The role of iNOS in alcohol-dependent hepatotoxicity and mitochondrial dysfunction in mice



Nitric oxide (NO) is now known to control both mitochondrial respiration and organelle biogenesis. Under conditions of ethanol-dependent hepatic dysfunction, steatosis is increased, and this is associated with increased expression of inducible nitric oxide synthase (iNOS). We have previously shown that after chronic exposure to ethanol, the sensitivity of mitochondrial respiration to inhibition by NO is enhanced, and we have proposed that this contributes to ethanol-dependent hypoxia. This study examines the role of iNOS in controlling the NO-dependent modification of mitochondrial function. Mitochondria were isolated from the livers of both wild-type (WT) and iNOS knockout (iNOS−/−) mice that were fed an isocaloric ethanol-containing diet for a period of 5 weeks. All animals that consumed ethanol showed some evidence of fatty liver; however, this was to a lesser extent in the iNOS−/− mice compared to controls. At this early stage in ethanol-dependent hepatic dysfunction, infiltration of inflammatory cells and the formation of nitrated proteins was also decreased in response to ethanol feeding in the iNOS−/− animals. Mitochondria isolated from wild-type ethanol-fed mice showed a significant decrease in respiratory control ratio and an increased sensitivity to NO-dependent inhibition of respiration relative to their pair-fed controls. In contrast, liver mitochondria isolated from iNOS−/− mice fed ethanol showed no change in the sensitivity to NO-dependent inhibition of respiration. In conclusion, the hepatic response to chronic alcohol-dependent cytotoxicity involves a change in mitochondrial function dependent on the induction of iNOS. (HEPATOLOGY 2004;40:565–573.)

The importance of nitric oxide (NO) in regulating several aspects of mitochondrial function, including control of respiration and mitochondrial biogenesis, is now emerging as a major new area of cell signaling.1–3 The activity of the enzyme cytochrome-c oxidase is controlled by the binding of NO to the oxygen binding site; this is reversed by the competing reaction with oxygen in the mitochondrial inner membrane.4–6 It has been suggested that one function of the NO–cytochrome-c oxidase signaling pathway is to control oxygen gradients in the cell through limiting oxygen consumption in the most actively respiring mitochondria.7 This concept is supported by the recent finding that NO has a greater impact on state 3 respiration, a condition in which mitochondria are most metabolically active in synthesizing adenosine triphosphate (ATP).8 It is also known that NO binding to cytochrome-c oxidase increases the rate of superoxide production at other respiratory complexes.9 It is now evident that under controlled conditions, the formation of superoxide from the mitochondrion can be an important source of hydrogen peroxide capable of participating in cell signaling.10, 11 Recent studies have revealed a cross-talk between the soluble guanylate cyclase signaling pathway and mitochondrial biogenesis.3 Taken together, these findings indicate a complex interplay between mitochondrial respiration, redox cell signaling, and the production of NO in the cell. The consequence of loss of control of NO-dependent respiration in response to stress and its role in the development of specific pathological processes has not been studied in depth.

We propose that under conditions in which NO levels are increased in response to induction of inducible nitric oxide synthase (iNOS), the consequent loss of control of the NO–cytochrome-c oxidase pathway is deleterious. Increased NO will lead to secondary production of reactive nitrogen species (RNS), such as peroxynitrite, leading to irreversible modification of respiratory chain proteins and an increased sensitivity to NO.12 We have recently investigated this possibility in a model of chronic ethanol exposure in rats in which it is known that fatty liver is associated with a profound defect in mitochondrial metabolism.13–18 This is an interesting model in which to examine these concepts since it has already been demonstrated that NO and nitrotyrosine formation, a marker of RNS, is enhanced in the liver of ethanol-exposed animals.19–21 Furthermore, numerous studies have indicated that ethanol-dependent hepatotoxicity is associated with increased reactive oxygen species (ROS) formation.14, 19, 21–23 It has also been shown that in mice deficient in manganese superoxide dismutase (SOD1), liver and oxidative damage and nitrotyrosine formation were enhanced in response to ethanol feeding.21

