The contribution of metabolic factors to the severity of liver disease is not completely understood. In this study, apolipoprotein E-deficient (ApoE−/−) mice were evaluated to define potential effects of hypercholesterolemia on the severity of carbon tetrachloride (CCl4)-induced liver injury. Under baseline conditions, hypercholesterolemic ApoE−/− mice showed increased hepatic oxidative stress (SOD activity/4-hydroxy-2-nonenal immunostaining) and higher hepatic TGF-β1, MCP-1, and TIMP-1 expression than wild-type control mice. After CCl4 challenge, ApoE−/− mice exhibited exacerbated steatosis (Oil Red O staining), necroinflammation (hematoxylin-eosin staining), macrophage infiltration (F4/80 immunohistochemistry), and fibrosis (Sirius red staining and α-smooth muscle actin immunohistochemistry) and more severe liver injury [alanine aminotransferase (ALT) and aspartate aminotransferase] than wild-type controls. Direct correlations were identified between serum cholesterol and hepatic steatosis, fibrosis, and ALT levels. These changes did not reflect the usual progression of the disease in ApoE−/− mice, since exacerbated liver injury was not present in untreated age-paired ApoE−/− mice. Moreover, hepatic cytochrome P-450 expression was unchanged in ApoE−/− mice. To explore potential mechanisms, cell types relevant to liver pathophysiology were exposed to selected cholesterol-oxidized products. Incubation of hepatocytes with a mixture of oxysterols representative of those detected by GC-MS in livers from ApoE−/− mice resulted in a concentration-dependent increase in total lipoperoxides and SOD activity. In hepatic stellate cells, oxysterols increased IL-8 secretion through a NF-κB-independent mechanism and upregulated TIMP-1 expression. In macrophages, oxysterols increased TGF-β1 secretion and MCP-1 expression in a concentration-dependent manner. Oxysterols did not compromise cell viability. Taken together, these findings demonstrate that hypercholesterolemic mice are sensitized to liver injury and that cholesterol-derived products (i.e., oxysterols) are able to induce proinflammatory and profibrogenic mechanisms in liver cells.
- liver cells
- cholesterol-oxidized products
- apolipoprotein E
epidemiological studies suggest the existence of a close association between liver disease and the prevalence of metabolic disorders such as dyslipemia, insulin resistance, and hyperglycemia (30). Although several lines of evidence suggest that hyperlipidemia may not play a causal role in liver injury, they may affect the severity of tissue damage (22). Indeed, liver damage is more severe when there are multiple coincidental insults. For example, excess lipid accumulation in the form of triglycerides in the cytosol of hepatocytes, which per se does not appear to impair liver function, significantly increases the vulnerability of the liver to the deleterious effects of cytokines, oxidative agents, and viral infections, predisposing this organ to inflammation and advanced fibrosis (12, 29).
Although hypercholesterolemia is a prominent metabolic disorder and a major risk factor for coronary heart disease and atherosclerosis, its precise contribution to the progression of hepatic steatosis, inflammation, and fibrosis has been practically overlooked. In this regard, apart from epidemiological studies suggesting hypercholesterolemia as an independent risk factor for liver disease, a study suggesting that hypercholesterolemia increases susceptibility to virus-induced immunopathological liver disease and few experimental studies in rodents and rabbits demonstrating an association between the intake of high cholesterol and high-fat diets with liver steatosis and inflammation (9, 26, 32, 36), scarce information is available on the influence of increased cholesterol levels on the progression of liver disease.
To test the hypothesis that hypercholesterolemia may predispose the liver to sustained injury, we examined whether apolipoprotein E-deficient mice (ApoE−/− mice), an established model of hypercholesterolemia at the center of experimental lipid research (42), were more sensitized to carbon tetrachloride (CCl4)-induced liver fibrosis than wild-type control animals. In addition, we characterized the profile of oxidized derivatives of cholesterol (i.e., oxysterols) by gas chromatography-mass spectrometry (GC-MS) in livers from ApoE−/− mice. Finally, given that these oxidized cholesterol products have been shown to exert potent biological effects (3, 8, 24, 25, 31), we explored the potential involvement of oxysterols as one of the mechanisms contributing to the increased susceptibility to liver injury in ApoE−/− mice. For this purpose, we assessed the effects of a mixture of oxysterols representative of those found in liver samples from ApoE−/− mice on key mechanisms governing inflammatory and fibrogenic responses in three different cell types relevant to liver pathophysiology. Specifically, we assessed the effects of oxysterols on SOD activity and total lipoperoxides as markers of oxidative stress in hepatocytes. We also explored the effects of oxysterols on the expression of profibrogenic genes, the secretion of IL-8 and NF-κB activity in hepatic stellate cells (HSCs), the main profibrogenic hepatic cell type, and on the secretion of TGF-β1 and the expression of MCP-1 in macrophages, the major inflammatory cell type.
