Potential conflict of interest: Nothing to report.
The mechanisms responsible for low mitochondrial respiratory chain (MRC) activity in the liver of patients with nonalcoholic steatohepatitis are unknown. In this study, we examined the cause of this dysfunction in ob/ob mice. Forty-six mice were distributed in six groups: group I: C57BL/6J mice; group II: C57BL/6J Lep(−/−) mice (ob/ob); group III, ob/ob mice treated with manganese [III] tetrakis (5,10,15,20 benzoic acid) porphyrin (MnTBAP); group IV, ob/ob mice treated with IgG1 immunoglobulin; group V, ob/ob mice treated with anti-TNF antibody; group VI: ob/ob mice treated with uric acid. In liver tissue, we measured MRC activity, fatty acid β-oxidation, tumor necrosis factor (TNF), inducible nitric oxide synthase (iNOS), 3-tyrosine-nitrated proteins, 3-tyrosine-nitrated mitochondrial proteins, including cytochrome c and ND4 subunit of complex I. MRC activity was decreased in ob/ob mice. TNF levels, iNOS protein expression, and tyrosine nitrated proteins were markedly increased in the liver of ob/ob mice. In these animals, mitochondrial proteins were markedly tyrosine nitrated, particularly the ND4 subunit of complex I and cytochrome c. Treatment of these animals with uric acid, a peroxynitrite scavenger, anti-TNF antibody, or MnTBAP decreased tyrosine nitrated proteins, improved the activity of MRC complexes, and led to a marked regression of hepatic steatosis and inflammation. In conclusion, MRC dysfunction and liver lesions found in ob/ob mice are likely to reflect the tyrosine nitration of mitochondrial proteins by peroxynitrite or a peroxynitrite-derivate radical. Increased hepatic TNF and iNOS expression might enhance peroxynitrite formation and inhibition of MRC complexes. (HEPATOLOGY 2006;44:581–591.)
Nonalcoholic fatty liver disease (NAFLD) represents a spectrum of liver disease extending from pure fatty liver through nonalcoholic steatohepatitis (NASH) to cirrhosis that occurs in patients who do not consume significant amounts of alcohol.1 NAFLD is likely the most frequent histological finding in patients with abnormal liver function tests in the United States.2
Although the pathogenesis of NAFLD remains undefined, it is now well established that insulin resistance is usually present in patients with NAFLD3 and might be involved in the pathogenesis of this disease.4 Some investigators have proposed a “double hit” theory in the development of NASH. The “first hit” involves the accumulation of fat in the liver, and the “second hit” includes oxidative stress resulting in lipid peroxidation, stellate cell activation, and fibrogenesis.5, 6 Mitochondrial dysfunction might play a crucial role in the induction of both “hits,” because mitochondria are involved in the β-oxidation of free fatty acids and are the most important cellular source of reactive oxygen species (ROS).7 In a previous study, we have shown that the activity of mitochondrial respiratory chain (MRC) complexes is decreased in liver tissue of patients with NASH.8 Although this reduction correlated with serum levels of tumor necrosis factor alpha (TNF-α), insulin resistance, and body mass indexes, because of the limited amounts of tissue available from liver biopsy, mechanisms of this dysfunction were not studied. Therefore, an animal model of NAFLD was mandatory to investigate the mechanisms of the mitochondrial dysfunction found in NASH patients. Obese ob/ob mice, which have a defect in leptin synthesis, reproduce many of the metabolic disturbances present in NASH patients. These animals show an uncontrolled food intake leading to obesity, insulin resistance, hyperlipidemia, hyperinsulinemia, hyperglycemia, and fatty liver.9 These animals have been used to study the pathogenesis of NAFLD.10
