- Email: [email protected]
Interaction of quercetin and its metabolites with warfarin: Displacement of warfarin from serum albumin and inhibition of CYP2C9 enzyme
Biomedicine & Pharmacotherapy 88 (2017) 574–581
Available online at
ScienceDirect www.sciencedirect.com
Original article
Interaction of quercetin and its metabolites with warfarin: Displacement of warfarin from serum albumin and inhibition of CYP2C9 enzyme Miklós Poóra,* , Gabriella Bodaa , Paul W. Needsb , Paul A. Kroonb , Beáta Lemlic,d , Tímea Bencsike a
Department of Pharmacology, University of Pécs, Faculty of Pharmacy, Szigeti út 12, Pécs, H-7624, Hungary Institute of Food Research, Norwich Research Park, Norwich, NR4 7UA, UK c Department of General and Physical Chemistry, University of Pécs, Ifjúság útja 6, Pécs, H-7624, Hungary d János Szentágothai Research Center, Ifjúság útja 20, Pécs, H-7624, Hungary e Institute of Pharmacognosy, University of Pécs, Faculty of Pharmacy, Rókus utca 2, Pécs, H-7624, Hungary b
A R T I C L E I N F O
Article history: Received 7 November 2016 Received in revised form 14 January 2017 Accepted 15 January 2017 Keywords: Warfarin Quercetin metabolites Human serum albumin Food-drug interaction CYP2C9
A B S T R A C T
Flavonoids are ubiquitous molecules in nature with manifold pharmacological effects. Flavonoids interact with several proteins, and thus potentially interfere with the pharmacokinetics of various drugs. Though much is known about the protein binding characteristics of flavonoid aglycones, the behaviour of their metabolites, which are extensively formed in the human body has received little attention. In this study, the interactions of the flavonoid aglycone quercetin and its main metabolites with the albumin binding of the oral anticoagulant warfarin were investigated by fluorescence spectroscopy and ultrafiltration. Furthermore, the inhibitory effects of these flavonoids on CYP2C9 enzyme were tested because the metabolic elimination of warfarin is catalysed principally by this enzyme. Herein, we demonstrate that each tested flavonoid metabolite can bind to human serum albumin (HSA) with high affinity, some with similar or even higher affinity than quercetin itself. Quercetin metabolites are able to strongly displace warfarin from HSA suggesting that high quercetin doses can strongly interfere with warfarin therapy. On the other hand, tested flavonoids showed no or weaker inhibition of CYP2C9 compared to warfarin, making it very unlikely that quercetin or its metabolites can significantly inhibit the CYP2C9-mediated inactivation of warfarin. © 2017 Elsevier Masson SAS. All rights reserved.
1. Introduction Flavonoids are ubiquitous molecules in the nature. They are able to interact with many proteins (e.g., enzymes, transporters) in the human organism leading to their manifold biochemical and pharmacological effects in the body [1,2]. Flavonoids occur in several foods, drinks, herbal products, dietary supplements as well as some medications [2,3]. After oral consumption of flavonoidcontaining foods or products, flavonoid aglycones undergo extensive presystemic elimination, resulting in the relatively low oral bioavailability of these compounds [4]. The rest of the parent compound and/or its metabolites can then reach the systemic circulation. Therefore the plasma concentration of the orally consumed or administered flavonoid aglycones is relatively
* Corresponding author. E-mail address: [email protected] (M. Poór). http://dx.doi.org/10.1016/j.biopha.2017.01.092 0753-3322/© 2017 Elsevier Masson SAS. All rights reserved.
