In liver cirrhosis, elevated levels of NO and ROS (reactive oxygen species) might greatly favour the generation of peroxynitrite. Peroxynitrite is a highly reactive oxidant and it can potentially alter the vascular reactivity and the function of different organs. In the present study, we evaluated whether peroxynitrite levels are related to the progression of renal vascular and excretory dysfunction during experimental cirrhosis induced by chronic BDL (bile-duct ligation) in rats. Experiments were performed at 7, 15 and 21 days after BDL in rats and in rats 21 days post-BDL chronically treated with L-NAME (NG-nitro-L-arginine methyl ester). Sodium balance, BP (blood pressure), basal RPP (renal perfusion pressure) and the renal vascular response to PHE (phenylephrine) and ACh (acetylcholine) in isolated perfused kidneys were measured. NO levels were calculated as 24-h urinary excretion of nitrites, ROS as TBARS (thiobarbituric acid-reacting substances), and peroxynitrite formation as the renal expression of nitrotyrosine. BDL rats had progressive sodium retention, and decreased BP, RPP and renal vascular responses to PHE and ACh in the time following BDL. They also had increasing levels of NO and ROS, and renal nitrotyrosine accumulation, especially in the medulla. All of these changes were either prevented or significantly decreased by chronic L-NAME administration. In conclusion, these results suggest that the increasing levels of peroxynitrite might contribute to the altered renal vascular response and sodium retention in the development of the experimental biliary cirrhosis. Moreover, the beneficial effects of decreasing NO synthesis are, at least in part, mediated by anti-peroxinitrite-related effects.
Increased NO production contributes to the renal excretory and vascular alterations observed in experimental and human cirrhosis [1–6]. In fact, contrary to what happens under normal conditions, acute or chronic inhibition of NO synthesis in cirrhosis increases sodium excretion without changing BP (blood pressure) [7–10]. In addition, the cirrhotic kidney has a decreased excretory response to manoeuvres such as increased arterial pressure  or volume expansion , which normally induce NO-dependent renal vasodilation and increase sodium and water excretion. The mechanisms that account for this paradoxical renal effect of NO during cirrhosis still remain unclear.
Previous studies have also suggested that an increased generation of ROS (reactive oxygen species) participates in the systemic and renal functional alterations of experimental cirrhosis [13–18], and that antioxidant administration, which quenches O2− (superoxide anion), improves these alterations [19–23]. During cirrhosis, increased levels of NO and O2− could easily react to form peroxynitrite (ONOO−), a RNOS (reactive nitroxy species). This reaction was initially viewed as a route for NO inactivation and it is at least three times faster than the dismutation of superoxide dismutase [24,25]. Peroxynitrite is a highly reactive intermediate and one of the most potent oxidants known in biological systems. The mechanism of injury caused by peroxynitrite involves multiple factors, including initiation of lipid peroxidation, increasing ROS and RNOS further, and nitration of tyrosine-containing proteins (nitrotyrosine), a footprint left by peroxynitrite in vivo . In addition, peroxynitrite is itself a vasodilator and can induce tachyphylaxis, preventing a further response to its own vasodilator actions. It also causes long-lasting impairment of the vasoactive response to other vasodilators  and even catecholamines .
Therefore we hypothesized that during cirrhosis the progressive generation of peroxynitrite could perpetuate a high oxidative stress status and participate in the advance of renal vascular and excretory abnormalities of liver cirrhosis. Thus, in the present study, we investigated renal excretory function and vascular reactivity in isolated kidneys, and measured the levels of NO and ROS, as well as peroxynitrite kidney levels, during different phases of cirrhosis induced by chronic BDL (bile-duct ligation) in rats. Furthermore, we examined the effect of blocking NO synthesis with L-NAME (NG-nitro-L-arginine methyl ester) on peroxynitrite formation and renal alterations in this experimental model of cirrhosis.
MATERIALS AND METHODS
Animals and models
Male Sprague–Dawley rats (225–250 g) born and raised in the animal facilities of the Universidad de Murcia (Murcia, Spain) were used in all of the experimental protocols, according to the Spanish Ministerio de Agricultura, Pesca y Alimentación and the European Community guidelines for the use of experimental animals, and the study was approved by the Ethics Committee of the Universidad de Murcia.