A growing body of evidence implicates excessive generation of nitric oxide (NO) and other reactive nitrogen species in the dysfunction of MRC.11 Thus, NO reacts with cytochrome c oxidase (complex IV) and competitively blocks oxygen binding and electron transfer.12 Conversely, peroxynitrite (ONOO−), the product of reaction of NO with superoxide anion (O2−), inhibits several mitochondrial proteins, including MRC complexes.11, 13, 14 Peroxynitrite may exert these effects through a variety of interactions, including oxidation,15 DNA strand breakage,16 and nitration of tyrosine residues on proteins to generate 3-nitrotyrosine (3NT).17 3NT is considered a specific marker of protein nitration by peroxynitrite and has been widely used as an indicator of peroxynitrite-mediated tissue damage.18 Increased formation of 3NT has been demonstrated in a number of pathological conditions, including NASH.19
We conducted the current study to test the hypothesis that increased TNF levels in the liver of ob/ob mice induce inducible nitric oxide synthase (iNOS) protein expression and elevate the formation of peroxynitrite and peroxynitrite-derived radicals in the liver tissue. Tyrosine-nitrated mitochondrial proteins by these radicals account for the decreased activity of MRC complexes in ob/ob mice. Thus, the aims of this study were: (a) to assess the activity of MRC complexes in the liver of ob/ob mice, and (b) to demonstrate that hepatic TNF and peroxynitrite are implicated in the pathogenesis of both the decreased MRC activity and the liver lesions in these obese mice.
All procedures were carried out in accordance with the Spanish Guidelines for the Care and Use of Laboratory Animals. The 6-week-old male C57BL/6J mice and the 6-week-old obese C57BL/6J Lep (−/−) (ob/ob) male mice were purchased from Charles River Laboratory (Charles River Laboratories España, SA, Santa Perpetua de la Mogada, Spain). Animals were housed at constant room temperature (23°C) (n = 3 per cage) under 12-hour light/dark cycles with ad libitum access to water and standard laboratory mouse chow. Forty-six mice were distributed in six groups: group I (wild-type) included 12 C57BL/6J mice. Group II (ob/ob) was formed by 16 obese ob/ob mice treated with 200 μL 0.8% saline solution. Group III (MnTBAP) was composed of six ob/ob mice treated with 10 mg/kg/day manganese [III] tetrakis (5, 10, 15, 20 benzoic acid) porphyrin (MnTBAP) (Calbiochem. San Diego. CA), a mimic of manganese superoxide dismutase. Group IV (IgG1) was formed by six ob/ob mice treated with 10 mg/kg IgG1 immunoglobulin (Sigma-Aldrich Quimica SA, Tres Cantos, Spain) three times weekly (tiw). This group of animals was used as controls of anti-TNF–treated mice. Group V (anti-TNF) consisted of six ob/ob mice treated with 10 mg/kg/tiw anti-TNF antibody (Remicade, Schering Plough, Leiden, Holland). Group VI was six ob/ob mice treated with 20 mg/day uric acid (UA) (Sigma-Aldrich Quimica SA, Tres Cantos, Spain). UA was used as a suspension of 20 mg in 200 μL 0.8% saline solution. Saline, MnTBAP, IgG1 immunoglobulin, and anti-TNF solutions and UA suspension were administered intraperitoneally for 12 weeks.
After the treatment, animals were anesthetized and killed at 18 weeks of age, and the liver was rapidly harvested for further analysis. A portion of the liver tissue was placed in a 10% formaldehyde solution and routinely processed for histological assessment. Sections were stained with hematoxylin-eosin and Masson trichrome. Formation of 3NT in the liver was assessed using immunohistochemistry with 3NT-specific antibody (Upstate Cell Signaling Solutions, Lake Placid, NY) according to the manufacturer's indications. In some experiments, 3NT formation was assessed in hepatocytes isolated from ob/ob mice and cultured for 24 hours in the absence or presence of an iNOS inhibitor [3 mmol L-N6-(1-iminoethyl)lysine hydrochloride].
MRC Activity Assays.
