Posttranscriptional changes of serum albumin: clinical and prognostic significance in hospitalized patients with cirrhosis.

Posttranscriptional Changes of Serum Albumin: Clinical and Prognostic Significance in Hospitalized Patients With Cirrhosis Marco Domenicali,1,2* Maurizio Baldassarre,1,2* Ferdinando A Giannone,1,2 Marina Naldi,3 Marianna Mastroroberto,1 Maurizio Biselli,1 Maristella Laggetta,1,2 Daniela Patrono,4 Carlo Bertucci,3 Mauro Bernardi,1,2 and Paolo Caraceni1,2 Beside the regulation of fluid distribution, human serum albumin (HSA) carries other activities, such as binding, transport, and detoxification of many molecules. In patients with cirrhosis, HSA exhibits posttranscriptional alterations that likely affect its functions. This study aimed at identifying the structural HSA alterations occurring in cirrhosis and determining their relationship with specific clinical complications and patient survival. One hundred sixty-eight patients with cirrhosis, 35 with stable conditions and 133 hospitalized for acute clinical complications, and 94 healthy controls were enrolled. Posttranscriptional HSA molecular changes were identified and quantified by using a high-performance liquid chromatography/electrospray ionization mass spectrometry technique. Clinical and biochemical parameters were also recorded and hospitalized patients were followed for up to 1 year. Seven HSA isoforms carrying one or more posttranscriptional changes were identified. Altered HSA isoforms were significantly more represented in patients than in healthy controls. Conversely, the native, unchanged HSA isoform was significantly reduced in cirrhosis. Native HSA and most altered isoforms correlated with both Child-Pugh and Model for End-Stage Liver Disease scores. In hospitalized patients, oxidized and N-terminal truncated isoforms were independently associated with ascites, renal impairment, and bacterial infection. Finally, the native HSA and cysteinylated/N-terminal truncated isoforms were predictors of 1year survival, with greater prognostic accuracy than total serum albumin concentration. Conclusions: Extensive posttranscriptional changes of HSA, involving several molecular sites and increasing in parallel with disease severity, occur in patients with cirrhosis. Altered isoforms are independently associated with specific clinical complications, whereas the residual, native HSA isoform independently predicts patient survival. These findings support the concept of the “effective albumin concentration,” which implies that the global HSA function is related not only to its serum concentration, but also to the preservation of its structural integrity. (HEPATOLOGY 2014;60:1851-1860)

Abbreviations: 2D, two-dimensional; ACLF, acute on chronic liver failure; ANOVA, analysis of variance; DM, diabetes mellitus; ESI-MS, electrospray ionization mass spectrometry; HCC, hepatocellular carcinoma; HPLC, high-performance liquid chromatography; HSA, native form of human serum albumin; HSA1CYS, cysteinylated human serum albumin; HSA1CYS-DA, cysteinylated and N-terminal truncated (-Asp-Ala) human serum albumin; HSA1CYS1GLYC, cysteinylated and glycosylated human serum albumin; HSA-DA, N-terminal truncated (-Asp-Ala) human serum albumin; HSA1GLYC, glycosylated human serum albumin; HSA-L, C-terminal truncated (-Leu) human serum albumin; HSA1SO2H, sulfinilated human serum albumin; LC, liquid chromatography; MELD, Model for End-Stage Liver Disease; Q-TOF, quadruple time of flight; ROC, receiver operating characteristic; SD, standard deviation; TIC, total ion current; US, ultrasound. From the 1Department of Medical and Surgical Sciences, Alma Mater Studiorum University of Bologna, Bologna, Italy; 2Center for Applied Biomedical Research (C.R.B.A.), S.Orsola-Malpighi University Hospital, Bologna, Bologna, Italy; 3Department of Pharmacology and Biotechnology, Alma Mater Studiorum University of Bologna, Bologna, Italy; and 4Centralized Laboratory, S. Orsola-Malpighi University Hospital, Bologna, Italy. Received December 20, 2013; accepted July 13, 2014. Additional Supporting Information may be found at onlinelibrary.wiley.com/doi/10.1002/hep.27322/suppinfo. Dr. Ferdinando A. Giannone has been, in part, supported by the AISF 2011 Annual Fellowship. *These authors contributed equally to this work. 1851

1852

DOMENICALI, BALDASSARRE, ET AL.