MATERIALS AND METHODS
Experimental animals and study design.
Thirty ApoE−/− mice and 20 control mice (wild-type mice) in the C57BL/6J background (The Jackson Laboratory, Bar Harbor, ME) of 10 wk of age were randomly assigned into four groups of study: wild-type mice (n = 5); ApoE−/− mice (n = 15); wild-type mice submitted to liver injury (n = 15); and ApoE−/− mice submitted to liver injury (n = 15). Liver injury was induced by intraperitoneal injection of CCl4 (1 ml/kg body wt, 1:7 vol/vol in olive oil, twice a week) and mice were euthanized after 0, 4, 6, or 8 wk of treatment. Mice were anesthetized with an injection of ketamine-xylazine, blood was obtained by cardiac puncture, and the liver was excised, rinsed in Dulbecco's PBS, and fixed in 10% formalin for histological analysis. The remaining liver tissue was immediately snap frozen in liquid nitrogen for RNA extraction or cryopreserved in optimal cutting temperature compound (OCT) for further analysis. An additional group of 16 male C57BL/6J mice of 9 wk of age were randomly assigned into two groups of study and fed for 12 wk with either a commercial standard mouse diet (chow) or a high-fat diet (TD.06415, Harlan Teklad, Madison, WI) that contained 45% of calories from fat (19.5% by weight from lard; approximate fatty acid profile as percentage of total fat: 36% saturated, 47% monosaturated, and 17% polyunsaturated). Blood and liver samples were obtained from these animals as described above. All animals were housed in cages with a humidity level of 50–60% and a 12-h light-dark cycle. All studies were approved by the Animal Experimentation Ethical Committee of the University of Barcelona.
Liver samples fixed in 10% formalin were embedded in paraffin and cut in 2-μm sections for hematoxylin-eosin staining. Necroinflammation was analyzed by a registered pathologist unaware of the treatments according to the histological scoring system used on a routine basis in the Pathology Laboratory of the Hospital Clinic: grade 0 (absent), grade 1 (spotty necrosis; one or few necrotic hepatocytes), grade 2 (confluent necrosis), and grade 3 (bridging necrosis) (14). Liver fibrosis was assessed in paraffin sections by Sirius red staining. Briefly, liver sections were incubated for 10 min in 0.5% thiosemicarbazide and stained in 0.1% Sirius red F3B in saturated picric acid for 1 h and subsequently washed with an acetic acid solution (0.5%). Sections were visualized under a Nikon Eclipse E600 microscope (Kawasaki, Kanagawa, Japan) at a magnification of ×40, and relative fibrosis area (expressed as % of positive Sirius red staining) was quantified by histomorphometry using a computerized image analysis system (AnalySIS, Soft Imaging System, Munster, Germany). A minimum of four independent fields were quantified, and the results were expressed as percentage of area occupied by fibrous tissue. Hepatic steatosis was assessed in samples collected in OCT by Oil Red O staining. Briefly, liver cryosections were fixed for 10 min in 60% isopropanol and stained with 0.3% Oil Red O in 60% isopropanol for 30 min and subsequently washed with 60% isopropanol. Sections were counterstained with Gill's hematoxylin, washed with acetic acid solution (4%), and mounted with aqueous solution. Once stained, sections were quantified by histomorphometry.
Immunohistochemical analysis of α-SMA, F4/80, and 4-HNE.
Immunohistochemistry was performed in 2-μm paraffin-embedded liver sections. Briefly, tissue sections were deparaffinized in xylene and rehydrated in descending alcohols. For F4/80 and 4-hydroxy-2-nonenal (4-HNE) detection, antigen retrieval was performed by trypsin digestion (0.05% in 0.1% calcium chloride) for 20 min at 37°C and by heating sections in citrate buffer on a pressure cooker for 3 min, respectively. Subsequently, sections were blocked for endogenous peroxidase (3% H2O2 for 20 min) and nonspecific signals (2% BSA in PBS) and then incubated with either a mouse anti-human α-smooth muscle actin (α-SMA) antibody (Novocastra, Menarini, Florence, Italy) for 90 min at room temperature, a rat anti-mouse F4/80 antibody (Serotec, Oxford, UK) overnight at 4°C (1/100 dissolved with 1% BSA in PBS), or a mouse anti-4-HNE antibody (Genox, Baltimore, MD) for 120 min at room temperature. Afterward, slides were washed in PBS and incubated with specific biotinylated secondary antibodies for 90 min at room temperature, washed again, and further incubated with Vectastain ABC standard kit (Vector Laboratories, Burlingame, CA) for 45 min at room temperature. Finally, sections were washed twice with PBS and incubated with DAB substrate chromogen, rinsed with water, counterstained with Gill's hematoxylin, and mounted in DPX mounting medium. Stained sections were quantified by histomorphometry.