Frozen liver tissues (50-70 mg) were homogenized with 15 vol. of 20 mmol/L KP buffer, pH 7.4, and centrifuged at 800g for 10 minutes. Respiratory chain enzymes and citrate synthase (CS) activities were measured in a DU-650 spectrophotometer (Beckman Instruments, Palo Alto, CA). Incubation temperatures were 30°C for complexes I, II, III, V, and CS, and 38°C for complex IV. Enzyme activities were performed in supernatants as described elsewhere,8 expressed as nanomoles of substrate used per minute per milligram of protein and, to correct for the hepatic content of mitochondria, referred as a percentage of the specific activity of CS. Enzyme assays were performed in triplicate.
Determination of Mitochondrial and Peroxisomal Fatty Acid β-Oxidation.
Liver (80 mg) was homogenized in 400 μL homogenization buffer (0.25 mmol/L sucrose containing 1 mmol/L Tris (Cl-), pH 7.5, and 1 mmol/L EDTA) with one stroke of a loose-fitting hand-driven all-glass homogenizer and mitochondria, and peroxisomes were isolated from fresh liver homogenates as described by Crane et al..20 Oxidation of [1-14C]fatty acids to water-soluble products was measured as described by Watkins et al.21 [1-14C]lignoceric acid (C20:0) and [1-14C]palmitic acid (C16:0) were used to measure peroxisomal or mitochondrial fatty acid β-oxidation, respectively.
Determination of Lipid Peroxidation and Glutathione Content in Mitochondria.
Lipid peroxidation was determined by measuring thiobarbituric acid-reacting substances (TBARS) in 100 μL liver homogenate as described by Ohkawa et al.22 Mitochondrial glutathione was measured using the Eady et al. modification of the Tietze's assay.23
Extraction of Hepatic and Mitochondrial Proteins.
Total protein was extracted from liver tissue using a standard protocol in our laboratory.24 Liver mitochondria were isolated from liver homogenates by differential centrifugation as described by Turko et al.25
Hepatic concentrations of TNF, interferon gamma (IFN-γ), and interleukin 1β (IL-1β) were measured using high-sensitivity enzyme-linked immunosorbent assays according to the manufacturer's instructions (Mouse TNF ELISA, Mouse IFN-γ ELISA; Mouse IL-1β ELISA, BLK Diagnostics, Badalona, Spain). Triglyceride concentration in liver tissue was determined using a serum triglyceride determination kit (Sigma-Aldrich Química SA), following the manufacturer's indications.
Immunoprecipitation of Mitochondrial Proteins and Western Blot Analysis.
Mitochondrial proteins (200 μg) were incubated with specific mouse monoclonal antibody against 3-nitrotyrosine (Upstate Biotechnology, Lake Placid, NY). The immunoprecipitation assays were performed as previously described.26 Immune complexes were detected using specific antibodies (anticytochrome c, antiND4) (Santa Cruz Biotechnology, Santa Cruz, CA).
Hepatocytes were isolated from wild-type and ob/ob mice according to the method of Moldéus et al.27 Two-color immunophenotyping of liver cells was performed by flow cytometry using a FACScan flow cytometer (Becton Dickinson, San Jose, CA). For double staining of intracellular iNOS and albumin antigens, cells were fixed for 1 hour in 1% paraformaldehyde and permeabilized with 1% Tween 20 before labeling with anti-NOS2-fluorescein isothiocyanate (FITC) (Santa Cruz Biotechnology) or anti-albumin (Lab Clinics SA, Barcelona, Spain) and Alexa Fluor 594-labeled secondary antibody (Molecular Probes, Barcelona, Spain). Background staining was assessed with appropriate isotype- and fluorochrome-matched control antibody and subtracted. In some experiments, liver cells were treated with 15 ng/mL TNF (Quimigranel, San Fernando de Henares, Spain) for 24 hours before flow cytometric analysis.
Statistical analysis was aided by the SPSS Statistical Software for Windows, version 9 (SPSS Inc, Chicago, IL). The unpaired t test was used to assess the significance of differences between means. All results were expressed as mean ± SD unless otherwise mentioned. Spearman's correlation coefficient was used for correlation analysis between variables. P values of less than .05 were considered significant.