H

uman serum albumin (HSA) is the most abundant plasma protein, and its oncotic power largely determines the fluid distribution in the different compartments of the body. Besides its oncotic capacity, HSA presents other biological properties, such as antioxidant and scavenging activities,1 binding and transport of many endogenous and exogenous substances, and regulation of endothelial function and inflammatory response.1-4 These nononcotic properties are related to the dynamic HSA structure. Of the 35 cysteine residues, only that at position 34 (Cys-34) remains free, representing the main antioxidant site of the molecule.5 In healthy individuals, 70%-80% of HSA circulates as mercaptalbumin (HMA), characterized by a reduced Cys-34 with preserved antioxidant and scavenging activities, wheras 20%-30% presents the Cys-34 residue reversibly oxidized and bound to small thiol molecules (nonmercaptalbumin 1 [HNA1]).5 The remaining 5% of HSA circulates as nonmercaptalbumin 2 (HNA2), with the Cys-34 residue irreversibly oxidized, thus leading to the permanent loss of its function.5 Furthermore, antioxidant activity is also exerted by the N-terminal portion, which is able to chelate metal ions,6 whereas the C-terminal region contributes to the molecule stability.5 Under physiological conditions, the occurrence of low-level posttranscriptional changes, resulting from oxidation, glycosylation, and truncation involving the Cys-34 and other molecular sites, contributes to the microheterogeneity of the circulating molecule.6-10 In patients with advanced liver disease, HSA is immersed in a stressful microenvironment as a result of the elevated serum concentration of proinflammatory and pro-oxidant substances.2,11 Moreover, because diabetes is present in approximately one third of patients with cirrhosis, additional stress can also derive from hyperglycemia.12,13 Thus, besides the universal finding of hypoalbuminemia, a significant damage of the molecule can be expected in these patients. And, indeed, the reversible and irreversible oxidation of the Cys-34 residue occurs in advanced cirrhosis.14-17 Furthermore, the circulating level of ischemia modified albumin, which predominantly reflects the cobalt-

HEPATOLOGY, December 2014

chelating activity at the N-terminal portion,6 is significantly increased in patients with acute-on-chronic liver failure (ACLF).18 Despite these most interesting results, the information on structural changes also involving sites other than the Cys-34 residue, resulting from both oxidative and nonoxidative mechanisms, is lacking. High-performance liquid chromatography (HPLC), coupled to electrospray ionization mass spectrometry (ESI-MS), allows the identification of HSA isoforms carrying posttranscriptional changes (i.e., cysteinylation, sulfinylation, glycosylation, and truncation of N- or C-terminal portions), providing the advantage of the simultaneous recognition of multiple defects with high sensitivity and specificity. This allows discriminating such isoforms from the native, unchanged HSA molecule quite clearly.19,20 In the present study, we first analyzed the HSA posttranscriptional changes in a large cohort of patients with cirrhosis admitted to the hospital because of a clinical complication of the disease. Then, we assessed whether these HSA isoforms are associated with the severity of disease, specific clinical features, and patient prognosis.

Patients and Methods Patients From July 2011 to March 2012, all the patients with cirrhosis admitted to our department because of the onset of a complication were screened for the study enrolment. A group of patients with cirrhosis with stable clinical conditions attending our outpatient clinics for scheduled surveillance were also evaluated. The diagnosis of cirrhosis was based on clinical, biochemical, ultrasound (US), and endoscopic features. Exclusion criteria were age under 18 years, admission for a scheduled procedure, hepatocellular carcinoma (HCC) exceeding the Milan criteria,21 heart and respiratory failure, organic renal diseases, oncohematologic disorders, protein-losing syndromes, albumin infusion in the previous month, and ongoing immunosuppressive treatment. Healthy voluntaries served as the control population. Informed written consent was obtained from patients and healthy controls, and the study protocol

Address reprint requests to: Paolo Caraceni, M.D., Department of Medical and Surgical Sciences, Alma Mater Studiorum, University of Bologna, Via Massarenti 9, 40138 Bologna, Italy. E-mail: [email protected]; fax: 139-051-6362930. C 2014 by the American Association for the Study of Liver Diseases. Copyright V View this article online at wileyonlinelibrary.com. DOI 10.1002/hep.27322 Potential conflict of interest: Nothing to report.

