- Email: [email protected]
Oxidative albumin damage in chronic liver failure: Relation to albumin binding capacity, liver dysfunction and survival
Research Article
Oxidative albumin damage in chronic liver failure: Relation to albumin binding capacity, liver dysfunction and survival Karl Oettl1,⇑, Ruth Birner-Gruenberger2,3, Walter Spindelboeck4, Hans Peter Stueger5, Livia Dorn6, Vanessa Stadlbauer4, Csilla Putz-Bankuti4, Peter Krisper7, Ivo Graziadei6, Wolfgang Vogel6, Carolin Lackner3, Rudolf E. Stauber4 1
Institute of Physiological Chemistry, Medical University of Graz, Graz, Austria; 2Center for Medical Research, Core Facility Mass Spectrometry, Medical University of Graz, Graz, Austria; 3Institute of Pathology, Medical University of Graz, Graz, Austria; 4Division of Gastroenterology and Hepatology, Department of Internal Medicine, Medical University of Graz, Graz, Austria; 5Austrian Agency for Health and Food Safety, Department of Data, Statisics, Risk Assessment, Graz, Austria; 6Division of Gastroenterology and Hepatology, Department of Internal Medicine II, Medical University of Innsbruck, Innsbruck, Austria; 7Division of Nephrology and Hemodialysis, Department of Internal Medicine, Medical University of Graz, Graz, Austria See Editorial, pages 918–920
Background & Aims: Impaired binding function of albumin has been demonstrated in end-stage liver disease. This and other functional disturbances of albumin may be related to oxidative stress which is believed to play an important role in the pathogenesis of liver failure as well as sepsis. The aim of the present study was to relate oxidative modification of albumin to loss of albumin binding function in advanced chronic liver failure and in sepsis. Methods: Patients with decompensated cirrhosis or sepsis and healthy controls were investigated. Three fractions of albumin were separated by chromatography according to the redox state of cysteine-34: non-oxidized human mercaptalbumin, reversibly oxidized human non-mercaptalbumin-1, and irreversibly oxidized human non-mercaptalbumin-2 (HNA2). Binding properties of albumin site II were measured using dansylsarcosine as a ligand. Results: Both in cirrhotic and septic patients, fractions of oxidized albumin were increased and binding capacity for dansylsarcosine was decreased. Mass spectroscopy confirmed specific oxidation of cysteine-34. In cirrhotic patients, dansylsarcosine binding correlated strongly with liver function parameters and
Keywords: Oxidative stress; Cirrhosis; Mercaptalbumin; Non-mercaptalbumin; Dansylsarcosine. Received 20 August 2012; received in revised form 5 June 2013; accepted 13 June 2013; available online 25 June 2013 q DOI of original article: http://dx.doi.org/10.1016/j.jhep.2013.08.001. ⇑ Corresponding author. Address: Institute of Physiological Chemistry, Medical University of Graz, Harrachgasse 21, A-8010 Graz, Austria. Tel.: +43 316 380 7544; fax: +43 316 380 9610. E-mail address: [email protected] (K. Oettl). Abbreviations: HMA, human mercaptalbumin; HNA1, human non-mercaptalbumin1; HNA2, human non-mercaptalbumin2; ACLF, acute-on-chronic liver failure; INR, international normalized ratio; CRP, C-reactive protein; MELD, Model for End-stage Liver Disease; HPLC, high performance liquid chromatography; DS, dansylsarcosine; ROC, receiver operating characteristics.
moderately with HNA2. Baseline levels of HNA2 accurately predicted 30-day and 90-day survival in cirrhotic patients and this was confirmed in an external validation cohort. Conclusions: Our results suggest that oxidative damage impairs binding properties of albumin. In advanced liver disease, reduced binding capacity of albumin site II is mainly related to impaired liver function. The plasma level of HNA2 is closely related to survival and may represent a novel biomarker for liver failure. Ó 2013 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved.
Introduction Albumin, a 66.5 kDa protein (Fig. 1), is the most abundant plasma protein, the main determinant of colloid osmotic pressure and an important carrier for endogenous and exogenous substances [1]. Among its most important functions are fatty acid transport (on several binding areas), drug binding and transport, metal chelation, and free radical scavenging which is mediated by the anti-oxidant properties of its thiol moiety at cysteine-34 (reviewed in [2]). Albumin is the major extracellular source of reduced sulfhydryl groups, which are potent scavengers of reactive oxygen and nitrogen species [3]. Depending on the redox state, there are three major fractions of albumin: human mercaptalbumin (HMA), the non-oxidized form with a free thiol group on cysteine-34, and two different oxidized forms: (i) human non-mercaptalbumin-1 (HNA1) with cysteine, homocysteine or glutathione bound to cysteine-34 by a disulfide bond and (ii) human non-mercaptalbumin-2 (HNA2) with cysteine-34 irreversibly oxidized to sulfinic or sulfonic acid [4,5]. Oxidative stress is believed to play an important role in advanced liver failure [6] and may be reflected in oxidative modification of albumin. Decreased HMA and/or elevated HNA levels have been reported in chronic liver disease and these
Journal of Hepatology 2013 vol. 59 j 978–983
JOURNAL OF HEPATOLOGY changes have been shown to correlate with its severity as estimated by the Child-Pugh score [7,8]. Recently, we demonstrated a 4-fold increase of the irreversibly oxidized albumin fraction, HNA2, indicating a marked alteration in the redox state of circulating albumin, in patients with advanced liver disease [9]. Albumin harbours two specific binding sites described by Sudlow: site I, which binds large heterocyclic compounds and dicarboxylic acids (such as bilirubin), and site II, which binds aromatic carboxylic compounds (such as benzodiazepines) [10]. Oxidized albumin shows altered binding capacities for several substances used to assess albumin function [11,12]. Decreased binding of dansylsarcosine – a model ligand for the benzodiazepin binding site II – was found in patients with end-stage liver disease [13,14]. The pathogenesis of this impaired binding capacity and, specifically, its relation to oxidative albumin damage remain unknown. Several disturbances of albumin binding function are known to occur in cirrhotic patients [9,14,15]. Based on these pathophysiologic changes, albumin infusions are being used therapeutically to normalize the serum albumin level and/or restore albumin function in various conditions such as prevention and treatment of hepatorenal syndrome, prevention of postparacentesis circulatory dysfunction, and extracorporeal albumin dialysis for removal of potentially toxic substances accumulating in liver failure (reviewed in [16]). These beneficial effects are believed to be mediated at least in part by improved toxin binding and transport following infusion of fresh albumin or, in case of albumin dialysis, regeneration of toxin-laden albumin. Acute-on-chronic liver failure (ACLF), i.e., acute exacerbation of chronic liver failure, is frequently triggered by infection and its clinical features such as end-organ failure resemble those of sepsis [17]. Since ACLF is believed to be mediated by oxidative stress [6], oxidation of albumin is likely to occur. The specific aims of the present study were (i) to assess oxidative modification of albumin in these clinical conditions, (ii) to relate these changes to albumin dysfunction, specifically the disturbance of site II specific binding, and (iii) to determine the prognostic significance of the observed changes in chronic liver failure.
Lysine-190
Bilirubin
Cysteine-34
Binding site II Binding site I
Fig. 1. Structure of human albumin. Classical binding sites I and II, bilirubin bound to a binding site (lysine-190), where it is bound covalently, and cysteine34 are indicated by arrows. The figure was created using a PDB-file by Zunszain et al. [28], PDB-ID: 2VUE.
Patients were managed following local treatment protocols for liver failure and sepsis, including full intensive care support. Patients with suspected hepatorenal syndrome received an intravenous fluid challenge with albumin 1 g/kg according to current guidelines. The study protocol was approved by the Ethics Committee of the Medical University of Graz and informed consent was obtained in accordance with the Declaration of Helsinki. Study design Blood samples were collected within 24 h after admission, immediately centrifuged at 4 °C and plasma aliquots were stored at –70 °C until batch analysis. Bilirubin, albumin, creatinine, prothrombin time (international normalized ratio, INR), and C-reactive protein (CRP) were routinely assessed. To estimate severity of liver disease, the model for end-stage liver disease (MELD) was calculated [19,20]. Albumin analysis
Patients and methods Patients The study population comprised 67 consecutive cirrhotic patients hospitalized for acute decompensation at the Medical University of Graz, including 16 patients requiring ICU treatment, and 18 consecutive non-cirrhotic patients with sepsis without evidence for cirrhosis or underlying hematologic disease admitted at our ICU. In addition, 15 age- and sex-matched blood donors served as healthy controls. Patients with hepatocellular carcinoma, recent gastrointestinal bleeding (<7 days before enrolment) or liver transplantation within 90 days from enrolment were excluded. Sepsis was diagnosed according to international guidelines [18] as the presence of the systemic inflammatory response syndrome (SIRS) in response to a confirmed infectious process. Systemic inflammatory response syndrome was defined by the presence of two or more of the following criteria: temperature >38 °C or <36 °C; heart rate >90 beats/min; respiratory rate >20 per min or PaCO2 <32 mmHg; white blood cells >12,000 cells/mm3 or <4000 cells/mm3. Infection was defined as positive cultures of blood, ascites, urine, sputum or wounds, and/or clinical findings suggestive of infections. An additional cohort of 40 cirrhotic patients hospitalized for acute decompensation at the Medical University of Innsbruck, Austria, who fulfilled the same inclusion and exclusion criteria as the original cirrhosis cohort, was used as an external validation set for determining the prognostic value of HNA2.