Total SOD activity in hepatocyte lysates and liver tissue samples was determined with a SOD assay kit from Cayman (Ann Arbor, MI). The mitochondrial SOD activity was determined by adding potassium cyanide, an inhibitor of Cu/Zn SOD, to the reaction mixture, following manufacturer's instructions.
Serum was obtained by centrifugation of total blood. Serum cholesterol, triglyceride, and feeding glucose concentrations as well as alanine aminotransferase (ALT) and aspartate aminotransferase (AST) activities were determined by standard laboratory procedures.
Gene expression profiling with TaqMan low-density arrays.
TaqMan low-density arrays (Applied Biosystems) based on real-time quantitative RT-PCR were used to compare hepatic gene expression in hypercholesterolemic ApoE−/− mice with respect to wild-type mice, either under baseline conditions or after 8 wk of CCl4 treatment. Briefly, RNA samples from five ApoE−/− mice and five wild-type mice with or without CCl4 treatment were isolated by the RNAqueous kit, and 1 μg of sample was reversed transcribed by using the High-Capacity cDNA Archive kit. For each sample, 5 μl of RT were then mixed with 95 μl of diethylene pyrocarbonate water and 100 μl of TaqMan universal PCR master mix (Applied Biosystems) to form reaction mix. A 100-μl portion of this reaction mix was then put on a microfluidic card into 48 miniwells containing primers and probes of 48 genes (46 genes known to be involved in lipid metabolism and inflammation and two genes that served as endogenous controls: 18S and GAPDH). For each 384-well card, four RT samples (matched) in duplicate were included at the same time for real-time RT-PCR reaction and analysis. The real-time RT-PCR reaction and laser scanning was performed on ABI 7900HT with the Sequence Detector Software version 2.1. The amount of target gene was determined by the arithmetic formula 2−ΔΔCt described in the comparative cycle threshold (Ct) method (as reported above). Differentially expressed genes were identified based on a significant change P value of 0.05.
Determination of cholesterol oxidation products by GC-MS.
5α-cholestane, cholesterol, 19-hydroxycholesterol (19-HC), cholesterol-5α,6α-epoxide (α-CE), cholesterol-5β,6β-epoxide (β-CE), 7β-hydroxycholesterol (7β-HC), cholestanetriol (CT), 7-ketocholesterol (7-KC), and 25-hydroxycholesterol (25-HC) were purchased from Sigma (St. Louis, MO). 7α-Hydroxycholesterol (7α-HC) was from Steraloids (Newport, RI). The purity of these standards was checked by GC-flame ionization detection. Determination of cholesterol oxidation products was conducted according to Grau et al. (15) with some modifications. One microgram of 19-HC as internal standard was added to 0.25–0.5 g of liver, and lipids were extracted with three successive fractions (17, 14, and 5 ml) of chloroform-methanol (2:1 vol/vol) by use of a Polytron PT 3100 (Kinematica, Lucerne, Switzerland). Centrifugations at 700 g for 20 min and at 540 g for 3 min were applied to separate chloroform and diethyl ether phases, respectively. The final cholesterol oxidation products (COP) extract was quantitatively transferred to a glass tube containing 1 μg of 5α-cholestane (internal standard) and silanized as previously described by Guardiola et al. (17). GC-MS was performed in a Fisons GC 8000 equipped with a Fisons MD 800 mass selective detector and a fused silica capillary column (30 m × 0.25 mm ID) with a film thickness of 0.25 μm of 5% phenyl + 95% methylsilicone from Agilent Technologies (Santa Clara, CA). Helium was used as carrier gas, and the chromatographic conditions were as follows: oven temperature programmed 1 min at 210°C, increased at 2°C/min to 264°C, at 3.5°C/min to 290°C, and kept for 5 min at 290°C; injector temperature = 280°C; flow rate of the carrier gas: 1 ml/min; split ratio 1:5; and sample volume injected = 2 μl. The samples were injected in duplicate. The MS conditions were as follows: interface temperature = 260°C; ion source temperature = 200°C; and electron energy = 70 eV. The identification of COPs in the samples was done with the MS operated in full scan mode [mass-charge ratio (m/z) 100–650]. The quantification was carried operating the mass selective detector in selected ion monitoring mode. The characteristic ions used to quantify the COPs were distributed in four retention time windows: 20–22 min, 5α-cholestane, m/z 217; 28–30 min, 7α-HC, m/z 456; 30.5–34 min, 19-HC, m/z 353, α-CE and β-CE, m/z 384, 7β-HC, m/z 456; and 34–37.5 min, 25-HC, m/z 131, CT, m/z 403, 7-KC, m/z 472. To quantify COPs only 19-HC was used as internal standard because a high variability was observed for the 5α-cholestane area. This drawback could be overcome in this type of samples by using a phytosterol, such as β-sitosterol, as second internal standard, because this compound has a higher retention time and characteristic ions with higher m/z and appears in a cleaner area of the chromatogram.