HEPATOLOGY, Vol. 60, No. 6, 2014

was approved by the local ethical committee in accord with the 1975 Declaration of Helsinki. Study Design and Definitions At the time of inclusion for outpatients with cirrhosis and controls, and within 24 hours from admission for hospitalized patients, peripheral blood was withdrawn from the brachial vein into pyrogen-free tubes (Vacutainer EDTA tubes; Becton Dickinson Italia, Milan, Italy). Blood samples were immediately centrifuged at 3,0003g for 10 minutes, and plasma was aliquoted into cryotubes (Corning Inc., Corning BV, Amsterdam, The Netherlands) and stored at 280 C until analysis. A routine biochemical evaluation, including liver and renal function tests and coagulation parameters, was also performed, and the Model of End-stage Liver Disease (MELD)22 and Child-Pugh23 scores calculated to assess disease severity. Ascites was ascertained by US examination; hepatic encephalopathy was graded according to the WestHaven classification24; renal impairment was defined as a serum creatinine greater than 1.5 mg/dL; and upper gastrointestinal bleeding was confirmed by endoscopy. Bacterial infection was diagnosed as follows: (1) spontaneous bacterial peritonitis: polymorphonuclear cell count in ascites >250/mm3; (2) urinary infection: positive culture or >10 leukocytes per high-power field in urine associated with suggestive clinical symptoms and signs; (3) pneumonia: presence of new infiltrates on chest Xray; (4) skin/soft tissue infection: physical exam findings of swelling, erythema, and heat and tenderness in the skin; (5) spontaneous bacteremia: positive blood cultures and no evident cause of bacteremia; and (6) other infections (diverticulitis, colecystitis, and meningoencephalytis): according to specific findings at the laboratory, microbiological, and imaging assessment. Diabetes was diagnosed according to the American Diabetes Association guidelines.25 Patient survival was recorded during hospitalization and after discharge up to 1 year. Assessment of the HSA Isoforms HPLC/ESI-MS Measurements. Serum samples were diluted 1:10 with water/acetonitrile (98/2; v/v) mixture and filtered with an 0.22-mm filter (Merck KGaA, Darmstadt, Germany). A further dilution (1:100) with the same mixture was then performed before each injection. Nano-LC (liquid chromatography)/nano-ESI/QTOF (quadruple time-of-flight) analysis was carried out by using a nano-LC Agilent 1100 Series (Agilent, Walbronn, Germany) interfaced with a Q-TOF hybrid analyzer (Q-TOF Micro; Micromass, Manchester, UK) with a nano-Z-spray ion source. Reversed-phase chromatographic separations of HSA from other serum


1853

proteins was performed on a C8 (50 mm 3 75 lm; 3.5 lm) column, using an elution gradient from A (water: acetonitrile: FA [99:1:0.1] [v/v/v] / B (acetonitrile: water: FA [98:2:0.1] [v/v/v] 85/15 [v/v]) to A/B 20/80 (v/v), in 20 minutes, at a flow rate of 0.5 lL/ min; the system was equipped with an autosampler, and the injection volume was 1 lL. The column was equilibrated with the mobile-phase composition of the starting conditions for 10 minutes before the next injection. The ESI/Q-TOF source temperature was set at 100 C, the capillary voltage at 3.5 kV, and the cone voltage at 42 V. The scan time was set at 2.4 seconds and the interscan time at 0.1 sec. The mass chromatograms were recorded as total ion current (TIC), within 600 and 2,500 m/z (mass/charge). To characterize HSA isoforms by molecular-weight determination, multicharged mass spectra were acquired on the peak apex identified as HSA by Q-TOF. Deconvoluted ESI mass spectra of HSA were obtained by using MassLynx 4.1 software (Waters Corporation, Milford, MA). The peak averaged mass spectra were reconstructed and the mass of the HSA isoforms derived. HSA Isoforms Identification and Relative Quantification. HSA isoforms were identified as previously reported.19 Namely, by analyzing the 65.800-67.100-Da mass range, it is possible to identify, in addition to native HSA isoform, all potential posttranscriptional alterations. Isoform relative abundance was calculated according to the following formula: isoform absolute intensity 3100 sum of all isoforms absolute intnsicty where isoform absolute intensity is represented by the TIC generated by each isoforms, including the native HSA isoform. To confirm the correct identification of HSA structural alterations by HPLC/ESI-MS, plasma protein digestion followed by two-dimensional (2D) LC tandem MS analysis was performed (see Methods section of the Supporting Information). The repeatability and reproducibility of HPLC/ESI-MS has been previously assessed.19 Namely, the relative standard deviations (SDs) of the intra- and interday precision were lower than 0.01%. Statistical Analysis Variables were expressed as mean and SD or frequencies according to their distribution. Comparisons between categorical variables were made by means of the chi-square (v2) test. Differences in HSA isoform relative abundance between control subjects, outpatients, and hospitalized patients with cirrhosis were assessed by one-way analysis of variance

1854


(ANOVA) with Bonferroni’s correction for multiple comparisons. The relationship between relative abundance of HSA isoforms and MELD and Child-Pugh scores was evaluated with Spearman’s rho. The univariate analysis of the association between HSA isoforms relative amount, age, MELD and Child-Pugh scores, and specific clinical complication (ascites, renal impairment, and bacterial infection) in hospitalized patients was assessed by means of the Student t test. Any variable showing a P value 1.5 mg/dL. † Prothrombin time international normalized ratio.

tion of the last two aminoacid residues at the Nterminal portion (HSA-DA); truncation of the last aminoacid residue at the C-terminal portion (HSAL); cysteinylation of the Cys-34 residue (HSA1CYS); sulfinylation of the Cys-34 residue (HSA1SO2H); and glycosylation (HSA1GLYC). Furthermore, two additional HSA isoforms were generated from the combination of the cysteinylated with the N-terminal truncated form (HSA1CYSDA) or the glycosylated form (HSA1CYS1GLYC). Therefore, in addition to the native HSA isoform, we were able to identify seven isoforms with one or two alterations (Table 2). Although no differences in the relative isoform abundances were observed between genders (Supporting Table 1), age was negatively correlated with the relative abundance of the native unchanged HSA isoform and positively correlated with the reversibly oxidized isoforms, namely, HSA1CYS and HSA1CYS1GLIC. We have also found a negative correlation between age and HSA-DA and HSA1SO2H (Supporting Table 2).



