Clinical relevance of drug binding to plasma proteins

Journal of Molecular Structure 1077 (2014) 4–13

Contents lists available at ScienceDirect

Journal of Molecular Structure journal homepage: www.elsevier.com/locate/molstruc

Clinical relevance of drug binding to plasma proteins q Paolo Ascenzi a,⇑, Gabriella Fanali b, Mauro Fasano b, Valentina Pallottini c, Viviana Trezza c a

Interdepartmental Laboratory of Electron Microscopy, University Roma Tre, Via della Vasca Navale 79, I-00146 Rome, Italy Biomedical Research Division, Department of Theoretical and Applied Sciences, and Center of Neuroscience, University of Insubria, Via Alberto da Giussano 12, I-21052 Busto Arsizio (VA), Italy c Department of Sciences, University Roma Tre, Viale Marconi 446, I-00146 Rome, Italy b

h i g h l i g h t s

g r a p h i c a l a b s t r a c t

 Clinical relevance of drug binding to

plasma proteins.  Binding to plasma proteins highly

influences drug efficacy, distribution, and disposition.  Serum albumin displays an extraordinary ligand binding capacity.  a-1-Acid glycoprotein is the main carrier for basic and neutral drugs.  High- and low-density lipoproteins play a limited role in drug binding.

a r t i c l e

i n f o

Article history: Available online 4 October 2013 Keywords: Human serum albumin Human a-1-acid glycoprotein Human lipoproteins Ligand binding properties Drug binding

a b s t r a c t Binding to plasma proteins highly influences drug efficacy, distribution, and disposition. Serum albumin, the most abundant protein in plasma, is a monomeric multi-domain macromolecule that displays an extraordinary ligand binding capacity, providing a depot and carrier for many endogenous and exogenous compounds, such as fatty acids and most acidic drugs. a-1-Acid glycoprotein, the second main plasma protein, is a glycoprotein physiologically involved in the acute phase reaction and is the main carrier for basic and neutral drugs. High- and low-density lipoproteins play a limited role in drug binding and are natural drug delivery system only for few lipophilic drugs or lipid-based formulations. Several factors influence drug binding to plasma proteins, such as pathological conditions, concurrent administration of drugs, sex, and age. Any of these factors, in turn, influences drug efficacy and toxicity. Here, biochemical, biomedical, and biotechnological aspects of drug binding to plasma proteins are reviewed. Ó 2013 Elsevier B.V. All rights reserved.

1. Introduction Drug binding to plasma proteins is often the first step in drug distribution, action and disposition. Human serum albumin (HSA) is

the most important drug carrier in adult humans [1] and mainly binds acidic drugs. a-1-Acid glycoprotein (AGP) is the next important one [2–4] and binds basic and neutral drugs; less relevant in drug binding are high- and low-density lipoproteins (HDL and LDL, respectively).

Abbreviations: AGP, human a-1-acid glycoprotein; APO, apoprotein; CETP, cholesterol ester transfer protein; FA, fatty acid; HAART, highly active antiretroviral therapy; HDL, high-density lipoproteins; HPX, hemopexin; HPX-heme-Fe, Fe-heme-hemopexin; HSA, human serum albumin; HSA-heme, human serum heme-albumin; LCAT, lecithincholesterol acyl transferase; LDL, low-density lipoproteins; RCT, reverse cholesterol transport; SRB1, scavenger receptor B1; VLDL, very low-density lipoprotein. q This article is part of a special issue titled ‘‘Fluorescence studies of biomolecular association processes. Towards a detailed understanding of spectroscopic, thermodynamic and structural aspects’’. ⇑ Corresponding author. Tel.: +39 06 5733 3621; fax: +39 06 5733 6321. E-mail address: [email protected] (P. Ascenzi).

0022-2860/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.molstruc.2013.09.053

P. Ascenzi et al. / Journal of Molecular Structure 1077 (2014) 4–13

5

Drug binding to plasma proteins such as HSA and AGP is usually reversible, occurs at specific sites, and is a major determinant in drug disposition since it highly affects the pharmacokinetics and pharmacodynamics of most commonly prescribed drugs. For instance, it results in increased solubility for lipophilic drugs, thus allowing them to reach their site of action. However, since the protein-bound drugs cannot readily leave the capillaries, only the unbound fractions can be distributed to tissues therefore having pharmacological activity, as well as toxic effects [5–9]. Usually, a drug is considered highly-bound to plasma proteins when the bound fraction exceeds about 90% of the total drug concentration. Besides distribution, binding to plasma proteins also affects drug metabolism and elimination, since both hepatic uptake and glomerular filtration are directly proportional to the free drug fraction present in the plasma. The fraction of total drug that is bound to plasma proteins depends on drug concentration and affinity as well as on the number of available binding sites [9–11]. Several factors influence drug binding to plasma proteins. First of all, the extent of the bound drug is altered in certain diseases. For example, hypoalbuminemia secondary to severe liver disease results in reduced drug binding and therefore in increasing the unbound drug fraction. Also, pathological conditions that induce an acute phase reaction response such as arthritis, myocardial infarction, and cancer lead to elevated levels of AGP, thus enhancing binding of basic drugs [12]. Drug binding to plasma proteins can be altered by the concurrent administration of drugs with similar physicochemical characteristics that compete with each other and with endogenous substances for common or functionallylinked binding sites. Lastly, age and sex are two other factors that can affect drug binding to plasma proteins [7,9,13–15]. Here, biochemical and biomedical aspects of drug binding to plasma proteins are reviewed, highlighting the clinical relevance of this process. 2. Serum albumin HSA, the most abundant protein in plasma (about 7.5  10 4 M), is the main determinant of plasma oncotic pressure and the main modulator of fluid distribution between body compartments. Moreover, HSA displays an extraordinary ligand-binding capacity, providing a depot and carrier for many endogenous and exogenous compounds. In fact, HSA represents the main carrier for fatty acids (FA), affects pharmacokinetics of many drugs, provides the metabolic modification of some ligands, renders potential toxins harmless, accounts for most of the anti-oxidant capacity of human plasma, and displays (pseudo-)enzymatic properties [1,14–25]. HSA may be considered as a biomarker of many diseases, including cancer, rheumatoid arthritis, ischemia, post-menopausal obesity, severe acute graft-versus-host disease, and diseases that need monitoring of the glycemic control [1,26–28]. Moreover, HSA is widely used in the clinical practice to treat several diseases, including hypovolemia, shock, burns, surgical blood loss, trauma, hemorrhage, cardiopulmonary bypass, acute respiratory distress syndrome, hemodialysis, acute liver failure, chronic liver disease, nutrition support, resuscitation, and hypoalbuminemia [1,29–34]. Recently, biotechnological applications of HSA, including implantable biomaterials, surgical adhesives and sealants, biochromatography, ligand trapping, fusion proteins, nanotubes and nanoparticles have been reported [1,35–38]. HSA is a monomeric protein constituted by a single non-glycosylated all-a chain of 66 kDa arranged in a globular heart-shaped conformation containing three homologous domains (labeled I, II, and III). Each domain is made up by two separate helical subdomains (named A and B), connected by random coils [1,18,39,40] (Fig. 1, panel A).

Fig. 1. The HSA structure. Panel A: FA binding to HSA. HSA is rendered with ribbons, colored as follows: subdomain IA: blue; subdomain IB: cyan; subdomain IIA: violet green; subdomain IIB: pink; subdomain IIIA: orange; subdomain IIIB: red. The FA binding sites (FA1 to FA9) are occupied by capric acid (green). FA binding sites are numbered as reported in [42]. Atomic coordinates were taken from PDB entry 1E7E [42]. Panel B: Drug binding to HSA. The FA1 site is occupied by azapropazone (blue) (PDB entry: 2BX8, [23]). The FA2 site has been hypothesized to be the secondary site of warfarin and ibuprofen on the basis of solution studies [57,155]. The FA3-FA4 and FA6 sites are occupied by ibuprofen (magenta) (PDB entry: 2BXG, [23]). The FA7 site is occupied by warfarin (green) (PDB entry: 2BXD, [23]). Both pictures were drawn with Swiss-PDB-Viewer [156]. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

The multidomain organization of HSA is at the root of its extraordinary ligand binding properties. HSA is able to bind up to nine equivalents of fatty acids (FAs) at multiple sites labeled FA1 to FA9 (Fig. 1, panel A) with different affinity. Crystallographic analysis of different complexes of HSA with a wide variety of exogenous ligands (e.g., drugs) revealed the architecture of the primary drug-binding sites (i.e., the FA3–FA4 cleft and the FA7 site). These cavities, constituted by distinct sub-compartments, are highly adaptable for the binding of several ligands. Moreover, structural data show a variety of secondary binding sites distributed across the protein, overlapping to a variable extent with FA sites [1,23]. The FA1 site (located in subdomain IB) has been evolved to bind the heme with high affinity contributing to the homeostasis and possibly the receptor-mediated endocytosis of the heme by hepatocytes. The FA2 site (located between subdomains IA and IIA) has been reported to be the ibuprofen and warfarin secondary site. Ligand (e.g., FAs) binding to the FA2 site stabilizes the HSA B-conformation. The FA3 and FA4 sites are located in a large cavity in subdomain IIIA that as a whole composes the so-called Sudlow’s

