A comparative study of the interaction of warfarin with human alpha 1-acid glycoprotein and human albumin

The interaction of warfarin with human alpha 1-acid glycoprotein (alpha 1-AGP) and human albumin (HA) has been investigated using fluorescence and circular dichroism techniques. The fluorescence of warfarin is greatly enhanced following binding to alpha 1-AGP or HA, the binding constant for a single site being estimated by the Scatchard method. The binding constants for the two serum proteins are similar, but the thermodynamic parameters differ. The binding constants increase as the pH is raised to 9.0. Various basic drugs, such as chlorpromazine, propranolol and imipramine, markedly inhibited the binding of warfarin to alpha 1-AGP. But, some acidic drugs, including phenylbutazone, effectively displaced warfarin bound to HA. The difference in CD spectra observed for alpha 1-AGP and HA indicated that the drug-binding sites of the two proteins might have different asymmetries. It thus appears that the mode of interaction of warfarin with the two proteins differs.     

Prazosin binding to human alpha 1-acid glycoprotein (orosomucoid), human serum albumin, and human serum. Further characterization of the 'single drug binding site' of orosomucoid

The plasma protein binding of the alpha 1-adrenergic blocking agent prazosin was investigated by means of circular dichroism (CD) and equilibrium dialysis (ED) measurements. The interaction of prazosin with human alpha 1-acid glycoprotein (alpha 1-AGP) results in pronounced negative extrinsic Cotton effects at 255 nm and a smaller negative band at 285 nm which are associated with the binding of prazosin to only one site of the protein. Various basic drugs, and warfarin also, at 50 microM displace prazosin 10 microM from its binding site on alpha 1-AGP and reduce the CD-spectra at 255 nm by 26% (disopyramide), 52% (mepivacaine), about 70% (verapamil, biperiden), and 90-100% (trihexyphenidyl, warfarin). (+/-)-Propranolol reduces the CD-spectra by 76%, its (-)-isomer by 89%, and the (+)-isomer by 65%. ED experiments indicated that the binding of prazosin to alpha 1-AGP is saturable with an association constant of 48 000 M-1 and 0.85 binding sites per protein molecule. Displacement of prazosin from alpha 1-AGP by the same drug as used for the CD experiments at displacer/prazosin ratios of 5 resulted in comparable reductions of the fraction bound as obtained by the CD experiments. Prazosin was also highly bound to human serum albumin (600 microM) with about 80-85% bound at prazosin concentrations from 1-100 microM. Since prazosin binding to human serum is only slightly higher (80-90%) it is concluded that prazosin binding in serum is largely mediated by the albumin fraction.(ABSTRACT TRUNCATED AT 250 WORDS)     

The binding characteristics of some adrenergic beta-receptor antagonists to human serum proteins

Binding of pindolol and 8 related compounds was studied in vitro by equilibrium dialysis. The overall binding in serum was compared with the binding to the main, isolated, serum proteins. Most substances show both saturable and non-saturable binding in serum. The saturable and main binding is to alpha 1-AGP, the low non-saturable binding corresponds to albumin and lipoprotein binding. The binding to alpha 1-AGP is characterized by approximately one binding site and association constants K ranging from 10(4) to 10(6) M-1. The binding of pindolol to alpha 1-AGP is strongly inhibited by propranolol, lidocaine, erythromycin, imipramine and TBEP. Significant correlations were found between log NK and log partition coefficient octanol-phosphate buffer suggesting that the protein binding of the 9 adrenergic beta-receptor antagonists to all serum proteins, including alpha 1-AGP, is predominantly hydrophobic in nature.     

Stereoselective binding of propranolol enantiomers to human alpha 1-acid glycoprotein and human plasma.

The binding of propranolol enantiomers to human albumin (ALB), alpha 1-acid glycoprotein (alpha 1-AGP) and plasma was studied. (-) propranolol is more bound than (+)propranolol to alpha 1-AGP (P less than 0.001) and to plasma (P less than 0.05). In solutions containing ALB at a constant concentration (580 mumol/l) and alpha 1-AGP at increasing concentrations, the binding of both isomers increases but the stereo selectivity is evident throughout the alpha 1-AGP concentration range examined (25-100 mumol/l).     