1855

Table 3. Relative Abundance of HSA Isoforms in Healthy Controls, Outpatients, and Hospitalized Patients With Cirrhosis

Isoform

HSA-DA (%) HSA-L (%) HSA1CYS-DA (%) Native HSA (%) HSA1SO2H (%) HSA1CYS (%) HSA1GLYC (%) HSA1CYS1GLYC (%)

Healthy Controls (n 5 67)

2.5 3.1 1.9 52.4 7.3 19.9 9.1 3.8

6 6 6 6 6 6 6 6

0.7 1.2 0.3 2.6 1.6 4.4 1.3 0.8

Outpatients With Cirrhosis (n 5 35)

2.4 2.4 2.3 45.9 8.0 24.3 9.6 5.1

6 6 6 6 6 6 6 6

0.7 1.3 0.4* 5.6* 1.1 7.3* 2.6 1.6*

Hospitalized Patients With Cirrhosis (n 5 133)

2.3 2.3 2.1 43.2 7.9 26.0 9.8 6.3

6 6 6 6 6 6 6 6

0.9 1.9 0.8* 6.2* 2.4 6.3* 3.1* 2.1*†

Data are presented as mean 6 SD. Each isoform relative abundance was calculated as percent over total HSA isoforms; ANOVA posthoc analysis. *P < 0.05 versus healthy controls. † P < 0.05 versus outpatients with cirrhosis.

Fig. 1. Representative deconvoluted ESI-MS spectra from a control subject (A) and a patient with cirrhosis (B). In addition to native HSA, seven HSA isoforms carrying the following structural alterations were detected: truncation of the last two amino acid residues at the Nterminal portion (HSA-DA); truncation of the last amino acid residue at the C-terminal portion (HSA-L); cysteinylation of the Cys-34 residue (HSA1CYS); sulfinylation of the Cys-34 residue (HSA1SO2H); and glycosylation (HSA1GLYC). Two additional HSA isoforms were generated from the combination of the cysteinylated with the N-terminal truncated form (HSA1CYS-DA) or the glycosylated form (HSA1CYS1GLYC).

Because of the described relationships between age and the relative HSA isoform abundances, we selected 67 healthy subjects with the same age distribution (mean age: 57 6 9 years; 39 [58%] male) of both groups of patients with cirrhosis to perform statistical comparisons. In this healthy subject subgroup, the lack of differences between genders was confirmed.

Relative Abundance of HSA Isoforms in Patients With Cirrhosis The same HSA isoforms identified in plasma samples from healthy subjects were detected in patients with cirrhosis (Fig. 1B). The analysis of the mass spectra revealed that the extent of most structural alterations were significantly greater than healthy controls (Table 3). The relatively most abundant HSA isoforms observed in patients were represented by the cysteinylated (HSA1CYS, HSA1CYS-DA, and HSA1CYS1GLYC) and glycosylated (HSA1GLYC) isoforms, whereas no differences were observed in the relative abundance of the sulfinylated (HSA1SO2H), N-terminal truncated (HSA-DA), and C-terminal truncated (HSA-L) isoforms. As a result, the relative abundance of the native, unchanged HSA isoform was significantly reduced in patients with cirrhosis (Fig. 2). Although the extent of HSA alterations tended to be higher in hospitalized patients with cirrhosis, the differences between out- and inpatients did not reach statistical significance, with the exception of the HSA1CYS1GLYC isoform (Table 3; Fig. 2).

Table 2. HSA Isoforms Identified in Plasma Samples: Molecular Weight, Structural Alteration, and Functional Consequences MW (Da)

Isoforms

2180 HSA-DA 2110 260 66439 130 1120 1160 1280

Structural Alteration

N-terminal truncated (-Asp-Ala)

Functional Consequence

Reference(s)

Loss of chelating function, loss of free radical Bar-Or D. et al. scavenging activity HSA-L C-terminal truncated (-Leu) Impaired protein stability, shorter half-life Fanali G. et al. HSA1CYS-DA Cysteinylated and N-terminal truncated Loss of antioxidant activity, loss of chelating function Bar-Or D. et al.; Carballal S. et al. (-Asp-Ala) Native HSA Native Form of HSA With Fully Preserved Nononcotic Activity HSA1SO2H Sulfinylated Irreversible oxidation of Cys-34 residue Fanali G. et al.; Carballal S. et al. HSA1CYS Cysteinylated Loss of antioxidant activity Bar-Or D. et al.; Carballal S. et al. HSA1GLY Glycosylated Impaired binding activity, conformational alteration Rondeau P. and Bourdon E. HSA1CYS1GLYCysteinylated and glycosylated Impaired antioxidant and binding activity, conformational alteration Rondeau P. and Bourdon E.; Bar-Or D. et al.; Carballal S. et al.