The iNOS−/− model has been widely used to examine the role of this enzyme in the inflammatory response.24–26 The induction of iNOS has been associated with damage to mitochondria and the nitration of several mitochondrial proteins.27–30 Recently, it has been shown that in iNOS−/− animals, ethanol hepatotoxicity is significantly prevented through a mechanism that involves a decreased inflammatory response, tumor necrosis factor α (TNF-α) formation, and fatty liver.26 These studies complement previous findings in which it was shown that a nitric oxide synthase (NOS) inhibitor prevented ethanol-dependent hepatotoxicity26 and that arginine reversed ethanol-induced inflammatory changes.31 However, the impact on mitochondrial function and its control by NO is unknown.

Recently, we demonstrated that after ethanol consumption, liver mitochondria are much more susceptible to NO-dependent inhibition of respiration.13 This finding led to the hypothesis that the early induction of iNOS contributes to the development of alcohol-dependent hepatotoxicity by contributing to the decrease in oxygen supply that leads to centrilobular hypoxia.32, 33 It has recently been shown in the TNF-α receptor 1−/− mouse that ethanol hepatotoxicity is markedly decreased, and this is associated with substantially lower levels of tyrosine residue nitration.34 In the present study, we have used the Lieber-De Carli alcohol-containing liquid diet, which is well known to lead to steatosis, oxidative stress, and mild liver injury, the earliest stages of alcoholic liver disease.35 We hypothesize that the initial stages of ethanol-dependent hepatic dysfunction, which are accompanied by hypoxia, would amplify the formation of ROS/RNS by the mitochondrion, leading to increased protein modification and mitochondrial damage. This scenario implies a possible linkage between increased NO formation from iNOS, enhanced sensitivity of mitochondrial respiration to NO, and the formation of inflammation and steatosis. This hypothesis was tested in a model of chronic alcohol toxicity using C57BL/6 and the iNOS−/− mice.


NO, nitric oxide; ATP, adenosine triphosphate; iNOS, inducible nitric oxide synthase; RNS, reactive nitrogen species; ROS, reactive oxygen species; SOD1, manganese superoxide synthase; TNF-α, tumor necrosis factor α; NOS, nitric oxide synthase; ADP, adenosine diphosphate; Hb, oxyhemoglobin; LPS, lypopolysaccharide.

Materials and Methods


All chemicals were from Sigma-Aldrich-Fluka (St. Louis, MO) except PAPANONOate, which was obtained from Alexis (San Diego, CA).

Animal Treatments.

Male wild-type (C57BL/6) or iNOS knockout (B6.129P2-NOS2 tm/lau) mice, 5 to 7 weeks old, were obtained from Jackson Laboratory (Bar Harbor, ME) and fed laboratory chow and water ad libitum for 1 week. Mice were fed a 4% ethanol-containing liquid diet or pair-fed control diets in which maltose-dextrin was substituted for ethanol calories as described by Lieber-De Carli.27, 36 The ethanol diet provides 28.8% of the total daily caloric intake as ethanol, 18.2% as carbohydrate, 18% as protein, and 35% as fat. Mice were maintained on the diets for 5 weeks and weighed each week. After feeding, livers were excised and prepared for mitochondrial isolation or pathological examination. All animals were handled in accordance with recommendations in The Guide for the Care and Use of Laboratory Animals (USDHHS, NIH Publication no 86-23, 1996). Human hemoglobin and liver mitochondria were prepared as previously described.5

Liver Histology and Immunohistochemistry.