Albumin was fractionated by high performance liquid chromatography to give three peaks according to cysteine-34 in the free sulfhydryl form (HMA), as a mixed disulfide (HNA1) or in a higher oxidation state (HNA2), as previously described [21]. Briefly, 20 lL of diluted plasma were injected into the HPLC system using a Shodex Asahipak ES-502N 7C anion exchange column (7.5 100 mm, Bartelt Labor- & Datentechnik, Graz, Austria) and 50 mM sodium acetate, 400 mM sodium sulfate, pH 4.85 as mobile phase. For elution, a gradient of 0–6% ethanol and a flow rate of 1 mL/min were used. The column was kept at 35 °C. Detection was carried out by fluorescence at 280/340 nm. Quantification was based on peak heights determined by EZ Chrom Elite chromatography software (VWR, Vienna, Austria). Mass spectrometry In order to further determine oxidative modification of HNA2, mass spectrometry was performed in a plasma sample of three additional patients with ACLF. Disulfide bridges of plasma proteins were reduced by incubation with 5 mM DTT and then alkylated by incubation with 10 mM iodoacetamide. Subsequently, protein was digested by adding modified trypsin (purchased from Promega, Mannheim, Germany; trypsin to plasma protein 1:50 by mass) and shaking overnight at 550 rpm and 37 °C. The peptide solution was acidified by adding 0.05% trifluoracetic acid (TFA, final concentration) and diluted in solvent A to a theoretical final total peptide concentration of 25 ng/lL. Digests (500 ng of sample) were separated by nano-HPLC (Dionex UltiMate 3000 RSLC system, Vienna, Austria)
Journal of Hepatology 2013 vol. 59 j 978–983
979
Research Article Table 1. Patient characteristics.
Age (yr) Sex (M:F) Etiology: alcohol Infectiona Albumin (g/dL) Bilirubin (mg/dL) INR Creatinine (mg/dL) Leucocytes (x10-9/L) CRP (mg/L) MELDb 30-day mortality 90-day mortality
Control (n = 15) 55 ± 5 9:6 n.a. n.a. 4.8 ± 0.2 0.6 ± 0.3 n.a. 0.8 ± 0.1 n.a. 2±2 n.a. n.a. n.a.
Cirrhosis (n = 67) 54 ± 12 53:14 87% 27% 3.2 ± 0.6 9.3 ± 9.3 1.8 ± 0.8 1.4 ± 1.1 9.0 ± 5.9 35 ± 35 21 ± 9 24% 33%
Sepsis (n = 18) 62 ± 15 13:5 n.a. 100% 2.3 ± 0.4 3.6 ± 5.2 1.4 ± 0.8 2.1 ± 2.1 13.5 ± 9.3 165 ± 77 n.a. 44% 44%
Cirrhosis-ext* (n = 40) 54 ± 11 31:9 83% 50% 3.2 ± 0.8 8.3 ± 10.8 1.8 ± 0.5 1.3 ± 0.9 8.1 ± 5.1 29 ± 33 21 ± 8 10% 28%
Means ± SD are given unless indicated otherwise. INR, international normalized ratio; CRP, C-reactive protein; n.a., not available; MELD, Model for End-Stage Liver Disease. ⁄ Denotes external validation set. a Presence of infectious focus and/or positive blood culture. b MELD score (UNOS modification) was obtained using the MELD calculator at the Mayo Clinic website (http://www.mayoclinic.org/meld/mayomodel6.html).
equipped with a Zorbax 300SB-C18 enrichment column (5 lm, 5 0.3 mm) and a Zorbax 300SB-C18 nanocolumn (3.5 lm, 150 0.075 mm). Separation was performed using the following gradient, where solvent A is 0.05% TFA in water and solvent B is a mixture of 80% acetonitrile in water containing 0.05% TFA; 0–2 min: 4% B; 2–180 min: 4–28% B; 180–255 min: 28–50% B, 255–260 min: 50–95% B; 260–279 min: 95% B; 279–280 min: 95–4% B; 280–300 min: re-equilibration at 4% B. The sample was ionized in the nanospray source equipped with nanospray tips and analyzed in a Thermo Orbitrap velos pro mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) operated in positive ion mode, applying alternating full scan MS (m/z 200 to 2000) in the ion cyclotron and MS/MS by collision induced dissociation of the 20 most intense peaks in the ion trap with dynamic exclusion enabled. The LC-MS/MS data were analyzed by searching the human NCBI protein database (http://www.ncbi.nlm.nih.gov/protein/) with Proteome Discoverer 1.3 (Thermo Fisher Scientific, Waltham, MA, USA). Oxidation to sulfonic acid at Cys (amino acid) residues and oxidation on Met residues were included as variable modifications next to carbamidomethylation on Cys residues as fixed modification. Binding of dansylsarcosine The capacity of the albumin binding site II was assessed using dansylsarcosine (DS) binding and determined by means of HPLC according to the method described by Watanabe et al. [22]. A dilution series with final concentrations of 20 lM, 10 lM, and 5 lM albumin in PBS was made. Each of these samples was mixed with the same volume of a 10 lM DS (Sigma-Aldrich, Vienna, Austria) solution. Unbound DS molecules were separated from the plasma components via ultrafiltration (Ultrafree-MC reg. Cellulose 30K, Millipore, Vienna, Austria) at 500g for 10 min at room temperature. Twenty ll of the filtrate were injected into the HPLC system with a Waters-Spherisorb 3 ODS2 column (Waters, Vienna, Austria), using acetonitrile and 100 mM sodium acetate (pH 4.5) 1:1 (v/v) as the mobile phase at a flow rate of 500 lL/min. DS was detected by fluorescence at 350/482 nm using a Hitachi F1050 (Merck) detector. Quantification was done using an external standard curve. From the straight line obtained by the graph of ln (free DS) vs. albumin, the albumin concentration binding half of the ligand was calculated and given as IC50 (lM). An increase in IC50 therefore indicates an impaired binding capacity of albumin binding site II.
Statistical analysis Results are given as mean ± SD unless indicated otherwise. For all inferential statistical calculations SPSS 19.0 was used. The relationship between the several blood parameters is described with Spearman’s rank correlation coefficient. Mul-
980
tiple linear regression was used to account for the intercorrelation structure between independent variables and DS binding. Group means were compared by analysis of variance (ANOVA). The diagnostic accuracy of prognostic variables was examined by receiver operating characteristic (ROC) analysis and the areas under the ROC curves (AUROCs) were compared according to Hanley & McNeil [23]. Optimal cut-offs for prognostic variables were determined by Youden index. Survival probability of individual cohorts was evaluated by the Kaplan-Meier method and compared by log-rank test.
Results The characteristics of the study population including the external validation set are shown in Table 1. Within the cirrhosis cohort, the main reasons for admission were ascites (40%), acute kidney injury (22%), gastrointestinal bleeding (19%), and hepatic encephalopathy (8%). As expected, both in cirrhotic and septic patients, bilirubin and CRP were increased and albumin decreased compared to controls. Septic patients presented with the highest leukocyte counts and CRP values while bilirubin levels were highest in cirrhotic patients. In all patient groups, oxidized albumin fractions were significantly increased as compared to controls (Table 2). HNA2 levels were markedly elevated both in cirrhosis and sepsis. Mass spectrometric analysis in plasma samples of three additional patients with ACLF, in order to determine the oxidative modification of HNA2, revealed the specific oxidation of cysteine-34 to sulfonic acid confirming that the allocated HNA2 peak in the HPLC chromatogram can be attributed to oxidative modification of HNA2 (Supplementary Fig. 1). The MS data were analyzed by searching the human NCBI database using Proteome Discoverer 1.3 software. Human serum albumin was identified as top hit with scores of 47,393.96, 53,135.28, and 63,965.13 and amino acid sequence coverages of 83%, 82%, and 97% in the three samples, respectively. Oxidation of cysteine-34 to sulfonic acid was detected in the tryptic peptide ALVLIAFAQYLQQCPFEDHVK (cysteine-34 shown in bold) in all samples with high confidence. An exemplary spectrum of one patient with ion score of 108 and expect of 1.1e-07 is shown in Supplementary Fig. 1.