1856



Fig. 2. Individual values and box plots representing mean, range, and 95% confidence interval of the relative amount of native HSA isoform in healthy controls, outpatients, and hospitalized patients with cirrhosis.

Relationships Between HSA Isoforms and Disease Severity Although no correlations were found between HSADA, HSA-L, and disease severity, all the HSA isoforms with structural changes at the Cys-34 (HSA1CYS, HSA1CYS-DA, HSA1CYS-GLYC, and HSA1SO2H) directly correlated with the MELD score (Table 4). The same was true with the ChildPugh score, with the exception of HSA1CYS and HSA1CYS1GLYC isoforms. Contrariwise, the correlation between HSA1GLYC isoform and MELD or Child-Pugh scores was inverse. As a result of the above abnormalities, a strong negative correlation was found between the relative abundance of native HSA isoform and both prognostic scores, indicating that the more severe cirrhosis is, the lower the circulating HSA is with a preserved molecular structure.

Table 4. Correlation Between Relative Abundance of HSA Isoforms and MELD and Child-Pugh Scores in the Whole Population of Patients With Cirrhosis* MELD

HSA-DA (%) HSA-L (%) HSA1CYS-DA (%) Native HSA (%) HSA1SO2H (%) HSA1CYS (%) HSA1GLYC (%) HSA1CYS1GLYC (%) *n 5 168.

Correlation Coefficient

0.045 20.040 0.266 20.318 0.259 0.219 20.220 0.181

Child-Pugh P Value

Correlation Coefficient

P Value

0.584 0.627 0.001 44.38% had a significantly longer survival than their counterparts with lower values (10.30 6 0.48 vs. 8.55 6 0.50 months; P 5 0.002; Fig. 3B). Finally, we also categorized the serum albumin concentration according to its best cutoff, as described above. Patients with serum albumin concentration 4.05 g/dL had a worse survival than those with higher values (10.16 6 0.65 vs. 9.06 6 0.43 months; P 5 0.063), but, contrary to HSA isoforms, the difference did not reach statistical significance (Fig. 3C).

Discussion In this study, posttranscriptional structural changes of albumin in cirrhosis were assessed by HPLC/ESIMS for the first time. This proteomic technique provides a spectrum of molecules with high resolution, discriminating between isoforms whose molecular

weight differs by a few Daltons. Moreover, this approach allows the simultaneous determination of all the changes occurring in the molecule, which is of particular relevance in the case of circulating HSA that is characterized by a high microheterogeneity.19,20 The reliability of this technique has been demonstrated by us and others in previous studies.19,20,26 Moreover, the identification of HSA isoforms by plasma protein digestion followed by 2D LC, which we performed in the present study, confirmed the nature of HSA structural alterations identified by HPLC/ESI-MS, thus providing evidence for its validity. This technique carries important advantages on those used in previous studies in patients with cirrhosis, when only the structural alterations at the Cys-34 site were addressed.14-17 Indeed, HPLC/ESI-MS enabled us to identify and quantify the relative proportion of both the native, unmodified HSA and the different HSA molecular isoforms resulting from posttranscriptional structural changes not only occurring at the Cys-34 residue, but also at the other sites of the molecule.19,20 Moreover, the higher throughput performance of HPLC/ESI-MS allows to envision its use in the clinical laboratory setting. Despite these additional benefits, some limitations of this technique deserve comments. First, in order to investigate all potential albumin posttranscriptional alterations, we focused our attention on the mass range of 65.800-67.100 Da. This range is around the expected native HSA molecular weight (66.439 Da) and includes well-characterized19,20 HSA isoforms— truncated as well as oxidized—which account for almost all circulating HSA. Even though we could not exclude the existence of additional HSA isoforms with molecular weight outside this mass range, it should not constitute a significant amount. Second, in the mass range we analyzed, some peaks representing not typified isoforms were observed in addition to the eight wellcharacterized isoforms. These peaks were excluded from analysis because they were poorly represented in both

1858


Fig. 3. Kaplan-Meier’s survival curves for HSA1CYS-DA (A), native HSA (B), and serum albumin concentration (C) in hospitalized patients with cirrhosis dichotomized according their best cut-off values assessed by ROC curve analysis.