Isolation and oxidation of LDL.
LDL (d = 1.019–1.063) was isolated from a pool of normal human plasma from 15–20 subjects by sequential preparative ultracentrifugation in a Kontron 45.6 rotor (Kontron Instruments, Milan, Italy). Prior to the oxidation, EDTA was removed by dialysis against PBS, pH 7.4, at 4°C in the dark. The oxidation was performed in the presence of copper chloride at 37°C in the dark for 24 h (20 μg copper chloride/mg protein). The oxidation process was stopped by the addition of 10 μl butylated hydroxytoluene. Samples were stored at 4°C under argon atmosphere and filtered through 0.22-μm pores. The relative degree of lipid peroxidation was determined by measuring thiobarbituric acid-reactive substances, using malondialdehyde (MDA) as standard, and the values were expressed as nanomoles of MDA per milligram of protein. Protein concentrations were measured by standard procedures, using BSA as standard.
Cell incubations and assays.
HSCs were isolated as previously described (7). HSCs were cultured in DMEM media supplemented with 15% FCS, l-glutamine (5,28 mM), pyruvate (1 mM), nonessential amino acids (NEAA) (1%), insulin (2 U/l), penicillin (25 U/ml), and streptomycin (25 μg/ml) at 37°C in a CO2 atmosphere. Experiments were performed after the second serial passage when cells showed phenotypic and immunocytochemical characteristics of myofibroblasts. These culture-activated myofibroblastic HSC closely mimic the phenotypical changes observed in HSC activated during liver fibrogenesis (7). Cells were grown at a density of 0.1 × 106/ml in six-well plates and incubated overnight in serum-depleted media. Thereafter, cells were incubated with vehicle (<0.5% ethanol) or increasing concentrations (1, 10, and 50 μM) of a mixture of oxysterols composed of 7-KC, α-CE, and β-CE at a proportion of 2:1:1, in representation of the oxysterols with proven biological properties more commonly generated in mouse tissues (2). These oxysterols are representative of the cholesterol oxidation products detected by GC-MS in our liver samples from hypercholesterolemic ApoE−/− mice (Table 1). The concentrations used in these experiments were comparable with physiological levels, which are in the micromolar range (8). A similar combination of oxysterols and similar concentrations have been used in previous studies assessing the effects of these oxidized cholesterol products in cell culture experiments (3, 23, 24, 30). HSCs were also exposed to oxidized LDL (oxLDL; 0.1, 1, and 5 μg/ml) and native LDL (nLDL; LDL not submitted to the oxidation protocol) (5 μg/ml). After 24 h of incubation, the expression of selected profibrogenic genes was assessed by real-time PCR, NF-κB activity was measured in a transactivation assay and IL-8 secretion was measured in unextracted supernatants by enzyme immunoassay (EIA) (Biosource, Camarillo, CA).