Activity of the MRC in ob/ob Mice.
We measured the activity of the MRC complexes in wild-type and ob/ob mice. The activity of complex I in liver tissue, which accepts electrons from NADH and transfers them to ubiquinone (Supplementary Fig. 1; Supplementary material for this article is available at: http://interscience.wiley.com/jpages/0270-9139/suppmat/index.html), was decreased from 51.3 ± 2.0 [(nmol × min−1 × mg protein−1/nmol × min−1 × mg protein−1 CS) × 100] (complex 1/CS) in wild-type mice to 36.0 ± 2.8 complex 1/CS (P < .001) (70.3% ± 5.5%) in ob/ob mice (Fig. 1A).
In ob/ob mice, the activity of complex II (succinate dehydrogenase complex), which passes electrons directly to ubiquinone (Supplementary Fig. 1), was also significantly reduced to 52.1% ± 5.3% of control activity (wild-type mice, 51.4 ± 6.0 complex II/CS; ob/ob mice, 25.7 ± 4,1 complex II/CS; P < .001) (Fig. 1A).
Ubiquinone passes electrons from complex I and II to the b-c1 complex (complex III), which transfers them to cytochrome c (Supplementary Fig. 1). In ob/ob mice, the activity of complex III was decreased to 43.2% ± 15.7% of control values (wild-type mice, 73.8 ± 2.3 complex III/CS; ob/ob mice, 31.9 ± 11.6 complex III/CS; p<0.01) (Fig. 1A).
Cytochrome c is involved in carrying electrons from the b-c1 complex to the cytochrome c oxidase complex (complex IV), which finally transfers these electrons to oxygen (Supplementary Fig. 1). Measurement of the activity of this complex in the liver tissue of ob/ob mice showed that it was decreased to 60.1 ± 7.1% of the control activity (wild-type mice, 56.5 ± 5 complex IV/CS; ob/ob mice, 34.0 ± 4.0 complex IV/CS; P < .001) (Fig. 1A).
Transport of electrons through the MRC is coupled to the pumping of protons from the mitochondrial matrix to the intermembrane space. Complex V converts adenosine diphosphate in adenosine triphosphate when protons flow back from the intermembrane space into the matrix (Supplementary Fig. 1). The activity of complex V (adenosine triphosphate synthase) was markedly reduced to 63.9% ± 8.4% of the control activity in ob/ob mice. Thus, although complex V activity was 272.1 ± 27.3 complex V/CS in wild-type mice, this activity was 185.2 ± 22.7 complex V/CS in ob/ob mice (P < .01) (Fig. 1A).
Specific activities of CS were 1,240 ± 271 nmol/min/mg protein for wild-type mice and 1,138 ± 249 nmol/min/mg protein for ob/ob mice, indicating no proliferation of mitochondria in ob/ob mice. There were no significant differences between both values.
Mitochondrial and Peroxisomal Fatty Acid β-Oxidation.
Because fatty acid β-oxidation has been found to be increased in patients with NASH,19 we measured oxidation of [1-14C]palmitic acid and [1-14C]lignoceric acid to water-soluble products by mitochondrion and peroxisome preparations obtained from liver tissue of wild-type and ob/ob mice. As Fig. 1B shows, β-oxidation of both fatty acids by the liver tissue was markedly increased in ob/ob mice.
Lipid Peroxidation in Liver Tissue.
Because oxidative stress is believed to play an important role in the pathogenesis of NASH,4, 7 and defective MRC is an important source of ROS,7 we measured TBARS in liver homogenates and glutathione concentration in mitochondria and cytosol prepared from fresh livers of wild-type and ob/ob mice. As Fig. 2A-B and Table 1 show, TBARS concentration was increased and mitochondrial glutathione was decreased in the livers of ob/ob mice. Treatment of these mice with MnTBAP, intraperitoneally for 3 months, a mimic of superoxide dismutase, normalized TBARS and glutathione levels. Moreover, treatment with this antioxidant increased the activity of MRC complexes I, III, and IV significantly, but decreased the activity of complex II (Fig. 2C).