6

P. Ascenzi et al. / Journal of Molecular Structure 1077 (2014) 4–13

Fig. 2. Ibuprofen and warfarin binding to the FA3-FA4 cleft (panel A) and the FA7 site (panel B) of HSA. Drug carbon atoms are colored in cyan. Atomic coordinates were taken from the PDB entries 2BXD and 2BXG [23]. The picture was drawn with UCSF-Chimera [157]. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

site II. This cleft is preferred by aromatic carboxylates with an extended conformation, the non-steroidal anti-inflammatory drug ibuprofen representing the prototypical ligand (Fig. 1, panel B). Ibuprofen binds in the center of the FA3–FA4 cleft, interacting with the hydroxyl group of Tyr411; Arg410 and Ser489 residues also contribute salt-bridge and hydrogen-bond interactions to drug binding. The presence of a single basic polar patch located at one end of the apolar binding pocket is at the root of the specificity for drugs with a peripherally located electronegative group (Fig. 2, panel A). The FA5 site is formed by a hydrophobic channel located in subdomain IIIB (Fig. 1, panel B). The FA6 site (located at the interface between subdomains IIA and IIB) has been reported to be the ibuprofen secondary site (Fig. 1, panel B). The FA7 site (located in subdomain IIA) corresponds to the Sudlow’s site I. This site binds preferentially bulky heterocyclic anions, the prototypical ligand being the coumarinic anticoagulant drug warfarin (Fig. 1, panel B). The R-(+) and S-( ) enantiomers of warfarin bind to the FA7

site in essentially the same position as one another and make three hydrogen bonds with Tyr150, His242, and either Lys199 or Arg222. The similarity of the binding environments for the enantiomers agrees with the poor stereoselectivity of HSA for warfarin (Fig. 2, panel B). The FA8 and FA9 sites are located at the base and in the upper region of the gap, respectively, between subdomains IA– IB–IIA on one side and subdomains IIB–IIIA–IIIB on the other side. Due to volume restrictions, the FA8 site can only accommodate short-chain FA molecules. The FA9 site binds thyroxine in the presence of saturating amounts of FAs; on the other hand, defatted HSA binds thyroxine at FA3–FA4, FA5, and FA7 sites [1,14,15,18,23,40– 49]. HSA binds metal ions at multiple sites, including the N-terminus, the Cys34 site, and the I–II interdomain contact region, and undergoes ligand-dependent chemical modifications (e.g., acetylation and cysteinylation) [1]. Many plasma-proteins have been found to be associated to HSA including angiotensinogen, apolipoproteins, ceruloplasmin, clusterin, hemoglobin, plasminogen, prothrombin, and transferrin. The fraction of peptides and proteins bound to HSA is defined as ‘‘albuminome’’ [1,50–52]. Moreover, HSA binds the albumin binding domains of the bacterial protein G at domain II [1,50,51,53,54]. The complex mechanism modulating allosterically and competitively ligand binding to HSA represents one of the most important structure–function correlations ever reported for monomeric proteins. In fact, the conformational adaptability of HSA involves more than the immediate vicinity of the binding site(s), affecting both the structure and the ligand binding properties of the whole molecule [14,15]. The pH-induced neutral-to-basic (N-to-B) transition of HSA (occurring in the 6–9 pH range) replicates that induced by endogenous (e.g., FAs) and exogenous (e.g., drugs) ligands [1,14,15,47]. However, only FA1, FA2, FA6, and FA7 sites of HSA located in domains I and II are allosterically coupled. In particular, the affinity of heme for FA1 decreases by about one order of magnitude upon FA7 ligand (e.g., warfarin) binding and vice versa, according to the linked function [1]. The same effect has been reported for the FA6 binding site [55]. On the contrary, ligand binding to the FA2 site (e.g., myristate) increases the affinity for FA1 binding site by one order of magnitude [55,56]. Indeed, the recombinant Asp1Glu382 truncated HSA, encompassing only domains I and II, displays spectroscopic and reactivity properties superimposable to those of the full-length protein [55,57]. In general, even if HSA is a monomeric protein, it shows a complex mechanism modulated allosterically and competitively. The relevance of drug binding to HSA has been investigated in some diseases that require multiple drug administration. Competitive displacement from the same binding site or allosteric displacement following the microenvironmental change at the binding site are responsible of a decrease of plasma drug binding in the presence of coadministered drugs [1,46,59,60]. In rheumatoid arthritis patients, an increased pharmacological activity of diclofenac is observed with the concomitant administration of nabumetone, a non-steroidal anti-inflammatory drug [58]. The mutual interaction(s) of binding equilibria with plasma proteins and in particular with HSA is important for the therapeutic efficacy of anti-HIV drugs in HAART (highly active antiretroviral therapy) [46]. Anti-HIV drugs at concentration used in HAART induce heterotropic interactions that influence reciprocally the heme-Fe(III) and FA binding properties of HSA [46,59,60]. Moreover, anti-HIV drugs have been reported to modulate allosterically the heme-based reactivity of human serum heme-albumin (HSA-heme). In particular, abacavir modulates allosterically kinetics of NO dissociation from HSA-hemeFe(II)-NO and of peroxynitrite-mediated oxidation of HSA-hemeFe(II)-NO [61,62].

P. Ascenzi et al. / Journal of Molecular Structure 1077 (2014) 4–13

In the treatment of Parkinson’s disease (PD), a multidrug therapy is required to alleviate PD symptoms, in the absence of suitable diagnostic tools that permit a neuroprotective treatment. So, monoamine oxydase type B (MAO-B) inhibitors, catechol-O-methyltransferase (COMT) inhibitors and dopamine receptor agonists (e.g., apomorphine) may be simultaneously administered in the anti-PD therapy. As previously reported, the anti-PD drug concentration required to achieve the therapeutic plasma level in the presence of plasma proteins appears to be at least one-order of magnitude higher than that required in the absence. Moreover, anti-PD drug binding to HSA is impaired by heme association and vice versa. Under some pathological conditions (e.g., severe hemolytic anemia, crash syndrome and post-ischemic reperfusion), the increase of the heme level in plasma could facilitate drug release and the concomitant intoxication of the patient. Accordingly, the toxic plasma heme concentration may increase in patients under anti-PD drug therapy [63]. On the other hand, the observed drug displacement may represent an advantageous phenomenon to strengthen the effect of some drugs, in order to reach therapeutic efficacy. In nephrotic syndrome patients, bucolome, a barbiturate derivative, inhibits the binding of furosemide, a diuretic, to HSA and increases the free fraction of furosemide at the target site [58]. More than 80 genetic variants of HSA accountable for the presence of two circulating forms of the protein (i.e., bisalbuminemia or alloalbuminemia) or the virtual absence of the protein from the blood (i.e., analbuminemia) have been reported. HSA multiple activities suggest that mutations localized in critical residues for specific ligand binding could significantly compromise tissue and organ physiopathology, with consequences on human health [1,16]. Remarkably, two Arg218His and Arg218Pro HSA mutants (i.e., located in subdomain IIA) are responsible of familial dysalbuminemic hyperthyroxinemia and show a very high affinity for thyroxine. This disease causes an increase of thyroxine synthesis and the decrease of the thyroxine availability in plasma; moreover, it may affect therapeutical treatment with thyroxine [1,44,64–66].

3. a-1-Acid glycoprotein For about half a century, HSA has been recognized as one of the major determinants of drug action, distribution, and disposition, and the contribution of AGP has been underestimated, mainly because the AGP plasma concentration (about 5  10 5 M) is much lower than that of HSA. However, in the last two decades, binding of endogenous and exogenous compounds to AGP has received increased attention [2,12]. AGP is a glycoprotein of about 42 kDa consisting of about 45% carbohydrate. AGP is a single polypeptide chain of 183 amino acids, 22 amino acid residues differing between the two F1/S and A variants which are encoded by ORM1 and ORM2 genes, respectively. Moreover, mutations at positions 32 and 47 occur, probably reflecting polymorphisms in the human population. Different forms of AGP can be distinguished in serum depending on the type of glycosylation and multiple amino acid substitutions. Remarkably, more than of 100 forms of AGP differing in carbohydrate composition have been reported. Changes in AGP glycosylation occur in several physiological and pathological conditions (including pregnancy, acute inflammatory conditions, severe rheumatoid arthritis, alcoholic liver cirrhosis and hepatitis) affecting its biological properties [3,4]. The F1/S and A variants of AGP display the typical lipocalin fold comprising an eight-stranded b-barrel flanked by one a-helix. Loops connecting b-strands A/B (loop 1), C/D (loop 2), E/F (loop 3), and G/H (loop 4) form the entrance to the ligand binding site at the open end of the b-barrel, the closed end of the b-barrel being covered by the N-terminal polypeptide segment (Fig. 3) [67,68].