CC-1 cells, a murine hepatocyte cell line, and RAW 264.7 cells, a murine macrophage cell line, were obtained from the European Collection of Cell Cultures (Salisbury, UK). These cell lines were necessary to ensure a sufficient number of cells for performing concentration-dependent studies with the different compounds. CC-1 cells were cultured in Eagle's minimum essential medium supplemented with 10% FBS, HEPES (20 mM), l-glutamine (2 mM), NEAA (1%), penicillin (50 U/ml), and streptomycin (50 μg/ml) at 37°C in a 5% CO2 atmosphere (35). Cells were grown at a density of 0.4 × 106/ml on 12-well plates and incubated overnight in serum-free medium. To test the effects of compounds on oxidative stress, cells were incubated with vehicle (<0.5% ethanol) or increasing concentrations of oxysterols (1, 10, and 50 μM). Because of the relatively high degree of lipid peroxidation products present in oxLDL as a consequence of the oxidation process, we did not include this lipid product in these experiments. After 24 h of incubation, cells were harvested and lysed by repeated cycles of freeze-thaw in acetone-water for the measurement of total lipoperoxidation products by the PerOx Colorimetric Test System (Immunodiagnostik, Bensheim, Germany) and SOD activity (see SOD activity). Murine RAW 264.7 macrophages were cultured in RPMI-1640 media supplemented with 10% FBS, l-glutamine (2 mM), penicillin (50 U/ml), and streptomycin (50 μg/ml) at 37°C in a 5% CO2 atmosphere. Cells were grown at a density of 0.5 × 106/ml in six-well plates and incubated overnight in serum-free medium. Thereafter, cells were incubated with vehicle (<0.5% ethanol) or increasing concentrations of oxysterols (1, 10, and 50 μM). After 24 or 48 h of incubation, supernatants and RNA from cell extracts were collected and TGF-β1 levels were measured by EIA (Biosource) and gene expression by real-time PCR. Cells were also exposed to increasing concentrations of oxLDL (0.1, 1, and 5 μg/ml) and nLDL (5 μg/ml).
Cell viability assay.
Cell viability was assessed by the dimethylthiazolyl-diphenyltetrazolium bromide assay as previously described (32). Absorbance was measured at 570 nm in a multiwell plate reader (Molecular Devices, Menlo Park, CA), and cell number was calculated from a standard curve.
NF-κB activity assay.
NF-κB activity in HSCs was measured as previously described (39). Briefly, HSCs were transfected with a recombinant adenovirus vector containing a luciferase reporter gene driven by NF-κB transcriptional activation (Ad5NF-κBLuc, multiplicity of infection 500) for 12 h. Medium was replaced and cells were incubated with the compounds for 24 h. NF-κB-mediated transcriptional induction was assessed with the Luciferase Assay System (BD Pharmigen). The signal emitted by cell lysates was measured in a Lumat LB 9507 luminometer (Berthold, Bad Wildbad, Germany).
RNA isolation and reverse transcription and real-time PCR.
RNA was isolated by the RNAqueous kit (Ambion, Austin, TX). RNA concentration was assessed in a UV spectrophotometer and its integrity was tested in a 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA). RT was performed by using the High-Capacity cDNA Archive Kit (Applied Biosystems, Foster City, CA). Validated and predesigned TaqMan primers and probes from Assays-on-Demand were used to quantify MCP-1, 5-lipoxygenase (5-LO), tissue inhibitor of matrix metalloprotease-1 (TIMP-1), collagen 1α1, TGF-β1, and matrix metalloproteinase-2 (MMP-2) gene expression with β-actin used as an endogenous control. Real-time PCR amplifications were carried out in an ABI Prism 7900HT and analyzed with the Sequence Detector Software version 2.1 (Applied Biosystems). Relative quantification of gene expression was performed using the comparative Ct method (User Bulletin no. 2; http://docs.appliedbiosystems.com/pebiodocs/04303859.pdf).
Statistical analyses were performed with the SPSS 13.0 statistical package (SPSS, Chicago, IL). Multiple comparisons between groups were performed by ANOVA. Differences between means were analyzed by the Student's t-test (parametric) and Mann-Whitney's U-test (nonparametric). Data are expressed as means ± SE.
ApoE−/− and wild-type mice had comparable body weight gains throughout the study (data not shown). Compared with wild-type, ApoE−/− mice at 10 wk of age had increased serum cholesterol levels without changes in other parameters of serum biochemistry (triglycerides, feeding glucose, ALT, and AST) (Table 2). In these animals, hypercholesterolemia was associated with increased hepatic steatosis and increased hepatic necroinflammation (Table 2). These findings were confirmed in mice fed a high-fat diet for 12 wk, which also showed increased serum cholesterol concentrations associated with increased hepatic steatosis and necroinflammation (Table 2). Interestingly, hypercholesterolemic ApoE−/− mice showed increased hepatic oxidative stress levels, as revealed by increased SOD activity and 4-HNE immunostaining, as well as increased hepatic TGF-β1, MCP-1, and TIMP-1 expression compared with liver samples obtained from wild-type animals (Fig. 1). No changes in hepatic MMP-2 expression were observed in ApoE−/− mice (Fig. 1C). Moreover, a group of 14 genes known to participate in lipid metabolism and inflammation were differentially regulated in the liver of hypercholesterolemic ApoE−/− mice as revealed by gene expression analysis by TaqMan Low-Density Arrays (Table 3). Among the genes downregulated, we identified two genes involved in cholesterol elimination and reverse transport (Cyp27a1 and Lcat), one gene related to apoptosis (Tnfrsf6), and one nuclear transcription factor PPARδ (Table 3). Among the genes upregulated in hypercholesterolemic ApoE−/− mice, we identified a number of genes related to inflammation [RANTES receptor (Cx3cr1) and IL-1α (II1a)] and enzymes and receptors of the arachidonic acid cascade (Table 3).