Liver samples were fixed in 10% buffered formalin and embedded in paraffin. Five-μm sections were stained with hemotoxylin-eosin and evaluated by a pathologist who was blinded from the experimental protocol. The severity of the liver pathology was assessed as steatosis (the percentage of liver cells containing fat) 0 (≤5%); 1+ (6%-25%); and 2+ (26%-50%). Intra-acinar inflammation was defined as the presence of aggregates of inflammatory cells including polymorphonuclear leukocytes, lymphocytes, and other mononuclear cells, as well as microgranulomas. In the grading system, the absence of any intra-acinar inflammation was scored 0, 1 foci or less was scored 1, 2 to 4 foci was scored 2, 5 to 10 foci was scored 3, and the presence of 10 or more foci per 10× objective was scored 4. Necrosis was evaluated as the number of necrotic foci per 10× objective field. Immunohistochemistry for 3-nitrotyrosine was performed on sections deparaffinized in xylene and rehydrated through graded ethanol concentrations. To block endogenous peroxidase activity, the sections were immersed in methanol containing 0.3% hydrogen peroxide at −20°C for 30 minutes. Sections were stained with primary antibody to nitrotyrosine 1:500 (a kind gift from Dr. Alvaro Estevez, University of Alabama, Birmingham) and were developed using Vectors ABC Kit (Vector Laboratories, Burlingame, CA) and counterstained with hematoxylin.

Blood Alcohol Measurement.

Ethanol concentration in the serum was measured using the alcohol dehydrogenase kit from Diagnostic Chemical Ltd. (Oxford, CT).

Western Blotting.

Protein samples from liver homogenate were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis under reducing conditions, transferred to nitrocellulose membranes, and blocked for 1 hour. For immunodetection of TNF-α, blots were incubated with primary antibody 1:100 dilution anti–TNF-α (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) overnight at 4°C. This was followed by incubation with secondary antibody for peroxidase-conjugated anti-mouse immunoglobulin G (Amersham Bioscience, Little Chalfont, UK). Immunoreactive proteins were detected using enhanced chemiluminescence (SuperSignal, WestDura Pierce, Rockford, IL).


Deparaffinized sections were rinsed with PBS-buffer (PBS) (0.1 mol, pH 7.2) and incubated for 1 hour with 10% goat serum, followed by overnight incubation at 4°C in a humidified chamber with rabbit anti-iNOS antibody (Transduction Laboratories, Lexington, KY) at a dilution of 1:500. Sections were blocked again for 10 minutes with 5% bovine serum albumin in PBS, washed, and then incubated with for 1 hour with the secondary antibody Alexa Fluor 350 conjugated goat anti-rabbit (Molecular Probes, Eugene, OR). Sections were washed several times with PBS, and the nuclei were counterstained with Hoechst 33258 (20 μg/mL; Sigma-Aldrich-Fluka) for 10 minutes. After washing in PBS, images were taken.

Respiration and NO Control Measurements.

Oxygen and NO concentrations were measured simultaneously using Clark-type electrodes (Instech, Plymouth Meeting, PA; Harvard Apparatus, Holliston, MA) in a 1.0 mL sealed Perspex chamber, magnetically stirred at 37°C.36 Mice liver mitochondria (0.5-1mg/mL) were incubated in respiration buffer comprising 120 mmol KCl, 3 mmol HEPES, 1 mmol EGTA, 25 mmol sucrose, 5 mmol MgCl2, and 5 mmol KH2PO4 at pH 7.4. Data was collected using a digital recording device (Dataq, Akron OH) connected to a PC. The relationship between NO concentration and inhibition of respiration was established by measuring the rate of oxygen consumption at 30 second increments before and after the addition of the NO donor. The NO concentration for a given rate of respiration was determined from the NO traces at the midpoint for the time window used for the measurement of oxygen consumption.13 The NO inhibition curves for mitochondrial respiration were then constructed as previously described.36


All experiments were performed a minimum of 4 to 6 times, and data are presented as means ± SEM. The experiment was performed with 6 pair-fed controls and ethanol-containing diets for wild-type and iNOS−/− animals with 2 deaths in the ethanol-fed groups over the period of the experiment. Statistical significance was determined using the paired Student t test, with P less than .05 taken as significantly different. For comparison of pathological scores, the Mann-Whitney rank sum test was used. A P value less than .05 were taken as the level of significance.


Feeding Characteristics and Liver Pathology in Ethanol-Consuming Wild-Type and iNOS−/−- Mice.