7

Fig. 3. The structure of the F1/S variant of AGP. The secondary structural elements are colored yellow (b-strands), pink (a-helix), and gray (coils). Loop 1 (A/B), loop 2 (C/D), loop 3 (E/F), and loop 4 (G/H) at the open end of the AGP are highlighted in cyan. The Asn side chains of the five N-linked glycosylation sites (Asn15, Asn38, Asn54, Asn75, and Asn85) are shown as blue sticks. The amitriptyline molecule is highlighted in red as ball and sticks. Atomic coordinates were taken from the PDB entry 3APV [68]. The picture was drawn with UCSF-Chimera [157]. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

The opening of the ligand binding site of the F1/S and A variants of AGP measures 8–12 Å in diameter and exhibits the positively charged residues Arg68 and Arg90. The essentially non-polar ligand binding site of the F1/S variant is approximately 18 Å deep, the binding region of the A variant being narrower [67,68]. The ligand binding cavity of the A variant of AGP displays three distinct lobes. The deep main lobe (lobe I) is very large and non-polar, providing space to accommodate a hydrophobic ligand. On each of its sides, there are the small negatively-charged lobes II and III; of note, lobe III displays a small entrance [67]. However, the F1/S variant of AGP displays lobes I and II only [68]. This structural organization agrees with the occurrence of partially overlapping subsites as hypothesized on the basis of comparative ligand binding observations [69]. As a whole, the geometry of the AGP binding pocket [67,68] is in agreement with the broad spectrum of basic and neutral compounds that are recognized by this member of the lipocalin superfamily [4]. The binding mode of disopyramide, amitriptyline, and chlorpromazine to the F1/S variant of AGP has been determined by X-ray crystallography [68], whereas diazepam and progesterone recognition by the A variant of AGP has been investigated by docking simulations [67]. The two aromatic rings of the S-form of disopyramide, which is preferentially-bound by AGP, are in direct contact with the Phe49 and Phe112 residues; in addition, van der Waals contacts with Glu64 and Arg90 occur. In the S-form of disopyramide, the nitrogen atom of the pyridinyl ring can be located near the guanidium group of Arg90 contributing to the selective binding of the S-form of disopyramide by the A variant of AGP. Moreover, the amide group of disopyramide forms a hydrogen bond with the OH group of Tyr127. Furthermore, disopyramide is stabilized by hydrogen bonding to a water molecule, which is held in position by hydrogen bonding with Ser114 and Ser125. Finally, the alkyl chain of disopyramide makes van der Waals interactions with the Tyr27, Tyr37, Val41, and Ile44 residues [68].

8

P. Ascenzi et al. / Journal of Molecular Structure 1077 (2014) 4–13

Similarly to disopyramide, the aromatic rings of amitriptyline contact the Phe49 and Phe112 residues and make van der Waals interactions with the Leu62 and Arg90 side chains of AGP. The tip of the alkyl chain of amitriptyline makes van der Waals contacts with Tyr37 and Val41 [68]. The aromatic ring system of chlorpromazine interacts with Phe112, makes contacts with the Phe49 and Ala99 residues, and forms van der Waals contacts with Phe51, Val88, and Arg90 of AGP. Moreover, the chlorine atom attached to one end of the aromatic ring system in chlorpromazine is accommodated at the end of the binding pocket formed by the Thr47, Phe9, Glu64, and Tyr127 residues [68].The planar phenyl ring of diazepam fits into the apolar lobe I of AGP, the polar diazepine ring fits into the charged lobe II, and the chlorobenzene moiety is buried in the charged lobe III. Contrasting with HSA [70], the F1/S variant of AGP binds diazepam preferentially in the P-conformation [67]. Progesterone fits into the lobe I of AGP with the acetyl substituent deeply buried in the binding pocket and the keto-group of the steroid ring A is solvent-exposed at the entrance of the b-barrel [67]. Different AGP variants display distinct specificity patterns. In particular, disopyramide, imipramine, and methadone bind selectively to the A variant, whereas binedaline, dipyridamole, and warfarin are preferentially bound by the F1/S variant. In contrast, progesterone, flunitrazepam, propranolol, and chlorpromazine are recognized by both the APG variants [71,72]. Many biological properties of AGP have been described [2–4]. First of all, AGP is one of the acute-phase proteins, whose levels increase in response to inflammation or tissue injury. The increased synthesis and secretion of AGP may be the response of the tissue to pro-inflammatory stimuli, as indicated by: (i) immunosuppressive action [73,74], (ii) immunomodulation activity [75], (iii) inhibitory effect on interleukin-2 secretion by lymphocytes [74], (iv) inhibition of lymphocyte proliferation [73,74], (v) binding of both exogenous and endogenous inflammatory mediators, including histamine [76], (vi) inhibition of platelet aggregation [77], (vii) inhibition of neutrophil activation [78], and (viii) induction of the synthesis of interleukin-1 receptor antagonist [79]. Furthermore, AGP is required to maintain capillary permeability probably by increasing the polyanionic charge selectivity of the endothelial barrier [80], and displays protective effects against neonatal sepsis [81]. AGP mainly binds basic and neutral endogenous and exogenous ligands. Endogenous molecules that bind AGP include vanilloids, heparin, serotonin, catecholamines, platelet activating factor, melatonin, histamine and steroid hormones [82–88]. Drugs that bind avidly AGP and that are currently used in the clinical practice include a- and ß-adrenergic receptor antagonists (i.e., alprenolol, carvedilol, oxprenolol, penbutolol, pindolol, propranolol, tertatolol, and timolol), analgesics (i.e., alfentanil, fentanyl, lofentanil, and sufentanil), local anesthetics (i.e., bupivacaine, etidocaine, lidocaine, and mepivacaine), several psychoactive drugs including antidepressant and neuroleptic compounds (i.e., amitriptyline, imipramine, desipramine, nortriptyline, chlorpromazine, olanzepine, and triazolam) and protease inhibitors used in the treatment of patients affected by AIDS [2–4]. Of note, HSA and AGP contribute almost equally in binding and transport of some drugs, including amitriptyline, chloroquine, dipyridamole, taxol, and paclitaxel [4]. It is generally assumed that in plasma, acidic drugs are mainly bound to HSA. However, binding to AGP may contribute significantly to the total plasma binding of these drugs, especially in diseases in which the concentration of AGP increases and/or that of HSA decreases [2]. It should be noted that drug binding to AGP is different in many aspects from that to HSA [12]. In fact, AGP plasma levels increase during a variety of pathological conditions, including cancer,

trauma, infections, and inflammation, while hyperalbuminemia is very rarely observed [1,3,12]. Furthermore, intra- and inter-individual variability in plasma levels of AGP depending on age, gender, hormonal status, or other physiological conditions occur in healthy individuals as well. Conversely, the decrease of HSA plasma level is observed only in severe conditions such as cirrhosis, nephrotic syndrome and malnourishment [1]. Last, the plasma level of AGP is lower than that of HSA, and AGP displays only one drug binding site per molecule [67]. As a result: (i) drug binding to AGP is saturable while binding to HSA is generally not saturable [8,12], and (ii) ligand (e.g., drug) binding to HSA is modulated competitively and allosterically [1,14,15]. 4. Lipoproteins Lipoproteins are a class of nanoparticles that are present in most species and are essential in human beings to control lipid metabolism and homeostasis. These endogenous nanoparticles are transporters of lipids and other hydrophobic molecules in the mammalian circulation. They are composed by phospholipids, cholesterol, apoproteins (APO), and enzymes. The lipoprotein particles have hydrophilic groups of phospholipids, cholesterol, and APO directed outward. Such characteristics make them soluble in the salt water-based blood pool. Triglyceride-fats and cholesterol esters are carried inside, shielded from the water by the phospholipid monolayer and the APOs [89] (Fig. 4). The human lipoproteins consist of several kinds of nanoparticles that have variable structure and function, differentiated by their size, density, and the APO bound on their surfaces. On the basis of their density, lipoproteins can be divided into four groups: (i) chylomicrons, which display a diameter of 75–1200 nm and have a density <0.95 g/ml; (ii) very low density lipoproteins (VLDL) which show a diameter of 30–80 nm and have a density ranging from 0.95 to 1.006 g/ml; (iii) low density lipoproteins (LDL) which present a diameter of 18–25 nm and have a density ranging from 1.006 to 1.063 g/ml; and (iv) high density lipoproteins (HDL) which display a diameter of 5–12 nm and have a density ranging from 1.063 to 1.21 g/ml [90,91]. The first characterization of APO divided them into three major groups: A, B, and C [92]; more recently, APO D, E, H, and J have been characterized [93]. Besides their involvement in the lipid transport, they also act as enzyme cofactors or activators. APOA-1 (mw 28 kDa; plasma concentration 100–150 mg/dl) is associated with HDL and chylomicrons, is a cofactor of lecithin-cholesterol acyl transferase (LCAT), and binds the scavenger receptor B1 (SRB1)

Fig. 4. Schematic lipoprotein structure. Lipoproteins show an external polar surface and a lipophilic core. A phospholipid monolayer and APOs form the external polar surface, while the lipophilic core is mainly composed by triglycerides and cholesteryl esters. Lipophilic drugs bind to the lipoprotein core.