To assess whether hypercholesterolemia is associated with an increased susceptibility to liver damage, we submitted hypercholesterolemic ApoE−/− mice to CCl4-induced liver fibrosis. As shown in Fig. 2 (A and B), histomorphometrical analysis of Oil Red O-stained liver sections revealed that hypercholesterolemic ApoE−/− mice had a higher susceptibility to accumulate hepatic lipid droplets than wild-type mice. At the sixth week of treatment, in addition to accelerated steatosis, ApoE−/− mice showed an increased immunostaining for F4/80, a specific macrophage marker (Fig. 2, A and B). Increased hepatic necroinflammation was confirmed by histological analysis of hematoxylin and eosin-stained sections (necroinflammation score: 1.4 ± 0.2 vs. 0.6 ± 0.2; P < 0.05). Two weeks later (at the eighth week of treatment) hypercholesterolemic ApoE−/− mice presented exacerbated collagen deposition, as detected by increased Sirius red staining and increased immunostaining for α-SMA, an established marker of activated HSCs (Fig. 2, A and B). At this particular week of CCl4 treatment, serum ALT and AST levels were significantly higher in ApoE−/− than in wild-type mice (ALT: 80.8 ± 13.5 vs. 43.4 ± 10.0 U/l; P < 0.05; AST: 398.0 ± 121.7 vs. 116.4 ± 17.9 U/l; P < 0.05). Interestingly, significant direct correlations were found between serum total cholesterol levels and serum ALT levels (r = 0.525; P < 0.001), Oil Red O staining (r = 0.4; P < 0.01) and α-SMA immunostaining (r = 0.52; P < 0.0001) in these hypercholesterolemic ApoE−/− mice. These changes did not appear to reflect the usual progression of the disease in ApoE−/− mice, since exacerbated steatosis, inflammation, and fibrosis were not present in a parallel age-paired group of ApoE−/− mice that did not receive CCl4 (Fig. 2C). Furthermore, exacerbated liver injury in ApoE−/− mice was not due to differential metabolism of CCl4 in ApoE−/− mice, since the hepatic expression of cytochrome P-450 remained unchanged compared with wild-type controls after CCl4 treatment (Fig. 2D). Moreover, the presence of advanced liver fibrosis in hypercholesterolemic ApoE−/− mice was associated with further changes in the expression of three genes involved in cholesterol synthesis and transport (Lipin 1, Scarb1, Abcb1b), one gene coding for an antioxidant enzyme (Pon1), a chemokine (Cx3cl1), and two synthases involved in prostaglandin synthesis (Ptges and Ptgis) (Table 4). In addition, following the induction of liver fibrosis, the expression of the apoptotic gene fas (Tnfsf6) and the nuclear transcription factor PPARδ became upregulated in ApoE−/− mice (Table 4).
To assess the potential involvement of oxidized cholesterol products as one of the mechanisms contributing to the increased susceptibility to liver injury present in hypercholesterolemic ApoE−/− mice, we performed additional experiments in vitro in cells related to liver pathophysiology. In these experiments, we employed a mixture of oxysterols that according to our GC-MS analysis are representative of the most common cholesterol oxidation products found in liver samples from hypercholesterolemic ApoE−/− mice (Table 1). Incubation of hepatocytes with micromolar concentrations of these oxysterols resulted in a concentration-dependent increase in both total and mitochondrial SOD activities, selected as established markers of oxidative stress status (Fig. 3). Moreover, oxysterols induced a sustained increase in total lipoperoxidation products in hepatocyte lysates (from 5.0 ± 0.9 to 5.6 ± 1.4, 7.8 ± 1.3, and 6.3 ± 0.3 μmol/l for vehicle and 1, 10, and 50 μM oxysterols, respectively). In these cells, oxysterols did not affect cell viability at any of the concentrations tested (Fig. 3). In these experiments, the effects of oxysterols were not compared with those of oxLDL, since oxLDL had a relative high degree of lipid peroxidation (MDA formation ∼64.5 nmol/mg protein).