Table 1. Liver Content in Glutathione and Triglycerides and Serum Aminotransferase Levels
Group of Mice
Cytosolic Glutathione (nmol/mg protein)
Mitochondrial Glutathione (nmol/mg protein)
Liver Triglycerides (mg/g liver)
Serum AST (IU/L)
Serum ALT (IU/L)
Serum Uric Acid (mg/dL)
NOTE. Cytosolic and mitochondrial glutathione and liver content in triglycerides were measured as described under “Material and Methods.” Serum AST, ALT, and uric acid were determined using automated techniques (Hitachi 747 Roche Diagnostic Corp., Indianapolis. IN). NS, not significant vs. wild-type mice.
TNF, IFN-γ, and IL-1β Concentrations in Liver Tissue.
To determine whether TNF, IFN-γ, and IL-1β are increased in the liver tissue of ob/ob mice, we measured levels of these cytokines in protein extracts from liver tissue using ELISA kits. As Fig. 3 shows, hepatic levels of these three cytokines were significantly increased in ob/ob mice (P < .001). To investigate whether TNF was involved in the decreased activity of MRC complexes, we administered 10 mg/kg anti-TNF antibody tiw by intraperitoneal injection for 3 months as described by Zwerina et al.28 Treatment of ob/ob mice with anti-TNF antibody resulted in a marked decrease in hepatic levels of TNF, IFN-γ and IL-1β (Fig. 3A-B) and in a significant improvement in the activity of the MRC complexes (Fig. 3C), particularly of complexes I, II, III, and V. Activity of complex IV remained unchanged. A negative correlation was found between hepatic TNF levels and the activity of MRC complexes (complex I, r = −0.57, P < .05; complex II, r = −0.56, P < .05; complex III, r = −0.56, P < .05; complex IV, r = −0.64, P < .01; complex V, r = −0.57, P < .05). This treatment increased cytosolic and mitochondrial glutathione levels in the liver of ob/ob mice (Table 1).
iNOS Protein Expression in Hepatic Mitochondria.
Because TNF is known as a strong inducer of iNOS expression in a variety of cell types, including hepatocytes,29 we examined whether iNOS protein expression was up-regulated in the hepatic mitochondria of ob/ob mice. Western blot analysis of mitochondrial proteins showed that iNOS expression was very low in wild-type mice; however, this expression was markedly (5.29 ± 0.6-fold; P < .001) increased in the mitochondria of ob/ob mice. Furthermore, flow cytometry of cells isolated from the liver of ob/ob mice showed that 48.2% ± 2.2% of liver cells expressing albumin also express iNOS. This percentage is significantly higher than the 11.4% ± 1.8% found in the liver of lean mice (P < .001) (Fig. 4B). Moreover, expression of iNOS was strikingly increased in cultured hepatocytes from lean mice treated with 15 ng/mL TNF for 24 hours (Fig. 4C). Treatment of mice with anti-TNF antibody for 3 months resulted in a decrease in the iNOS protein expression to control levels (Fig. 4A). Likewise, treatment of mice with MnTBAP for the same period down-regulated iNOS expression significantly (Fig. 4A).
3-Tyrosine Nitrated Proteins in Liver Tissue and the Effect of Uric Acid.
NO, which is produced endogenously by iNOS, can rapidly react with superoxide anion to produce peroxynitrite. This anion reacts with complexes I, II, V, and cytochrome c and leads to their inactivation.12, 13 Mechanisms of these effects are not well known, but nitration of critical tyrosine residues may alter protein function. We therefore assessed the presence of 3-tyrosine nitrated proteins in the liver by immunohistochemical staining for 3NT. Figure 5 shows that although a small amount of 3NT staining was detectable in wild-type mice (Fig. 5A), considerable immunofluorescence for 3NT was noted in the liver of ob/ob mice (Fig. 5B) Treatment of cultured hepatocytes isolated from ob/ob mice for 24 hours with 3 mmol L-N6-(1-iminoethyl)lysine hydrochloride resulted in a striking decrease in the 3NT staining of these cells (Fig. 5C-D).