Differences in the serum protein binding of prazosin in man and rat

The serum protein binding of prazosin in man and rat has been studied in vitro by equilibrium dialysis. Prazosin was more extensively bound in human serum than in rat serum with binding ratios (B/F) of 14.3 +/- 3.4 and 4.4 +/- 0.2 (corresponding to 93.4 and 81.4% bound), respectively. This difference in binding between the species was partly due to qualitative differences between human and rat serum albumin, but also to the lower concentration of albumin in rat serum. Rat serum albumin (RSA) apparently showed two different classes of binding sites for prazosin, one with high (KD = 5.78 X 10(-6) M) and one with low (KD = 1.1 X 10(-4) M) affinity; the former is suggested as representing alpha 1-acid glycoprotein (alpha 1-AGP) with one binding site for prazosin per molecule, the latter as representing RSA with 0.28 binding sites per molecule. Human serum albumin (HSA) and human alpha 1-AGP both showed one class of binding sites with KD values of 2.7 X 10(-5) and 1.95 X 10(-6) M, respectively. HSA possessed 0.5 and human alpha 1-AGP 1 binding site for prazosin per molecule. The binding parameters obtained for the isolated serum proteins overestimated to some degree the total serum protein binding of prazosin in man. This was explained by a specific deviation from the law of mass action. HSA was the major binding protein in human serum at therapeutic concentrations, with ca. 60% of the total binding, the remaining 40% being bound to alpha 1-AGP. Anticipating that the high affinity binding site on the RSA preparation represents the binding of prazosin to alpha 1-AGP, then this protein accounts for 70% of the binding in rat serum, while rat serum albumin accounts for approximately 23%. The binding of prazosin to lipoproteins was insignificant in both species. The observed differences between man and rat in the serum protein binding of prazosin implicate differences in the two species with respect to prazosin pharmacokinetics and the pharmacological effect.     

Investigations into the binding of phenprocoumon to albumin using fluorescence spectroscopy

The fluorescence of phenprocoumon is enhanced following interaction with bovine and human serum albumin, allowing the binding parameters for two binding sites to be estimated by a modified Bjerrum method and the Scatchard method. The fluorescence is increased as the pH is raised to 9·0 and also on addition of the drugs ibuprofen and fenoprofen. Chloride ions, phenylbutazone and acenocoumarin seem to displace phenprocoumon from its primary binding site on human serum albumin. Investigations with the enantiomers of phenprocoumon show that the binding sites have some stereospecificity.     

The interaction of human and bovine serum proteins with CYP3A in human liver microsomes: inhibition of testosterone 6beta-hydroxylation by albumin,alpha-globulins,alpha(1)-acid glycoprotein and gamma-globulins

The effects of human and bovine serum proteins on CYP3A activity, using testosterone as the probe substrate, were investigated in human liver microsomes. Serum albumin, alpha-globulins, and alpha(1)-acid glycoprotein (alpha(1)-AGP) of both species significantly inhibited testosterone 6beta-hydroxylation. When the inhibitory effects of serum proteins were compared with serum protein binding data, human alpha-globulins, with a ratio (relative metabolic activity/unbound fraction) of 0.3, showed higher, and bovine alpha(1)-AGP, with the ratio of 1.4, showed lower inhibitory effects than those expected from protein binding of testosterone. The effects of the other serum proteins were close to those expected from protein binding, according to the free drug hypothesis. The K(i) values obtained from the Dixon plots were 0.32% (w/v, 48 microM) for human serum albumin (HSA), 0.48% for human alpha-globulins, and 0.23% (52 microM) for human alpha(1)-AGP. K(i) values of bovine serum albumin, bovine alpha-globulins and bovine alpha(1)-AGP were 3-5 times higher than those of the respective human proteins. The results suggest a direct interaction of some of these serum proteins with the active site of the CYP3A isoform. Since the bovine serum proteins showed weaker inhibitory effects than human serum proteins, the wide use of BSA, which is viewed as interchangeable with HSA, needs to be cautioned.     

Inhibitory effect of phenothiazines on the binding of [3H]perazine to α1-acid glycoprotein

A system is described which allows the determination of the afffinity constant of unlabelled drugs to α₁-acid glycoprotein (α₁-AGP) by displacing [³H]perazine from the binding protein with equilibrium dialysis. All drugs investigated appear to bind to only one site at the α₁-AGP molecule. From experiments, in which the chemical structure of the displacers was varied, the fragment 21–31 of the amino acid sequence appears to be a candidate for hydrophobic interactions. The glutamic acids 177 and 178 of the α₁-AGP molecule could be involved in ionic interactions with the side chain of phenothiazine derivatives. The relevance of α₁-AGP for drug binding, distribution, and the possible reasons for insufficient correlation between psychotropic plasma concentration and therapeutic response is discussed.     

