Potential conflict of interest: Nothing to report.
Cholestasis is associated with changes in NO metabolism and thiol oxidation. Thioredoxin contributes to regulate vascular tone and intracellular redox status by cleaving nitrosothiols and maintaining −SH groups. This study investigated the changes in circulating thioredoxin and nitrosothiols and the relationship with protein sulfhydryls (PSH), hepatic concentrations, hyaluronate, and histology in patients with primary biliary cirrhosis (PBC) and in rats with bile duct ligation (BDL). PSH in erythrocytes were significantly decreased in stage III and IV PBC and at day 10 after BDL. Compared with controls, erythrocyte thioredoxin levels were higher in stage I through III PBC and lower in stage IV patients. Serum thioredoxin levels were significantly higher in PBC stages I and II and lower in stages III and IV. Serum nitrosothiols were higher in all PBC patients and inversely related to thioredoxin and hyaluronate. In rats, serum, hepatic, and mitochondrial thioredoxin had initially increased after BDL (day 1-3) and then decreased. After day 7 BDL, nitrosothiols were 10-fold increased in serum and liver, and even higher in mitochondria. In the liver, thioredoxin was inversely related to both nitrosothiols and PSH. In rats, the difference in time average changes from baseline among serum, hepatic, and erythrocyte thioredoxin suggests that most of circulating thioredoxin originates from the liver. Conclusion: Our findings indicate that cholestasis is associated with significant mutual and interrelated changes between circulating and hepatic thioredoxin and nitrosothiols. The increase of hepatic, mitochondrial, and circulating nitrosothiols with ongoing cholestasis suggests an active participation of NO in both liver injury and extrahepatic changes. (HEPATOLOGY 2007;45:331–339.)
Although several observational studies have been performed in patients with primary biliary cirrhosis (PBC), why the disease evolution is faster in some patients and exhibits a lower pace in others remains unknown. Thus, early indicators of prognostic value are actively searched. However, no serum parameter has been proposed as predictor of evolution in this liver condition, and liver biopsy remains the best diagnostic tool for diagnosis and for providing prognostic information.
A number of biochemical abnormalities including toxic bile salt accumulation and oxidative mitochondrial alterations1 are supposed to be implicated in the onset and evolution of cholestatic liver damage. Changes in hepatic and extrahepatic membrane lipid composition and protein redox status are also major features of cholestatic liver disease and are associated with more progressive forms.1, 2 Under cholestasis, erythrocyte membrane alterations parallel those found in hepatocytes3 and include a decreased content of protein sulfhydryls (PSH), especially in more advanced stages.4 PSH maintenance is fundamental for cell membrane structure and function and confers resistance to damaging insults.5
Several intracellular and extracellular factors including nitric oxide (NO), a potential free radical intermediate, modulate cell membrane activity and cholestatic disease progression. In cholestatic conditions, hepatic NO is mainly produced by a hepatocyte-inducible NO synthase6, 7 as the result of enhanced inflow of gut-derived endotoxins to the liver.7, 8 The consequent excess generation of the hyperreactive NO• may thus explain the accumulation of nitrotyrosine in the liver of PBC patients and of bile duct–ligated (BDL) rats.9, 10 Nitrotyrosine accumulation, already at an early stage of PBC,11 likely reflects an increased peroxynitrite formation and suggests a NO-mediated liver injury.9, 10
Thioredoxin, a redox-active protein with several biological activities,12 including maintenance of protein −SH groups,13 regulation of redox-sensitive molecules,14 and membrane permeability protection,15 is induced by oxidative stress.14 Extracellular thioredoxin is increased in patients with a high rate of free radical generation16 and also exerts specific activities for the modulation of the immunological response.17 Thioredoxin is actively involved as well in the regulation of some NO activities through the cleavage of nitrosothiols.18, 19 Nitrosothiols are formed by conjugation of NO• with thiols and glutathione in particular, and thereby oppose dangerous side reactions such as peroxynitrite formation.12 Nitrosothiols also act as intracellular messengers and control cellular and mitochondrial functions20 through the nitrosylation of proteins and enzymes.21, 22 Conversely, circulating nitrosothiols mainly act as reserve for NO, whose release after molecule decomposition serves for the vascular tone modulation.23
Although serum thioredoxin levels discriminate more aggressively from rather benign forms of chronic non-cholestatic liver disease,24 and circulating NO derivatives are enhanced in cirrhotic conditions,25 little is known about the behavior of these parameters in conditions of chronic cholestasis, and whether their circulating levels are in relation with the hepatic concentrations and with the severity of liver injury.