Because UA has been proposed as a natural scavenger of peroxynitrite and peroxynitrite-derived radicals,30 we assessed for the presence of 3NT in the liver of ob/ob mice treated with 20 mg/day UA intraperitoneally for 3 months. As shown in Fig. 5E, in animals treated with uric acid, the degree of immunofluorescence decreased to the level in wild-type mice. Likewise, treatment of ob/ob mice with anti-TNF antibody or MnTBAP decreased significantly the presence of 3NT in the liver (Fig. 5F-G). On the contrary, treatment of mice with IgG1 immunoglobulin did not reduce the accumulation of 3NT residues in the liver of ob/ob mice (Fig. 5H).
Moreover, intraperitoneal administration of UA normalized the activity of MRC complexes I and V and improved significantly the activity of complexes II and III (Fig. 6). Complex IV activity did not change significantly after UA treatment. We also found that UA reduced TBARS concentration in the liver significantly (Fig. 6B) but did not change TNF concentrations in liver tissue (Fig. 6C).
Tyrosine-Nitrated Mitochondrial Proteins.
Considering the intramitochondrial location of the MCR complexes, we analyzed 100 μg mitochondrial proteins by Western blotting. Membranes were probed with specific antibody against 3-NT. Analysis of mitochondrial proteins shows several 3NT-positive bands, whose intensity was clearly increased in ob/ob and IgG1 mice (Fig. 7A). Treatment of ob/ob mice with uric acid, anti-TNF antibody, or MnTBAP, but not with IgG1 immunoglobulin, decreased the amount of 3NT proteins in mitochondria of ob/ob mice.
To determine the identity of tyrosine-nitrated mitochondrial proteins, 200 μg of these proteins was immunoprecipitated with specific anti-3NT antibody and resolved by SDS-PAGE electrophoresis. Western blot analysis performed using specific anti-cytochrome c or anti-ND4 antibodies showed that ND4 and cytochrome c were clearly increased in ob/ob mice and ob/ob mice treated with IgG1 immunoglobulin. Treatment of these animals with MnTBAP, anti-TNF antibody, or UA decreased markedly nitration of these mitochondrial proteins (Fig. 7C-D).
The liver of ob/ob mice showed a marked accumulation of fat droplets in about 50% of hepatocytes. In 10% of liver cells, fat was seen mainly as macrovesicular droplets, whereas, in the remaining 40% of hepatocytes, fat caused microvesicular steatosis (Fig. 8B). Focal areas of necrosis and collections of mononuclear cells and neutrophils were also present in the parenchyma and in the portal tracts (Fig. 8B inset). Treatment of obese mice with antioxidant doses of MnTBAP (Fig. 8C-D), anti-TNF antibody (Fig. 8E-F), or UA (Fig. 8G-H) for 3 months led to a complete resolution or decreased markedly hepatic steatosis and inflammation. Table 1 shows triglyceride concentrations in liver tissue and serum aminotransferase levels.
In a previous study, we showed that the activity of MRC complexes is decreased in liver tissue from patients with NASH8 In the current study, we measured MRC activity in obese ob/ob mice, which reproduce many of the metabolic disturbances present in NASH patients.9 In these mice, we confirm that the activity of MRC complexes is reduced by 30% to 50% of the activity in wild-type animals (Fig. 1A). These results were similar to those we found in NASH patients.8 and indicate that this animal model can be used to investigate the mechanisms of the MRC dysfunction.
Although no other studies have been published concerning the function of MRC in patients with NASH, our results concur with those reported by Spahr et al.31 and Cortez-Pinto et al.32 showing that mitochondrial function is impaired in these patients.