Liver cirrhosis was induced by chronic BDL as described previously [28,29]. Briefly, rats were subjected to either BDL or sham surgery. Rats were anaesthetized with an intramuscular injection of a mixture of xylazine (50 mg/kg of body weight; Rompun®; Bayer) and ketamine (100 mg/kg of body weight; Imalgène; Merial). The bile duct was then exposed under aseptic conditions by a midline incision under the sternum, doubly ligated with a 2-0 silk suture and excised between the ligatures. For sham or control surgery, the bile duct was exposed and dissected, but not ligated or cut. All rats were maintained under comparable conditions with an ad libitum diet and free access to drinking water.
Rats were randomized into six groups: (i) control (rats without any surgical intervention; n=6); (ii) sham (rats studied 7 days after sham surgery; n=6); (iii) BDL7 (rats studied 7 days after BDL; n=7); (iv) BDL15 (rats studied at 15 days after BDL; n=8); (v) BDL21 (rats studied at 21 days after BDL; n=6); and (vi) BDL21+L-NAME [rats received chronic L-NAME (10 mg·kg−1 of body weight·day−1) in the drinking water 7 days before the BDL surgery and were then studied after 21 days of BDL; n=6]. The sham group was considered the control for the BDL7 and BDL15 groups to evaluate any effect of the surgery itself; and the control group (healthy rats) served as the control for the BDL21 and BDL21+L-NAME groups. During the time course study, rats were progressively acclimatized to individual metabolic cages, which provide a good separation of the urine and faeces. After 2 days of adaptation, urine samples were collected over 2 consecutive days (days −2 and −1) before surgery (day 0) and for 2 days before the experimental day (days 6 and 7, 14 and 15, or 20 and 21 depending on the experimental group). Urine samples were centrifuged to remove solid matter before analysis, and the values of urine volume obtained on each experimental day were averaged for statistical purposes.
Preparation of isolated perfused kidneys and functional procedures
The rats were anaesthetized and placed on a heated table to maintain a body temperature of 37 °C. A polyethylene cannula (PE-50) was placed in the right femoral artery to measure BP and to obtain plasma samples. Subsequently, the kidney was isolated and perfused as described previously [30,31]. Briefly, the left kidney was exposed by a midline laparotomy and the renal artery was cannulated via the suprarenal aorta to prevent or minimize the interruption of blood flow. The kidney was perfused in situ with warm oxygenated Krebs buffer (118 mmol/l NaCl, 4.7 mmol/l KCl, 2.5 mmol/l CaCl2, 1.2 mmol/l MgSO4, 25 mmol/l NaHCO3, 1.2 mmol/l KH2PO4, 0.026 mmol/l EDTA and 5.6 mmol/l glucose, pH 7.4) at 37 °C and at a constant rate of 5 ml·min−1·g−1 of kidney weight with a peristaltic pump (Master-flex 7518-00; Coler-Parmer Instrument Company). The left renal vein was then cut and the ureter transected to allow the exit of the perfusate. Finally, the kidney was excised from the surrounding tissues, decapsulated and placed in a chamber containing Krebs solution at 37 °C. Renal vascular responses were recorded through a transducer connected to a Macintosh LCII computer and were analysed with MacLab software (AD Instruments) as changes in RPP (renal perfusion pressure) downstream from the pump. The right kidney was then removed and weighed (as an index of the left kidney weight). Samples of kidney and liver tissue were either frozen or fixed with a 10% formalin solution for later studies.
Renal responses to PHE (phenylephrine) and ACh (acetylcholine)
After a stabilization period of 30 min, a basal measurement of RPP was taken and a dose–response curve to increasing doses of PHE (10−7, 5×10−7, 10−6, 5×10−6, 10−6, 5×10−5, 10−5, 5×10−4 and 10−4 mol/l; administered as a 0.1 ml bolus) was determined. The approximate time between doses was 2 min, the estimated time to reach the plateau with each dose. After obtaining the PHE dose–response curve and a second stabilization period of 30 min, the renal vasculature was pre-constricted with Krebs solution containing PHE (10−6–10−4 mol/l) to reach 75% of the maximal constriction observed in the previous PHE curve. Once a stable elevated perfusion pressure was reached, the vasodilator responses to ACh (10−8–10−4 mol/l) and SNP (sodium nitroprusside; 10−6–10−4 mol/l), in this same order, were determined in all of the groups to assess the NO endothelium-dependent or -independent vasodilator responses respectively. Changes in RPP in response to vasodilators are expressed as a percentage of the vasoconstriction obtained with PHE.