P. Ascenzi et al. / Journal of Molecular Structure 1077 (2014) 4–13

on cell surface [94]. APOA-2 (mw 17 kDa; plasma concentration 30–50 mg/dl) is associated with HDL, its physiological role being is still debated [95]. APOA-4 (mw 46 kDa; plasma concentration 15–18 mg/dl) is bound to HDL, VLDL, LDL, and chylomicrons, is an activator of LCAT, and promotes triglyceride secretion from liver [96]. APOB-48 (mw 265 kDa; plasma concentration 5–7 mg/dl) is bound to chylomicrons, and acts as a carrier of triglycerides [97]. APOB-100 (mw 549 kDa; plasma concentration 80–100 mg/dl) is the carrier of LDL and transports cholesterol from liver to peripheral tissues [94]. APOCs are transported by chylomicron, HDL, and VLDL and are fundamentally enzyme activators and cofactors. In particular, APOC-1 (mw 6.5 kDa; plasma concentration 4–7 mg/ dl) promotes LCAT activation, APOC-2 (mw 8.5 kDa; plasma concentration 3–8 mg/dl) is a lipoprotein lipase cofactors, and APOC3 (mw 8.75 kDa; plasma concentration 8–15 mg/dl) inhibits APOC-2 activating lipoprotein lipase [97–99]. APOD (mw 22 kDa; plasma concentration 10–11 mg/dl) is principally expressed in HDL; little is known about its functions in lipid metabolism but recent evidence shows that the APOD level is significantly modified in various pathologies [100]. APOE (mw 35–39 kDa; plasma concentration 3–6 mg/dl) is present in chylomicrons, HDL, and VLDL, facilitating lipoprotein uptake [101]. Lipoproteins may act as drug-delivery systems due to their small size, long half-life in the circulation and their hydrophobic core that facilitates the incorporation of non-soluble drugs. However, it cannot be excluded that drugs might bind to specific sites of APO (Fig. 4). Lipoprotein ligands are highly lipophilic and are the main candidates for drug metabolism as the main path of removal from the body. Examples of drugs that are known to bind significantly to lipoproteins include cyclosporine A (binding to HDL and LDL) [102], amiodarone (binding to all types of lipoproteins) [103], halofantrine (binding to HDL and LDL) [104], amphotericin B (binding to LDL) [105], and eritoran (binding to HDL) [106]. Consequently, lipoproteins and synthetic-reconstituted lipoprotein preparations have been evaluated with great interest towards clinical applications [107]. The structures of lipoproteins have a critical role; as an example, some studies demonstrated that the extracellular matrix of some cancer tumors could impair the diffusion of nanoparticles [108], and some measurements obtained by electron microscopy showed that the opening between collagen fibrils of this matrix are less than 40 nm. Therefore, since LDL and HDL display diameters less than 40 nm, they could be easily used for cancer drug delivery. Moreover, lipoprotein-based nanoparticles can be generated by conjugating tumor-homing molecules to the protein components of naturally occurring lipoproteins redirecting them from their normal lipoprotein receptors to other selected cancer-associated receptors. Multiple or different copies of targeting moieties may be attached to the same nanoparticles. Thus, various sets of tumor-homing molecules could be used to generate a group of conjugated lipoproteins as biocompatible nanoparticles with a wide-ranging application for the treatment of cancer [109]. In a normolipidaemic plasma profile, the presence of lipids and proteins is ideal for the distribution of hydrophobic lipids through the aqueous environment. When this profile is deregulated, drug transport and delivery may be affected. Thus, in the presence of dyslipidaemia, lipophilic drugs should be administrated under careful control, since changes in plasma lipoprotein content and the following variations in the cellular uptake of drugs could provoke a modification in the drug plasma concentration/efficacy relationship [110]. 4.1. Human serum high-density lipoproteins HDL play an important role in reverse cholesterol transport (RCT), which is involved in the removal of cholesterol from

9

peripheral tissues and its delivery to the hepatic tissue and steroidogenic cells for catabolism [111]. Excessive amount of cholesterol in peripheral tissues can be up-taken by nascent HDL (discoidal) that are forced to mature into more spherical HDL by LCAT during their translocation to the liver. APOA-1 is the principal protein constituent of HDL, it is synthesized in the hepatic tissue and intestine and is secreted into the bloodstream in a lipid-free state; then, it associates with phospholipid and cholesterol. During circulation, nascent HDL are remodeled by LCAT that is able to esterify free cholesterol into cholesteryl ester. APOA-1 is a very important modulator of this reaction that makes these particles more spherical and lipophilic. Moreover, APOA-1 plays an important role in the interaction of the HDL SRB1 which binds the lipoproteins to the cells [112]. APOA-2 is the second most important constituent of HDL and seems to be involved in the development of coronary artery disease. Indeed, an increase of this protein on HDL surface displaces paraoxonase, decreasing HDL anti-oxidant activity [113]. Paraoxonase is an HDL-associated enzyme that catalyzes the hydrolysis of oxidized fatty acids from phospholipids [114]. Typically, HDL have been studied from a physiological point of view as antiatherosclerotic particles; nevertheless hydrophobic biomolecules can be incorporated into HDL. In fact, APOA1 can be potentially used to target drugs to tissues expressing SRB1 on the surface of their cell membranes. In particular, since HDL diameter ranges from 5 to 12 nm, they can assemble in caveolae domains, which are described as 50–100 nm invaginations rich in SRB1 [115]. Actually, anticancer, antifungal, and antiviral drugs have already been incorporated into synthetic HDL [116–118]. Of note, three molecules of taxol, an anticancer drug, could be incorporated into HDL particles [116]. Moreover, paclitaxel incorporated into HDL nanoparticles displays superior cytotoxicity against several cancer cell lines (MCF7, DU145, OV1063 and OVCAR-3), the half maximal inhibitory concentration (IC50) having been found to be 5–20 times lower than that of the free drug. Studies with mice showed that the paclitaxel/HDL nanoparticles were substantially well tolerated [119]. The antibiotic amphotericin B binds to HDL, elicits significant antifungal effects both in vitro and in vivo without any toxicity [117]. a-Tocopherol can also bind to HDL. Remarkably, SRB1 at the inner blood brain barrier is responsible for a-tocopherol uptake from the circulating blood and plays an important role in maintaining a-tocopherol levels in the neural retina [120]. Very recently, it has been reported that a purified fraction of HDL, deriving from healthy human plasma, contains miRNA [121]. MiRNA are small non-coding RNA that can influence many gene transcript exerting modulations in cell physiology [122,123]. Circulating miRNA may represent biomarkers for assessing and monitoring pathophysiological conditions, since they function as paracrine or endocrine signals to communicate pathophysiological changes. They derive from multiple sources, such as mast cells, tumors, and urine, then miRNA are transported to several tissues and organs where they act as modulators [124]. These molecules are generally transported in microvescicles and exosomes that protect them from RNase digestion [125]. The transport mediated by exosomes and microvescicles is a general process; on the contrary HDL-miRNA has specific receptors, as well as SRB1 [121], and therefore the tissues expressing SRB1 can be considered good candidates for HDL-miRNA therapy. 4.2. Human serum low-density lipoproteins LDL are the principal carriers for cholesterol delivery to cells. Their most important lipid component are cholesteryl esters and free cholesterol. The principal structural APO is APOB-100. These particles derive from VLDL catabolism even if different in vivo and in vitro experiments demonstrated that hepatic tissue is able to produce APOB-100 containing lipoproteins [126]. LDL are

10

P. Ascenzi et al. / Journal of Molecular Structure 1077 (2014) 4–13

up-taken by cells through LDL receptor-mediated endocytosis. Once into the cells, LDL are degraded and cholesteryl ester hydrolyzed to release free cholesterol. Cholesterol surplus with the respect of cellular needs is re-esterified by acetyl-CoA acetyl-transferase and stored within the cells [127]. The process by which cholesterol esters are transferred into cells is different between HDL and LDL. In fact, LDL cholesterol displays a receptor-mediated endocytosis pathway; on the contrary, HDL bind to SRB1 and transfer lipids (but not proteins), thus releasing a cholesterol-depleted HDL particle from the receptor. Therefore, drugs that preferentially associate with the HDL or LDL fraction might show differences in the pharmacodynamics owing to different receptor-mediated uptake [110]. Oral administered lipophilic drugs could primarily be linked to HDL and chylomicron as these nanoparticles are formed in the enterocytes during fasting and after feeding, respectively [128]. The distribution of drugs between plasma lipoproteins is greatly related to the specific apolar core of lipids and proteins of individual plasma lipoprotein fractions and is also affected by cholesterol ester transfer protein (CETP) activity [104,129,130]. CETP is a plasma enzyme that promotes the redistribution and equilibration of hydrophobic lipids packaged within the lipoprotein core between HDL and LDL, VLDL, and chylomicrons, it circulates in plasma bound to lipoproteins and displays a lipid transfer activity. The CETP effect on lipoproteins is the depletion of cholesteryl esters, the enrichment with triglycerides producing a net reduction of lipoprotein size. Beside shifting lipids, CETP is able to transfer drugs from a lipoprotein to another [131]; thus, an increase of the plasma content of this enzyme could result in a transfer of the drug from HDL to LDL [110]. Amphotericin B is an example: it is one of the most effective agents in the treatment of systemic fungal infections [132], although its clinical use has been limited by dose-dependent renal toxicity [132,133]. There is ample evidence that amphotericin B association with serum LDL is regulated by an increase in CETP plasma concentration and activity. It has been postulated that this association might be involved in the development of amphotericin B-induced kidney toxicity [110]. Indeed, in vivo experiments demonstrated that, if amphotericin B is administered in a lipid complex unable to bind LDL, amphotericin B-induced renal toxicity decreases [131]. Another example of a LDL-bound drug is cyclosporin A, an active immunosuppressant used to treat patients who have undergone organ transplantation. The majority of these patients experience a significant increase in total serum cholesterol levels. Variation in the lipid profile can modify the cyclosporin A-lipoprotein association, and thus may modify the pharmacological activity, the pharmacokinetic properties as well as the toxicity of this drug. An increased antiproliferative effect of cyclosporin A was observed when the drug was bound to LDL [134], this effect was not evident when the drug was bound to either VLDL or HDL. Moreover, also cyclosporine A was partially influenced by CETP [135–137]. In conclusion, lipoproteins are natural drug delivery system for lipophilic drugs or lipid-based formulations, but the association of these drugs with lipoproteins might have consequences on their pharmacological and toxicological activities, as well as their pharmacokinetic properties. Thus, a better comprehension of the mechanisms underlying the lipoprotein distribution of drugs could increase the therapeutic profiles of these molecules. 5. Binding to plasma proteins influences the pharmacokinetics and pharmacodynamics of drugs Binding to plasma proteins affects the pharmacokinetic and pharmacodynamics parameters of most commonly prescribed drugs [6,7,9]. Binding to plasma proteins facilitates drug distribution. First of all, plasma proteins act as delivery system for lipophilic drugs otherwise insoluble in plasma, thus allowing them