low. Most of the flavonoids can bind to serum albumin with high affinity resulting in their strong albumin binding property [5–7]. Quercetin (Q) is one of the most commonly occurring flavonoid in the nature, it is contained by many fruits, vegetables, and grains [8,9]. Therefore, Q is part of the normal diet; furthermore, there are several dietary supplements containing high doses of Q (even 500–1000 mg in a tablet). During the normal diet, human plasma concentrations of Q and its metabolites are in the nanomolar range [9], while continuous supplementation with high doses of Q (500–1000 mg) can result in few micromolar levels in the circulation [10,11]. Like other flavonoid aglycones, Q is also highly metabolised [9,12]. Methylation of Q by catechol-O-methyltransferase (COMT) leads to the formation of isorhamnetin (IR; 30 -Omethylquercetin) and tamarixetin (TAM; 40 -O-methylquercetin); as COMT prefers more the 30 -O-methylation vs. 40 -O-methylation, considerably higher amounts of IR are formed [13–15]. Furthermore, sulfate and glucuronide conjugates of Q also appear [9,12];
M. Poór et al. / Biomedicine & Pharmacotherapy 88 (2017) 574–581
the dominant circulating metabolites of Q are quercetin-30 -sulfate (Q30 S), quercetin-3-glucuronide (Q3G), and isorhamnetin-3-glucuronide (I3G) [16]. Q binds to human serum albumin (HSA) with high affinity [6]. Recent in vitro studies demonstrate that Q is able to effectively displace strongly albumin-bound molecules from HSA causing the disruption of their albumin binding [17,18]. Previous investigations also suggest that some Q metabolites can bind to HSA with lower, similar or even higher affinity compared to the parent compound [19]; however, displacing abilities of the most important circulating Q metabolites (Q30 S, Q3G, and I3G) have not been reported. Warfarin (WAR) is an orally-administered anticoagulant drug which is commonly used to prevent thrombosis and thromboembolism [20]. Because WAR is a drug with a narrow therapeutic window, the interaction of WAR with other drugs or dietary supplements can result in thrombus formation or bleedings in WAR-treated patients [21,22]. Therefore, the pharmacokinetic interactions of WAR are of high pharmacological as well as toxicological importance. More than 99% of WAR in the human circulation is albumin-bound form [23,24]. Thus, the displacement of only a few percent of WAR from HSA can dramatically increase the free concentration of WAR in the circulation. Previous in vitro studies demonstrated the strong displacing ability of flavonoids vs. WAR [18,25]. The main process responsible for the elimination of WAR is its biotransformation by CYP2C9 enzyme [26]. If the free concentration of WAR increases in the blood, and consequently in the liver cells, it is plausible to hypothesize that the metabolic inactivation of WAR by CYP2C9 becomes elevated as well, which can partly compensate the increased pharmacological effects of WAR. On the other hand, it was previously reported that the displacement of WAR from HSA by other drugs can cause bleeding [21,27]. These observations suggest that the strong displacement of WAR from albumin can result in serious consequences despite of the compensatory effects in the body. Furthermore, Q and IR are able to competitively inhibit the CYP2C9 enzyme [28,29], and the inhibition of CYP2C9-mediated metabolic inactivation of WAR by flavonoids can further aggravate the toxic consequences of the increased free (not HSA-bound) concentration of WAR. In this study, we examined the albumin binding abilities of Q and its metabolites (Q30 S, Q3G, I3G, IR, and TAM; see in Fig. 1) as well as the potential inhibitory effects of these compounds on CYP2C9 enzyme. After the binding constants of flavonoid-HSA complexes were quantified by fluorescence quenching method, displacing abilities of Q and Q metabolites vs. WAR were demonstrated using fluorescence spectroscopic and ultrafiltration
techniques. The effects of flavonoids on CYP2C9-catalysed hydroxylation of diclofenac were evaluated, and compared to the influence of the same amounts of WAR. 2. Materials and methods 2.1. Reagents All reagents were of spectroscopic or analytical grade. Quercetin (Q) was purchased from Fluka. Isorhamnetin (IR) and tamarixetin (TAM) were purchased from Extrasynthese. Quercetin-30 -sulfate (Q30 S), quercetin-3-glucuronide (Q3G), and isorhamnetin-3-glucuronide (I3G) were synthetized as described previously [30]. Warfarin (WAR), human serum albumin (HSA) and CypExpressTM 2C9 (Cytochrome P450 human) kit were purchased from Sigma. Diclofenac (free acid) and 40 -hydroxydiclofenac were purchased from Carbosynth. 2.2. Fluorescence spectroscopic measurements Fluorescence measurements were carried out employing Hitachi F-4500 fluorescence spectrophotometer. All measurements were performed at 25 C. In order to mimic extracellular physiological conditions, albumin-ligand interactions and competitive interaction of flavonoids with WAR were investigated in phosphate buffered saline (PBS; pH 7.4). Binding constants of Q and its metabolites with HSA were quantified with fluorescence quenching experiments. Fluorescence emission spectra of 2 mM HSA in the absence and presence of increasing flavonoid concentrations (0.5, 1.0, 2.0, 3.0, 4.0, 5.0, and 6.0 mM) were recorded. Assuming 1:1 stoichiometry, binding constants were calculated by non-linear fitting with Hyperquad2006 (Protonic Software) using 295 nm and 340 nm as excitation and emission wavelengths, respectively [31,32]: ðIHG I0 Þ I ¼ I0 þ 2 ½H0 0 1 @½H0 þ ½G0 þ K
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi1 1 2 ½H0 þ ½G0 þ 4 ½H0 ½G0 A K
ð1Þ
where I and I0 denote the fluorescence emission intensity of HSA with and without flavonoids, respectively; IHG is the fluorescence emission intensity of pure flavonoid-HSA complex (calculated by the Hyperquad2006); K denotes the binding constant (with the unit of dm3/mol); while [H]0 and [G]0 are the total concentrations of HSA and flavonoids, respectively. Evaluation of spectroscopic data with the graphical application of Stern-Volmer equation was also performed: I0 ¼ 1 þ K SV ½Q I
Fig. 1. Chemical structure of quercetin and its metabolites.