In a subgroup of four rats from the control and BDL21+L-NAME (acute) groups, the experiments were performed in the presence of a high dose of L-NAME (10−4 mol/l) to inhibit NO synthesis. L-NAME was added to the Krebs solution 30 min before the initiation of the PHE dose–response curve and was maintained during the rest of the experiment.
TBARS (thiobarbituric acid-reacting substances) were determined in plasma and kidney tissue as reported previously . Briefly, 0.5 ml of PBS was added to 25 μl of plasma sample or 50 μl of kidney lysate (between 1 and 3 mg of protein). After mixing, 1 ml of reagent solution [1 mmol/l deferoxamine mesylate, 7.5% (w/v) trichloroacetic acid, 0.25 mol/l HCl and 0.37% thiobarbituric acid] was added and the mixture was vortex-mixed, covered with aluminium foil and heated for 45 min in boiling water. TBARS from standards (prepared from 1,1,3,3-tetraethoxypropane), samples and lysates were extracted into 1 ml of 1-butanol. The colour layer was read at A532 in a spectrophotometer (Varian 634) after a vigorous vortex-mixing and a brief centrifugation (1000 g for 5 min). The values are expressed as μg of TBARS/ml of plasma or μg of TBARS/mg of kidney protein.
The urinary excretion of nitrites was determined using the Griess reaction . Briefly, samples (100 μl) were mixed with 50 μl of 1% sulfanilamide in 5% potassium phosphate. Then, 50 μl of 0.1% N-(1-naphthyl) ethyl-enediamine dihydrochloride was added and incubated for 15 min. The nitrite concentration was quantified in a spectrophotometer at A540 against standards and after subtracting a blank from each individual sample. The final concentration is expressed in μg/day.
Sodium balance was assessed as the difference between sodium intake and excretion, and factored by body weight. The sodium intake (mmol/day) was obtained by multiplying the consumption of food per day (g/day) by the sodium content of the diet (0.104 mmol/g). The sodium concentration was measured by a sodium electrode (Orion Research), and urinary sodium excretion (mmol/day) was determined as the product of sodium concentration and 24-h urinary volume (ml/day).
For nitrotyrosine immunohistochemistry, a peroxidase–antiperoxidase technique was used. Briefly, kidney sections were rinsed and endogenous peroxidase was inhibited. Sections were blocked in TBS (Tris-buffered saline)/0.1% Tween 20, 10% FCS (fetal calf serum) and 5% (w/v) BSA for 30 min at room temperature (25 °C) and was incubated with a rabbit anti-nitrotyrosine polyclonal antibody (1:100 dilution; Cayman Chemical) overnight at 4 °C. After incubation, the sections were washed three times in TBS/0.1% Tween 20 and were incubated with secondary antibody [biotin-labelled anti-(rabbit IgG); 1:250 dilution; Vector Labs] for 30 min at room temperature. The signal was amplified with an ABC Kit (Vector Labs). The sections were then washed twice in TBS/0.1% Tween 20, once with TB (TBS without NaCl), rinsed, developed with 3,3′-diaminobenzidine (0.5 mg/ml) in 50 mmol/l TB (pH 7.6), counterstained with haematoxylin, dehydrated and mounted.
Mean nitrotyrosine-labelled protein expression was evaluated by assigning subjective values to every single specimen after screening six fields (magnification, 40×) of three different kidney regions (glomerular, outer and inner medulla) for every experimental animal as follows: 0, no expression; 1, low expression (only cytoplasmic accumulation of nitrotyrosine-labelled proteins in certain cells; 25%); 2, medium expression (medium expression of nitrotyrosine-labelled protein-containing cells in the field; 25–50%) with some cells with intracellular nitrotyrosine-labelled protein accumulation; 3, medium-to-high expression (50–75%); 4, high expression (most cells containing clustered-like accumulation of nitrotyrosine-labelled proteins throughout the whole field (>75%), and 5, most cells containing large intracellular clustered-like accumulation of nitrotyrosine-labelled proteins. All of the immunohistological studies were carried out by an expert pathologist blinded to the different experimental groups.