to reach their site of action. Furthermore, binding to plasma proteins reduces the unbound drug fraction in plasma, thus maintaining a concentration gradient between the intestinal lumen and plasma, which results in the increased intestinal absorption of the drug. The apparent volume of distribution of a drug (Vd) gives an indication of the extent of drug binding and distribution [9,10]. Drugs highly bound to plasma proteins usually have low Vd, because their strong association with plasma proteins confines them to the vascular space. Conversely, drugs that are largely free in plasma are generally available for distribution to tissues. However, the Vd is also affected by the drug affinity for tissue proteins [7,9]. Several drugs, such as amiodarone, digoxin, and tricyclic antidepressants, although highly bound to plasma proteins, bind with greater affinity to tissue proteins, resulting in a large Vd value [7]. Thus, drug distribution depends on binding to both plasma and tissue proteins. Much pharmaceutical research is aimed at designing drugs that bind to only their target proteins with high affinity (typically 0.1–10 nM) or at improving the affinity between a particular drug and its in vivo protein target. Besides distribution, binding to plasma proteins also affects other pharmacokinetic processes, such as drug metabolism and elimination; indeed, both hepatic uptake and glomerular filtration are directly proportional to the free drug fraction present in the plasma [9,10]. If a drug is highly bound to plasma proteins and has a high extraction ratio from the liver, plasma proteins act as delivery systems. Indeed, the rate of metabolism of some highly bound drugs such as propranolol depends on the rate of delivery to the liver via the bloodstream (the so-called flow-dependent hepatic elimination). Alternatively, if a drug has a low extraction ratio by the liver, metabolism may be decreased by high plasma protein binding. Drugs bound to plasma proteins cannot undergo glomerular filtration with the result that the half-life of drugs that are neither actively secreted by the renal tubules nor rapidly metabolized by the liver is increased. Conversely, similar to hepatic uptake, if a drug has a high renal extraction ratio due to active tubular secretion, plasma protein binding will promote drug elimination by transporting the drug to the kidney [9,10]. Binding to plasma proteins also affects pharmacodynamics of drugs. Indeed, once at the site of action, the interaction of the drug with its molecular target is related to the availability of the free drug and of its unoccupied target, as well as on the affinity for each other. In general, despite the influence that plasma protein binding has on the pharmacokinetic and pharmacodynamics parameters of drugs, the risk of clinically relevant interactions via changes in plasma protein binding is low. Nevertheless, the possibility of displacement interactions of drugs known to be highly bound to plasma proteins should be considered. This may be of particular importance for highly bound drugs (unbound fraction <1%) having a narrow therapeutic window, a high hepatic extraction ratio (if administered intravenously), or a high renal extraction ratio. In these cases, the risk of interaction should be addressed by in vitro displacement studies using therapeutically relevant concentrations [9,11]. To summarize, the drug-plasma protein complexes act as dynamic depots or reservoir systems from which drugs are released: as soon as the unbound drugs start to be metabolized and excreted, more drug molecules are released from the plasma proteins. Therefore, drug binding to plasma proteins reduces the intensity of the pharmacological effect, but prolongs the duration of the drug action. In drug discovery, the information on drugplasma protein binding is pivotal to evaluate and to understand the pharmacokinetic profile of drug candidates. Plasma protein binding data are useful to design optimal dose regimens for efficacy studies and to estimate safety margins during drug development.

P. Ascenzi et al. / Journal of Molecular Structure 1077 (2014) 4–13

6. Factors affecting drug binding to plasma proteins Pathological conditions, the concurrent administration of drugs, age, and sex are the major factors that can affect drug binding to plasma proteins [1,7,9]. Several pathological conditions are characterized by changes in plasma levels of HSA and AGP which, in turn, may significantly alter the pharmacokinetic and pharmacodynamics properties of drugs. Thus, the fraction of drugs bound to HSA can be decreased in patients suffering of analbuminemia, hypoalbuminemia, uremia, and hyperbilirubinemia, with consequent increase in the free drug concentrations [1,9]. Similarly, lower than normal levels of AGP have been found in patients with pancreatic cancer, hepatic cirrhosis, hepatitis, hyperthyroidism, nephrotic syndrome, malnutrition, and cachexia [4]. Any reduction in drug binding, per se, is usually unlikely to give rise to adverse effects in patients. This happens because since the unbound drug can normally easily distribute out of the plasma compartment, the increase of plasma free drug concentration is usually transient and too small to affect the total pool of the free drug in the body [1]. Thus, under pathological conditions where plasma AGP or HSA levels are expected to be lower than normal, therapeutic drug monitoring is suggested only for drugs which are highly bound to plasma proteins and have a narrow therapeutic window. Pathological states can also increase the levels of plasma proteins. Thus, while hyperalbuminemia is a rare condition, pathological states resulting in the acute-phase reaction response, such as cancer, arthritis and myocardial infarction, lead to increased levels of AGP and therefore enhanced binding of basic drugs [2]. AGP levels are higher also in obese individuals and in patients with injury, trauma, and severe burns and in recipients of bone marrow and organ transplants [4]. In most of these cases, the plasma AGP level increases after stimulation by the inflammatory cytokines, and the unbound fraction of drugs carried by AGP decreases significantly, resulting in lower efficacy. As an example, an increase in plasma AGP levels in patients affected by AIDS with an underlying infection could result in increasing drug binding; thus, decreased efficacy of the protease inhibitors administrated can be observed. In these situations, dose adjustment could be necessary to obtain the desired therapeutic effect [4]. Pathological states may also affect drug binding to lipoproteins. Changes in the free fraction of lipoprotein-bound drugs such as cyclosporine A, amiodarone, halofantrine and amphotericin B have been observed in patients with dyslipidaemias [110]. Allosteric and competitive mechanisms modulate drug binding to plasma proteins. Drug-drug and drug-endogenous ligand interactions at one or more binding sites on the plasma proteins may occur when a drug or endogenous substance with high affinity displaces another drug with a low affinity, thus increasing the unbound fraction of the latter, with enhanced pharmacological activity (either therapeutic and toxic). These interactions are clinically relevant for drugs highly bound to plasma proteins that have low clearance and narrow therapeutic index, or when a displacing drug is co-administered to the patient via rapid intravenous injection [1]. Age is an important physiological factor that influences plasma protein binding properties. Because of the low levels of plasma proteins in fetal and neonatal serum, a high free fraction of drugs is expected in the fetus and in neonates [138]. Therefore, therapeutic monitoring of the unbound drugs in plasma of mothers may be necessary to avoid indirect drug toxicity to the fetus or the neonate, particularly in the case of highly bound drugs which cross the placenta and/or are secreted into milk, or drugs with a low therapeutic window [1,4,138]. Changes in drug binding to plasma proteins also occur in the elderly, the HSA levels being approximately 19% lower than in young populations because of reduced renal function and altered capacity of the liver to synthesize