575
ð2Þ
where I0 and I are fluorescence emission intensities of HSA in the absence and presence of flavonoids, respectively. KSV is the SternVolmer quenching constant while [Q] is the concentration of the quencher. To investigate the displacement of WAR from HSA by flavonoids, our previously described method was applied [31,33]. The complex formation of WAR with HSA results in the strong increase of its fluorescence (lexc = 317 nm, lem = 379 nm), therefore the complex formation as well as the displacement of WAR from HSA can be precisely followed. The increasing concentrations of HSA result in the gradual elevation of the fluorescence signal of WAR at 379 nm (Fig. S1) thus the concentration of the free and HSA-bound WAR can be calculated, as described previously [18]. In the presence of 1 mM WAR and 3.5 mM HSA, approximately 70% of the WAR molecules are presented in albumin-bound form. To achieve a
576
M. Poór et al. / Biomedicine & Pharmacotherapy 88 (2017) 574–581
further significant increase of the bound fraction of WAR, relatively high amounts of HSA are necessary (Fig. S1), which would make our model less sensitive regarding the displacement. To examine the displacing ability of Q and its metabolites, increasing flavonoid concentrations (0.5, 1.0, 2.0, 3.0, 4.0, and 5.0 mM) were added to 1 mM WAR and 3.5 mM HSA. 2.3. Ultrafiltration Ultrafiltration studies were performed as described elsewhere [33]. After ultrafiltration of 2.5 mL samples (7500 g, 10 min, 25 C) with Amicon Ultra-4 centrifugal filter units (10 kDa molecular weight cut-off value; from Merck Millipore), the concentration of WAR in the filtrate was determined by fluorescence spectroscopy using a calibration curve (0.1–1.0 mM). Emission intensities of the filtrates were measured at 389 nm (lexc = 309 nm). Under the applied circumstances, no interferences were observed between the spectra of WAR and the examined flavonoids. The constant recovery of WAR was certified regardless of its applied concentration (0.2–1.0 mM) [15]. 2.4. CYP2C9 assay In order to investigate a potential inhibitory action of Q and/or its metabolites on CYP2C9 enzyme, CypExpressTM 2C9 kit was used. Following the guide of this assay kit, the FDA-recommended substrate of CYP2C9 diclofenac was applied. The reaction of diclofenac with CYP2C9 results in the formation of its 40 -hydroxy derivative. Samples contained 8 mg/mL CypExpressTM 2C9 as well as 15 mM diclofenac in 50 mM potassium phosphate buffer (pH 7.5). Under these conditions, approximately 90 percent (89.3 3.0%) of diclofenac was converted to 40 -hydroxydiclofenac during 2-h incubation. The complete volume of incubates was 200 mL. The inhibitory effects of 25 and 50 mM concentrations of WAR and flavonoids were tested. Each sample (including the control) contained the same amount of DMSO (approximately 2 v/v%) which was the solvent of diclofenac, WAR, and flavonoids. Incubations were started by adding the enzyme. Samples were incubated in a Thermomixer (Eppendorf) for 120 min at 30 C and 700 rpm. The incubation was stopped by 100 mL HPLC eluent (see below), after which samples were vortexed and the enzyme suspension was centrifuged for 10 min at room temperature and 14,000g. Then the supernatant was carefully removed and the