Journal of Hepatology 2013 vol. 59 j 978–983
JOURNAL OF HEPATOLOGY Table 2. Albumin fractions and binding properties in patients and controls. Albumin was fractionated according to the redox state of cysteine-34 into HMA, HNA1, and HNA2. Binding properties for dansylsarcosine are given as albumin concentration necessary for 50% binding of 5 lM of the ligand (IC50).
HMA (%) HNA1 (%) HNA2 (%) IC50 (µmol/L)**
Control (n = 15) 65 ± 4 31 ± 4 4±1 12 ± 2
Cirrhosis (n = 67) 46 ± 13a 44 ± 12a 9 ± 4a 16 ± 7a
Sepsis (n = 18) 49 ± 17a 44 ± 17a 8 ± 5a 18 ± 7a
Cirrhosis-ext* (n = 40) 38 ± 18a 50 ± 17a 12 ± 8a n.a.
⁄
Denotes external validation set. IC50 was obtained in a subset of the cirrhosis cohort (n = 29). a p <0.05 compared to controls, by ANOVA. n.a., not available.
Receiver operating characteristic (ROC) analysis revealed excellent diagnostic accuracy of HNA2 for 30-day and 90-day survival, both in the original and external cirrhosis cohort (Table 4). For 30-day survival, the diagnostic accuracy of HNA2 tended to be superior to that of MELD; however, when compared by the method of Hanley & McNeil, AUROCs for HNA2, and MELD did not differ significantly. Youden index identified a value of 12% as an optimal cut-off for HNA2. Kaplan-Meier analysis demonstrated a significantly higher 90-day mortality for cirrhotic patients with baseline HNA2 levels >12%, both in the test set (Fig. 2) and in the external validation set (data not shown).
⁄⁄
Discussion
Table 3. Correlation analysis within cirrhotic patients enrolled at the Medical University of Graz in whom both HNA2 and IC50 values were available. Significant correlations of HNA2, dansylsarcosine binding (IC50), liver function parameters (bilirubin, INR), CRP as an inflammation parameter, and age were found in cirrhotic patients. Only correlations demonstrating at least moderate association (R >0.50) are shown.
Variable 1 HNA2
IC50
Variable 2 CRP Bilirubin INR INR Bilirubin HNA2 Age
R 0.68 0.63 0.59 0.82 0.82 0.57 -0.54
p value <0.001 <0.001 0.001 <0.001 <0.001 0.001 0.002
Albumin, total serum albumin; CRP, C-reactive protein; INR, international normalized ratio.
DS binding studies were performed in a subset of the cirrhotic patients (n = 29) and in all septic patients (n = 18). Similarly, and paralleling the changes observed in HNA2, IC50 values were increased indicating reduced DS binding capacity in both cirrhotic and septic patients (Table 2). Within our cirrhotic patients, HNA2 correlated significantly with MELD, IC50 (data not shown), bilirubin, INR, and CRP (Table 3), indicating a possible influence of both liver dysfunction and inflammation on the generation of HNA2. DS binding was related to age, bilirubin, INR, HNA2 (Table 3), and MELD. Using a stepwise regression model, we found that only bilirubin (p = 0.008) and INR (p = 0.038) had an independent impact on DS binding, while albumin fractions showed no further significant contribution to impaired DS binding. In septic patients, no such correlations were found.
In the present study, we observed marked alterations of the redox state of albumin and significant impairment of albumin binding capacity (site II) in patients with chronic liver failure as well as in septic patients without liver failure. This is in line with our previous finding of decreased binding of bilirubin to the oxidized albumin fractions, HNA1 and HNA2 [9]. Univariate analysis showed correlations of DS binding with liver functions parameters as well as HNA2 in cirrhotic but not in septic patients. However, multivariate analysis in cirrhotic patients yielded liver dysfunction parameters (bilirubin, INR) as the only independent factors for impaired albumin binding capacity for DS, indicating that altered redox state of albumin does not play a major role for the loss of albumin function in chronic liver failure. In sepsis, a condition with systemic inflammation but no major liver dysfunction (in the absence of sepsis-associated cholestasis), we similarly found increased albumin oxidation and decreased DS binding. However, no correlation of oxidative modification with binding capacity was found in this group. In chronic liver failure, high bilirubin levels seem to have a major impact on albumin’s binding properties. The possibility of allosteric effects of ligands leading to changes in binding of other ligands has been reported [24]. It should be noted that the high content of the antioxidant bilirubin and the close spatial vicinity of cysteine-34 to the bilirubin binding sites for high affinity as well as covalent binding (lysine-190, Fig. 1) apparently cannot protect cysteine-34 from oxidation [25]. On the other hand, binding of fatty acids to albumin has been shown to enhance the susceptibility of cysteine-34 to oxidation [26]. Bilirubin can bind to albumin at various sites [27]. The primary site of high-affinity binding is still controversial [28]. While drug binding site I is generally believed to represent the major high-affinity binding site for bilirubin, evidence is accumulating that other sites are important as well. The spatial vicinity of cysteine-34 and the putative high-affinity bilirubin binding site (site I) as well as covalently bound bilirubin at lysine-190 (Fig. 1) may explain the previously reported influence of the redox state of
Table 4. Prognostic value of HNA2 and MELD in chronic liver failure. Using receiver operating characteristic analysis, the diagnostic accuracy of HNA2 and MELD for prediction of 30-day and 90-day survival was assessed within the cirrhosis group and the external validation set. Values in parentheses denote 95% confidence intervals.
HNA2 MELD
Cirrhosis (n = 67) 0.85 (0.75-0.95) 0.83 (0.70-0.96)
AUROC 30 d-survival Cirrhosis-ext* (n = 40) 0.86 (0.72-1.00) 0.75 (0.46-1.00)
Cirrhosis (n = 67) 0.82 (0.70-0.94) 0.82 (0.69-0.94)
AUROC 90 d-survival Cirrhosis-ext* (n = 40) 0.81 (0.68-0.94) 0.79 (0.64-0.95)
AUROC, area under the receiver operating characteristic curve; HNA2, human non-mercaptalbumin 2; MELD, Model for End-Stage Liver Disease. ⁄ Denotes external validation set.
Journal of Hepatology 2013 vol. 59 j 978–983
981
Research Article 1.0
Cumulative survival
0.8
HNA2 <12%
0.6 0.4 0.2 HNA2 >12% 0.0 0
360 Days
720
Fig. 2. Kaplan–Meier plot of survival probability according to baseline HNA2 levels in the cirrhosis cohort (n = 67). An HNA2 value >12% indicates a markedly increased 90-day mortality (p <0.001 by Log-rank test).
cysteine-34 on bilirubin binding [28]. Likewise, our finding that DS binding is more closely related to bilirubin than to HNA2 is consistent with the closer vicinity of site II to bilirubin binding sites than to cysteine-34 (Fig. 1). Further lines of evidence for impaired albumin function in advanced liver failure were recently presented by Jalan et al. in a study of 34 patients with chronic liver failure, including 20 patients with ACLF [15]. These authors demonstrated a reduced affinity of albumin fatty acid binding sites using electron paramagnetic resonance spectroscopy as well as a relative increase in ischemia modified albumin indicated by a reduced cobalt binding capacity [15]. The reduced albumin binding capacity in advanced liver failure and in sepsis has important clinical implications. Higher unbound fractions of ligands may increase their toxicity thus deteriorating the clinical course. Impaired binding, transport and delivery of drugs (e.g., antibiotics) may hinder specific treatment. Finally, the interaction of drugs and/or toxins with endogenous substances accumulating in liver failure may play an important pathogenic role as they compete for a reduced number of functional binding sites on the damaged albumin molecules. It should be noted that the fractions of irreversibly oxidized albumin (HNA2) observed in our cohorts of cirrhotic patients represent the highest values among all patient groups studied so far [29]. Nevertheless, apart from irreversibly oxidized HNA2, the major proportion of circulating albumin is present in the unoxidized HMA or reversibly oxidized HNA1 fractions which still may provide adequate binding function and are accessible to therapeutic interventions such as albumin dialysis [30]. With respect to prognosis, the short-term mortality of our cirrhosis cohorts (Table 1) lies within the range of recently published data from the CANONIC trial, which included a large multicenter cohort of 1343 cirrhotic patients hospitalized for acute decompensation, with overall 28-day mortality ranging between 7.5% and 17% and overall 90-day mortality between 19.8% and 27.6% at various centers [31]. Interestingly, analysis of the prognostic value of HNA2 with respect to short-term survival indicated a trend for superior diagnostic accuracy as compared to MELD, the current standard for assessment of 982
prognosis in chronic liver failure, which could be confirmed in the external validation set. However, the prognostic value of HNA2 has to be validated in larger cohorts of cirrhotic patients. Based on our findings, we conclude that impaired albumin site II binding capacity observed in chronic liver failure is mainly related to the severity of liver dysfunction. Concurrent oxidative stress seems to enhance the loss of albumin function in chronic liver failure but not in sepsis. Irreversible oxidation of albumin confers bad prognosis and thus may represent a novel biomarker for chronic liver failure.