Protein binding of chloroquine enantiomers and desethylchloroquine.

The protein binding of racemic chloroquine, its enantiomers and desethylchloroquine to plasma, purified human albumin, and alpha 1-acid glycoprotein (alpha 1-AGP) was determined by equilibrium dialysis. The binding was not concentration dependent. (+)-Chloroquine bound more to plasma (66.6 +/- 1.9%) and albumin (45.9 +/- 0.8%) than (-)-chloroquine (48.5 +/- 2.4% and 35.3 +/- 0.6%, respectively). These differences were statistically significant. (-)-Chloroquine bound more to alpha 1-AGP (47.5 +/- 0.7%) than (+)-chloroquine (34.5 +/- 0.5%). The binding of desethylchloroquine to alpha 1-AGP is higher than to albumin (38.9 +/- 0.9% and 21.1 +/- 0.4%, respectively.     

Interaction of plasma proteins with cytochromes P450 mediated metabolic reactions: inhibition by human serum albumin and alpha-globulins of the debrisoquine 4-hydroxylation (CYP2D) in liver microsomes of human, hamster and rat

The effect of human serum albumin (HSA), alpha1-acid glycoprotein (alpha1-AGP), and alpha- and gamma-globulins on the in vitro metabolism of debrisoquine in human, hamster and rat liver microsomes was studied. Interaction of albumin with cytochrome P450 mediated phenytoin metabolism has been reported. Since plasma protein binding of phenytoin is high, in the present study a weakly protein bound drug, debrisoquine, was studied. Debrisoquine is a substrate of CYP2D6. The debrisoquine 4-hydroxylation was measured using a radio-TLC method. Among the four plasma proteins, alpha-globulins had the strongest inhibitory effect on the debrisoquine 4-hydroxylase activity. The inhibition with 2% alpha-globulins was 42+/-18% for human and higher for hamster and rat liver microsomes (65-71%). HSA had less effect than alpha-globulins. In the presence of HSA, the decrease in activity was between 18 and 35% for all liver microsomes studied. The debrisoquine 4-hydroxylase activity was not significantly changed by alpha1-AGP or gamma-globulins. Using an ultra-filtration method, the protein binding of debrisoquine to 4% HSA, 0.5% alpha1-AGP, 2% alpha-globulins and 2% gamma-globulins was found to be 22, 20, 22 and 5%, respectively. Since the observed inhibition is inconsistent with level of protein binding, it appears, particularly in the case of alpha-globulins, that the plasma proteins interact with CYP2D directly.     

Spectroscopic study of the interaction of coumarin anticoagulant drugs with polyvinylpyrrolidone

The interaction of coumarin anticoagulants with polyvinylpyrrolidone (PVP) was investigated using a fluorescence technique. The fluorescence intensities of warfarin and phenprocoumon were greatly enhanced following binding to PVP, while the fluorescence of 4-hydroxycoumarin was little enhanced in the presence of PVP. The enhanced fluorescence of warfarin and phenprocoumon bound to PVP can be explained by their incorporation into the hydrophobic environment in the PVP and by a decrease in the internal rotation of the alpha-substituted benzyl group in the drugs. The binding parameters of warfarin and phenprocoumon were estimated by the Klotz method; the binding constants for phenprocoumon and warfarin were found to be 2.6 X 10(4) and 2.2 X 10(4) M-1, respectively. The 13C-NMR measurements suggest the lactone moiety in the 4-hydroxycoumarin and the substituted benzene ring play an important role in the binding to PVP.     

Distribution of moricizine in human blood: binding to plasma proteins and erythrocytes.