Liver tissue samples were fixed in 10% buffered formaldehyde, and then processed and embedded in paraffin and sectioned (approx. 5 μm), as reported previously . The sections were then stained with H&E (haematoxylin and eosin), and viewed by light microscopy. Morphological evaluation on randomized sections of the liver tissue was performed by a pathologist blinded to the different experimental groups.
All the drugs were purchased from Sigma, except where indicated. Stock solutions of PHE, ACh and SNP were dissolved in distilled water and maintained frozen. Working solutions were freshly prepared in Krebs solution and the concentrations expressed are final concentrations.
Results are means±S.E.M. Differences between groups were compared by one-way repeated measures ANOVA. For the dose–response curves, a two-way repeated measures ANOVA was applied and, when significantly different, means were compared further by the Student–Neuman–Keuls' test. The ED50 (concentration of agonist producing half of its maximal response) is derived from log-transformation and regression analysis of each individual concentration–response curve. In the immunohistochemical study, an unpaired Student's t test was used for the statistical purposes. A P value <0.05 was considered as statistically significant.
BDL-induced experimental cirrhosis produced a time-dependent increase in inflammatory parameters in the liver. By 7 days after BDL, a mild inflammatory infiltrate was widespread throughout the liver parenchyma [high numbers of PMNs (polymorphonuclear cells) and lower number of MNs (mononuclear cells)]. At this time, hepatocyte degeneration was minimal. This infiltrate became larger and clusters of PMNs and MNs (mainly monocytes) became evident at the periportal levels by 15 days after BDL. Hepatocytes started to become eosinophilic and to accumulate biliary pigments. Although, signs of hepatocyte vacuolar degeneration were present, cell apoptosis was not that frequent. Fibrosis relieved inflammation by 21 days after BDL. Therefore, although the diffuse inflammatory infiltrate was still evident and almost granulomatous formations were present at the portal level, especially surrounding biliary conducts, larger areas of the hepatic parenchyma had been substituted for fibrotic tissue and the typical hepatic sinusoidal structure had been almost completely disrupted. Hepatocytes became very eosinophilic, as a sign of cytoplasmic degeneration, and contained larger amounts of biliary pigments within their cytoplasms. At this stage, hepatocyte vacuolar degeneration and apoptosis appeared throughout the whole parenchyma.
However, there was almost no evidence of fibrosis by 7 days after BDL. Accumulation of fibrocytes and collagen fibres were almost exclusively restricted to the environment near the bile ducts. By 15 days, fibrosis had expanded to the surrounding areas of the hepatic lobule. By 21 days after BDL, fibrosis was widespread, even disrupting the normal architecture of most hepatic lobules (Figure 1).
Representative photomicrographs of liver sections stained with H&E from control, BDL21 and BDL21+
L-NAME administration partly attenuated liver cirrhotic changes by 21 days after BDL (Figure 1). Inflammatory infiltrates tended to be more diffuse, although some clusters of cells were still formed. Hepatocytes were damaged and had signs of cytoplasm degeneration and increased eosinophilia, but did not suffer as much of an insult as those in livers from non-treated BDL rats. Therefore these cells had less features of cell degeneration and apoptosis. Finally, fibrosis, although still present, was more restricted to periportal areas, rather than spread throughout the whole parenchyma.
As shown in Table 1, BDL rats had lower body weight and haematocrit values than the control group. Treatment with L-NAME normalized body weight and increased haematocrit. In contrast, all BDL rats had higher spleen weights and kidney-to-body weight ratios than the control group.
|Group .||Body weight (g) .||Spleen weight (g) .||Kidney-to-body weight (%) .||Haematocrit (%) .|
|Group .||Body weight (g) .||Spleen weight (g) .||Kidney-to-body weight (%) .||Haematocrit (%) .|
BDL rats had a progressive retention of sodium (Figure 2A) with advancement of the disease and decreases in BP (Figure 2B) and RPP (Figure 2C), with statistically significant effects being observed at 15 and 21 days after BDL. The BDL21+L-NAME group had similar values to the control group with regard to sodium balance, BP and RPP.