The results of incubating HSCs with oxysterols are shown in Fig. 4. In these cells, oxysterols induced a concentration-dependent increase in the release of IL-8 (Fig. 4A). The induction of IL-8 secretion by oxysterols was independent of NF-κB activation, since these oxidized cholesterol products downregulated the activity of this transcription factor in a cell-based reporter assay (Fig. 4B). Moreover, the expression of the profibrogenic factor TIMP-1 was also significantly upregulated by oxysterols (Fig. 4C). No changes in IL-8 secretion and the expression of profibrogenic genes were observed when HSCs were exposed to oxLDL or nLDL, the native nonoxidized LDL (Fig. 4, A and C). Similar to oxysterols, oxLDL significantly downregulated NF-κB activity in HSCs (Fig. 4B). However, none of the concentrations of oxysterols and oxLDL tested in HSCs affected cell viability (Fig. 4D).
The results of incubating macrophages with micromolar concentrations of oxysterols are shown in Fig. 5. In these inflammatory cells, increasing concentrations of oxysterols triggered the secretion of the profibrogenic cytokine TGF-β1 (Fig. 5A). The addition of oxLDL to macrophage cultures produced a comparable concentration-dependent increase in TGF-β1 release, whereas no changes in the secretion of this cytokine were observed when cells were exposed to nLDL (Fig. 5B). TGF-β1 release was induced by oxysterols and oxLDL to a similar extent as LPS (Fig. 5, A and B). Moreover, oxysterols significantly increased the expression of MCP-1, a potent chemoattractant protein that contributes to the maintenance of the inflammatory infiltrate during liver injury, without changes in the expression of 5-LO, a major inflammatory pathway within the arachidonic acid cascade in macrophages (Fig. 5C). In these cell culture experiments, oxysterols did not affect cell viability at any of the concentrations tested, whereas oxLDL reduced cell viability only at concentrations higher than those used in our study (Fig. 5D). nLDL had no effect on cell viability (Fig. 5D).
The results of the present study demonstrate that hypercholesterolemic ApoE−/− mice are sensitized to liver injury since these mice show under baseline conditions increased hepatic oxidative stress levels, increased expression of proinflammatory and profibrogenic factors such as MCP-1, TGF-β1, and TIMP-1, increased lipid deposition, and increased hepatocellular damage. Despite the existence of a prominent inflammatory component in the liver of ApoE−/− mice, these animals did not spontaneously develop hepatic fibrosis. Exacerbated liver injury was more evident when hypercholesterolemic ApoE−/− mice were challenged with a potent hepatotoxic agent such as CCl4, which induces massive hepatic lipoperoxidation (14, 33). Indeed, during the time course of CCl4-induced liver injury ApoE−/− mice exhibited exacerbated steatosis, necroinflammation, and advanced fibrosis compared with wild-type control mice. In our study, important shifts in the inflammatory and fibrogenic responses were noted during the time course of CCl4-induced liver injury. In this regard, compared with controls, ApoE-deficient mice showed increased inflammation at the sixth week of treatment, whereas increased fibrosis, assessed by Sirius red staining and α-SMA, was detected 2 wk later, at the eighth week of treatment. A similar peak in the inflammatory response at the sixth week of treatment followed by a decline at the eighth week, when fibrosis is more evident, has previously been described (18). This temporal dissociation is consistent with the natural course of liver disease (1) and could reflect active inflammation and necrosis accompanied by diffuse deposition of extracellular matrix in the sinusoids in early stages of the disease, followed by the replacement of the damaged tissue by extracellular matrix proteins that begin to form fibrotic tracts and bridge fibrosis at more advanced stages.
Our study also provides evidence that cholesterol-derived products (i.e., oxysterols) are able to induce proinflammatory and profibrogenic mechanisms in liver cells. In our experiments, we employed a mixture of 7-KC, α-CE, and β-CE, which according to our GC-MS analysis are representative of the most common cholesterol oxidation products found in liver samples from hypercholesterolemic ApoE−/− mice. A similar combination of oxysterols and a similar range of concentrations have previously been used in cell culture studies by other laboratories (3, 24, 25, 31). Indeed, these cholesterol-oxidized products triggered a number of cellular and molecular mechanisms in liver cells, including a direct induction by oxysterols of oxidative stress in hepatocytes and the induction of proinflammatory and/or profibrogenic factors (IL-8, MCP-1, and TGF-β1) in HSCs and macrophages. These particular outcomes were selected for scrutiny because they are key proinflammatory and profibrogenic factors involved in the pathogenesis of liver disease (6, 13, 20, 22). For example, MCP-1 is a potent chemoattractant protein that contributes to the maintenance of the inflammatory infiltrate during liver injury and its expression has been shown to be elevated in patients with chronic viral hepatitis as well as in experimental models of liver injury (10, 41). Therefore, increased MCP-1 expression in livers from ApoE−/− mice could in theory contribute to hepatic accumulation of macrophages in these mice. Although we found an increased immunostaining for F4/80, a specific macrophage marker, this antibody does not distinguish recruited from resident macrophages. Similarly, IL-8 is a member of the C-X-C family of chemokines with potent chemotactic activity toward neutrophils, which serum and hepatic levels have been found to closely correlate with the severity of chronic viral hepatitis (4, 28). On the other hand, TGF-β1 is considered one of the main cytokines involved in liver fibrosis since secretion of this cytokine by resident or infiltrated macrophages has been shown to favor the transition of HSCs from a quiescent state to a myofibroblast-like activated phenotype, thus promoting the increased synthesis of extracellular matrix components (16).