Financial support This study was supported by the Franz Lanyar Stiftung (project # 314) and by BioPersMed (COMET K-project 825329), which is funded by the Austrian Federal Ministry of Transport, Innovation and Technology (BMVIT) and the Austrian Federal Ministry of Economics and Labour/the Federal Ministry of Economy, Family and Youth (BMWA/BMWFJ) and the Styrian Business Promotion Agency (SFG). V.S. was the recipient of an Erwin-Schrödinger fellowship (J2547) from the Austrian Science Foundation.
Conflict of interest The authors who have taken part in this study declared that they do not have anything to disclose regarding funding or conflict of interest with respect to this manuscript. Acknowledgements The technical assistance of Martina Brtnik, Doris Payerl, and Stefan Spoerk is gratefully acknowledged.
Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jhep.2013.06.013.
References [1] Rothschild MA, Oratz M, Schreiber SS. Serum albumin. Hepatology 1988;8:385–401. [2] Evans TW. Review article: albumin as a drug-biological effects of albumin unrelated to oncotic pressure. Aliment Pharmacol Ther 2002;16:6–11. [3] Quinlan GJ, Margarson MP, Mumby S, Evans TW, Gutteridge JM. Administration of albumin to patients with sepsis syndrome: a possible beneficial role in plasma thiol repletion. Clin Sci (Lond) 1998;95:459–465. [4] Sogami M, Nagoka S, Era S, Honda M, Noguchi K. Resolution of human mercapt- and nonmercaptalbumin by high-performance liquid chromatography. Int J Pept Protein Res 1984;24:96–103. [5] Hughes WL, Dintzis HM. Crystallization of the mercury dimers of human and bovine mercaptalbumin. J Biol Chem 1964;239:845–849. [6] Sen S, Williams R, Jalan R. The pathophysiological basis of acute-on-chronic liver failure. Liver 2002;22:5–13. [7] Sogami M, Era S, Nagaoka S, Kuwata K, Kida K, Shigemi J, et al. Highperformance liquid chromatographic studies on non-mercapt in equilibrium with mercapt conversion of human serum albumin. II. J Chromatogr 1985;332:19–27. [8] Watanabe A, Matsuzaki S, Moriwaki H, Suzuki K, Nishiguchi S. Problems in serum albumin measurement and clinical significance of albumin microheterogeneity in cirrhotics. Nutrition 2004;20:351–357.
Journal of Hepatology 2013 vol. 59 j 978–983
JOURNAL OF HEPATOLOGY [9] Oettl K, Stadlbauer V, Petter F, Greilberger J, Putz-Bankuti C, Hallstrom S, et al. Oxidative damage of albumin in advanced liver disease. Biochim Biophys Acta 2008;1782:469–473. [10] Sudlow G, Birkett DJ, Wade DN. The characterization of two specific drug binding sites on human serum albumin. Mol Pharmacol 1975;11:824–832. [11] Oettl K, Stauber RE. Physiological and pathological changes in the redox state of human serum albumin critically influence its binding properties. Br J Pharmacol 2007;151:580–590. [12] Meucci E, Mordente A, Martorana GE. Metal-catalyzed oxidation of human serum albumin: conformational and functional changes. Implications in protein aging. J Biol Chem 1991;266:4692–4699. [13] Klammt S, Mitzner S, Stange J, Brinkmann B, Drewelow B, Emmrich J, et al. Albumin-binding function is reduced in patients with decompensated cirrhosis and correlates inversely with severity of liver disease assessed by model for end-stage liver disease. Eur J Gastroenterol Hepatol 2007;19: 257–263. [14] Klammt S, Brinkmann B, Mitzner S, Munzert E, Loock J, Stange J, et al. Albumin binding capacity (ABiC) is reduced in commercially available human serum albumin preparations with stabilizers. Z Gastroenterol 2001;39:24–27. [15] Jalan R, Schnurr K, Mookerjee RP, Sen S, Cheshire L, Hodges S, et al. Alterations in the functional capacity of albumin in patients with decompensated cirrhosis is associated with increased mortality. Hepatology 2009;50:555–564. [16] Arroyo V. Review article: albumin in the treatment of liver diseases – new features of a classical treatment. Aliment Pharmacol Ther 2002;16:1–5. [17] Jalan R, Williams R. Acute-on-chronic liver failure: pathophysiological basis of therapeutic options. Blood Purif 2002;20:252–261. [18] Bone RC, Balk RA, Cerra FB, Dellinger RP, Fein AM, Knaus WA, et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/SCCM Consensus Conference Committee. American College of Chest Physicians/Society of Critical Care Medicine. Chest 1992;101:1644–1655. [19] Kamath PS, Wiesner RH, Malinchoc M, Kremers W, Therneau TM, Kosberg CL, et al. A model to predict survival in patients with end-stage liver disease. Hepatology 2001;33:464–470.
[20] Wiesner R, Edwards E, Freeman R, Harper A, Kim R, Kamath P, et al. United Network for Organ Sharing Liver Disease Severity Score Committee. Model for end-stage liver disease (MELD) and allocation of donor livers. Gastroenterology 2003;124:91–96. [21] Kawai K, Yoh M, Hayashi T, Imai H, Negawa T, Tomida M, et al. Effect of diabetic retinopathy on redox state of aqueous humor and serum albumin in patients with senile cataract. Tokai J Exp Clin Med 2001;26:93–99. [22] Watanabe H, Yamasaki K, Kragh-Hansen U, Tanase S, Harada K, Suenaga A, et al. In vitro and in vivo properties of recombinant human serum albumin from Pichia pastoris purified by a method of short processing time. Pharm Res 2001;18:1775–1781. [23] Hanley JA, McNeil BJ. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 1983;148:839–843. [24] Otagiri M. A molecular functional study on the interactions of drugs with plasma proteins. Drug Metab Pharmacokinet 2005;20:309–323. [25] Adachi Y, Murata N, Tsunasawa S, Kamisako T, Yamamoto T. Lysine 190 is a covalent bilirubin-binding amino acid in human delta bilirubin. Int Hepatol Commun 1996;5:282–288. [26] Gryzunov YA, Arroyo A, Vigne JL, Zhao Q, Tyurin VA, Hubel CA, et al. Binding of fatty acids facilitates oxidation of cysteine-34 and converts copperalbumin complexes from antioxidants to prooxidants. Arch Biochem Biophys 2003;413:53–66. [27] Vitek L, Ostrow JD. Bilirubin chemistry and metabolism; harmful and protective aspects. Curr Pharm Des 2009;15:2869–2883. [28] Zunszain PA, Ghuman J, McDonagh AF, Curry S. Crystallographic analysis of human serum albumin complexed with 4Z,15E-bilirubin-IXalpha. J Mol Biol 2008;381:394–406. [29] Oettl K, Marsche G. Redox state of human serum albumin in terms of cysteine-34 in health and disease. Methods Enzymol 2010;474:181–195. [30] Oettl K, Stadlbauer V, Krisper P, Stauber RE. Effect of extracorporeal liver support by molecular adsorbents recirculating system and Prometheus on redox state of albumin in acute-on-chronic liver failure. Ther Apher Dial 2009;13:431–436. [31] Moreau R, Jalan R, Gines P, Pavesi M, Angeli P, Cordoba J, et al. Acute-onchronic liver failure is a distinct syndrome that develops in patients with acute decompensation of cirrhosis. Gastroenterology 2013;144:1426–1437.