controls and patients with cirrhosis. Finally, HPLC/ESIMS does not allow, at present, to assess the plasma concentration of the HSA isoforms. However, this does not hinder to assess the effect of their relative abundance on the variables evaluated in this study. The first important finding of the present study is that the molecular structure of HSA is extensively altered in patients with cirrhosis. Indeed, the extent of many posttranscriptional molecular changes, which can also be detected at a low, physiological level in healthy subjects, was significantly increased in patients. Cysteinylation was the more frequent alteration, which can occur alone or in combination with other molecular changes, including truncation of N-terminal portion and glycosylation.19,20 In a pro-oxidant environment, as it develops in the plasma of patients with advanced cirrhosis, circulating cysteine is oxidized to cysteine disulfide, which interacts with the free sulfhydryl group at the albumin Cys-34 site undergoing a disulfide exchange.27,28 As a result of this reversible oxidation, a cysteinylated protein takes form. Because reduced Cys-34 accounts for the vast majority of the total plasma free sulfhydryl groups, a significant impairment of the circulating antioxidant system can be expected in the case of extensive oxidation.27 This likely worsens in parallel with the severity of cirrhosis. In fact, we found that these alterations were more evident in hospitalized patients with complications than in outpatients with stable clinical conditions and were significantly correlated with markers of disease severity, such as Child-Pugh and MELD scores. Such relationships were observed with all the HSA isoforms with oxidized Cys-34, alone or in combination with other defects, with the exception of glycosylated HSA, a finding that will be discussed later. Conversely, an inverse correlation was noted by plotting Child-Pugh and MELD scores versus native HSA. The sulfinylated HSA, resulting from the irreversible oxidation of the Cys-34 residue,27 was also increased in patients, even though the difference with healthy controls did not reach the statistical significance. This may appear at variance with previous studies showing a significant elevation of HNA2 in cirrhosis.15-17 However, it has to be outlined that many patients enrolled in the present study had a less-severe disease than those included in the above-mentioned studies, where most cases had ACLF.1517 In fact, the relative proportion of the sulfinylated isoform correlated with the markers of disease severity. Thus, the irreversible oxidation of HSA could be considered a late event in the course of cirrhosis. Glycosylated HSA also significantly increased in patients, but this was a result of the fact that


approximately one third of cases had DM. In fact, whereas HSA1GLYC was only increased in diabetic patients, the HSA1CYS1GLYC, an isoform also carrying the oxidation of the Cys-34 residue, was also higher in nondiabetic patients. It is well known that sustained hyperglycemia produces a nonenzymatic glycosylation of several circulating proteins, including HSA.13 This process affects both structure and function of the molecule, impairing the binding and detoxification capacity of HSA.8 As an example, the capacity of binding bilirubin of glycosylated HSA is reduced by up to 50%.29 The finding that HSA1GLYC was inversely correlated with the markers of disease severity is at variance with respect to what we found with most HSA isoforms and suggests that diabetes, rather than the progression of cirrhosis, favors this defect. Another novel finding of the present study is the association existing between specific HSA isoforms and the presence of complications of cirrhosis. In fact, HSA1CYS-DA was associated with the presence of ascites and bacterial infection, whereas HSA-SO2H and HSA1CYS was associated with renal impairment. Interestingly, multivariate analyses showed that these associations were independent from Child-Pugh and MELD scores, suggesting that they did not merely reflect the severity of cirrhosis. The exact mechanisms underlying these associations cannot be ascertained by the present study. However, it is reasonable to hypothesize that the pro-oxidant and -inflammatory status occurring during these complications, characterized by increased protease activity, proinflammatory cytokine release, and free radical generation,30,31 may favor the multiple alterations of HSA demonstrated by this study. Given the striking extent of the HSA molecular damages observed in patients with cirrhosis, it emerges quite clearly that the amount of structural preserved “healthy” HSA is far lower than the actual albumin serum concentration. Becasue altered HSA isoforms increase in parallel with the progression of cirrhosis, it means that, in patients with an advanced disease, the antioxidant, scavenging, and binding activities of albumin are dramatically reduced. Thus, it is not surprising that HSA1CYS-DA isoform was a strong predictor of 1-year survival. Interestingly, the residual proportion of the native, unmodified HSA, which was not associated with any specific complication, but mirrors all the occurring molecular abnormalities, was also a survival predictor, far better than total serum albumin concentration, an established prognostic factor in patients with cirrhosis. It should be noted that the


1859

prognostic power of serum albumin concentration has usually been assessed over longer periods than the 1year follow-up of our study.32 Based on our results in hospitalized patients, it would appear that the native HSA isoform more precisely predicts survival in the medium term, as it is mostly influenced by the occurrence of complications, rather than by reduced liver synthetic activity per se. Thus, at least in the clinical context evaluated in the present study, the assessment of native HSA isoform would provide an incremental prognostic value over total albumin concentration. In conclusion, this study demonstrated that extensive posttranscriptional changes in the albumin molecule occur in patients with cirrhosis. These are not limited to the Cys-34 residue, but also involve several other sites of the molecule and increase in parallel of the severity of the disease. HSA alterations are independently associated with major complications of cirrhosis, such as ascites, renal impairment, and bacterial infection, likely contributing to their ominous prognosis. In fact, residual native unchanged HSA emerged as a more potent predictor of survival than total serum albumin concentration. This finding, far from being intended to identify a new prognostic indicator, strengthens the concept of “effective albumin concentration,”2,33 which implies that the global functions of albumin, resulting from both oncotic and nononcotic properties, are not only related to its absolute circulating level, but also, and perhaps mainly, to the preservation of its structural integrity.