The finding that oxysterols increase the release of IL-8 from HSCs also has broad implications in liver fibrosis. However, contrary to what has been reported in previous investigations (34), in our hands upregulation of IL-8 secretion in HSCs by oxysterols did not appear to be mediated by NF-κB, since the activity of this transcription factor was rather decreased by oxysterols. Similar effects on IL-8 secretion, independent of NF-κB activation, have been reported elsewhere (5, 21). The finding that oxysterols did not affect HSC viability despite a reduction in NF-κB activity is quite intriguing, because a number of studies have provided evidence that NF-κB is involved in resistance to apoptosis in activated HSCs (11). However, similar results to ours have been reported previously in response to oxysterols and oxLDL in macrophages and vascular smooth muscle cells (3, 31). Although a clear explanation for this phenomenon is not given, it is important to note that a variety of mechanisms modulate or counterregulate NF-κB activity and thereby modulate the decision of life or death in a cell (23). For example, Ares and collaborators (3) have provided evidence that the absence of apoptosis in vascular smooth muscle cells after oxysterol-induced inhibition of NF-κB is compensated by a parallel activation of AP-1. Other compensatory mechanisms, including the fact that oxysterols are natural ligands of liver X receptors (LXRs) and retinoid X receptors (RXRs), two nuclear receptors with clear antiapoptotic actions, cannot be excluded (37). On the other hand, oxysterol treatment in HSCs was accompanied by significant upregulation of profibrogenic genes related to extracellular matrix synthesis and degradation. In particular, in HSCs, oxysterols upregulated the expression of TIMP-1, a tissue metalloproteinase inhibitor that promotes liver fibrogenesis (19, 40). Consistent with this finding, we observed that TIMP-1 expression was also upregulated in vivo in livers from hypercholesterolemic ApoE−/− mice. Although changes in the expression of profibrogenic genes in HSCs were less than twofold, changes of a similar magnitude and with biological significance have been described previously in cultured cells (25, 27, 38).
In summary, our study provides evidence that hypercholesterolemic ApoE−/− mice are more susceptible to develop severe liver injury, suggesting that in addition to vascular disease, increased cholesterol products also play a contributory role in accelerating the progression of inflammatory and fibrogenic responses in the liver. Our study also provides evidence that oxidized cholesterol products (i.e., oxysterols) representative of those detected in the liver from hypercholesterolemic ApoE−/− mice trigger key molecular events involved in hepatic inflammation and fibrogenesis in liver cells that were consistent with those found in vivo in hypercholesterolemic ApoE−/− mice. These include the activation of potent proinflammatory and/or profibrogenic factors such as MCP-1, TGF-β1, and IL-8 in macrophages and HSCs and the induction of oxidative stress in hepatocytes. Additional studies are needed to clearly delineate the mechanisms by which oxysterols induce these effects on liver cells, but the fact that oxysterols are the natural ligands of LXRs is a subject that merits further investigation. Finally, although we cannot directly prove whether increased susceptibility to liver injury of hypercholesterolemic ApoE-deficient mice is the result of accumulation of oxysterols, our findings are of pathophysiological relevance because it is known that fatty livers are particularly sensitive to develop severe liver injury (e.g., after ischemia-reperfusion) and that hepatic steatosis accelerates the progression of several types of chronic liver diseases (e.g., chronic hepatitis C) (12, 29).
Supported by grants from the Ministerio de Educación y Ciencia (MEC) (SAF 06/03191 to J. Clària and SAF 05/06245 to R. Battaler) and Instituto de Salud Carlos III (ISCIII) (CIBEREHD). N. Ferré is under a Juan de la Cierva contract (MEC), and M. López-Parra has a contract with ISCIII. A. González-Périz and M. Martínez-Clemente are supported by MEC. R. Horrillo is supported by Generalitat de Catalunya-European Social Funds (2006FI-00091).
We are indebted to A. Rull and J. Marsillach for assistance in the studies involving ApoE knockout mice and immunohistochemistry.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2009 the American Physiological Society