Clinically, Human Serum Albumin (HSA) is used for the restoration of blood volume, emergency treatment of shock, acute management of burns, and other situations associated with hypovolemia 1 . Many studies have focused on the use of albumin as a resuscitation fluid and its effects on patient outcomes in Intensive Care Units (ICU). Some of them have revealed that use of either albumin or normal saline for resuscitation fluid results in similar outcomes 2 . Others have shown negative impacts: for example, in a prospective cohort study, hyperoncotic albumin infusion was associated with harmful effects on renal function and outcome 3 . Conversely, other studies have demonstrated positive effects or beneficial trends of albumin infusion 4 . In a large controlled, randomized trial, the administration of albumin may have decreased the risk of death in patients with severe sepsis compared with saline 5 . Furthermore, a randomized, controlled trial in patients with cirrhosis and spontaneous bacterial peritonitis has shown that the addition of an albumin infusion to an antibiotic treatment significantly reduced the incidence of renal impairment and death 6 . The several meta-analyses assessing the effectiveness of albumin solutions for resuscitation in critically i1l patients failed to clarify this issue 7 8 9 10 . In this context, the choice of a resuscitation fluid in ICU patients is still a conflicting matter under debate, and it is not the scope of the present review to provide an answer to this issue. Our objective is rather to highlight the mechanisms by which albumin might exert its potential beneficial effects in the critical care and hepatological settings.

Among the hypotheses that have been proposed to explain the positive effects of albumin, its antioxidant properties seem to be of paramount importance 11 12 13 14 . This hypothesis is supported by the central role of oxidative stress in critical pathologies, such as sepsis or liver failure 1 15 . Indeed, the pathophysiology of sepsis may be characterized by the negative role of free radicals during the onset, progression, and outcome of sepsis 16 17 . Reactive Oxygen Species (ROS) and Reactive Nitrogen Species (RNS) exert their detrimental effects at least partially, through endothelial dysfunction with alterations of vascular tone, increased cell adhesion, and vascular permeability 18 19 . The activation of endothelial cells contributes to maintain the oxidant-rich environment at the inflammatory locus 20 . Moreover, high levels of ROS and RNS can result in organ damages 21 , the ROS amount being related to the severity of sepsis and mortality 20 . In the same way, oxidative stress is very high in cirrhotic patients and plays an important role in the pathophysiological mechanisms involved in the hemodynamic disturbances observed 16 22 . Indeed, excessive systemic Nitric Oxide (NO) production is involved in peripheral vasoplegia, which induces portal hypertension, an important complication of cirrhosis 23 . The detrimental role of oxidative stress occurs during both the onset of hepatic alterations and during progression stages of the disease, in correlation with the severity of cirrhosis 24 .

The detrimental role of oxidative damage in critical pathologies, added to the demonstration of potential positive effects of albumin infusion in these patients, constitute a strong rationale that justifies for addressing the specific antioxidant capacities of HSA 11 12 13 14 . Our review summarizes the state of the art on this topic and focuses on the antioxidant properties of HSA related to (i) first its native form, (ii) the chemically modified-HSA (iii) the commercial HSA. The implication of this recent literature for future clinical trials using albumin as a drug is finally discussed.

Albumin is a non-glycosylated protein of 66 kDa 25 26 ; its normal plasma concentration is between 35 and 50 g/l constituting up to 60% of total plasma proteins 27 . Its half-life is ~ 20 days in normal conditions. The HSA structure consists of a single-chain polypeptide of 585 amino acid residues and approximately 67% alpha-helix and no beta-sheet 28 29 . Native HSA contains 6 methionines and 35 cysteine residues involved in the formation of 17 disulfide bonds. The Cys-34 residue is the only free cysteine in the whole molecule. HSA exerts specific antioxidant functions due to its multiple ligand-binding capacities and free radical-trapping properties, both closely related to its structure 16 29 .

The various structural modifications that HSA can undergo during its in vivo lifetime or during the processes employed to isolate therapeutic HSA from plasma, modify not only its conformation and hence its binding properties but also its redox state 16 50 51 . Kawakami et al. have clearly shown that reduced HSA and oxidized HSA have different ligand-binding properties 50 . These authors also have investigated the potential effect of oxidation on the antioxidant capacity of HSA by comparing the radical scavenging activities of HSA in various states of oxidation 50 . By using electron spin resonance, they have observed that the hydroxyl radical signal is reduced from 68.7% with highly oxidized cysteinylated HSA (proportion of HSA reduced = 8%) to 54.4% with non-oxidized HSA (proportion of HSA reduced = 73%). They have then demonstrated that the radical scavenging activity of reduced HSA is greater than that of cysteinylated HSA 50 . Those results are in accordance with other reports establishing that oxidized HSA decreases scavenging ability against highly ROS (hydroxyl radicals). Similarly, HSA nitrosylation leads to a significant loss of its buffering capacity 52 . Moreover, the loss of aspartate-alanine from the N-terminus of albumin completely abolishes the ability of albumin to chelate free copper and is associated with the loss of its free radical scavenging ability 52 . Finally, Iwao et al. (2006) demonstrated that oxidation of several amino acids of albumin had an impact on its pharmacokinetics by decreasing its half-life 53 .

Beyond oxidation and nitrosylation, HSA may undergo other chemical modifications affecting its structure, which is closely related to its specific antioxidant properties 27 . One of them is a non-enzymatic glycation consisting in the attachment of free carbohydrate (glucose, galactose, fructose…) on an amine residue leading to the formation of a stable fructosamine residue 54 . Accumulation of numerous studies has permitted to identify up to 29 sites that can be glycated in vivo in the HSA 34 . Due to their high nucleophile properties, lysine, arginine and cysteine are the main residues prone to glycation and the two principal sites are Arg410 and Lys525 55 56 . Glycation-induced modifications have an important impact on HSA functional properties mainly related to alteration of its conformation. These conformational changes affect HSA binding properties because affinity of glycated HSA for different ligands (long chain fatty acids, bilirubin, copper) undergoes an important decrease 57 . Through, these various processes, the antioxidant capacities of glycated HSA are dramatically reduced 54 58 59 .