References 1. Anraku M, Chuang VTG, Maruyama T, Otagiri M. Redox properties of serum albumin. Biochim Biophys Acta 2013;1830:5465-5472. 2. Garcia-Martinez R, Caraceni P, Bernardi M, Gines P, Arroyo V, Jalan R. Albumin: pathophysiologic basis of its role in the treatment of cirrhosis and its complications. HEPATOLOGY 2013;:1836-1846. 3. Evans T. Review article: albumin as a drug—biological effects of albumin unrelated to oncotic pressure. Aliment Pharmacol Ther 2002;16:611. 4. Fanali G, Di Masi A, Trezza V, Marino M, Fasano M, Ascenzi P. Human serum albumin: from bench to bedside. Mol Aspects Med 2012;33:209-290. 5. Peters T. All About Albumin: Biochemistry, Genetics, and Medical Applications. San Diego, CA: Academic; 1996. 6. Bar-Or D, Curtis G, Rao N, Bampos N, Lau E. Characterization of the Co(21) and Ni(21) binding amino-acid residues of the Nterminus of human albumin. An insight into the mechanism of a new assay for myocardial ischemia. Eur J Biochem 2001;268:42-47. 7. Carballal S, Alvarez B, Turell L, Botti H, Freeman BA, Radi R. Sulfenic acid in human serum albumin. Amino Acids 2007;32:543-551. 8. Rondeau P, Bourdon E. The glycation of albumin: structural and functional impacts. Biochimie 2011;93:645-658. 9. Chan B, Dodsworth N, Woodrow J, Tucker A, Harris R. Site-specific N-terminal auto-degradation of human serum albumin. Eur J Biochem 1995;227:524-528.

1860


10. Kawakami A, Kubota K, Yamada N, Tagami U, Takehana K, Sonaka I, et al. Identification and characterization of oxidized human serum albumin. A slight structural change impairs its ligand-binding and antioxidant functions. FEBS J 2006;273:3346-3357. 11. 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. 12. Gentile S, Loguercio C, Marmo R, Carbone L, Del Vecchio Blanco C. Incidence of altered glucose tolerance in liver cirrhosis. Diabetes Res Clin Pract 1993;22:37-44. 13. Iberg N, Fl€uckiger R. Nonenzymatic glycosylation of albumin in vivo. Identification of multiple glycosylated sites. J Biol Chem 1986;261: 13542-13545. 14. 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. 15. Oettl K, Stadlbauer V, Petter F, Greilberger J, Putz-Bankuti C, Hallstr€om S, et al. Oxidative damage of albumin in advanced liver disease. Biochim Biophys Acta 2008;1782:469-473. 16. Stauber RE, Spindelboeck W, Haas J, Putz-Bankuti C, Stadlbauer V, Lackner C, et al. Human nonmercaptalbumin-2: a novel prognostic marker in chronic liver failure. Ther Apher Dial 2014;18:74-78. 17. Oettl K, Birner-Gruenberger R, Spindelboeck W, Stueger HP, Dorn L, Stadlbauer V, et al. Oxidative albumin damage in chronic liver failure: relation to albumin binding capacity, liver dysfunction and survival. J Hepatol 2013;59:978-983. 18. 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. 19. Naldi M, Giannone FA, Baldassarre M, Domenicali M, Caraceni P, Bernardi M, Bertucci C. A fast and validated mass spectrometry method for the evaluation of human serum albumin structural modifications in the clinical field. Eur J Mass Spectrom (Chichester, Eng) 2013;19:491-496. 20. Bar-Or D, Bar-Or R, Rael LT, Gardner DK, Slone DS, Craun ML. Heterogeneity and oxidation status of commercial human albumin preparations in clinical use. Crit Care Med 2005;33:1638-1641. 21. Mazzaferro V, Regalia E, Doci R, Andreola S, Pulvirenti A, Bozzetti F, et al. Liver transplantation for the treatment of small hepatocellular carcinomas in patients with cirrhosis. N Engl J Med 1996;334:693-699. 22. Kamath PS, Wiesner RH, Malinchoc M, Kremers W, Therneau TM, Kosberg CL, et al. A model to predict survival in patients with endstage liver disease. HEPATOLOGY 2001;33:464-470.