11

plasma proteins [139,140]. Conversely, age per se does not alter AGP levels [7]. For most drugs, age-related changes in protein binding, although statistically significant, are not likely clinically relevant. The free fraction is changed by more than 50% in the elderly for only a few drugs, e.g. acetazolamide, diflunisal, etomidate, naproxen, salicylate, valproate, and zimeldine [7]. Sex can also affect drug binding to plasma proteins [141,142]. As an example, the prototype b-adrenoceptor antagonist propranolol is highly bound to AGP and is marketed as an equal mixture of the two enantiomers. The affinity of the b-adrenoceptor for the S-enantiomer of propranolol is at least 100 times higher than that for the R-isomer. Although no age-related difference for the binding of both enantiomers of propranolol to AGP has been reported, there is a sex-related difference with the females having significantly greater binding of the S-enantiomer than the males [142]. This sex effect might lead to an apparent resistance to the drug action in females because a higher drug concentration would be necessary to obtain the same concentration of the active Senantiomer. However, the clinical relevance of this finding needs to be further addressed [143]. The effect of age and sex on plasma protein binding of both acidic (salicylic acid, phenytoin) and basic (chlorpromazine, meperidine, propranolol, and desipramine) drugs has been investigated [141]. In this study, only the correlation of the free plasma fraction of salicylic acid with age was statistically different between males and females. However, the difference in the free plasma fraction of salicylic acid between sexes was too small to be clinically significant [141]. The effect of age and sex on the protein binding of different benzodiazepines has also been reported, and it has been found that the binding of these compounds to plasma proteins is more affected by age than by sex [144–146]. 7. Conclusion and perspectives Here, the biochemical, biotechnological, and biomedical aspects of drug binding to plasma proteins have been outlined. It should be mentioned, however, that in recent years red blood cells have also emerged as potentially biocompatible carriers for a number of bioactive substances, such as glucocorticoids, antiviral and anticancer drugs [147]. Interestingly, drugs binding to red cells have been reported to act as allosteric effectors of hemoglobin affecting O2 uptake, transport, and release [148]. Thus, although in the past years binding to red blood cells has been underestimated in pharmacokinetic and pharmacodynamics studies [149], the use of red blood cells as drug delivery systems may represent an attractive and versatile technology, and more research in this field is warranted. A striking example of ligand transport involving multiple plasma proteins is represented by the heme-scavenging mechanism, providing protection against free heme oxidative damage, limiting the access of pathogens to heme, and contributing to iron homeostasis by recycling the heme-Fe-atom. In fact, during the first seconds after heme-Fe appearance in plasma, more than 80% of this powerful oxidizer binds to HDL and LDL, and only the remaining 20% binds to HSA and hemopexin (HPX). Although HDL and LDL are the most oxidatively intolerant plasma components and bind heme with high affinity, kinetics of heme transfer from HDL and LDL to HSA is faster than the heme-induced lipoprotein oxidation. Then, HSA removes most of the heme from HDL and LDL. Afterwards, heme transits from HSA to HPX, that releases it into hepatic parenchymal cells after internalization of the HPX-heme-Fe complex by CD91 receptor-mediated endocytosis. After delivering the heme intracellularly, HPX is released intact into the bloodstream and the heme is degraded [1,21,62,150–154]. As a whole, the modulation of ligand binding to plasma proteins is relevant not only under physiological conditions but also in the

12

P. Ascenzi et al. / Journal of Molecular Structure 1077 (2014) 4–13

pharmacological therapy management. Indeed, drug binding to plasma proteins is an important determinant for their biological efficacy; in turn drug availability to the intended target is modulated by binding to and dissociation from blood components. As an example, the increase of plasma levels of the heme-Fe under pathological conditions (e.g., severe hemolytic anemia, crash syndrome, and post-ischemic reperfusion) may induce the release of HSA-bound drugs with the concomitant intoxication of the patient. As expected, the plasma heme-Fe concentration may increase in patients after drug administration [15,24]. References [1] G. Fanali, A. di Masi, V. Trezza, M. Marino, M. Fasano, P. Ascenzi, Mol. Aspects Med. 33 (2012) 209–290. [2] J.M. Kremer, J. Wilting, L.H. Janssen, Pharmacol. Rev. 40 (1988) 1–47. [3] T. Fournier, N.N. Medjoubi, D. Porquet, Biochim. Biophys. Acta 1482 (2000) 157–171. [4] Z.H. Israili, P.G. Dayton, Drug Metab. Rev. 33 (2001) 161–235. [5] M.C. Meyer, D.E. Guttman, J. Pharm. Sci. 57 (1968) 895–918. [6] P.F. D’Arcy, J.C. McElnay, Pharmacol. Ther. 17 (1982) 211–220. [7] M.K. Grandison, F.D. Boudinot, Clin. Pharmacokinet. 38 (2000) 271–290. [8] M. Otagiri, Drug Metab. Pharmacokinet. 20 (2005) 309–323. [9] S. Schmidt, D. Gonzalez, H. Derendorf, J. Pharm. Sci. 99 (2010) 1107–1122. [10] W.E. Lindup, M.C. Orme, Br. Med. J. (Clin. Res. Ed.) 282 (1981) 212–214. [11] M. Pellegatti, S. Pagliarusco, L. Solazzo, D. Colato, Expert Opin. Drug Metab. Toxicol. 7 (2011) 1009–1020. [12] Z. Huang, T. Ung, Curr. Drug Metab. 14 (2013) 226–238. [13] K. Turnheim, Exp. Gerontol. 39 (2004) 1731–1738. [14] P. Ascenzi, M. Fasano, IUBMB Life 61 (2009) 1118–1122. [15] P. Ascenzi, M. Fasano, Biophys. Chem. 148 (2010) 16–22. [16] T. Peters Jr., All About Albumin: Biochemistry, Genetics and Medical Applications, San Diego, London, 1996. [17] C. Bertucci, E. Domenici, Curr. Med. Chem. 9 (2002) 1463–1481. [18] S. Curry, Drug Metab. Pharmacokinet. 24 (2009) 342–357. [19] U. Kragh-Hansen, V.T. Chuang, M. Otagiri, Biol. Pharm. Bull. 25 (2002) 695– 704. [20] Y. Sakurai, S.F. Ma, H. Watanabe, N. Yamaotsu, S. Hirono, Y. Kurono, U. KraghHansen, M. Otagiri, Pharm. Res. 21 (2004) 285–292. [21] P. Ascenzi, A. Bocedi, P. Visca, F. Altruda, E. Tolosano, T. Beringhelli, M. Fasano, IUBMB Life 57 (2005) 749–759. [22] M. Fasano, S. Curry, E. Terreno, M. Galliano, G. Fanali, P. Narciso, S. Notari, P. Ascenzi, IUBMB Life 57 (2005) 787–796. [23] J. Ghuman, P.A. Zunszain, I. Petitpas, A.A. Bhattacharya, M. Otagiri, S. Curry, J. Mol. Biol. 353 (2005) 38–52. [24] P. Ascenzi, A. Bocedi, S. Notari, G. Fanali, R. Fesce, M. Fasano, Mini. Rev. Med. Chem. 6 (2006) 483–489. [25] F. Yang, C. Bian, L. Zhu, G. Zhao, Z. Huang, M. Huang, J. Struct. Biol. 157 (2007) 348–355. [26] D. Gupta, C.G. Lis, Nutr. J. 9 (2010) 69. [27] M. Koga, S. Kasayama, Endocr. J. 57 (2010) 751–762. [28] E. Sbarouni, P. Georgiadou, V. Voudris, Clin. Chem. Lab. Med. 49 (2011) 177– 184. [29] M.R. Alexander, B. Alexander, A.L. Mustion, R. Spector, C.B. Wright, JAMA 247 (1982) 831–833. [30] B.L. Erstad, B.J. Gales, W.D. Rappaport, Arch. Intern. Med. 151 (1991) 901– 911. [31] G.E. Hastings, P.G. Wolf, Arch. Fam. Med. 1 (1992) 281–287. [32] G.R. Haynes, R.J. Navickis, M.M. Wilkes, Eur. J. Anaesthesiol. 20 (2003) 771– 793. [33] P. Alderson, F. Bunn, C. Lefebvre, W.P. Li, L. Li, I. Roberts, G. Schierhout, Cochrane Database Syst. Rev. (2004) CD001208. [34] A. Liberati, L. Moja, I. Moschetti, G.F. Gensini, R. Gusinu, Intern. Emerg. Med. 1 (2006) 243–245. [35] C. Bertucci, M. Pistolozzi, A. De Simone, Anal. Bioanal. Chem. 398 (2010) 155– 166. [36] M. Furukawa, R. Tanaka, V.T. Chuang, Y. Ishima, K. Taguchi, H. Watanabe, T. Maruyama, M. Otagiri, J. Control Release 154 (2011) 189–195. [37] T. Komatsu, X. Qu, H. Ihara, M. Fujihara, H. Azuma, H. Ikeda, J. Am. Chem. Soc. 133 (2011) 3246–3248. [38] A. Chopra, Molecular Imaging and Contrast Agent Database (MICAD), National Center for Biotechnology Information (US), Bethesda (MD), 2013, pp 2004–2013. [39] S. Sugio, A. Kashima, S. Mochizuki, M. Noda, K. Kobayashi, Protein Eng. 12 (1999) 439–446. [40] M. Wardell, Z. Wang, J.X. Ho, J. Robert, F. Ruker, J. Ruble, D.C. Carter, Biochem. Biophys. Res. Commun. 291 (2002) 813–819. [41] A.A. Bhattacharya, S. Curry, N.P. Franks, J. Biol. Chem. 275 (2000) 38731– 38738. [42] A.A. Bhattacharya, T. Grune, S. Curry, J. Mol. Biol. 303 (2000) 721–732. [43] I. Petitpas, A.A. Bhattacharya, S. Twine, M. East, S. Curry, J. Biol. Chem. 276 (2001) 22804–22809.