Therefore, this study was designed to assess erythrocyte and serum thioredoxin and nitrosothiol levels in PBC patients at different disease stages and to relate circulating with hepatic and mitochondrial changes and with PSH levels in BDL rats.
BDL, bile duct ligation; NEDD, N-1-naphthylethylenediamine hydrochloride; NO, nitric oxide; PBC, primary biliary cirrhosis; PSH, protein sulfhydryls; SULF, sulfanilamide.
Patients and Methods
This study was carried out on 30 female patients with PBC (27-68 years old) and 16 age-matched healthy women. PBC patients were all admitted at our department (demographic and clinical characteristics are reported in Table 1). Diagnosis of PBC was made according to clinical, biochemical, and histological parameters. Liver disease was histologically staged according to Scheuer26 and patients subsequently divided (stage I: n = 5, stage II: n = 13, stage III: n = 8, stage IV: n = 4). Twenty-three patients showed serum positivity for anti-mitochondrial antibody, seven were also affected by systemic scleroderma and three by the Sjögren-like syndrome. All were negative for hepatitis virus infection and reported no history of alcohol abuse. None was under ursodeoxycholic acid, corticosteroid, or other immunosuppressive therapy at the time of enrollment. Written consent was obtained from all participants. From overnight fasted subjects, blood was collected in the morning and immediately centrifuged to separate erythrocytes from plasma.
Table 1. Characteristics of Patients with Primary Biliary Cirrhosis (PBC) and Healthy Controls
Data are means ± SD.
Significantly different compared with healthy control subjects.
Experiments with animals needed to investigate the intrahepatic changes occurring during cholestasis that, conversely, were not performable in the small specimens obtained by ultrasonography-guided liver biopsy in humans.
Male Wistar rats (body weight 250-280 g, Harlan, S. Pietro al Natisone, Italy), kept in individual cages under controlled conditions of temperature and humidity and on a 12-hour dark/light cycle, were subjected to double ligation and transection of the common bile duct under anesthesia (2 ml Ketamine plus 0.75 ml 2% Xylazine intramuscularly). Sham operations were performed by exposing but not ligating the common bile duct. Surgical procedures were carried out under sterile conditions. Sham-operated rats were pair-fed individually to BDL rats that had free access to standard rat chow and tap water. Overall, mortality was less than 15% in the BDL group and less than 10% in control animals. The protocol was conducted according to the Guide Principles for the Care and Use of Laboratory Animals.
After overnight fasting, five to nine rats per group were killed by decapitation at days 1, 3, 7, and 10 after surgery. Blood was collected and immediately centrifuged to separate erythrocytes from plasma. Livers were removed immediately and homogenized in ice-cold MSM buffer (440 M mannitol, 70 M sucrose, 5 mM 3-morpholinopropanesulfonic acid, pH 7.4) for mitochondria isolation.27 Liver homogenate and mitochondrial fractions were assayed for succinate dehydrogenase: the recovery in the mitochondrial pellet averaged 85% to 92%. General and liver parameters are reported in Table 2.
Table 2. Characteristics of Bile Duct Ligated (BDL) and Control Sham-operated (SHAM) Rats Before (Day 0) and After Surgery (Day 10)
NOTE. Data are mean ± SD.
Significantly different from baseline and control rats at the same time point.