Journal of Hepatology 2013 vol. 59 j 978–983
983
Oxidative albumin damage in chronic liver failure: Relation to albumin binding capacity, liver dysfunction and survival Karl Oettl1,⇑, Ruth Birner-Gruenberger2,3, Walter Spindelboeck4, Hans Peter Stueger5, Livia Dorn6, Vanessa Stadlbauer4, Csilla Putz-Bankuti4, Peter Krisper7, Ivo Graziadei6, Wolfgang Vogel6, Carolin Lackner3, Rudolf E. Stauber4 1
Institute of Physiological Chemistry, Medical University of Graz, Graz, Austria; 2Center for Medical Research, Core Facility Mass Spectrometry, Medical University of Graz, Graz, Austria; 3Institute of Pathology, Medical University of Graz, Graz, Austria; 4Division of Gastroenterology and Hepatology, Department of Internal Medicine, Medical University of Graz, Graz, Austria; 5Austrian Agency for Health and Food Safety, Department of Data, Statisics, Risk Assessment, Graz, Austria; 6Division of Gastroenterology and Hepatology, Department of Internal Medicine II, Medical University of Innsbruck, Innsbruck, Austria; 7Division of Nephrology and Hemodialysis, Department of Internal Medicine, Medical University of Graz, Graz, Austria See Editorial, pages 918–920
Background & Aims: Impaired binding function of albumin has been demonstrated in end-stage liver disease. This and other functional disturbances of albumin may be related to oxidative stress which is believed to play an important role in the pathogenesis of liver failure as well as sepsis. The aim of the present study was to relate oxidative modification of albumin to loss of albumin binding function in advanced chronic liver failure and in sepsis. Methods: Patients with decompensated cirrhosis or sepsis and healthy controls were investigated. Three fractions of albumin were separated by chromatography according to the redox state of cysteine-34: non-oxidized human mercaptalbumin, reversibly oxidized human non-mercaptalbumin-1, and irreversibly oxidized human non-mercaptalbumin-2 (HNA2). Binding properties of albumin site II were measured using dansylsarcosine as a ligand. Results: Both in cirrhotic and septic patients, fractions of oxidized albumin were increased and binding capacity for dansylsarcosine was decreased. Mass spectroscopy confirmed specific oxidation of cysteine-34. In cirrhotic patients, dansylsarcosine binding correlated strongly with liver function parameters and
Keywords: Oxidative stress; Cirrhosis; Mercaptalbumin; Non-mercaptalbumin; Dansylsarcosine. Received 20 August 2012; received in revised form 5 June 2013; accepted 13 June 2013; available online 25 June 2013 q DOI of original article: http://dx.doi.org/10.1016/j.jhep.2013.08.001. ⇑ Corresponding author. Address: Institute of Physiological Chemistry, Medical University of Graz, Harrachgasse 21, A-8010 Graz, Austria. Tel.: +43 316 380 7544; fax: +43 316 380 9610. E-mail address: [email protected] (K. Oettl). Abbreviations: HMA, human mercaptalbumin; HNA1, human non-mercaptalbumin1; HNA2, human non-mercaptalbumin2; ACLF, acute-on-chronic liver failure; INR, international normalized ratio; CRP, C-reactive protein; MELD, Model for End-stage Liver Disease; HPLC, high performance liquid chromatography; DS, dansylsarcosine; ROC, receiver operating characteristics.
moderately with HNA2. Baseline levels of HNA2 accurately predicted 30-day and 90-day survival in cirrhotic patients and this was confirmed in an external validation cohort. Conclusions: Our results suggest that oxidative damage impairs binding properties of albumin. In advanced liver disease, reduced binding capacity of albumin site II is mainly related to impaired liver function. The plasma level of HNA2 is closely related to survival and may represent a novel biomarker for liver failure. Ó 2013 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved.
Introduction Albumin, a 66.5 kDa protein (Fig. 1), is the most abundant plasma protein, the main determinant of colloid osmotic pressure and an important carrier for endogenous and exogenous substances [1]. Among its most important functions are fatty acid transport (on several binding areas), drug binding and transport, metal chelation, and free radical scavenging which is mediated by the anti-oxidant properties of its thiol moiety at cysteine-34 (reviewed in [2]). Albumin is the major extracellular source of reduced sulfhydryl groups, which are potent scavengers of reactive oxygen and nitrogen species [3]. Depending on the redox state, there are three major fractions of albumin: human mercaptalbumin (HMA), the non-oxidized form with a free thiol group on cysteine-34, and two different oxidized forms: (i) human non-mercaptalbumin-1 (HNA1) with cysteine, homocysteine or glutathione bound to cysteine-34 by a disulfide bond and (ii) human non-mercaptalbumin-2 (HNA2) with cysteine-34 irreversibly oxidized to sulfinic or sulfonic acid [4,5]. Oxidative stress is believed to play an important role in advanced liver failure [6] and may be reflected in oxidative modification of albumin. Decreased HMA and/or elevated HNA levels have been reported in chronic liver disease and these
Journal of Hepatology 2013 vol. 59 j 978–983
JOURNAL OF HEPATOLOGY changes have been shown to correlate with its severity as estimated by the Child-Pugh score [7,8]. Recently, we demonstrated a 4-fold increase of the irreversibly oxidized albumin fraction, HNA2, indicating a marked alteration in the redox state of circulating albumin, in patients with advanced liver disease [9]. Albumin harbours two specific binding sites described by Sudlow: site I, which binds large heterocyclic compounds and dicarboxylic acids (such as bilirubin), and site II, which binds aromatic carboxylic compounds (such as benzodiazepines) [10]. Oxidized albumin shows altered binding capacities for several substances used to assess albumin function [11,12]. Decreased binding of dansylsarcosine – a model ligand for the benzodiazepin binding site II – was found in patients with end-stage liver disease [13,14]. The pathogenesis of this impaired binding capacity and, specifically, its relation to oxidative albumin damage remain unknown. Several disturbances of albumin binding function are known to occur in cirrhotic patients [9,14,15]. Based on these pathophysiologic changes, albumin infusions are being used therapeutically to normalize the serum albumin level and/or restore albumin function in various conditions such as prevention and treatment of hepatorenal syndrome, prevention of postparacentesis circulatory dysfunction, and extracorporeal albumin dialysis for removal of potentially toxic substances accumulating in liver failure (reviewed in [16]). These beneficial effects are believed to be mediated at least in part by improved toxin binding and transport following infusion of fresh albumin or, in case of albumin dialysis, regeneration of toxin-laden albumin. Acute-on-chronic liver failure (ACLF), i.e., acute exacerbation of chronic liver failure, is frequently triggered by infection and its clinical features such as end-organ failure resemble those of sepsis [17]. Since ACLF is believed to be mediated by oxidative stress [6], oxidation of albumin is likely to occur. The specific aims of the present study were (i) to assess oxidative modification of albumin in these clinical conditions, (ii) to relate these changes to albumin dysfunction, specifically the disturbance of site II specific binding, and (iii) to determine the prognostic significance of the observed changes in chronic liver failure.
Lysine-190
Bilirubin
Cysteine-34
Binding site II Binding site I
Fig. 1. Structure of human albumin. Classical binding sites I and II, bilirubin bound to a binding site (lysine-190), where it is bound covalently, and cysteine34 are indicated by arrows. The figure was created using a PDB-file by Zunszain et al. [28], PDB-ID: 2VUE.
Patients were managed following local treatment protocols for liver failure and sepsis, including full intensive care support. Patients with suspected hepatorenal syndrome received an intravenous fluid challenge with albumin 1 g/kg according to current guidelines. The study protocol was approved by the Ethics Committee of the Medical University of Graz and informed consent was obtained in accordance with the Declaration of Helsinki. Study design Blood samples were collected within 24 h after admission, immediately centrifuged at 4 °C and plasma aliquots were stored at –70 °C until batch analysis. Bilirubin, albumin, creatinine, prothrombin time (international normalized ratio, INR), and C-reactive protein (CRP) were routinely assessed. To estimate severity of liver disease, the model for end-stage liver disease (MELD) was calculated [19,20]. Albumin analysis
Patients and methods Patients The study population comprised 67 consecutive cirrhotic patients hospitalized for acute decompensation at the Medical University of Graz, including 16 patients requiring ICU treatment, and 18 consecutive non-cirrhotic patients with sepsis without evidence for cirrhosis or underlying hematologic disease admitted at our ICU. In addition, 15 age- and sex-matched blood donors served as healthy controls. Patients with hepatocellular carcinoma, recent gastrointestinal bleeding (<7 days before enrolment) or liver transplantation within 90 days from enrolment were excluded. Sepsis was diagnosed according to international guidelines [18] as the presence of the systemic inflammatory response syndrome (SIRS) in response to a confirmed infectious process. Systemic inflammatory response syndrome was defined by the presence of two or more of the following criteria: temperature >38 °C or <36 °C; heart rate >90 beats/min; respiratory rate >20 per min or PaCO2 <32 mmHg; white blood cells >12,000 cells/mm3 or <4000 cells/mm3. Infection was defined as positive cultures of blood, ascites, urine, sputum or wounds, and/or clinical findings suggestive of infections. An additional cohort of 40 cirrhotic patients hospitalized for acute decompensation at the Medical University of Innsbruck, Austria, who fulfilled the same inclusion and exclusion criteria as the original cirrhosis cohort, was used as an external validation set for determining the prognostic value of HNA2.