Using equilibrium dialysis and incubation experiments, we determined the binding of moricizine to human plasma, isolated plasma proteins, and erythrocytes. The mean (% +/- SD) plasma protein binding at various moricizine concentrations ranged from 81.2 +/- 2.1 to 89.9 +/- 2.1%. There was no apparent relationship between drug concentration and extent of binding in pooled plasma over the concentration range tested. However, protein concentration-dependent binding was observed with albumin and alpha 1-acid glycoprotein (alpha 1-AGP). The unbound fraction of moricizine fell from 61 to 19% and from 70 to 17% with increasing albumin (5 and 50 g/L, respectively) and alpha 1-AGP (0.2 and 1.2 g/L) concentrations. The binding of moricizine to beta-lipoprotein (5 g/L) was 70.6 +/- 3.1% and to gamma-globulin (12 g/L) was 13.6 +/- 3.3%. Moricizine partitioned into erythrocytes, showing an erythrocyte/plasma drug concentration ratio of 1.325 +/- 0.070 and erythrocyte/buffer ratio of 8.561 +/- 0.620. An estimation could be made that 57% of total drug in whole blood was associated with erythrocytes, 39% bound to plasma proteins, and 4% was free. The results of this study demonstrated that erythrocytes, albumin, and alpha 1-AGP were the major binding components in blood.     

Binding of vinca alkaloid analogues to human serum albumin and to alpha 1-acid glycoprotein

The binding of a series of vinca alkaloid analogues having eburnane or indolo[2,3-a]quinolizidine skeletons was studied with human serum albumin (HSA) by affinity chromatography and with alpha 1-acid glycoprotein by means of competition experiments. On HSA the binding occurs at the benzodiazepine-indole binding site via hydrophobic interaction and shows slight stereoselectivity preferring the trans isomers. The binding to alpha 1-AGP proved to be highly stereoselective in favour of the trans isomers having 3(S),16(R)eburnane or 1(R),12b(S)indolo[2,3-a]quinolizidine absolute configurations.     

Auramine O as a fluorescent probe for the binding of basic drugs to human alpha 1-acid glycoprotein (alpha 1-AG). The development of a simple fluorometric method for the determination of alpha 1-AG in human serum

A cationic fluorescent dye, auramine O (AO), exhibited an intense increase in fluorescence after binding to human alpha 1-acid glycoprotein (alpha 1-AG). The interaction between AO and the protein was studied by fluorescence spectroscopy and by equilibrium dialysis. AO binds to the protein via a single site with a dissociation constant of 24 microM. Various basic drugs such as chlorpromazine, imipramine, desipramine, quinidine, propranolol and lidocaine, which are known to bind to the protein, competitively inhibited the AO binding to the protein. The dissociation constants of these basic drugs obtained from such inhibitory experiments were comparable to those obtained with other methods (equilibrium dialysis, quenching of protein intrinsic fluorescence, and the difference spectrophotometric method) and from the literature. It is concluded that AO may be a useful fluorescent probe that binds to a single basic drug binding site on alpha 1-AG. In addition, a simple fluorometric method for the determination of alpha 1-AG in serum was developed using AO, and the validity of this method was confirmed by comparing it with the conventional radial immunodiffusion method.     

Characterization of a binding site of UCN-01, a novel anticancer drug on alpha-acid glycoprotein

The binding site of 7-hydroxystaurosporine (UCN-01) on alpha-acid glycoprotein (AGP) was studied by fluorescence and ultracentrifugation experiments. Three ligands, propranolol, warfarin and progesterone were employed as marker ligands and quinaldine red was employed as a fluorescent probe. The presence of UCN-01, pro- pranolol, warfarin and progesterone resulted in a significant quenching of the fluorescence of quinaldine red, when bound to AGP, depending upon the potency of the binding to AGP. The construction of Klotz plots indicated that the displacement effects of propranolol, warfarin and progesterone on UCN-01-AGP binding were competitive in nature. These data suggest that the binding site of UCN-01 on the AGP partly overlaps the binding site for basic drugs, acidic drugs, as well as steroid hormones.     

Decrease in penbutolol central response as a cause of changes in its serum protein binding

Penbutolol is a beta-adrenoceptor antagonist that is extensively bound to alpha 1-acid glycoprotein (alpha 1-AGP), a protein that increases in inflammatory diseases thereby binding more drug in such conditions. Changes in serum binding can lead to modifications in the pharmacokinetics and pharmacodynamics of a drug, therefore, the central effect (as the anticonvulsant response) and brain uptake of penbutolol given intravenously to mice with experimental inflammation have been measured. A significant decrease of the central effect of penbutolol and its brain uptake was seen in diseased when compared with control animals (P less than 0.01). A parallel decrease in free fraction of penbutolol in diseased vs normal animals was detected. These results suggest that there is an increase in serum binding of basic drugs related to increments in alpha 1-AGP concentration, which reduces their central pharmacological effect.     