concentrations of 40 -hydroxydiclofenac as well as diclofenac were quantified by HPLC. In order to quantify diclofenac and 40 -hydroxydiclofenac, the recommended HPLC method of CypExpressTM 2C9 kit was used with some modifications. Integrated HPLC system (Thermo Scientific Dionex UltiMate 3000) was equipped with a pump (LPG-3400SD), an autosampler column compartment (ACC-3000; with integrated autosampler and column oven), and a diode array detector (DAD-3000). Separation was performed on a LiChroCART 125-4 RP-8 (5 mm) analytical column with a guard column (Teknokroma TR-C-160K1) using isocratic mobile phase containing methanol (Promochem), water (Promochem), and acetic acid (69.5:30:0.5 v/v%) as a mobile phase with 1 mL/min flow rate. The column temperature was 35 C, peak areas were monitored at 275 nm, and the injected volume was 20 mL. Data were recorded and evaluated using Dionex Chromeleon 7 software. 2.5. Statistical analyses Mean data SEM values were shown on figures. A One-Way ANOVA test (IBM SPSS Statistics, Version 21) was applied for statistical analyses, where the level of significance was set at a minimum of p < 0.05 with a maximum of p < 0.01. 3. Results and discussion 3.1. Interaction of quercetin metabolites with human serum albumin In order to investigate the binding ability of quercetin metabolites to HSA, fluorescence quenching method was applied. During these experiments, increasing flavonoid concentrations (0–6 mM) were added to standard amount of HSA (2 mM). Under the applied circumstances (lexc = 295 nm, lem = 340 nm), the presence of each tested compound resulted in significant reduction of the fluorescence signal of HSA (Figs. 2 and 3). Fluorescence and absorption spectra of flavonoids were also recorded, no spectroscopic interferences were observed. Q30 S, IR, and TAM caused similar while the presence of glucuronide metabolites of quercetin (Q3G and I3G) led to lower fluorescence quenching compared to Q (Fig. 2). Based on these data, the Stern-Volmer quenching constants (KSV) as well as the binding constants (K) of flavonoidHSA complexes were calculated by applying the Stern-Volmer equation and Hyperquad2006 program package, respectively (see details in 2.2.). Both methods suggested 1:1 stoichiometry
Fig. 2. Fluorescence quenching of emission intensities of 2 mM human serum albumin (lexc = 295 nm, lem = 340 nm) in the presence of increasing quercetin (Q; left), quercetin-30 -sulfate (Q30 S; middle), and quercetin-3-glucuronide (Q3G; right) concentrations (0–6 mM) in PBS (pH 7.4).
M. Poór et al. / Biomedicine & Pharmacotherapy 88 (2017) 574–581
577
Fig. 3. Fluorescence emission intensities (lexc = 295 nm, lem = 340 nm) of human serum albumin (HSA; 2 mM) in the presence of increasing flavonoid concentrations (0.5, 1.0, 2.0, 3.0, 4.0, 5.0, and 6.0 mM) in PBS (pH 7.4; left); and Stern-Volmer plots of the same flavonoid-HSA interactions (right). [Q: quercetin, Q30 S: quercetin-30 -sulfate, Q3G: quercetin-3-glucuronide, I3G: isorhamnetin-3-glucuronide, IR: isorhamnetin, and TAM: tamarixetin].