To test the role of iNOS induction in ethanol-dependent hepatoxicity, wild-type and iNOS−/− mice were fed the Lieber-De Carli 4% ethanol-containing diet for 5 weeks. The ethanol concentration in the diet was gradually increased from 7.2% to 28.8% of the total caloric intake per day during the first week of feeding and then maintained for 4 weeks. The pair-fed controls received an isocaloric diet in which ethanol calories were replaced by maltose-dextrin. Alcohol consumption was not different in wild-type and iNOS−/− mice (Table 1) and in the case of wild-type mice was associated with a statistically significant increase in liver weight and the liver/body weight ratio. In addition, there was no difference in body weight in all 4 groups of mice studied during the entire period of feeding. Although the weight gain for the wild-type and iNOS−/− mice on the ethanol-containing diet appeared to be slightly higher, this was not significantly different when compared with controls. Interestingly, the increase in liver weight associated with ethanol consumption in the wild-type mice did not occur in iNOS−/− mice consuming ethanol.

Table 1. Effect of Chronic Ethanol Feeding on General Features in Wild-Type and iNOS−/− Mice
 iNOS+/+ ControliNOS+/+ EthanoliNOS−/− ControliNOS−/− Ethanol
  • NOTE. Data represent the mean ± SEM for 5 pairs of wild-type mice on an ethanol-containing or control diet and iNOS−/− in mice on an ethanol-containing or control diet.

  • *

    P < .001 for corresponding controls. For pathological evaluation, Mann-Whitney rank sum test was performed.

  • P < .05 compared with wild-type mice on ethanol-containing diet.

  • P < .01 compared with wild-type mice on control diet.

Ethanol consumption (g/kg body weight/d) 13.7 ± 0.80 13.7 ± 0.70
Initial body weight (g)23.1 ± 0.4822.1 ± 0.4022.4 ± 0.4222.2 ± 0.36
Final body weight (g)23.8 ± 0.5324.2 ± 0.622.6 ± 0.3623.3 ± 0.53
Liver weight (g)1.26 ± 0.041.47 ± 0.04*1.13 ± 0.081.28 ± 0.06
Liver/body weight ratio (%)5.35 ± 0.186.14 ± 0.26*4.98 ± 0.395.27 ± 0.26
Blood alcohol (mmol/L)8.33 ± 1.538.37 ± 2.25
TNF-α (fold change)1.80 ± 0.233.85 ± 0.401.0 ± 0.41.75 ± 0.0
No. of inflammatory cells (per 10× power field)0.0 ± 0.010 ± 2.71.2 ± 0.83.4 ± 1.8
Steatosis score0.0 ± 0.01.2 ± 0.30.0 ± 0.00.2 ± 0.2
Necrosis and apoptosis score0.0 ± 0.01.2 ± 0.20.4 ± 0.20.2 ± 0.2
Inflammation score0.0 ± 0.01.0 ± 0.00.0 ± 0.00.4 ± 0.2
Total score0.0 ± 0.01.13 ± 0.160.1 ± 0.080.26 ± 0.1

To assess the pathological responses to ethanol, liver tissues were fixed in 10% buffered formalin and sections of the paraffin-embedded tissue then stained with hematoxylin-eosin. As seen in the photomicrographs presented in Figs. 1A and B, no pathological changes were observed in livers from wild-type and iNOS−/− mice fed control diets. Consumption of ethanol in the wild-type mice resulted in the development of mild to moderate steatosis (Fig. 1C, Table 1). The distribution of the steatotic hepatocytes was predominantly in zones 1 and 2 with relative sparing of zone 3. Wild-type mice on ethanol diet showed mild necrosis and inflammation that were significantly decreased in iNOS−/− mice on an ethanol diet (Fig. 1D and Table 1). A slight increase in serum alanine aminotransferase and myeloperoxidase was observed in mice on ethanol diet (data not shown); however, these changes were not statistically significant when compared to mice on control diet. These results are in accordance with earlier reports.21, 37, 38 Taken together with the changes in liver weight in response to ethanol, these data support the finding that ethanol consumption using the Lieber-De Carli diet induces liver dysfunction. These data are similar to those reported by Mckim et al.26 and confirm the finding that the iNOS−/− attenuates the majority of these pathological changes.

Figure 1.