Liver tissue and mitochondrial suspension were homogenized in 50 mM Tris-HCl buffer pH 7.5 containing 1 mM EDTA. Erythrocytes were lysed 1:10 with bi-distilled water and then treated as the other tissues. Serum samples were diluted 1:5 with saline. All biological materials were centrifuged at 70,000g for 15 minutes. The supernatant was first adjusted at pH 5.0 with ice-cold 1.0 M acetic acid and after centrifugation to pH 7.5 with 1.0 M ammonium hydroxide. Next, the supernatant was incubated at 60°C for 15 minutes and subsequently at 37°C for 15 minutes in the presence of 2 μL 2 mM dithiothreitol buffer (50 mM HEPES, 1 mM EDTA, 2 mM dithiothreitol, 1 mg/ml bovine serum albumin). A reaction mixture (40 μl) containing 200 μl 1.0 M HEPES buffer, 40 μl; 500 mM EDTA, 40 μl 48 mM NADPH, and 500 μl insulin was added to 70 μl biological sample and 10 μl water. The reaction was started by the addition of 10 μl mammalian thioredoxin reductase (Sigma T 9698). After incubation at 37°C for 20 minutes, the reaction was terminated by the addition of 500 μl 0.4 mg/ml dithio-nitrobenzoic acid/6 M guanidine hydrochloride in 0.2 M Tris-HCl (pH 8.0). The absorbance of samples treated and lacking thioredoxin reductase was measured at 412 nm and was compared with a standard curve of thioredoxin.28
The method described by Cook et al.,29 which uses a mixture of SULF/NEDD (sulfanilamide/N-1-naphthylethylenediamine hydrochloride) as reagents, was used. Plasma, erythrocytes, and liver fractions were suspended 1:4 in PBS (pH 7.4) containing 10 mM N-ethylmaleimide and 4 mM potassium ferricyanide, acidified with 25% sulfosalicylic acid and centrifuged at 10,000g for 10 minutes. The supernatant was added to 40 μl 1% ammonium sulfamate, 200 μl 0.4N HCl containing 0.3% HgCl2 and 4.6% sulfanilamide, and 300 μl 0.4 N HCl containing 0.2% NEDD. After 30 minutes of incubation at 25°C, the samples were spectrophotometrically analyzed at 544 nm. Standards were prepared by reacting equal molar reduced glutathione and nitrite in water.
A modification of Elmann's procedure was used.30 Sulfosalicylic acid precipitated proteins were resuspended in 700 μl of 6 M guanidine (pH 6.0); optical density was read spectrophotometrically at 412 and 530 nm before and after incubation (30 minutes) with 50 μl 10 mM 5,5-dithiobis 2-nitrobenzoic acid. Reduced glutathione was used to prepare a standard curve.
Nitrites Plus Nitrates (NOx).
Liver homogenates [1:5 with phosphate-buffered saline (PBS), pH 7.4] were added with NADPH, FAD, and nitrate reductase (50 and 5 μM, 0.160 U/ml), incubated at 37°C (20 minutes), and mixed with lactate dehydrogenase (10 mg/ml) and 10 mM sodium pyruvate. After incubation (37°C, 5 minutes), samples were deproteinized by adding 0.05 volumes of zinc sulfate (15 g/l). After centrifugation, the supernatant was added to 500 μl Griess reagent (2 g/l SULF, 50 g/l ortho-phosphoric acid, 0.2 g/l NEDD) and incubated (15 minutes, 25°C). Total nitrites were measured at 540 nm and compared with samples treated only with orthophosphoric acid.31
Quantitative serum hyaluronate levels, an indirect index of liver fibrosis, were determined by a standardized ELISA method (Corgenix Inc., Westminster, CO). Serum samples were incubated in coated microwells containing immobilized binding proteins; after unbound serum removal, horseradish peroxidase solution was added. After a washing step, tetramethylbenzidine and H2O2 were added to develop a colored reaction, whose optical density was measured at 450 nm. Hyaluronate concentrations were calculated by using a reference curve with known standard amounts. Normal range was defined between 0 and 50 μg/l.
Protein concentration was measured by using a Bio-Rad kit for the assay of proteins (Bio-Rad GmbH, Munich, Germany).
Unless otherwise indicated, all the materials used were purchased from Sigma-Aldrich Chemical Co. (Milan, Italy) or were of the highest purity grade commercially available.
