moreover, combinations of COX-2 and MEK inhibitors synergistically increase apoptosis in human HCC.37 Nevertheless, the in vitro anti-HCC effect of nonsteroidal antiinflammatory drugs and COX-2–specific inhibitors needs to be confirmed in in vivo animal models.
The role of COX-2 as inhibitor of apoptosis also has been demonstrated in vivo in Tg mouse models of COX-2 expression in several tissues. Tg mice expressing COX-2 in mammary gland developed tumors, and these animals had reduced levels of the proapoptotic proteins Bax and Bcl-XL and elevated levels of Bcl-2.22 However, Tg mice expressing COX-2 in neurons developed age-dependent cognitive deficiencies because of an increase in neuronal apoptosis. An increase in apoptotic neurons beginning at middle age and astrocytic activation at old age was detected, and the authors concluded that neuronal COX-2 expression over time promotes neuronal injury and dysfunction.38 A recent study demonstrated hyperplastic gastric tumors in Tg mice expressing COX-2 and mPGES. This study relates PGE2 with the recruitment, infiltration, and activation of mucosal macrophages establishing a relationship between inflammation and tumor cell growth.25
Our previous results obtained in vivo from mice hydrodynamically transfected with a GFP-COX-2 expression vector clearly demonstrated that PGs produced by COX-2 protected the liver against Fas-mediated apoptosis.39 These results are completely confirmed in the Tg mice that we have developed in this work. In fact, the levels of hepatic transaminases and the activity of caspase-3 are decreased in the Tg animals per se, without apoptotic stimuli.
Several signaling pathways have been proposed as mediators of COX-2–dependent inhibition of apoptosis. PGE2 inhibits apoptosis in gastric mucosa cells via the mitochondrial pathway and PKA activation.40 In cholangiocarcinoma cells, celecoxib suppressed Akt phosphorylation and favored the execution of apoptosis.19 Few reports describe the direct effect of COX-2 expression on apoptosis in liver or HCC cells. Leng et al.41 reported an enhanced phosphorylation of Akt in Hep 3B cells transiently or permanently transfected with a COX-2 expression vector, but Bcl-2 was not induced. Celecoxib reduced Akt activation and induced caspase-9 and caspase-3 activation with a concomitant release of cytochrome c. Our results demonstrate that PGs produced by COX-2 in liver inhibit the activation of caspase-3, decrease Bax protein levels, and increase the levels of Bcl-2 protein, all these hallmarks of apoptosis.
In conclusion, we generated a novel transgenic model that may provide new clues to explore the role of COX-2 liver disease. Although we did not find significant liver alterations just by the expression of the transgene, we cannot discard the occurrence of minor alterations. Further studies in old animals or by inducing HCC in the COX-2-Tg mice might help to elucidate whether COX-2 overexpression inhibits apoptosis and protects or sensitizes mouse liver for carcinogenesis and senescence.