Representative photomicrographs of livers from wild-type and iNOS−/− mice. Liver sections from ethanol-fed and control mice were fixed in 10% buffered formalin, embedded in paraffin, and sectioned and stained with hematoxylin-eosin. Liver sections from (A) wild-type mice and (B) iNOS−/− mice fed control diets showed no signs of pathological changes or fat deposition in hepatocytes. (C) Wild-type mice given an ethanol-containing diet show moderate fatty liver changes, whereas (D) iNOS−/− mice on the ethanol-containing diet show only mild fatty liver changes (original magnification, ×40).

Changes in Protein Nitration and iNOS Expression.

The nitration of tyrosine by RNS generated from either peroxynitrite or myeloperoxidase and nitrite has been reported to be increased in the rat liver in response to chronic ethanol consumption.20, 39 We examined the effect of ethanol consumption on liver sections of wild-type and iNOS−/− mice immunohistochemically for the presence of nitrotyrosine. In agreement with the literature on rats,20 ethanol consumption in wild-type mice resulted in enhanced formation of 3-nitrotyrosine protein adducts (Fig. 2C) compared to their pair-fed controls (Fig. 2A). The immunoreactivity was more pronounced in the perivenular hepatocytes. In contrast, iNOS−/− mice fed ethanol did not show positive staining for nitrotyrosine (Fig. 2D). Consistent with earlier reported observations, ethanol consumption enhanced iNOS expression in the centrilobular region in liver in wild-type mice (Fig. 3B). However, as expected, there was essentially background staining of iNOS protein in wild-type mice on control diet and iNOS−/− mice with or without the ethanol diet (Fig. 3).

Figure 2.

Nitrotyrosine staining in liver from wild-type and iNOS−/− mice fed control and ethanol diets. Immunohistochemical staining of liver sections from individual mice were performed using a polyclonal 3-nitrotyrosine antibody and counterstained with hematoxylin. No increase in staining was found in (A) wild-type on control diet, (B) iNOS−/− mice on control diet, and (D) iNOS−/− mice on ethanol-containing diet. (C) An intense staining for 3-nitrotyrosine can be seen around the centrilobular region of the liver (arrows) of wild-type mice fed ethanol.

Figure 3.

iNOS staining in liver from wild-type and iNOS−/− mice fed control and ethanol diets. Immunofluorescence analysis of iNOS in liver specimens obtained from (A) wild-type mice on control diet, (C) iNOS−/− on control diet, and (D) iNOS−/− on ethanol-containing diet showed very minimal staining of iNOS protein. A dramatic increase in iNOS staining, particularly in the centrilobular area, was evident in (B) wild-type mice on ethanol containing diet. (E) Quantitation of the immunofluorescence is shown. WT, wild-type.

Changes in Mitochondrial Function in Response to Ethanol Consumption.

Mitochondrial dysfunction is an early characteristic of ethanol-dependent hepatotoxicity. The yield of total mitochondrial protein was not changed when normalized to the wet weight of livers from either wild-type or iNOS−/− mice in response to ethanol consumption. In addition, the specific activity of citrate synthase, a mitochondrial enzyme resistant to oxidative modification, did not show any significant difference between mitochondria from control and ethanol-fed mice (result not shown). In the first series of experiments, oxygen consumption in the presence of substrates and adenosine diphosphate (ADP) was measured in mitochondria isolated from ethanol-fed mice and their pair-fed controls. This maximally stimulated ADP-dependent respiration is known as state 3 and is the condition under which NO-dependent inhibition is most pronounced.40 In contrast to published studies in the rat,13, 15 the mice did not show the same decrease in activity of components of the electron transport chain. Respiration in the absence of ADP is known as state 4 and represents a quiescent state for the organelle, although increased ADP-independent respiration may also indicate mitochondrial dysfunction via uncoupling. Although a statistically significant increase in state-4 respiration occurred in both groups of mice consuming ethanol with either succinate or glutamate-malate as substrates, there was no change in state-3 respiration (Table 2). Due to increased state-4 respiration, the respiratory control ratio was lower in mice fed ethanol chronically. There was no significant difference in state-4 or state-3 respiration between wild-type and iNOS−/− mice fed control diets.