All data are expressed as mean ± SD. The Mann-Whitney rank sum test was used to compare groups. For multiple comparisons, the ANOVA on ranks analysis of variance followed by Dunn's method was used. The nonparametric Spearman rank order correlation and the linear regression analysis were used to relate biochemical parameters. P less than 0.05 defined significance.
PSH in erythrocytes were significantly lower in PBC patients with histological stages II to IV, as compared with controls, and in patients at stages III and IV compared with stage I (Table 3). Erythrocyte thioredoxin concentrations were higher in stage I to III PBC than in healthy subjects (1.62 ± 0.29 versus 1.44 ± 0.17 nmol/g Hb, P = 0.02) but lower in stage IV patients (1.30 ± 0.34 nmol/g Hb). No significant differences were noted in erythrocyte nitrosothiols among patients and between patients and healthy controls.
Table 3. Concentrations of Protein Sulfhydryls (PSH) in Erythrocytes of Healthy Controls and of Patients with Primary Biliary Cirrhosis (PBC) Divided by Disease Stages, and in Erythrocytes, Liver Homogenate and Mitochondria of Control Sham-Operated (SHAM) and of Bile Duct–ligated (BDL) Rats at Days 1-10 After Surgery
PBC Stage I
PBC Stage II
PBC Stage III
PBC Stage IV
NOTE. Results are mean ± SD. PSH concentrations are reported as nmol/mg protein.
P < 0.05 versus control subjects and SHAM rats at baseline.
P < 0.05 versus stage I patients and BDL day 1 and day 3.
Serum thioredoxin levels were not significantly different between healthy subjects and PBC patients considered at all stages (1.90 ± 0.33 versus 1.91 ± 0.81 nM, P = NS). However, by dividing patients according to histology, the concentrations of thioredoxin resulted significantly different among PBC subgroups (Fig. 1A). In particular, thioredoxin was found to be significantly higher in patients at early stages (2.27 ± 0.39 and 2.55 ± 0.45 nM, stages I and II respectively, P < 0.05 and P < 0.001 versus controls) and significantly lower in patients with more advanced stages (1.29 ± 0.34 and 0.71 ± 0.36 nM, stages III and IV respectively, P < 0.001 versus controls). In PBC patients with stages III and IV, serum thioredoxin was inversely related with serum nitrosothiols (r = −0.609, P < 0.001; Fig. 2A) and hyaluronate (r = −0.432; P < 0.001) and with histology (r = −0.757; P < 0.001).
Serum nitrosothiols were significantly higher in PBC patients than in healthy controls (403.9 ± 284.3 versus 92.2 ± 50.8 nM) (Fig. 1B). The increase was directly proportional to the disease severity (r = 0.879, P = 0.002) and was linearly and significantly related with serum hyaluronate (r = 0.531, P < 0.001) (Fig. 2B).
Overall, serum hyaluronate was higher in PBC patients than in control subjects (106 ± 120 versus 22 ± 13 μg/l, P < 0.001), with the difference reaching the statistical significance in patients at stages II to IV (51 ± 35, 119 ± 72, and 374 ± 57 μg/l in patients with stages II, III, and IV, respectively). As expected, hyaluronate levels correlated well with histology (r = 0.775, P = 0.002).
A significant decrease of PSH concentrations in erythrocytes was observed at day 10 after BDL (Table 3). Also, significant decreases in liver and mitochondrial PSH were observed at days 7 and 10 of BDL. At day 10, mitochondrial PSH were significantly decreased also compared with days 1 to 3 of BDL (Table 3). No significant changes in erythrocyte thioredoxin and nitrosothiols were observed in BDL rats.
As compared with sham-operated rats (Fig. 3), a significant increase in serum thioredoxin levels was observed at day 3 of BDL, which was followed by a significant decrease. Similar changes were observed in liver tissue (Fig. 3) and in mitochondria (Table 4). No significant changes in thioredoxin concentrations were observed at erythrocyte level in these non-cirrhotic BDL rats; therefore, the difference in the time average changes from baseline among serum, hepatic, and erythrocyte thioredoxin levels suggests a hepatic origin for circulating thioredoxin.