[44] I. Petitpas, C.E. Petersen, C.E. Ha, A.A. Bhattacharya, P.A. Zunszain, J. Ghuman, N.V. Bhagavan, S. Curry, Proc. Natl. Acad. Sci. USA 100 (2003) 6440–6445. [45] P.A. Zunszain, J. Ghuman, T. Komatsu, E. Tsuchida, S. Curry, BMC Struct. Biol. 3 (2003) 6. [46] G. Fanali, A. Bocedi, P. Ascenzi, M. Fasano, FEBS J. 274 (2007) 4491–4502. [47] G.J. van der Vusse, Drug Metab. Pharmacokinet. 24 (2009) 300–307. [48] A. Varshney, P. Sen, E. Ahmad, M. Rehan, N. Subbarao, R.H. Khan, Chirality 22 (2010) 77–87. [49] P. Ascenzi, A. di Masi, L. Leboffe, T. Alberio, G. Fanali, M. Fasano, IUBMB Life (2013). [50] R.L. Gundry, Q. Fu, C.A. Jelinek, J.E. Van Eyk, R.J. Cotter, Proteomics Clin. Appl. 1 (2007) 73–88. [51] R.L. Gundry, M.Y. White, J. Nogee, I. Tchernyshyov, J.E. Van Eyk, Proteomics 9 (2009) 2021–2028. [52] D. Scumaci, M. Gaspari, M. Saccomanno, G. Argiro, B. Quaresima, C.M. Faniello, P. Ricci, F. Costanzo, G. Cuda, Anal. Biochem. 418 (2011) 161–163. [53] S. Lejon, J.F. Cramer, P. Nordberg, Acta Crystallogr., Sect. F: Struct. Biol. Cryst. Commun. 64 (2008) 64–69. [54] F. Polticelli, S. Caprari, S. Gianni, P. Ascenzi, IUBMB Life 64 (2012) 885–888. [55] G. Fanali, G. Pariani, P. Ascenzi, M. Fasano, FEBS J. 276 (2009) 2241–2250. [56] G. Fanali, Y. Cao, P. Ascenzi, V. Trezza, T. Rubino, D. Parolaro, M. Fasano, IUBMB Life 63 (2011) 446–451. [57] A. di Masi, F. Gullotta, A. Bolli, G. Fanali, M. Fasano, P. Ascenzi, FEBS J. 278 (2011) 654–662. [58] K. Yamasaki, V.T. Chuang, T. Maruyama, M. Otagiri, Biochim. Biophys. Acta (2013). [59] A. Bocedi, S. Notari, P. Narciso, A. Bolli, M. Fasano, P. Ascenzi, IUBMB Life 56 (2004) 609–614. [60] A. Bocedi, S. Notari, E. Menegatti, G. Fanali, M. Fasano, P. Ascenzi, FEBS J. 272 (2005) 6287–6296. [61] P. Ascenzi, F. Imperi, M. Coletta, M. Fasano, Biochem. Biophys. Res. Commun. 369 (2008) 686–691. [62] P. Ascenzi, M. Fasano, Biochem. Biophys. Res. Commun. 353 (2007) 469–474. [63] G. Fanali, V. Rampoldi, A. di Masi, A. Bolli, L. Lopiano, P. Ascenzi, M. Fasano, IUBMB Life 62 (2010) 371–376. [64] C.E. Petersen, A.G. Scottolini, L.R. Cody, M. Mandel, N. Reimer, N.V. Bhagavan, J. Med. Genet. 31 (1994) 355–359. [65] T. Sunthornthepvarakul, P. Angkeow, R.E. Weiss, Y. Hayashi, S. Refetoff, Biochem. Biophys. Res. Commun. 202 (1994) 781–787. [66] N. Wada, H. Chiba, C. Shimizu, H. Kijima, M. Kubo, T. Koike, J. Clin. Endocrinol. Metab. 82 (1997) 3246–3250. [67] D.L. Schonfeld, R.B. Ravelli, U. Mueller, A. Skerra, J. Mol. Biol. 384 (2008) 393– 405. [68] K. Nishi, T. Ono, T. Nakamura, N. Fukunaga, M. Izumi, H. Watanabe, A. Suenaga, T. Maruyama, Y. Yamagata, S. Curry, M. Otagiri, J. Biol. Chem. 286 (2011) 14427–14434. [69] K. Matsumoto, K. Sukimoto, K. Nishi, T. Maruyama, A. Suenaga, M. Otagiri, Drug Metab. Pharmacokinet. 17 (2002) 300–306. [70] I. Fitos, J. Visy, M. Simonyi, J. Biochem. Biophys. Methods 54 (2002) 71–84. [71] F. Herve, G. Caron, J.C. Duche, P. Gaillard, N. Abd Rahman, A. TsantiliKakoulidou, P.A. Carrupt, P. d’Athis, J.P. Tillement, B. Testa, Mol. Pharmacol. 54 (1998) 129–138. [72] V.T. Chuang, M. Hijioka, M. Katsuki, K. Nishi, T. Hara, K. Kaneko, M. Ueno, A. Kuniyasu, H. Nakayama, M. Otagiri, Biochim. Biophys. Acta 1725 (2005) 385– 393. [73] D.A. Cheresh, D.H. Haynes, J.A. Distasio, Immunology 51 (1984) 541–548. [74] S.A. Elg, A.R. Mayer, L.F. Carson, L.B. Twiggs, R.B. Hill, S. Ramakrishnan, Cancer 80 (1997) 1448–1456. [75] S.D. Shiyan, N.V. Bovin, Glycoconj. J. 14 (1997) 631–638. [76] W. van Dijk, Adv. Exp. Med. Biol. 376 (1995) 223–229. [77] M. Costello, B.A. Fiedel, H. Gewurz, Nature 281 (1979) 677–678. [78] A.V. Timoshenko, N.V. Bovin, S.D. Shiyan, S.Y. Vakhrushev, S. Andre, H.J. Gabius, Biochemistry (Mosc) 63 (1998) 546–550. [79] H. Tilg, E. Vannier, G. Vachino, C.A. Dinarello, J.W. Mier, J. Exp. Med. 178 (1993) 1629–1636. [80] B. Haraldsson, B. Rippe, Acta Physiol. Scand. 129 (1987) 127–135. [81] A.G. Philip, Pediatrie 39 (1984) 371–378. [82] W. Chachaj, Z. Bartecka, J. Malolepszy, Arch. Immunol. Ther. Exp. (Warsz) 28 (1980) 947–951. [83] P.J. McNamara, R.C. Jewell, M.N. Gillespie, Prog. Clin. Biol. Res. 300 (1989) 307–319. [84] N. de Vera, R.M. Cristofol, E. Rodriguez Farre, Life Sci. 43 (1988) 1277–1286. [85] B. Grimaldi, C. Hamberger, D. Tremblay, J. Barre, J.P. Tillement, Prog. Clin. Biol. Res. 300 (1989) 445–448. [86] A. Szallasi, N.E. Lewin, P.M. Blumberg, J. Pharmacol. Exp. Ther. 262 (1992) 883–888. [87] D. Morin, N. Simon, P. Depres-Brummer, F. Levi, J.P. Tillement, S. Urien, Pharmacology 54 (1997) 271–275. [88] G. Sager, H. Bratlid, C. Little, Biochem. Pharmacol. 36 (1987) 3607–3612. [89] M.J. Chapman, J. Lipid Res. 21 (1980) 789–853. [90] J.T. Sparrow, A.M. Gotto Jr., Ann. NY Acad. Sci. 348 (1980) 187–211. [91] K.K. Ng, J.F. Lovell, G. Zheng, Acc. Chem. Res. 44 (2011) 1105–1113. [92] G.M. Kostner, Adv. Lipid Res. 20 (1983) 1–43. [93] M. Irshad, R. Dubey, Indian J. Biochem. Biophys. 42 (2005) 73–80. [94] S.W. Law, K.J. Lackner, S.S. Fojo, A. Hospattankar, J.C. Monge, H.B. Brewer Jr., Adv. Exp. Med. Biol. 201 (1986) 151–162.