Table 2. Mitochondrial Respiration in Response to Ethanol Feeding
  • NOTE. Data represent mean ± SEM for 6 pairs of wild-type mice on ethanol-containing diet (WT ETOH) or control diet (WT) and iNOS−/− mice on ethanol-containing diet (iNOS−/− ETOH) or control diet (iNOS−/−).

  • *

    P < .001 for corresponding group of mice on control diet.

State 4 (succinate)0.121 ± .0100.212 ± .018*0.123 ± .0110.227 ± .018*
State 3 (succinate)0.514 ± .0460.672 ± .0450.582 ± .0280.740 ± .039
State 4 (glutamate-malate)0.082 ± .0270.139 ± .031*0.052 ± .0050.103 ± .007*
State 3 (glutamate-malate)0.434 ± .027.416 ± .055.391 ± .0250.495 ± .015
RCR (succinate)4.28 ± 0.383.342 ± 0.464.86 ± 0.393.38 ± 0.32
RCR (glutamate-malate)8.17 ± 1.854.04 ± 1.17*7.82 ± 0.715.15 ± 0.47*

In an earlier study, we observed that rats that consumed ethanol had increased sensitivity to NO-mediated inhibition of mitochondrial respiration.13 Based on this finding, we hypothesized that this change in sensitivity could be due to increased expression of iNOS. To test this, the effect of ethanol consumption on the inhibition of mitochondrial respiration by NO was determined. Oxygen uptake and NO release from an NO donor (PAPANONOate) were monitored simultaneously in mitochondria isolated from animals on ethanol and control diets. Figures 4A and B show a typical respiration profile of mitochondria isolated from ethanol-fed mice and their pair-fed controls. When state-3 respiration was initiated by the addition of succinate and ADP, an increase in oxygen uptake was evident. A constant rate of oxygen consumption was observed until the samples became anoxic (result not shown). Upon addition of the NO donor, a slow release of NO was detected that resulted in a progressive inhibition of mitochondrial respiration. To assess the extent of change in the sensitivity of mitochondrial respiration to inhibition by NO in mitochondria isolated from control and ethanol fed-mice, state-3 respiration was determined as a function of NO concentration (Fig. 4C). In agreement with earlier studies, at all concentrations of NO there was a significantly greater inhibition of respiration in mitochondria isolated from wild-type ethanol-fed mice than those from corresponding control-fed animals. However, iNOS−/− mice fed ethanol did not show this enhanced sensitivity to NO-dependent inhibition of respiration. To test for a possible relationship between the total pathology score and the altered sensitivity of the mitochondria to inhibition by NO, these parameters for individual animals in the wild- type group were plotted against each other. The resulting relationship showed a significant negative correlation (r2 = 0.819, n = 6, P < .003), such that the greater the degree of inflammation the greater the sensitivity to NO-dependent inhibition of respiration. This was accompanied by a 2-fold increase in expression of TNF-α in wild- type mice fed ethanol; however, levels of TNF-α in iNOS−/− mice on ethanol diet were comparable to that of animals on control diet (Table 1).

Figure 4.

Effect of ethanol consumption on NO-mediated inhibition of respiration in mitochondria isolated from wild-type and iNOS−/− mice. Mitochondrial respiration and NO production were monitored simultaneously in mitochondria (0.5mg/mL) isolated from the liver of wild-type and iNOS−/− mice on ethanol-containing diet and their pair-fed controls in the presence of succinate (15 mmol) and ADP (0.5 mmol). PAPANONOate (5.0 μmol) was added to the chamber as indicated by the double-headed arrow. Representative traces for respiration for (A) wild- type control and (B) ethanol-fed micein the presence of PAPANONOate, and the corresponding trace for NO formation (dotted line) are shown. (C) Quantification of the data obtained with succinate and ADP, in which the measured respiration rate is plotted against NO concentration. *Significantly different from control (P < .05, mean ± SEM, n = 6).