Table 4. Mitochondrial Concentrations of Thioredoxin and Nitrosothiols in Control Sham-operated (SHAM) and in Bile Duct–ligated (BDL) Rats at Days 1-10 After Surgery
BDL Day 1
BDL Day 3
BDL Day 7
BDL Day 10
NOTE. Results are mean ± SD. Concentrations are given as nmol/mg protein.
Differently from erythrocytes, serum and hepatic nitrosothiol concentrations significantly and progressively increased with ongoing cholestasis, reaching more than 10-fold values at day 10 (Fig. 4). Mitochondrial concentrations of nitrosothiols had also increased after BDL even to a higher extent than that observed in the extra-mitochondrial compartment (Table 4). After BDL, NOx content was significantly increased in liver homogenate (478 ± 57 versus 204 ± 49 nmoles/g liver, P < 0.01, day 10 BDL and basaline values, respectively).
A significant increase of serum hyaluronate followed BDL: at day 10, hyaluronate levels were more than 20 times higher than baseline (Fig. 5A). A positive correlation (Fig. 5B) was observed between serum nitrosothiols and hyaluronate values (r = 0.862, P = 0.003).
So far, no early non-invasive prognostic indicator has been identified in PBC, and it is currently difficult to predict the rate of disease progression in these patients without performing repeated liver biopsies and histology.
By investigating redox active proteins and NO vehicle molecules, this integrated and translational study in animals and humans has shown that cholestasis results in early and significant interrelated changes between thioredoxin and nitrosothiols both at circulating level and in the liver. The close relation between thioredoxin and nitrosothiols levels with liver histology in PBC patients points to potential implication of these molecules in the progression of the disease itself. In particular, from our results it emerges that the circulating levels of thioredoxin are increased in patients with stage I to II PBC, indicating that thioredoxin is likely induced to counteract increased ongoing oxidative stress in PBC. This interpretation is in line with the increased levels of thioredoxin observed in other chronic inflammatory conditions.24 An additional hypothesis is that the higher levels of thioredoxin in PBC patients with earlier hepatic disease might be an adaptive measure to regulate NO metabolism, nitrosothiols decomposition, and to maintain PSH in the reduced form. Conversely, the decreased circulating level of thioredoxin in patients with more advanced stages of PBC likely reflects a diminished hepatic synthesis and delivery into vasculature. An effect of thioredoxin decrease in stage IV patients is the increased serum levels of nitrosothiols and the increased oxidation of erythrocyte PSH. As a consequence, our results strengthen the idea that circulating thioredoxin serves to protect surface PSH from oxidation and to regulate nitrosothiol levels through decomposition.
Although a specific pathway for protein denitrosation has not been identified yet, it is generally accepted that nitrosothiols can be decomposed either by enzymatic homolytic cleavage in a NADPH-dependent reaction involving the thioredoxin system and a redox-sensitive metalloprotein or via a nonenzymatic one-electron reduction.19, 32, 33 In support of a role for thioredoxin, one recent study has reported that thioredoxin-deficient cells denitrosate nitrosothiols less efficiently.19 Another study, however, also suggests the participation of other molecules.32
Differently from hepatic microcirculation in which a decreased NO production is likely due to a decreased expression of endothelial NOS (eNOS), systemic vasodilation in patients with cirrhosis underscores a systemic eNOS upregulation, NO overproduction, and increased formation of nitrosothiols.34 Together with thioredoxin decreases, this may explain the higher levels of nitrosothiols in stage IV PBC.
Excess NO causes a nitrosative stress that has been implicated in several hepatic fibrogenic conditions such as non-alcoholic steatohepatitis and ethanol-induced liver fibrosis.24, 35 Our findings demonstrate a progressive enhancement of serum nitrosothiol levels in PBC patients by showing significant differences among disease stages and suggest that increased nitrosothiol formation is associated with disease progression. The positive correlation found between circulating levels of nitrosothiols, histology, and hyaluronate levels in PBC patients suggests that the changes in serum nitrosothiol levels might be used as a prognostic indicator in these patients.