Albumin was fractionated by high performance liquid chromatography to give three peaks according to cysteine-34 in the free sulfhydryl form (HMA), as a mixed disulfide (HNA1) or in a higher oxidation state (HNA2), as previously described [21]. Briefly, 20 lL of diluted plasma were injected into the HPLC system using a Shodex Asahipak ES-502N 7C anion exchange column (7.5 100 mm, Bartelt Labor- & Datentechnik, Graz, Austria) and 50 mM sodium acetate, 400 mM sodium sulfate, pH 4.85 as mobile phase. For elution, a gradient of 0–6% ethanol and a flow rate of 1 mL/min were used. The column was kept at 35 °C. Detection was carried out by fluorescence at 280/340 nm. Quantification was based on peak heights determined by EZ Chrom Elite chromatography software (VWR, Vienna, Austria). Mass spectrometry In order to further determine oxidative modification of HNA2, mass spectrometry was performed in a plasma sample of three additional patients with ACLF. Disulfide bridges of plasma proteins were reduced by incubation with 5 mM DTT and then alkylated by incubation with 10 mM iodoacetamide. Subsequently, protein was digested by adding modified trypsin (purchased from Promega, Mannheim, Germany; trypsin to plasma protein 1:50 by mass) and shaking overnight at 550 rpm and 37 °C. The peptide solution was acidified by adding 0.05% trifluoracetic acid (TFA, final concentration) and diluted in solvent A to a theoretical final total peptide concentration of 25 ng/lL. Digests (500 ng of sample) were separated by nano-HPLC (Dionex UltiMate 3000 RSLC system, Vienna, Austria)
Journal of Hepatology 2013 vol. 59 j 978–983
979
Research Article Table 1. Patient characteristics.
Age (yr) Sex (M:F) Etiology: alcohol Infectiona Albumin (g/dL) Bilirubin (mg/dL) INR Creatinine (mg/dL) Leucocytes (x10-9/L) CRP (mg/L) MELDb 30-day mortality 90-day mortality
Control (n = 15) 55 ± 5 9:6 n.a. n.a. 4.8 ± 0.2 0.6 ± 0.3 n.a. 0.8 ± 0.1 n.a. 2±2 n.a. n.a. n.a.
Cirrhosis (n = 67) 54 ± 12 53:14 87% 27% 3.2 ± 0.6 9.3 ± 9.3 1.8 ± 0.8 1.4 ± 1.1 9.0 ± 5.9 35 ± 35 21 ± 9 24% 33%
Sepsis (n = 18) 62 ± 15 13:5 n.a. 100% 2.3 ± 0.4 3.6 ± 5.2 1.4 ± 0.8 2.1 ± 2.1 13.5 ± 9.3 165 ± 77 n.a. 44% 44%
Cirrhosis-ext* (n = 40) 54 ± 11 31:9 83% 50% 3.2 ± 0.8 8.3 ± 10.8 1.8 ± 0.5 1.3 ± 0.9 8.1 ± 5.1 29 ± 33 21 ± 8 10% 28%
Means ± SD are given unless indicated otherwise. INR, international normalized ratio; CRP, C-reactive protein; n.a., not available; MELD, Model for End-Stage Liver Disease. ⁄ Denotes external validation set. a Presence of infectious focus and/or positive blood culture. b MELD score (UNOS modification) was obtained using the MELD calculator at the Mayo Clinic website (http://www.mayoclinic.org/meld/mayomodel6.html).
equipped with a Zorbax 300SB-C18 enrichment column (5 lm, 5 0.3 mm) and a Zorbax 300SB-C18 nanocolumn (3.5 lm, 150 0.075 mm). Separation was performed using the following gradient, where solvent A is 0.05% TFA in water and solvent B is a mixture of 80% acetonitrile in water containing 0.05% TFA; 0–2 min: 4% B; 2–180 min: 4–28% B; 180–255 min: 28–50% B, 255–260 min: 50–95% B; 260–279 min: 95% B; 279–280 min: 95–4% B; 280–300 min: re-equilibration at 4% B. The sample was ionized in the nanospray source equipped with nanospray tips and analyzed in a Thermo Orbitrap velos pro mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) operated in positive ion mode, applying alternating full scan MS (m/z 200 to 2000) in the ion cyclotron and MS/MS by collision induced dissociation of the 20 most intense peaks in the ion trap with dynamic exclusion enabled. The LC-MS/MS data were analyzed by searching the human NCBI protein database (http://www.ncbi.nlm.nih.gov/protein/) with Proteome Discoverer 1.3 (Thermo Fisher Scientific, Waltham, MA, USA). Oxidation to sulfonic acid at Cys (amino acid) residues and oxidation on Met residues were included as variable modifications next to carbamidomethylation on Cys residues as fixed modification. Binding of dansylsarcosine The capacity of the albumin binding site II was assessed using dansylsarcosine (DS) binding and determined by means of HPLC according to the method described by Watanabe et al. [22]. A dilution series with final concentrations of 20 lM, 10 lM, and 5 lM albumin in PBS was made. Each of these samples was mixed with the same volume of a 10 lM DS (Sigma-Aldrich, Vienna, Austria) solution. Unbound DS molecules were separated from the plasma components via ultrafiltration (Ultrafree-MC reg. Cellulose 30K, Millipore, Vienna, Austria) at 500g for 10 min at room temperature. Twenty ll of the filtrate were injected into the HPLC system with a Waters-Spherisorb 3 ODS2 column (Waters, Vienna, Austria), using acetonitrile and 100 mM sodium acetate (pH 4.5) 1:1 (v/v) as the mobile phase at a flow rate of 500 lL/min. DS was detected by fluorescence at 350/482 nm using a Hitachi F1050 (Merck) detector. Quantification was done using an external standard curve. From the straight line obtained by the graph of ln (free DS) vs. albumin, the albumin concentration binding half of the ligand was calculated and given as IC50 (lM). An increase in IC50 therefore indicates an impaired binding capacity of albumin binding site II.
Statistical analysis Results are given as mean ± SD unless indicated otherwise. For all inferential statistical calculations SPSS 19.0 was used. The relationship between the several blood parameters is described with Spearman’s rank correlation coefficient. Mul-
980
tiple linear regression was used to account for the intercorrelation structure between independent variables and DS binding. Group means were compared by analysis of variance (ANOVA). The diagnostic accuracy of prognostic variables was examined by receiver operating characteristic (ROC) analysis and the areas under the ROC curves (AUROCs) were compared according to Hanley & McNeil [23]. Optimal cut-offs for prognostic variables were determined by Youden index. Survival probability of individual cohorts was evaluated by the Kaplan-Meier method and compared by log-rank test.
Results The characteristics of the study population including the external validation set are shown in Table 1. Within the cirrhosis cohort, the main reasons for admission were ascites (40%), acute kidney injury (22%), gastrointestinal bleeding (19%), and hepatic encephalopathy (8%). As expected, both in cirrhotic and septic patients, bilirubin and CRP were increased and albumin decreased compared to controls. Septic patients presented with the highest leukocyte counts and CRP values while bilirubin levels were highest in cirrhotic patients. In all patient groups, oxidized albumin fractions were significantly increased as compared to controls (Table 2). HNA2 levels were markedly elevated both in cirrhosis and sepsis. Mass spectrometric analysis in plasma samples of three additional patients with ACLF, in order to determine the oxidative modification of HNA2, revealed the specific oxidation of cysteine-34 to sulfonic acid confirming that the allocated HNA2 peak in the HPLC chromatogram can be attributed to oxidative modification of HNA2 (Supplementary Fig. 1). The MS data were analyzed by searching the human NCBI database using Proteome Discoverer 1.3 software. Human serum albumin was identified as top hit with scores of 47,393.96, 53,135.28, and 63,965.13 and amino acid sequence coverages of 83%, 82%, and 97% in the three samples, respectively. Oxidation of cysteine-34 to sulfonic acid was detected in the tryptic peptide ALVLIAFAQYLQQCPFEDHVK (cysteine-34 shown in bold) in all samples with high confidence. An exemplary spectrum of one patient with ion score of 108 and expect of 1.1e-07 is shown in Supplementary Fig. 1.