Free fatty acids are β-oxidized either in the mitochondria or in the peroxisomes. Although palmitic acid is oxidized mainly in the mitochondria, lignoceric acid, a very-long-chain fatty acid, is initially chain-shortened in the peroxisomes, and their products enter into the mitochondria for completion of the β-oxidation.21 Our study shows that oxidation of both palmitic and lignoceric acids are significantly increased in the liver of ob/ob mice (Fig. 1B). These results concur with those reported by others in ob/ob mice33 or in patients with NASH19 and have been ascribed to the insulin resistance.19
A number of mechanisms can be considered to explain the mitochondrial dysfunction found in NAFLD patients and ob/ob mice. One of them is oxidative stress. Malondialdehyde and 4-hydroxynonenal, two byproducts of lipid peroxidation, can inhibit mitochondrial cytochrome c oxidase by forming adducts with this enzyme.34 ROS-induced depletion in mitochondrial DNA (mtDNA) can severely affect mitochondrial function and induces steatosis and liver lesion.35 Such depletion can impair the synthesis of complexes I, III, IV, and V of the MRC, because mtDNA encodes for 13 of the polypeptides of these complexes. In NASH patients, Haque et al. have reported depletion in mtDNA.36 Our study demonstrates that TBARS, a marker of oxidative stress, was markedly increased in the liver of ob/ob mice. Likewise, cytosolic and mitochondrial glutathione was decreased in these animals. Evidence of oxidative stress has also been found in patients with NASH.19 Decreased activity of the MRC in ob/ob mice is in part attributable to the oxidative stress, because treatment of ob/ob mice with antioxidant doses of MnTBAP, a mimic of manganese superoxide dismutase, improved the activity of several complexes of the MRC. Surprisingly, activity of complex II, which polypeptides are only encoded by nuclear DNA, decreased even more in ob/ob mice treated with MnTBAP. The cause of this decrease deserves further study; however, we could speculate that MnTBAP might inhibit this complex directly. Furthermore, liver histology improved markedly in mice treated with MnTBAP (Fig. 8C-D).
TNF is another factor that can interfere with mitochondrial function. In fact, we and others detected impaired electron flow at the level of complexes I and III in cells treated with TNF.37, 38 Raised serum TNF levels have been demonstrated in patients with fatty liver or NASH.8, 39 In NASH patients, we found a significant negative correlation between the serum levels of TNF and the activity of the MRC complexes.8 The current study indicates that concentrations of TNF, IFN-γ, and IL-1β were markedly elevated in the liver tissue of ob/ob mice (Fig. 3A-B). Consistent with these results, Li et al.33 found that hepatic TNF-mRNA levels were increased in ob/ob mice. Adipose tissue, hepatocytes, and Kupffer cells are important sources of TNF production.40, 41 A number of studies have demonstrated that free fatty acids, which blood levels are increased in NAFLD patients, induce TNF expression in hepatocytes42 and adipocytes.43. The role of this cytokine in the pathogenesis of the mitochondrial dysfunction is strongly supported by the results of the current study, because intraperitoneal administration of anti-TNF antibody to ob/ob mice resulted in a marked increase in the MRC activity (Fig. 3C) and in an almost completely regression of hepatic lesions (Fig. 8E-F). These effects of anti-TNF antibody on liver histology have also been reported by Li et al.33
iNOS catalyze the oxidation of L-arginine to form the free radical NO. Under normal conditions, only constitutive endothelial NOS is present in the liver. However, under certain conditions, including TNF treatment, iNOS is induced, and large amounts of NO are generated in the liver.44 Flow cytometric analysis demonstrates that iNOS expression is markedly increased in hepatocytes from ob/ob mice and that this expression can be induced by the treatment of cultured hepatocytes from lean mice with TNF (Fig. 4B-C). TNF is a potent inducer of nuclear factor-kappaB (NF-κB). NF-κB activity is markedly increased in ob/ob mice,33 and iNOS is an NF-κB–dependent gene.29 The current study shows that iNOS protein expression is upregulated in the liver mitochondria of ob/ob mice compared with wild-type mice. Evidence indicating the existence of a mitochondrial isoform of iNOS has been presented by various groups.11, 13, 45 Consistent with these results are those reported by Laurent et al.,46 who showed that the concentrations of nitrites and nitrates were significantly increased in liver homogenates of ob/ob mice.46 Our study also indicates that TNF plays a major role in the induction of iNOS protein expression, because intraperitoneal administration of anti-TNF antibody decreased iNOS expression to control levels (Fig. 4A).