It is known that the mitochondrial inner membrane consumes NO, and in the presence of mitochondria the rate of NO consumption was far more rapid than in buffer alone (Fig. 5A).5 This is due to NO consumption in the lipid phase of the mitochondrial inner membrane and plays an important role in the regulation of respiration.2, 8 Moreover, there was no statistical difference in NO consumption by mitochondria isolated from wild- type mice and iNOS−/− mice fed ethanol or their corresponding pair-fed controls (Figs. 5A and B). Inhibition of respiration was reversible because the addition of oxyhemoglobin (Hb) decreased NO levels and restored the rate of respiration to the level prior to NO addition in all groups (Fig. 5C).

Figure 5.

NO consumption in mitochondria from control and ethanol-fed mice. Mitochondria were incubated in the chamber with succinate (15 mmol). PAPANONOate (5.0 μmol) was added to the chamber and NO production measured. (A) Representative traces for NO production by PAPANONOate in the presence of buffer alone, mitochondria isolated from wild-type mice, and iNOS−/− mice on ethanol-containing diet, and their respective controls. (B) Quantitation of the same. (C) Reversibility of NO-dependent inhibition of mitochondrial respiration in mitochondria from control and ethanol-treated animals by the addition of oxyhemoglobin (Hb). Succinate (15 mmol) and ADP (0.5 mmol) were added to mitochondria (0.5 mg/mL) to initiate state-3 respiration, and PAPANONOate (5.0μmol) was added to the chamber at 80% O2. Oxyhemoglobin (5.0 μmol) was added to scavenge NO once inhibition of respiration was observed. Respiration and NO concentration were measured simultaneously. Respiration rate, as a percent of the initial value, before and after the addition of 5.0 μmol PAPANONOate, and after the addition of oxyhemoglobin (5.0 μmol) is shown. Data are expressed as the average of 6 pairs of control and ethanol-fed mice ± SEM. *P < .05 compared to mice on ethanol-fed diet and their respective controls.


The development of mitochondrial dysfunction is a characteristic of chronic ethanol consumption in humans and experimental animal models.41, 42 In rats, the later stages of alcohol-dependent hepatotoxicity are associated with a decrease in the activity of the major proteins of oxidative phosphorylation.18, 43 The effects of alcohol on liver mitochondrial function in mice have not been reported previously. In contrast to ethanol feeding in the rat, no significant decrease in the activity of respiratory complexes was observed in the mice. This may be due to a number of factors, including the species difference, the use of a shorter time course for ethanol exposure, and a lower dose of ethanol (28.8% of total calories) in the diet. In the present studies, a decrease in respiratory control ratio was observed; this can be ascribed to an increased rate of respiration in state 4.

The role of NO in controlling cell signaling has now been extended to the modulation of mitochondrial function.2, 3 The NO–cytochrome-c oxidase signaling pathway in which NO binds reversibly to the binuclear oxygen-binding site of the enzyme results in the control of respiration.1–5 We and others have characterized this pathway in detail and demonstrated that it can prevent release of cytochrome c when the mitochondrion is exposed to pro-apoptotic stimuli, differentially controls respiration in state 3, and has the potential to control the formation of ROS from the organelle.2, 8, 9 In addition, it has recently been shown that mitochondrial biogenesis is controlled through the activation of the soluble guanylate cyclase-cGMP signaling pathway.3 However, little is known of the chronic effects of NO exposure on mitochondria regulation of respiration and how this may change in response to a pathological or toxicological stress.

In our recent study,13 we found that the mitochondria isolated from ethanol-fed rats are much more sensitive to NO-dependent inhibition of respiration. It is likely this response results from the interaction of NO with a site other than cytochrome-c oxidase, or chronic alcohol exposure results in redistribution of the control of each individual enzyme over respiration. We hypothesize that the centrilobular hypoxia associated with chronic alcohol consumption has a contribution from this enhanced sensitivity of mitochondrial respiration to NO. In this model it is likely that the NO-dependent contribution to hepatotoxicity is exacerbated by increased formation of NO from iNOS.34 Taken together, these concepts provide a mechanism for the NO-dependent amplification of hepatotoxicity and mitochondrial dysfunction.