P. Ascenzi et al. / Journal of Molecular Structure 1077 (2014) 4–13 [95] A.T. Remaley, Arterioscler. Thromb. Vasc. Biol. 33 (2013) 166–167. [96] F.R. Silvia, M. Jeffery, A. Gred, H. Brian, LCAT Deficiency and Fish Eye Disease, eighth ed., vol. 118, McGrow Hill, London UK, 2001, pp. 2817-2833. [97] J. Scott, J. R. Coll. Physicians Lond. 24 (1990) 101–106. [98] G.D. Calvert, M. Abbey, Adv. Clin. Chem. 24 (1985) 217–298. [99] K.S. Montine, C.N. Bassett, J.J. Ou, W.R. Markesbery, L.L. Swift, T.J. Montine, J. Lipid Res. 39 (1998) 2443–2451. [100] G. Perdomo, H. Henry Dong, Aging (Albany NY) 1 (2009) 17–27. [101] P.W. Wilson, Clin. Chem. 41 (1995) 165–169. [102] J.C. Yau, L.J. Brunner, G. Lopez-Berestein, C.F. LeMaistre, D.R. Luke, Pharmacotherapy 11 (1991) 291–295. [103] A. Shayeganpour, S.D. Lee, K.M. Wasan, D.R. Brocks, Pharm. Res. 24 (2007) 672–678. [104] M.P. McIntosh, C.J. Porter, K.M. Wasan, M. Ramaswamy, W.N. Charman, J. Pharm. Sci. 88 (1999) 378–384. [105] K.M. Wasan, G.A. Brazeau, A. Keyhani, A.C. Hayman, G. Lopez-Berestein, Antimicrob. Agents Chemother. 37 (1993) 246–250. [106] D.P. Rossignol, K.M. Wasan, E. Choo, E. Yau, N. Wong, J. Rose, J. Moran, M. Lynn, Antimicrob. Agents Chemother. 48 (2004) 3233–3240. [107] N. Sabnis, A.G. Lacko, Ther. Deliv. 3 (2012) 599–608. [108] A. Pluen, Y. Boucher, S. Ramanujan, T.D. McKee, T. Gohongi, E. di Tomaso, E.B. Brown, Y. Izumi, R.B. Campbell, D.A. Berk, R.K. Jain, Proc. Natl. Acad. Sci. USA 98 (2001) 4628–4633. [109] G. Zheng, J. Chen, H. Li, J.D. Glickson, Proc. Natl. Acad. Sci. USA 102 (2005) 17757–17762. [110] K.M. Wasan, D.R. Brocks, S.D. Lee, K. Sachs-Barrable, S.J. Thornton, Nat. Rev. Drug Discov. 7 (2008) 84–99. [111] C.J. Fielding, P.E. Fielding, J. Lipid Res. 36 (1995) 211–228. [112] P.G. Yancey, A.E. Bortnick, G. Kellner-Weibel, M. de la Llera-Moya, M.C. Phillips, G.H. Rothblat, Arterioscler. Thromb. Vasc. Biol. 23 (2003) 712–719. [113] J.M. Martin-Campos, J.C. Escola-Gil, V. Ribas, F. Blanco-Vaca, Curr. Opin. Lipidol. 15 (2004) 247–253. [114] H.W. Eckerson, C.M. Wyte, B.N. La Du, Am. J. Hum. Genet. 35 (1983) 1126– 1138. [115] J. Babitt, B. Trigatti, A. Rigotti, E.J. Smart, R.G. Anderson, S. Xu, M. Krieger, J. Biol. Chem. 272 (1997) 13242–13249. [116] A.G. Lacko, M. Nair, S. Paranjape, S. Johnso, W.J. McConathy, Anticancer Res. 22 (2002) 2045–2049. [117] M.N. Oda, P.L. Hargreaves, J.A. Beckstead, K.A. Redmond, R. van Antwerpen, R.O. Ryan, J. Lipid Res. 47 (2006) 260–267. [118] M.K. Bijsterbosch, H. van de Bilt, T.J. van Berkel, Biochem. Pharmacol. 52 (1996) 113–121. [119] W.J. McConathy, M.P. Nair, S. Paranjape, L. Mooberry, A.G. Lacko, Anticancer Drugs 19 (2008) 183–188. [120] M. Tachikawa, S. Okayasu, K. Hosoya, Mol. Vis. 13 (2007) 2041–2047. [121] K.C. Vickers, B.T. Palmisano, B.M. Shoucri, R.D. Shamburek, A.T. Remaley, Nat. Cell Biol. 13 (2011) 423–433. [122] D. Baek, J. Villen, C. Shin, F.D. Camargo, S.P. Gygi, D.P. Bartel, Nature 455 (2008) 64–71. [123] M. Selbach, B. Schwanhausser, N. Thierfelder, Z. Fang, R. Khanin, N. Rajewsky, Nature 455 (2008) 58–63. [124] X. Liu, R. Suo, S.L. Xiong, Q.H. Zhang, G.H. Yi, Clin. Chim. Acta 415 (2013) 94– 100. [125] P.S. Mitchell, R.K. Parkin, E.M. Kroh, B.R. Fritz, S.K. Wyman, E.L. PogosovaAgadjanyan, A. Peterson, J. Noteboom, K.C. O’Briant, A. Allen, D.W. Lin, N.

[126] [127] [128] [129] [130] [131] [132] [133] [134] [135] [136] [137] [138] [139] [140] [141] [142] [143] [144] [145] [146] [147] [148] [149] [150] [151] [152] [153] [154] [155] [156] [157]

13

Urban, C.W. Drescher, B.S. Knudsen, D.L. Stirewalt, R. Gentleman, R.L. Vessella, P.S. Nelson, D.B. Martin, M. Tewari, Proc. Natl. Acad. Sci. USA 105 (2008) 10513–10518. R.J. Pease, J.M. Leiper, Curr. Opin. Lipidol. 7 (1996) 132–138. F.J. Field, S.N. Mathur, J. Lipid Res. 24 (1983) 1049–1059. J.A. Yanez, S.W. Wang, I.W. Knemeyer, M.A. Wirth, K.B. Alton, Adv. Drug Deliv. Rev. 63 (2011) 923–942. K.M. Wasan, M. Ramaswamy, M.P. McIntosh, C.J. Porter, W.N. Charman, J. Pharm. Sci. 88 (1999) 185–190. D.R. Brocks, M. Ramaswamy, A.I. MacInnes, K.M. Wasan, Pharm. Res. 17 (2000) 427–431. K.M. Wasan, R.E. Morton, M.G. Rosenblum, G. Lopez-Berestein, J. Pharm. Sci. 83 (1994) 1006–1010. R.D. Meyer, Clin. Infect. Dis. 14 (Suppl 1) (1992). S154-60. F. Meunier, J. Infect. 28 (Suppl 1) (1994) 51–56. M. Lemaire, W.M. Pardridge, G. Chaudhuri, J. Pharmacol. Exp. Ther. 244 (1988) 740–743. K.M. Wasan, M. Ramaswamy, W. Wong, P.H. Pritchard, J. Pharmacol. Exp. Ther. 284 (1998) 599–605. M. Kwong, O. Sivak, E.H. Kwong, K.M. Wasan, J. Pharm. Sci. 90 (2001) 1308– 1317. N.S. Chung, K.M. Wasan, Adv. Drug Deliv. Rev. 56 (2004) 1315–1334. L.J. Notarianni, Clin. Pharmacokinet. 18 (1990) 20–36. D.J. Greenblatt, J. Am. Geriatr. Soc. 27 (1979) 20–22. A. Corsonello, C. Pedone, R.A. Incalzi, Curr. Med. Chem. 17 (2010) 571– 584. R.K. Verbeeck, J.A. Cardinal, S.M. Wallace, Eur. J. Clin. Pharmacol. 27 (1984) 91–97. D.A. Gilmore, J. Gal, J.G. Gerber, A.S. Nies, J. Pharmacol. Exp. Ther. 261 (1992) 1181–1186. R. Mehvar, D.R. Brocks, J. Pharm. Pharm. Sci. 4 (2001) 185–200. D.J. Greenblatt, M. Divoll, J.S. Harmatz, R.I. Shader, J. Pharmacol. Exp. Ther. 215 (1980) 86–91. M. Divoll, D.J. Greenblatt, J.S. Harmatz, R.I. Shader, J. Pharm. Sci. 70 (1981) 1104–1107. D.J. Greenblatt, M. Divoll, J.S. Harmatz, D.S. MacLaughlin, R.I. Shader, Clin. Pharmacol. Ther. 30 (1981) 475–486. S. Biagiotti, M.F. Paoletti, A. Fraternale, L. Rossi, M. Magnani, IUBMB Life 63 (2011) 621–631. P. Ascenzi, A. Bellelli, M. Coletta, A. Colosimo, G. Falcioni, G.M. Giacometti, R. Ippoliti, L. Zolla, B. Giardina, IUBMB Life 59 (2007) 600–616. P.H. Hinderling, Pharmacol. Rev. 49 (1997) 279–295. Y.I. Miller, N. Shaklai, Biochim. Biophys. Acta 1454 (1999) 153–164. M.E. Conrad, J.N. Umbreit, Blood Cells Mol. Dis. 29 (2002) 336–355. T. Goswami, A. Rolfs, M.A. Hediger, Biochem. Cell Biol. 80 (2002) 679–689. V. Hvidberg, M.B. Maniecki, C. Jacobsen, P. Hojrup, H.J. Moller, S.K. Moestrup, Blood 106 (2005) 2572–2579. E. Tolosano, S. Fagoonee, N. Morello, F. Vinchi, V. Fiorito, Antioxid. Redox Signal. 12 (2010) 305–320. P. Ascenzi, Y. Cao, G.R. Tundo, M. Coletta, G. Fanali, M. Fasano, Biochem. Biophys. Res. Commun. 411 (2011) 185–189. N. Guex, M.C. Peitsch, T. Schwede, Electrophoresis 30 (suppl 1) (2009). S16273. E.F. Pettersen, T.D. Goddard, C.C. Huang, G.S. Couch, D.M. Greenblatt, E.C. Meng, T.E. Ferrin, J. Comput. Chem. 25 (2004) 1605–1612.