Sodium balance (A), BP (B) and basal RPP in the isolated kidneys (C) in the experimental groups
Renal vascular responses in the isolated and perfused kidneys
The administration of PHE induced a dose–dependent increase in RPP, which was significantly decreased after 15 (Figure 3A) and 21 (Figure 3B) days following BDL compared with the control group. A significant decrease in the maximum PHE response in the BDL15 and BDL21 groups was also observed. L-NAME administration (Figure 3B) normalized the dose–response curve and the maximum response to PHE in the BDL21 rats (BDL21+L-NAME group).
Renal vascular reactivity to PHE (A and B) and ACh (C and D) in the experimental groups
The dose–response curve and the maximum response to ACh were decreased in kidneys from untreated BDL rats in all of the experimental days (Figures 3C and 3D), whereas chronic L-NAME administration normalized the response in BDL21 rats (BDL21+L-NAME group).
The administration of acute L-NAME (10−4 mol/l) to kidneys from the control (results not shown) and BDL21+L-NAME (Figures 3B and 3D) groups did not modify the response to PHE with respect to that observed after chronic L-NAME administration (maximum response, 241±17 and 243±10 mmHg in the control and BDL21+L-NAME groups respectively), although the response to ACh was significantly attenuated in both experimental groups (maximum response, 18±4 and 12±3% in the control and BDL21+L-NAME groups respectively).
SNP induced similar responses in all experimental groups (results not shown).
Systemic and renal levels of ROS and NO
The plasma levels of TBARS were significantly increased in all of the BDL groups, which was significantly decreased in the BDL21+L-NAME group (Figure 4). On the other hand, TBARS in kidney tissue (Figure 4B) and the urinary excretion of nitrites (Figure 4C) increased progressively as the disease progressed. Both TBARS in kidney tissue and urinary excretion of nitrites were significantly decreased in the BDL21+L-NAME group (Figures 4B and 4C).
Levels of TBARS in plasma (A) and kidney (B), and 24-h urinary excretion of nitrites (C) in the experimental groups
Nitrotyrosine protein expression
Figures 5 and 6 show the expression of nitrotyrosine in the kidney samples from the experimental groups. Glomerular nitrotyrosine had a tendency to increase in the kidneys from the BDL groups, but no significant differences were observed between the groups (Figure 5A). Nevertheless, a significant accumulation of nitrotyrosine was observed in the renal outer and inner medulla with the progression of the disease (Figures 5B and 5C respectively), which was more pronounced in the inner medulla, where both vasa recta and tubules were affected from the earliest phase of BDL (Figure 6). The BDL21+L-NAME group had a similar expression of nitrotyrosine compared with the control group in all of the kidney regions (Figures 5 and 6). All comparisons for nitrotyrosine expression were performed with respect to the sham group, because the control group had ‘no expression’ and thus a value of zero in all of the kidney regions.
Semi-quantitative evaluation of renal nitrotyrosine in the glomerular (A), outer medulla (B) and inner medulla (C) regions in the experimental groups
Representative nitrotyrosine expression in the inner medulla of kidneys from the experimental groups
The main finding of the present study, reported for the first time, is that BDL induces a gradual increase in nitrotyrosine in the kidney, mainly in the renal medulla, and affects both vascular and tubular components of this region. This accumulation might contribute to the onset and progression of renal vascular and excretory derangement in this experimental model of liver cirrhosis, as L-NAME administration, which decreased NO synthesis and nitrotyrosine formation, also prevented the altered renal vascular responses and sodium retention in BDL rats.