23. Pugh RN, Murray-Lyon IM, Dawson JL, Pietroni MC, Williams R. Transection of the oesophagus for bleeding oesophageal varices. Br J Surg 1973;60:646-649. 24. Ferenci P, Lockwood A, Mullen K, Tarter R, Weissenborn K, Blei AT. Hepatic encephalopathy—definition, nomenclature, diagnosis, and quantification: final report of the working party at the 11th World Congresses of Gastroenterology, Vienna, 1998. HEPATOLOGY 2002;35: 716-721. 25. American Diabetes Association. Standards of medical care in diabetes— 2012. Diabetes Care 2012;35(Suppl. 1):S11-S63. 26. Kleinova M, Belgacem O, Pock K, Rizzi A, Buchacher A, Allmaier G. Characterization of cysteinylation of pharmaceutical-grade human serum albumin by electrospray ionization mass spectrometry and lowenergy collision-induced dissociation tandem mass spectrometry. Rapid Commun Mass Spectrom 2005;19:2965-2973. 27. Colombo G, Clerici M, Giustarini D, Rossi R, Milzani A, DalleDonne I. Redox albuminomics: oxidized albumin in human diseases. Antioxid Redox Signal 2012;17:1515-1527. 28. Carballal S, Radi R, Kirk MC, Barnes S, Freeman BA, Alvarez B. Sulfenic acid formation in human serum albumin by hydrogen peroxide and peroxynitrite. Biochemistry 2003;42:9906-9914. 29. Shaklai N, Garlick RL, Bunn HF. Nonenzymatic glycosylation of human serum albumin alters its conformation and function. J Biol Chem 1984;259:3812-3817. 30. Sen S, Williams R, Jalan R. The pathophysiological basis of acute-onchronic liver failure. Liver 2002;22(Suppl. 2):5-13. 31. Giron-Gonzalez JA, Martınez-Sierra C, Rodriguez-Ramos C, Macıas MA, Rendon P, Dıaz F, et al. Implication of inflammation-related cytokines in the natural history of liver cirrhosis. Liver Int 2004;24:437-445. 32. D’Amico G, Garcia-Tsao G, Pagliaro L. Natural history and prognostic indicators of survival in cirrhosis: a systematic review of 118 studies. J Hepatol 2006;44:217-231. 33. Jalan R, Bernardi M. Effective albumin concentration and cirrhosis mortality: from concept to reality. J Hepatol 2013;59:918-920.

Author names in bold designate shared co-first authorship.

Supporting Information Additional Supporting Information may be found at onlinelibrary.wiley.com/doi/10.1002/hep.27322/suppinfo.

Ischaemia-modified albumin: a marker of bacterial infection in hospitalized patients with cirrhosis.

Decreasing mortality in patients hospitalized with cirrhosis.

Dynamic changes and clinical significance of serum tryptase levels in STEMI patients treated with primary PCI.

Human serum albumin, systemic inflammation, and cirrhosis.

[Prognostic value of serum homocysteine levels in elderly hospitalized patients].

Serum Ammonia in Associated With Transplant-free Survival in Hospitalized Patients With Acutely Decompensated Cirrhosis.

Gastric clearance of serum albumin in patients with alcoholic liver cirrhosis.

Serum C-reactive protein predicts early mortality in hospitalized patients with HBV-related decompensated cirrhosis.

globulin ratio in patients with hepatocellular carcinoma.

Clinical significance of the serum levels of the 7S domain of type IV collagen in patients with primary biliary cirrhosis.

Prognostic significance of nonprotein respiratory quotient in patients with liver cirrhosis.

Prognostic Significance of Preoperative Albumin-Globulin Ratio in Patients with Upper Tract Urothelial Carcinoma.

Serum Albumin Can Identify Patients With Compensated Cirrhosis With a Good Prognosis.

The prevalence and clinical significance of anemia in patients hospitalized with acute heart failure.

The prevalence and clinical significance of anemia in patients hospitalized with acute heart failure.

Significance of technetium-99m human serum albumin diethylenetriamine pentaacetic acid scintigraphy in patients with nephrotic syndrome.

Prognostic Significance of 1-Year Serum Albumin Levels Within the Normal Range After Kidney Transplantation.

Clinical and prognostic significance of serum transforming growth factor-beta1 levels in patients with pancreatic ductal adenocarcinoma.

Prognostic significance of elevated troponin in non-cardiac hospitalized patients: a systematic review and meta-analysis.

Clinical Significance of Preoperative Albumin and Globulin Ratio in Patients with Gastric Cancer Undergoing Treatment.

Clinical characteristics and prognostic impact of bacterial infection in hospitalized patients with alcoholic liver disease.

Prognostic significance of serum galectin-3 levels in patients with hepatocellular cancer and chronic viral hepatitis.

Prognostic significance of serum squamous cell carcinoma antigen in patients with head and neck cancer.

Clinical significance of serum CA15-3 as a prognostic parameter during follow-up periods in patients with breast cancer.