In physiological conditions, due to the large amount of HSA in plasma, impairments in the protein molecule and its antioxidant properties have been considered as biologically insignificant 31 . However, antioxidant properties of modified HSA may be related to pathological conditions, in particular in septic and cirrhotic conditions in which the decline of HSA plasma antioxidant force also is mediated by the quantitative reduction of the HSA concentration 16 20 28 60 .

The finding that HSA exerts a plasma antioxidant force also is important regarding its possible therapeutic effects, because it may provide the opportunity to enhance endogenous antioxidant protection in pathological conditions by HSA infusion 61 . Commercially available HSA solutions are complex products that contain not only native HSA but also various species of HSA under different redox states, as well as several minor degradation products 46 62 . Significant HSA variability has been reported between the commercial HSA solutions (Table 1). Bioprocesses and storage conditions increase heterogeneity of HSA: truncated, cysteinylated, nitrosylated, and glycosylated forms or HSA-HSA dimers may be found in commercialized products 29 . These frequently encountered modifications can change HSA antioxidant properties and its binding capacity to endogenous or exogenous molecules 29 51 63 . The observed heterogeneity also can potentially influence clinical outcome and should be correlated with morbidity/mortality in randomized trials 47 63 64 . The use of an antioxidant for the treatment of sepsis has been considered a new, interesting, adjunctive therapy and constitutes a challenge in the clinical management of these patients 17 . However, it is still unclear whether the administration of commercially available HSA that is largely oxidized, increases or decreases oxidative stress burden in critically ill patients 52 .

Until now, several analytical techniques have been employed to monitor the heterogeneity or the degradation profiles of HSA. High-Performance Liquid Chromatography (HPLC), Mass Spectrometry (MS), Sodium Dodecyl Sulfate-PolyAcrylamide Gel Electrophoresis (SDS-PAGE) and Western blot analyses, or Capillary Zone Electrophoresis (CZE) have contributed to point out this heterogeneity of commercial HSA 51 62 . Aggregation or chemical degradation can be induced during sample preparation or purification. Procedures, such as temperature variations, freeze-thawing process, mechanical agitation, or lyophilization, may affect HSA structure and composition of therapeutic HSA. Anraku et al. studied the protective effect of sodium N-acetyl-L-tryptophanate against albumin oxidation, using HPLC. They partially resolved mercaptalbumin (reduced form of HSA) from two populations of non-mercaptalbumin (oxidized forms), demonstrating that it is possible to improve the quality of the solutions 65 . Ogasawara et al. used SDS-PAGE and Western blot analyses to detect albumin disulfide dimers in plasma, considered to be a biomarker of oxidative stress 66 . More recently, Qian et al. (2008) reported a size exclusion HPLC method to estimate the proportion of HSA dimers and oligomers suitable for a quality control 67 . MS and MS coupled to HPLC also have been employed to characterize different variants of HSA in commercial HSA preparations. In particular, six related proteins have been identified but the method provided only qualitative data and did not detect any dimer form of HSA 52 . Finally, Alahmad et al. developed a reproducible CZE method to separate HSA from most of its variants. This method proved to be useful in detecting quantitative differences in the proportion of native HSA present in batches produced according to different fractionation ways 51 . Because an increased percentage of oxidized HSA is responsible for impaired HSA functions 28 , the development of a reliable method providing qualitative and quantitative data on HSA variants in commercial preparations, especially the ratio of native HSA to degraded forms, is of paramount importance for optimizing the clinical use of HSA.

As we have previously mentioned, studies using HSA as a resuscitation fluid have shown conflicting results in septic and cirrhotic conditions. The relative contribution of resuscitation fluid capacity and antioxidant properties should be further investigated because it may influence the design of future trials. As a matter of fact, the rationale for infusing HSA was different in recent albumin trials which may account for different results. Should we rely on albuminemia and at what threshold: 30 g/l, or specific plasma concentrations (e.g. thiols, markers of ROS aggression)? The clinical studies have used different quantities of HSA infusion and various commercial forms of HSA. Regarding the HSA quality, it is well established that significant variability exists between the different commercial HSA solutions. They contain native HSA associated with various species of HSA in different redox states. They also contain several minor degradation products 47 63 . These modifications are known to change the antioxidant capacities of HSA and may account for the observed heterogeneity in the results of clinical trials 29 46 51 63 64 . It would be interesting to conduct an analysis correlating, for each clinical trial, the qualitative and quantitative characteristics of HSA infusion with the observed results in critically ill patients.

Moreover, the main characteristics of commercially available HSA should be tested in vitro, because they may differ from one product to another with different in vivo effect. The elucidation of the composition of commercialized HSA used in critically ill patients, with particular attention to the oxidized forms of HSA, is of great interest for the understanding of the observed variability in the results of clinical trials.

Finally, it is relevant to pursue the investigations concerning HSA antioxidant functions in physiological and pathophysiological conditions. Accumulative evidence suggests that HSA may interfere with microcirculation and endothelium function through an interaction with glycocalyx and specific antioxidant properties 68 . As a consequence, monitoring of endothelial function and microcirculation may guide HSA administration and contribute to the optimization of tissue perfusion in septic shock. The preliminary results of the Albios trial (presented orally at the ESICM meeting in Lisbon October 2012) suggest that HSA was effective in the subgroup of patients with the most severe form of sepsis which is septic shock. Future trials should probably focus on that specific high-risk population 69 .