Experimental chronic cholestasis in the rat progressively alters liver structure and function, leading to the development of portal hypertension and a hyperdynamic circulation, which is characterized by systemic vasodilation, high cardiac output and an altered response to vasoactive substances. As the disease advances, there is progressive sodium retention and eventual development of ascites, hepatorenal syndrome and death [5,6,28]. The time course of these events is variable depending on several factors, including the strain of rats, gender and even the laboratory [15,28,30,32–34]. The animals used in the present study, in the 4th week of BDL, had typical signs of cirrhosis, such as decreased growth, jaundice, coluria and haematocrit. After death, an abdominal inspection revealed a small amount of ascites (1–5 ml) in only two rats in the BDL group, but there was mesenteric oedema and enlarged liver and spleen (indirect evidence of portal hypertension) in all of them. Morphologically, the livers had marked ductal proliferation, fibrosis and signs of hepatocyte degeneration, with apoptosis appearing throughout the whole parenchyma. L-NAME treatment had a moderate effect on liver morphology, improving the inflammatory process, but not on portal hypertension, as suggested by a similar spleen weight. This is not entirely surprising as L-NAME therapy is not addressing the primary insult, which is the surgical cholestasis. Comparable results have been described in other studies, using this and other treatments, where a recovery in the systemic haemodynamic and renal changes was shown, although the improvement observed in liver function or portal pressure was small [20,22,23,35,36].
The incidence of renal dysfunction in patients and animals with obstructive jaundice is elevated in various clinical settings, such as bleeding or infection , as well as a particular predisposition to renal failure [13,35]. In the present study, rats had significant changes in renal parameters measured after 15 days of BDL. Thus we found a progressive sodium retention starting on day 15 after BDL, an increased spleen weight and a decrease in BP, basal RPP in the isolated kidneys and haematocrit. These results are similar to those obtained in our previous studies in the same model of liver disease , showing that the systemic haemodynamic and renal excretory parameters start to change between 12 and 14 days after the biliary occlusion. At this stage, the antinatriuretic neurohormonal systems are activated as a consequence of the decrease in BP, and sodium and water retention occurs through different vascular and tubular mechanisms along the different segments of the nephron [2,6].
The mechanisms leading to the early decrease in BP are probably due to an excess of vasodilator substances, mainly NO, as we found increasing urinary excretion of nitrites as early as 7 days after BDL. NO has been widely implicated in the decreased vascular response to endogenous vasoconstrictors, a generalized phenomenon observed in clinical and experimental cirrhosis, in the setting of elevated circulating levels of vasoconstrictors [29,31,33,37,38], although the mechanisms by which this happens are not completely elucidated. The present study also shows a decrease in renal response to PHE at 15 and 21 days after BDL, coexisting with an increase in urinary nitrite excretion. In this regard, it has been reported that serum from the cirrhotic rats progressively decreases the contractile responses to PHE in aortic rings of healthy animals, starting at the first week after BDL . This lower response was reversed by acute NO synthesis blockade, pointing to NO as the major factor for this vascular hyporesponse. In agreement with this, our present findings also demonstrate that chronic NOS (NO synthase) inhibition with L-NAME completely prevents the renal vascular hyporesponsiveness to PHE in BDL rats. It is possible that mechanisms derived from inhibition of systemic NO synthesis could contribute to the beneficial effects on renal vascular reactivity of BDL rats, such as the increase in BP and/or structural changes in the vessel wall. Indeed, it has been shown that NO-dependent vascular remodelling occurs during liver cirrhosis , affecting conductive vessels and contributing to the profound circulatory dysfunction developed by experimental animals with advanced liver disease. Because L-NAME treatment resulted in a similar increase in BP in control and cirrhotic rats, the reversal of the abnormal vascular remodelling in the cirrhotic animals was considered to be an effect of the inhibition of NO synthesis and not as a consequence of the increase in arterial pressure. In our present study, the BDL21+L-NAME group had similar arterial pressure compared with the control group. Nevertheless, the direct influence of BP in the recovery of renal vascular reactivity to PHE in those rats must be partial because (i) the kidneys from all of the groups were isolated from systemic influences; and (ii) different factors, other than blocking the excess of NO synthesis, appeared to be implicated as 24-h urinary excretion of nitrites was not completely normalized by chronic L-NAME treatment in our animals with BDL-induced cirrhosis.