Journal of Hepatology 2013 vol. 59 j 978–983
JOURNAL OF HEPATOLOGY Table 2. Albumin fractions and binding properties in patients and controls. Albumin was fractionated according to the redox state of cysteine-34 into HMA, HNA1, and HNA2. Binding properties for dansylsarcosine are given as albumin concentration necessary for 50% binding of 5 lM of the ligand (IC50).
HMA (%) HNA1 (%) HNA2 (%) IC50 (µmol/L)**
Control (n = 15) 65 ± 4 31 ± 4 4±1 12 ± 2
Cirrhosis (n = 67) 46 ± 13a 44 ± 12a 9 ± 4a 16 ± 7a
Sepsis (n = 18) 49 ± 17a 44 ± 17a 8 ± 5a 18 ± 7a
Cirrhosis-ext* (n = 40) 38 ± 18a 50 ± 17a 12 ± 8a n.a.
⁄
Denotes external validation set. IC50 was obtained in a subset of the cirrhosis cohort (n = 29). a p <0.05 compared to controls, by ANOVA. n.a., not available.
Receiver operating characteristic (ROC) analysis revealed excellent diagnostic accuracy of HNA2 for 30-day and 90-day survival, both in the original and external cirrhosis cohort (Table 4). For 30-day survival, the diagnostic accuracy of HNA2 tended to be superior to that of MELD; however, when compared by the method of Hanley & McNeil, AUROCs for HNA2, and MELD did not differ significantly. Youden index identified a value of 12% as an optimal cut-off for HNA2. Kaplan-Meier analysis demonstrated a significantly higher 90-day mortality for cirrhotic patients with baseline HNA2 levels >12%, both in the test set (Fig. 2) and in the external validation set (data not shown).
⁄⁄
Discussion
Table 3. Correlation analysis within cirrhotic patients enrolled at the Medical University of Graz in whom both HNA2 and IC50 values were available. Significant correlations of HNA2, dansylsarcosine binding (IC50), liver function parameters (bilirubin, INR), CRP as an inflammation parameter, and age were found in cirrhotic patients. Only correlations demonstrating at least moderate association (R >0.50) are shown.
Variable 1 HNA2
IC50
Variable 2 CRP Bilirubin INR INR Bilirubin HNA2 Age
R 0.68 0.63 0.59 0.82 0.82 0.57 -0.54
p value <0.001 <0.001 0.001 <0.001 <0.001 0.001 0.002
Albumin, total serum albumin; CRP, C-reactive protein; INR, international normalized ratio.
DS binding studies were performed in a subset of the cirrhotic patients (n = 29) and in all septic patients (n = 18). Similarly, and paralleling the changes observed in HNA2, IC50 values were increased indicating reduced DS binding capacity in both cirrhotic and septic patients (Table 2). Within our cirrhotic patients, HNA2 correlated significantly with MELD, IC50 (data not shown), bilirubin, INR, and CRP (Table 3), indicating a possible influence of both liver dysfunction and inflammation on the generation of HNA2. DS binding was related to age, bilirubin, INR, HNA2 (Table 3), and MELD. Using a stepwise regression model, we found that only bilirubin (p = 0.008) and INR (p = 0.038) had an independent impact on DS binding, while albumin fractions showed no further significant contribution to impaired DS binding. In septic patients, no such correlations were found.
In the present study, we observed marked alterations of the redox state of albumin and significant impairment of albumin binding capacity (site II) in patients with chronic liver failure as well as in septic patients without liver failure. This is in line with our previous finding of decreased binding of bilirubin to the oxidized albumin fractions, HNA1 and HNA2 [9]. Univariate analysis showed correlations of DS binding with liver functions parameters as well as HNA2 in cirrhotic but not in septic patients. However, multivariate analysis in cirrhotic patients yielded liver dysfunction parameters (bilirubin, INR) as the only independent factors for impaired albumin binding capacity for DS, indicating that altered redox state of albumin does not play a major role for the loss of albumin function in chronic liver failure. In sepsis, a condition with systemic inflammation but no major liver dysfunction (in the absence of sepsis-associated cholestasis), we similarly found increased albumin oxidation and decreased DS binding. However, no correlation of oxidative modification with binding capacity was found in this group. In chronic liver failure, high bilirubin levels seem to have a major impact on albumin’s binding properties. The possibility of allosteric effects of ligands leading to changes in binding of other ligands has been reported [24]. It should be noted that the high content of the antioxidant bilirubin and the close spatial vicinity of cysteine-34 to the bilirubin binding sites for high affinity as well as covalent binding (lysine-190, Fig. 1) apparently cannot protect cysteine-34 from oxidation [25]. On the other hand, binding of fatty acids to albumin has been shown to enhance the susceptibility of cysteine-34 to oxidation [26]. Bilirubin can bind to albumin at various sites [27]. The primary site of high-affinity binding is still controversial [28]. While drug binding site I is generally believed to represent the major high-affinity binding site for bilirubin, evidence is accumulating that other sites are important as well. The spatial vicinity of cysteine-34 and the putative high-affinity bilirubin binding site (site I) as well as covalently bound bilirubin at lysine-190 (Fig. 1) may explain the previously reported influence of the redox state of
Table 4. Prognostic value of HNA2 and MELD in chronic liver failure. Using receiver operating characteristic analysis, the diagnostic accuracy of HNA2 and MELD for prediction of 30-day and 90-day survival was assessed within the cirrhosis group and the external validation set. Values in parentheses denote 95% confidence intervals.
HNA2 MELD
Cirrhosis (n = 67) 0.85 (0.75-0.95) 0.83 (0.70-0.96)
AUROC 30 d-survival Cirrhosis-ext* (n = 40) 0.86 (0.72-1.00) 0.75 (0.46-1.00)
Cirrhosis (n = 67) 0.82 (0.70-0.94) 0.82 (0.69-0.94)
AUROC 90 d-survival Cirrhosis-ext* (n = 40) 0.81 (0.68-0.94) 0.79 (0.64-0.95)
AUROC, area under the receiver operating characteristic curve; HNA2, human non-mercaptalbumin 2; MELD, Model for End-Stage Liver Disease. ⁄ Denotes external validation set.
Journal of Hepatology 2013 vol. 59 j 978–983
981
Research Article 1.0
Cumulative survival
0.8
HNA2 <12%
0.6 0.4 0.2 HNA2 >12% 0.0 0
360 Days
720
Fig. 2. Kaplan–Meier plot of survival probability according to baseline HNA2 levels in the cirrhosis cohort (n = 67). An HNA2 value >12% indicates a markedly increased 90-day mortality (p <0.001 by Log-rank test).
cysteine-34 on bilirubin binding [28]. Likewise, our finding that DS binding is more closely related to bilirubin than to HNA2 is consistent with the closer vicinity of site II to bilirubin binding sites than to cysteine-34 (Fig. 1). Further lines of evidence for impaired albumin function in advanced liver failure were recently presented by Jalan et al. in a study of 34 patients with chronic liver failure, including 20 patients with ACLF [15]. These authors demonstrated a reduced affinity of albumin fatty acid binding sites using electron paramagnetic resonance spectroscopy as well as a relative increase in ischemia modified albumin indicated by a reduced cobalt binding capacity [15]. The reduced albumin binding capacity in advanced liver failure and in sepsis has important clinical implications. Higher unbound fractions of ligands may increase their toxicity thus deteriorating the clinical course. Impaired binding, transport and delivery of drugs (e.g., antibiotics) may hinder specific treatment. Finally, the interaction of drugs and/or toxins with endogenous substances accumulating in liver failure may play an important pathogenic role as they compete for a reduced number of functional binding sites on the damaged albumin molecules. It should be noted that the fractions of irreversibly oxidized albumin (HNA2) observed in our cohorts of cirrhotic patients represent the highest values among all patient groups studied so far [29]. Nevertheless, apart from irreversibly oxidized HNA2, the major proportion of circulating albumin is present in the unoxidized HMA or reversibly oxidized HNA1 fractions which still may provide adequate binding function and are accessible to therapeutic interventions such as albumin dialysis [30]. With respect to prognosis, the short-term mortality of our cirrhosis cohorts (Table 1) lies within the range of recently published data from the CANONIC trial, which included a large multicenter cohort of 1343 cirrhotic patients hospitalized for acute decompensation, with overall 28-day mortality ranging between 7.5% and 17% and overall 90-day mortality between 19.8% and 27.6% at various centers [31]. Interestingly, analysis of the prognostic value of HNA2 with respect to short-term survival indicated a trend for superior diagnostic accuracy as compared to MELD, the current standard for assessment of 982
prognosis in chronic liver failure, which could be confirmed in the external validation set. However, the prognostic value of HNA2 has to be validated in larger cohorts of cirrhotic patients. Based on our findings, we conclude that impaired albumin site II binding capacity observed in chronic liver failure is mainly related to the severity of liver dysfunction. Concurrent oxidative stress seems to enhance the loss of albumin function in chronic liver failure but not in sepsis. Irreversible oxidation of albumin confers bad prognosis and thus may represent a novel biomarker for chronic liver failure.