Our study shows that the liver of ob/ob mice contains large amount of tyrosine nitrated proteins (Fig. 5B,H) and that mitochondrial proteins were particularly nitrated (Fig. 7A). These nitrated proteins seem to be dependent from the iNOS activity in liver cells, because inhibition of this enzyme with L-N6-(1-iminoethyl)lysine hydrochloride for 24 hours reduced markedly the content of ob/ob hepatocytes in 3NT (Fig. 5C-D). Moreover, immunoprecipitation of mitochondrial proteins with anti-3NT antibody demonstrated that at least ND4, a subunit of complex I, and cytochrome c, were tyrosine nitrated (Fig. 7C-D). Because nitration of these mitochondrial enzymes has been found to be associated with a decrease in their catalytic activity,11, 13, 47 we wanted to assess whether peroxynitrite or peroxynitrite-derived intermediates30 are involved in the pathogenesis of liver lesions and MRC dysfunction. Therefore, we analyzed the effect of UA on the activity of the MRC complexes in ob/ob mice. UA reacts very rapidly with peroxynitrite, leads to the formation of nitrated urate derivatives, and has been proposed as a natural scavenger of peroxynitrite anion and peroxynitrite-derived intermediates.30, 48 In any case, it has been shown that UA inhibits peroxynitrite-mediated 3NT formation30, 48 and prevents progression of neurological lesions in an experimental model for multiple sclerosis.18 Our study shows that UA normalized activity of complexes I and V and improved activity of complexes II and III. These effects were associated with a marked reduction in the presence of abnormal 3NT in hepatic proteins, as assessed by immunohistochemistry (Fig. 5E), and in tyrosine nitrated specific mitochondrial proteins (Fig. 7). Improvement in MRC activity after UA treatment supports the concept that peroxynitrite or peroxynitrite-derived intermediates play a critical role in the pathogenesis of MRC dysfunction in ob/ob mice. That UA did not improve the activity of complex IV was hardly surprising, because this complex is unaffected by peroxynitrite.11, 13, 49 The low activity of complex IV in ob/ob mice requires further studies. However, the improvement in the activity of this complex observed after MnTBAP administration to ob/ob mice suggests that this decrease is attributable to the oxidative stress.
A striking finding in our study was that treatment with UA resulted in a nearly complete resolution of fatty liver in these animals (Fig. 8G-H), suggesting that a peroxynitrite-related mechanism is also involved in the pathogenesis of this lesion. A protective role of UA in peroxynitrite-mediated lesions has been shown by other authors in mice.18 Thus, some authors have speculated that UA may be an important agent in humans against peroxynitrite-mediated tissue damage.50
Finally, treatment of ob/ob with anti-TNF antibody has similar effects as those obtained with uric acid, including the decreased tyrosine nitrated mitochondrial proteins and the resolution of liver lesions, suggesting that TNF exerts its pathological effects, acting through peroxynitrite-dependent mechanisms and the tyrosine nitration of mitochondrial proteins.
In conclusion, these results suggest that MRC dysfunction and liver lesions found in ob/ob mice are likely attributable to the tyrosine nitration of mitochondrial proteins by peroxynitrite or peroxynitrite-derived substances. Increased hepatic TNF and induced iNOS might results in enhanced peroxynitrite formation and inhibition of MRC complexes.