Inverse stereoselectivity in the binding of acenocoumarol to human serum albumin and to alpha 1-acid glycoprotein

Stereoselective binding of rac-acenocoumarol to human serum albumin (HSA) and alpha 1-acid glycoprotein (alpha 1-AGP) was investigated by affinity chromatography and by combined ultrafiltration (UF) and circular dichroism (CD) methods. For HSA, the ratio of the enantiomeric constants was KR/KS = 2, while for alpha 1-AGP, KS/KR = 3.     

Investigation of the binding of various tricyclic neuroleptics and antidepressants to alpha 1-acid glycoprotein

Recent studies have shown that the interaction of various tricyclic neuroleptics and antidepressants with isolated alpha 1-acid glycoprotein occurs at one common binding site and with relatively high association constants. The aim of the present study was to find differences in the binding of some phenothiazines, thioxanthenes, and several other drugs reported to bind alpha 1-AGP. The findings suggest that the affinity of the phenothiazines and thioxanthenes depends primarily on the existence of the tricyclic skeleton and is generally increased by the basic side chain. Substituents at position 3 of the phenothiazine nucleus influence the affinity in a variable way. The losses of radioactivity by non-specific absorptions to the dialysing chambers were considered for the calculation of the association constants. No correlation between association constants and the antipsychotic potency of neuroleptic drugs could be detected.     

Displacement of phenprocoumon (Marcumar®) from albumin by sulfonylurea compounds,suramin, and ioglycamic acid

Summary The technique of Sephadex gel filtration was employed to characterize the effect of some sulfonylurea compounds, ioglycamic acid, and suramin on the binding of phenprocoumon to bovine serum albumin.All drugs under investigation markedly increased the free phenprocoumon concentration and decreased the binding energy and binding constant for phenprocoumon to albumin. The order of displacing potency of sulfonylurea compounds was: tolbutamide < chlorpropamide < carbutamide. Suramin and ioglycamic acid were the most effective displacers. Sulfonylurea compounds and ioglycamic acid competitively inhibited the binding of phenprocoumon to albumin.In view of the high plasma levels which may occur with these displacing drugs in clinical practice, it seems possible that these drugs may influence the distribution, elimination, and anticoagulant action of phenprocoumon.This study was supported by a grant from the Bundesministerium für Jugend, Familie und Gesundheit, Bonn.     

The stereoselectivity of the 'single drug binding site' of human alpha 1-acid glycoprotein (orosomucoid)

The stereoselective binding of six pairs of basic, one pair of acidic drug enantiomers, and one pair of diastereomers for human alpha 1-acid glycoprotein was investigated by means of competition experiments against [3H]propranolol- or [14C]nicardipine-labelled binding sites using equilibrium dialysis to separate free from bound marker ligand. The affinity constants (Ka) for association of [3H]propranolol and [14C]nicardipine with alpha 1-AGP were 1.2 +/- 0.6 X 10(5) M-1 and 3.4 +/- 1.4 X 10(5) M-1, respectively, and control binding amounted to 57 +/- 7 and 91 +/- 2%, respectively. The following selectivity factors, calculated as the ratio of the higher over the lower enantiomer concentrations displacing 15% of control radiomarker binding (IC15-value), were obtained against propranolol and nicardipine: (-)/(+) propranolol: 1.9 and 1.7.; (+)-/(-)-disopyramide: 2.8 and 1.4; (+)-/(-)-verapamil: 1.6 and 1.9; (+)-(S)-/(-)-(R)-202-791, a dihydropyridine derivative: 2.6 and 2.0; (-)-/(+)-asocainol: 1.7 and 3.0; (+)-/(-)-tilidine: 1.1 and approximately equal to 2; (-)-(S)-/(+)-(R)- warfarin: 1.6 and 2.4; (+/-)-cis/(+/-)-trans-trans-tilidine: 1.7 and 1.8. When the calculation of radioligand-free fractions is also taken into account, it is apparent that only the tilidine isomers show no selectivity at propranolol-marked, and the disopyramide isomers at nicardipine-marked alpha 1-AGP-binding sites, in all other cases, a weak selectivity is detectable, which is, however, far below the values obtained for most neurotransmitter receptors. It is concluded that the single drug binding site of alpha 1-AGP is only slightly stereoselective and that the stereoselective binding of the drugs investigated is probably of no clinical consequence.