Financial support This study was supported by the Franz Lanyar Stiftung (project # 314) and by BioPersMed (COMET K-project 825329), which is funded by the Austrian Federal Ministry of Transport, Innovation and Technology (BMVIT) and the Austrian Federal Ministry of Economics and Labour/the Federal Ministry of Economy, Family and Youth (BMWA/BMWFJ) and the Styrian Business Promotion Agency (SFG). V.S. was the recipient of an Erwin-Schrödinger fellowship (J2547) from the Austrian Science Foundation.
Conflict of interest The authors who have taken part in this study declared that they do not have anything to disclose regarding funding or conflict of interest with respect to this manuscript. Acknowledgements The technical assistance of Martina Brtnik, Doris Payerl, and Stefan Spoerk is gratefully acknowledged.
Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jhep.2013.06.013.
References [1] Rothschild MA, Oratz M, Schreiber SS. Serum albumin. Hepatology 1988;8:385–401. [2] Evans TW. Review article: albumin as a drug-biological effects of albumin unrelated to oncotic pressure. Aliment Pharmacol Ther 2002;16:6–11. [3] Quinlan GJ, Margarson MP, Mumby S, Evans TW, Gutteridge JM. Administration of albumin to patients with sepsis syndrome: a possible beneficial role in plasma thiol repletion. Clin Sci (Lond) 1998;95:459–465. [4] Sogami M, Nagoka S, Era S, Honda M, Noguchi K. Resolution of human mercapt- and nonmercaptalbumin by high-performance liquid chromatography. Int J Pept Protein Res 1984;24:96–103. [5] Hughes WL, Dintzis HM. Crystallization of the mercury dimers of human and bovine mercaptalbumin. J Biol Chem 1964;239:845–849. [6] Sen S, Williams R, Jalan R. The pathophysiological basis of acute-on-chronic liver failure. Liver 2002;22:5–13. [7] Sogami M, Era S, Nagaoka S, Kuwata K, Kida K, Shigemi J, et al. Highperformance liquid chromatographic studies on non-mercapt in equilibrium with mercapt conversion of human serum albumin. II. J Chromatogr 1985;332:19–27. [8] Watanabe A, Matsuzaki S, Moriwaki H, Suzuki K, Nishiguchi S. Problems in serum albumin measurement and clinical significance of albumin microheterogeneity in cirrhotics. Nutrition 2004;20:351–357.
Journal of Hepatology 2013 vol. 59 j 978–983
JOURNAL OF HEPATOLOGY [9] Oettl K, Stadlbauer V, Petter F, Greilberger J, Putz-Bankuti C, Hallstrom S, et al. Oxidative damage of albumin in advanced liver disease. Biochim Biophys Acta 2008;1782:469–473. [10] Sudlow G, Birkett DJ, Wade DN. The characterization of two specific drug binding sites on human serum albumin. Mol Pharmacol 1975;11:824–832. [11] Oettl K, Stauber RE. Physiological and pathological changes in the redox state of human serum albumin critically influence its binding properties. Br J Pharmacol 2007;151:580–590. [12] Meucci E, Mordente A, Martorana GE. Metal-catalyzed oxidation of human serum albumin: conformational and functional changes. Implications in protein aging. J Biol Chem 1991;266:4692–4699. [13] Klammt S, Mitzner S, Stange J, Brinkmann B, Drewelow B, Emmrich J, et al. Albumin-binding function is reduced in patients with decompensated cirrhosis and correlates inversely with severity of liver disease assessed by model for end-stage liver disease. Eur J Gastroenterol Hepatol 2007;19: 257–263. [14] Klammt S, Brinkmann B, Mitzner S, Munzert E, Loock J, Stange J, et al. Albumin binding capacity (ABiC) is reduced in commercially available human serum albumin preparations with stabilizers. Z Gastroenterol 2001;39:24–27. [15] Jalan R, Schnurr K, Mookerjee RP, Sen S, Cheshire L, Hodges S, et al. Alterations in the functional capacity of albumin in patients with decompensated cirrhosis is associated with increased mortality. Hepatology 2009;50:555–564. [16] Arroyo V. Review article: albumin in the treatment of liver diseases – new features of a classical treatment. Aliment Pharmacol Ther 2002;16:1–5. [17] Jalan R, Williams R. Acute-on-chronic liver failure: pathophysiological basis of therapeutic options. Blood Purif 2002;20:252–261. [18] Bone RC, Balk RA, Cerra FB, Dellinger RP, Fein AM, Knaus WA, et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/SCCM Consensus Conference Committee. American College of Chest Physicians/Society of Critical Care Medicine. Chest 1992;101:1644–1655. [19] Kamath PS, Wiesner RH, Malinchoc M, Kremers W, Therneau TM, Kosberg CL, et al. A model to predict survival in patients with end-stage liver disease. Hepatology 2001;33:464–470.
[20] Wiesner R, Edwards E, Freeman R, Harper A, Kim R, Kamath P, et al. United Network for Organ Sharing Liver Disease Severity Score Committee. Model for end-stage liver disease (MELD) and allocation of donor livers. Gastroenterology 2003;124:91–96. [21] Kawai K, Yoh M, Hayashi T, Imai H, Negawa T, Tomida M, et al. Effect of diabetic retinopathy on redox state of aqueous humor and serum albumin in patients with senile cataract. Tokai J Exp Clin Med 2001;26:93–99. [22] Watanabe H, Yamasaki K, Kragh-Hansen U, Tanase S, Harada K, Suenaga A, et al. In vitro and in vivo properties of recombinant human serum albumin from Pichia pastoris purified by a method of short processing time. Pharm Res 2001;18:1775–1781. [23] Hanley JA, McNeil BJ. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 1983;148:839–843. [24] Otagiri M. A molecular functional study on the interactions of drugs with plasma proteins. Drug Metab Pharmacokinet 2005;20:309–323. [25] Adachi Y, Murata N, Tsunasawa S, Kamisako T, Yamamoto T. Lysine 190 is a covalent bilirubin-binding amino acid in human delta bilirubin. Int Hepatol Commun 1996;5:282–288. [26] Gryzunov YA, Arroyo A, Vigne JL, Zhao Q, Tyurin VA, Hubel CA, et al. Binding of fatty acids facilitates oxidation of cysteine-34 and converts copperalbumin complexes from antioxidants to prooxidants. Arch Biochem Biophys 2003;413:53–66. [27] Vitek L, Ostrow JD. Bilirubin chemistry and metabolism; harmful and protective aspects. Curr Pharm Des 2009;15:2869–2883. [28] Zunszain PA, Ghuman J, McDonagh AF, Curry S. Crystallographic analysis of human serum albumin complexed with 4Z,15E-bilirubin-IXalpha. J Mol Biol 2008;381:394–406. [29] Oettl K, Marsche G. Redox state of human serum albumin in terms of cysteine-34 in health and disease. Methods Enzymol 2010;474:181–195. [30] Oettl K, Stadlbauer V, Krisper P, Stauber RE. Effect of extracorporeal liver support by molecular adsorbents recirculating system and Prometheus on redox state of albumin in acute-on-chronic liver failure. Ther Apher Dial 2009;13:431–436. [31] Moreau R, Jalan R, Gines P, Pavesi M, Angeli P, Cordoba J, et al. Acute-onchronic liver failure is a distinct syndrome that develops in patients with acute decompensation of cirrhosis. Gastroenterology 2013;144:1426–1437.
Journal of Hepatology 2013 vol. 59 j 978–983
983