Over the last few years, ROS have been the focus of many studies in diverse hepatic insults. Thus ROS levels (malondialdehyde, isoprostanes etc.) are elevated in different models of chronic liver pathologies [14,16,18] and the levels correlate with the severity of the disease [17,40]. In the present study, we detected a significant increase in ROS levels, as measured as TBARS (malondialdehyde), in plasma and kidney tissue from the earliest stages of BDL. In the kidney, the increase was progressive with the advance of the disease and was related to the haemodynamic and renal changes, whereas plasma TBARS remained elevated from the beginning of the disease. These results are in agreement with a previous study by Panozzo et al. , who found a correlation between the variation in hepatic and urinary TBARS, liver damage and functional glomerular and tubular renal alterations in rats following BDL, but not in plasma TBARS. It is possible that circulating TBARS represent a pool from several organs affected differently by cirrhosis; however, the administration of chronic L-NAME in the present study decreased levels of TBARS in plasma and kidney tissue, suggesting that, in some way, NO is involved in inducing ROS [24,25]. This is important as previous studies have shown that the increased generation of ROS contributes to altered renal vascular reactivity and participates in the systemic and renal functional alterations induced by BDL [19,22,23,30].
With respect to NO, under physiological conditions it is a vasodilator. Thus the progressive increase in NO production could explain the progressive decrease in the responses to PHE, as acute [30,31,33] or chronic inhibition of NO synthesis, as shown in the present study, reverses renal vascular hyporesponsiveness. However, this excess of NO should promote sodium and water excretion, which is opposite to what we found in the present study in BDL rats, which is a decrease in sodium excretion as BDL progress; and accordingly, the administration of NO inhibitors improved sodium and water excretion [7–10]. Additionally, the endothelium-dependent relaxation was lower than normal from the initial periods of BDL and chronic L-NAME fully restored the altered vasodilation in the BDL rats. This interesting finding suggests that this dose of L-NAME does not completely eliminate NO synthesis, which is supported by the fact that 24-h urinary excretion of nitrites was not completely normalized and that a high dose of acutely administered L-NAME almost abolished the ACh-dependent vasorelaxation in the BDL21 rats chronically treated with L-NAME. Finally, it is possible that the excess of NO, or some NO-derived substances, is harmful to endothelial and renal function. Regarding this, we have recently shown that chronic administration of an antioxidant (vitamin E) to BDL rats paralleled the effects of L-NAME on renal vascular reactivity to PHE and ACh, kidney TBARS and most of the parameters studied in the present study, with the exception of NO levels which were increased further . Taken together with our present study, these results suggest that during BDL the NO excess in the presence of elevated oxidative stress damages renal vascular and excretory function. Although different substances may potentially alter vascular reactivity in the kidney [15,41], in cirrhosis elevated levels of NO and O2− can greatly favour the formation of peroxynitrite. Peroxynitrite may alter the vasoactive responses to several substances in several ways. First, peroxynitrite formed endogenously induces vascular relaxation and this effect is subjected to rapid tachyphylaxis thus altering subsequent vasorelaxant responses. In fact, after the development of tachyphylaxis to peroxynitrite, the haemodynamic effects produced by the systemic administration of ACh (an endothelium-dependent vasorelaxant), but not to NO donors, were significantly attenuated . In addition, it produces a substantial and selective attenuation of the haemodynamic effects in vivo produced by α- and β-adrenoceptor agonists  and AngII (angiotensin II) . These effects appear to be due to nitration of extracellular tyrosine residues by peroxynitrite [27,42–44]. Despite its non-radical nature, peroxynitrite is more reactive than its parent molecules [24,25] and can alter intracellular signalling pathways involved in contractile vascular function [45–47]. In addition, it can promote the initiation of lipid peroxidation, yet again increasing oxidative-stress-derived products with the resultant consequences for vascular and renal function [13,48].