Using knockout and transgenic animals, a number of contributory mechanisms to inflammation, and ethanol-dependent hepatoxicity in particular, have been established.21, 24–26 The iNOS−/− model has been well characterized and has revealed a complex role for NO in the inflammatory response.24–26 The iNOS−/− is unable to mount an effective response to pathogens such as Listeria and shows variable responses to lipopolysaccharide (LPS) challenge but otherwise portrays no overt phenotype.19 As with all genetic models, the interpretation of the data must be constrained by the complexity of a biological response such as inflammation. For example, direct effects of NO on the mitochondrion are difficult to distinguish from an indirect effect of removing iNOS that is upstream of another independent mediator of mitochondrial dysfunction. Nevertheless, several lines of evidence suggest a direct interaction of the NO derived from iNOS in contributing to mitochondrial dysfunction. These are (1) the presence of multiple targets for NO and other RNS in the mitochondrion12; (2) the ability of RNS to elicit the posttranslational modification of mitochondrial proteins27–30; and (3) the established mitochondrial defect in ethanol-dependent hepatoxicity leading to increased sensitivity to NO-dependent inhibition of respiration.13

In the present study, we have tested this hypothesis in wild-type and iNOS−/− mice. The expected pathological changes in response to ethanol were observed in the wild- type mice: increased inflammation, steatosis, and nitrotyrosine staining. A major role for NO derived from iNOS in the development of these pathological changes is supported by the finding that these markers of hepatic dysfunction were markedly decreased in the iNOS−/− animals fed ethanol. As in our previous studies with a rat model of chronic alcohol consumption, it is clear that ethanol exposure results in much greater sensitivity to inhibition of respiration by NO. This response is not due to increased consumption of NO by the mitochondrial inner membrane in the organelle isolated from ethanol-treated animals or to a major change in the activity of the respiratory chain components. In the iNOS−/− animals, no effects of deficient iNOS on NO-dependent control of respiration was evident in the animals on the control diet. Moreover, the alcohol-mediated increase in the sensitivity of mitochondrial respiration to inhibition by NO was not observed in iNOS−/− mice. Several possibilities could explain these results. The increased expression of iNOS and generation of superoxide are induced by hypoxia and ethanol-dependent hepatotoxicity.19, 20, 34, 44 In addition, inflammation and ethanol hepatotoxicity have been associated with the nitration of tyrosine, which is a reaction not mediated by NO directly.39 The mechanism for nitration may involve either the reaction of NO with superoxide to form peroxynitrite or the reaction of nitrite with HOCl catalyzed by the enzyme myeloperoxidase.39 In the context of ethanol toxicity, either mechanism is plausible, and a role for both iNOS and superoxide derived from NADPH oxidase has been proposed.44 The substantial decrease in protein nitration observed in the iNOS−/− mice exposed to ethanol does however suggest that the majority of the NO leading to nitration is derived from this isoform of NOS.

In summary, our data suggest that the induction of iNOS makes a major contribution to the development of ethanol-dependent hepatic dysfunction. The finding that increased sensitivity to NO-dependent inhibition of respiration is iNOS-dependent suggests a highly dynamic interaction between fundamental aspects of mitochondrial function and the response to stress that has not been previously appreciated. It is clear that NO derived from iNOS or RNS formed in response to inflammation can mediate their effects through a number of mechanisms. For example, iNOS has been shown to increase the expression of adhesion molecules and orchestrate the transcriptional regulation of cytokines through its effects on NF-κB.25 For this reason, we cannot determine from this data whether the effects on mitochondrial function derive directly from NO produced by iNOS, from other RNS, or indeed from a downstream regulatory process mediated by other cytokines. Nevertheless, it is clear that in this early model of hepatic dysfunction, removing the effects of iNOS has a profound and novel impact on the acute sensitivity of the respiratory chain to NO. It is tempting to speculate that novel therapeutic pathways directed at the signaling pathways leading to increased NO formation and the control of mitochondrial respiration may be beneficial in ameliorating pathologies in which hypoxia is a major contributor to cellular dysfunction.