Our results in human PBC are also supported by the changes observed in rats undergoing BDL. In this animal model, in fact, hepatic thioredoxin levels increased immediately after operation (days 1-3), probably reflecting an upregulation mechanism in response to cholestasis-induced oxidative and nitrosative stress. Thereafter, with cholestasis progression, the decrement in hepatic thioredoxin levels may be due to a downregulation process associated with excess retention of hydrophobic bile salts and toxic molecules and may indirectly contribute to the progressive appearance of oxidative alterations, which promote, in turn, hepatic fibrosis. Adequate thioredoxin levels, in fact, protect against stellate cells activation and opposite collagen synthesis and hepatic fibrosis.36
Based on our findings, increased hepatic NOx and nitrosothiols after BDL may be indicative of an enhanced hepatocellular NO production. Indeed, controversy exists on NO production in cholestatic livers, and although some studies have shown no increase or even decreased hepatic NO production by endothelial cells and macrophages in BDL rats,37, 38 others have reported increased hepatic inducible nitric oxide synthase expression in PBC patients10 and in hepatocytes of BDL rats.6 In line with the latter hypothesis and with Esrefoglu et al.,39 we found a parallel increase of hepatic NOx and nitrosothiols levels in non-cirrhotic BDL rats. Differently, a previous study38 reported decreased hepatic NOx levels in pre-cirrhotic rats (3 weeks BDL).
In tissues, the formation of nitrosothiols adducts occurs under conditions of nitrosative stress and to counteract excess NO radicals.12 Therefore, increased levels of tissue nitrosothiols may result from an increased formation or a decreased decomposition,18, 33 such as in the presence of reduced thioredoxin availability.
The physiological role of nitrosothiols in biological systems has not been completely clarified. Nitrosothiols are unstable thioesters exerting different functions within the cell and in the extracellular compartment.33 Nitrosothiols are not formed only in cells generating NO and can be taken up into cells via amino acid transport systems.40 In a concentration-dependent manner, intracellular nitrosothiols seem to be actively involved either as redox signaling or as a nitrosative stress mediator.41 By representing a storage pool of thiols and NO, extracellular nitrosothiols may have importance also for the regulation of the protein redox status at the cell surface.42
In hepatocytes, NO is known to stimulate intracellular GMPc-dependent signaling pathways and in mitochondria it is an important physiological reactant. NO controls mitochondrial ATP synthase, gene expression, and permeability transition pore through protein nitrosylation and nitrosation.41, 43 At this level, a specific mitochondrially located NOS potentially generates the highest rate of intracellular NO•.41, 44 In addition, mitochondria are known to possess their own NO metabolism,45 and therefore the assessment of nitrosothiols may represent a useful and simple method to quantify the NO content in these organelles. Moreover, NO is a determining factor for mitochondrial proliferation.46, 47 In this regard, compensatory mitochondria proliferation is a characteristic of chronic cholestasis.48 Thus, high levels of hepatic nitrosothiols during cholestasis may provide beneficial effects by supporting cell survival and mitochondriogenesis. However, with the progression of cholestasis, the decreased availability of thioredoxin together with increased production of NO and with glutathione depletion49 may result in enhanced protein nitrosation9 and PSH oxidation. These events have negative consequences for hepatocyte survival and bile duct integrity.11, 50
In conclusion, the current integrated and translational study suggests that changes of circulating thioredoxin and nitrosothiol levels may be useful to follow disease progression in cholestatic conditions. With ongoing cholestasis, the decreased hepatic content of thioredoxin is associated with progressively higher intracellular concentrations of nitrosothiols. Interventions directed to sustain hepatic thioredoxin levels and to favorably support NO metabolism in the early phase of cholestasis may yield protection by favoring mitochondrial proliferation and allowing protein redox status maintenance.
The authors are indebted to Michele Persichella, Rosa De Venuto, Paola De Benedictis, Claudia Curci, and Laura Castellano for skillful technical support.