To date, only a few studies have reported the existence of increased nitrotyrosine in plasma, liver and heart during liver cirrhosis [49–51]. The present results now reveal, we believe for the first time, a progressive accumulation of nitrotyrosine in the kidneys of BDL rats during the advancement of the disease, preferentially in the endothelium of vasa recta and tubules of the inner medullary region, and more strikingly over 21 days after initiation of BDL. Thereby, these findings suggest that excessive NO, by way of its transformation into peroxynitrite, can be noxious for renal vascular and tubular function. Therefore preventing excess NO synthesis in BDL rats with L-NAME decreases the formation of peroxynitrite (and therefore nitrotyrosine) and reverses the renal functional alterations determined in the present study. According to this, it is tempting to speculate that the improvement observed in renal parameters after inhibiting NO synthesis in the present or previous studies might be also a consequence of an anti-peroxynitrite effect. For example, previous in vivo studies have demonstrated an inability of the cirrhotic kidney to adequately respond, increasing diuresis and natriuresis, to manoeuvres that vasodilate the kidney in an NO-dependent manner [11,12]. We have shown subsequently that administration of very low doses of L-NAME, having no noticeable systemic or renal haemodynamic consequences, restored the lowered diuresis and natriuresis of cirrhotic animals, and these effects were reversed by L-arginine . In addition, Mani et al.  have recently shown that decreasing the nitration of cardiac proteins with N-acetylcysteine and L-NAME led to the normalization of the in vitro cardiac chronotropic response of BDL animals. The results from our present study show that administration of a chronic dose of L-NAME, which avoids nitrotyrosine formation in the vasculature and tubules of kidneys from BDL rats, prevents the changes in renal vascular reactivity (which can alter peritubular haemodynamic and sodium handling) and returns sodium excretion to normal in BDL rats.
Limitations and unexplained findings
As mentioned in the Introduction, we intended to study some of the paradoxical effects of elevated NO production during liver cirrhosis. We consider that the present study might help to explain part of these effects, although some aspects require further investigation. One logical question that emerged concerns the existence of a decrease in the renal ACh response, in the context of a decreased response to PHE, due to increased NO production. In this respect, it is possible that the basal and stimulated production of NO in the renal vasculature of BDL animals is affected differently . Basal NO production might be increased, given that the lower response induced by PHE was reversed by L-NAME. In contrast, the agonist-stimulated ACh release of NO was affected in an opposite manner, suggesting a defective receptor or post-receptor activation of eNOS (endothelial NOS) or a down-regulation of this NO release mechanism. In this way, NO can be generated by iNOS (inducible NOS) from the endothelium and vascular smooth muscle cells, resulting in high NO output with potential damaging consequences, such as the impairment of eNOS-derived NO production observed in vessels treated with inflammatory mediators . Although both eNOS and iNOS appear to contribute to the high NO levels during BDL, the present study does not discriminate between the individual contributions of each isoform.
The local damaging effects of overproduced NO during BDL appear to be mediated by its reaction with ROS to form peroxynitrite, as we found that, when NO levels are decreased, the altered renal response to ACh and PHE is reversed. Therefore we consider that L-NAME inhibits peroxynitrite, as NO is an essential precursor of peroxynitrite [24,25] and a striking decrease in the BDL-induced increase in nitrotyrosine expression in renal tissue in BDL21+L-NAME rats was observed in the present study. Similar results have been obtained in BDL rats  and in rats with endotoxaemic shock (which causes comparable haemodynamic and vascular changes to those by BDL) using a peroxynitrite decomposition catalyst . In addition to the mechanisms described above, peroxynitrite and other ROS can alter agonist eNOS activity via caveolae or uncoupling-mediated mechanisms, resulting in less NO production [54,55]. Furthermore, ROS can affect iNOS expression which may increase NO output and later altering eNOS-derived NO production in vessels [45,53,55]. Thus elevated NO levels in the setting of high oxidative stress status are ineffective and engaged in a vicious circle where both NO and ROS are produced continuously.
In conclusion, the results of the present study suggest that the effects of peroxynitrite might account for the altered renal vascular responses and abnormal renal sodium excretion during chronic BDL and thus contribute to the progression of renal dysfunction observed during this liver injury. The beneficial effects of decreasing NO synthesis with L-NAME are two-fold: first avoiding peroxynitrite formation and subsequently decreasing the re-initiation of lipid peroxidation and thus oxidative stress.
- BDL7 etc.
7 days after BDL etc.
haematoxylin and eosin
NG-nitro-L-arginine methyl ester
reactive nitroxy species
reactive oxygen species
renal perfusion pressure
thiobarbituric acid-reacting substances
TBS without NaCl
This work was supported by grants from the Spanish Ministerio de Ciencia y Tecnología (SAF2003-07467), Fundación Séneca de Murcia (PB/45/FS/02), and Instituto de Salud Carlos III (RNIHG, CO3/02). A. A. is supported by a predoctoral fellowship grant from SAF2003-07467.