during the complex formation (Fig. 3). The values and tendencies of KSV and K values were in a good agreement and suggest that each metabolite of Q forms stable complex with HSA; however, the stabilities of these flavonoid-HSA complexes show some differences (Table 1). Q30 S and IR bind with slightly higher while TAM binds slightly lower affinity to HSA when compared to Q. Furthermore, the binding constants of the glucuronide metabolites (Q3G and I3G) were considerably lower than the other compounds. On the other hand, it is important to note that despite having a lower affinity for HSA than Q, the stabilities of the Q3G-HSA and I3G-HSA complexes (logK = 4.5–4.7) were still high enough for significant interaction with HSA. The logK value of the WAR-HSA complex is approximately 5.3 [34] demonstrating that Q, Q30 S, IR, and TAM bind with similar affinity while Q3G and I3G bind with lower affinity to HSA than WAR. Interaction of some of these Q metabolites with HSA has been investigated by fluorescence spectroscopy before, and the described results are in a good agreement with our observation that affinity order of the metabolites is Q30 S > Q > Q3G > I3G [35]. The significantly lower binding affinity of glucuronides toward HSA can be explained by the followings. The benzopyrone moiety of Q is located within the binding pocket of subdomain IIA, and the nitrogen atom of Lys195 forms intermolecular H-bond or ionic interaction with the 3-hydroxyl group of Q [36]. The presence of the bulky polar substituent in position 3 may decrease the binding affinity of glucuronides because of steric reasons; furthermore, the absence of the interaction of 3-hydroxyl moiety with Lys195 can result in the lower complex stabilities as well. Complex formations of Q glycosides with HSA follow a similar fashion, binding constants of rutin-HSA and hyperoside-HSA complexes are significantly lower compared to the Q-HSA complex resulted from the presence of the bulky polar rutinose and galactose substituents in position 3, respectively [37]. Previous studies suggest that most of the flavonoid aglycones that form stable complexes with HSA occupy the binding site called Sudlow’s Site I (or Drug Binding Site I; on subdomain IIA) [5–7]. Because the single tryptophan residue (Trp-214) of HSA is located in this part of the protein, binding of flavonoids to the HSA
molecule results in the strong quenching of its fluorescence [6]. For this reason the interaction of flavonoids with HSA can be precisely followed using the fluorescence quenching method. Q metabolites significantly decreased the fluorescence of HSA at 340 nm (Figs. 2 and 3), therefore it is plausible to hypothesize that the binding site of each tested metabolite is located on Sudlow’s Site I. 3.2. Displacement of warfarin from albumin by quercetin metabolites Because most of the Q metabolites (except the glucuronides) bind with similar affinity to HSA as Q, it is very likely that Q metabolites are also able to displace the Sudlow’s Site I ligand warfarin from HSA in the same way as Q and other flavonoids [18]. To test this hypothesis, our previously described model was applied [33]. Since WAR shows a much stronger fluorescence signal in HSA-bound form than free WAR, the interaction of WAR with HSA as well as the displacement of HSA-bound WAR results in significant changes to the fluorescence spectrum of WAR [18]. Using these principles, increasing flavonoid concentrations (0–5 mM) were added to standard WAR (1 mM) and HSA (3.5 mM), and fluorescence emission spectra of WAR were recorded (lexc = 317 nm). Applying our previously described method [18], the concentration of HSA-bound WAR was quantified based on the emission intensities of WAR (lem = 379 nm). Our results demonstrate that the presence of each tested metabolite resulted in a significant decrease of the fluorescence signal of WAR suggesting that all Q metabolites are able to displace WAR from HSA (Fig. 4). Flavonoids and HSA alone did not give a fluorescence signal at the wavelengths used under the applied conditions. Furthermore, to exclude the potential involvement of an internal filter effect, the absorption spectra of the flavonoids were also recorded. Only negligible absorbance values of Q and Q metabolites were observed. The strongest displacement effects were caused by IR and Q30 S, followed by Q and TAM, while considerably lower but significant impacts resulted from Q3G and I3G treatment. These data are in a very good agreement with our previous results (see in 3.1.) because the competitors which showed higher binding constants with HSA demonstrated stronger displacing abilities
Table 1 Stabilities of flavonoid-HSA complexes: Stern-Volmer quenching constants (KSV) and binding constants (K) (the unit of KSV and K is dm3/mol).
log KSV log K
QUE
Q30 S
Q3G
I3G
IR
TAM
5.24 0.03 5.37 0.00
5.35 0.02 5.49 0.01
4.72 0.03 4.73 0.00
4.54 0.08 4.55 0.00
5.27 0.04 5.43 0.01
5.10 0.01 5.19 0.00
578
M. Poór et al. / Biomedicine & Pharmacotherapy 88 (2017) 574–581
After oral consumption, the molar amount of Q itself gives only a few percent of the plasma concentration [38]. Approximately 50, 25, and 10% of the circulating Q is presented by Q30 S, Q3G, and I3G, respectively [16], indicating that Q30 S is the dominant metabolite of Q in the human plasma. Considering the observation that continuous administration of 1000 mg Q daily can result in even 2 mM plasma concentration of total Q (including Q as well as its metabolites) [11,39], it is plausible to hypothesize that the plasma concentration of Q30 S may exceed the 1 mM, while Q3G and I3G likely presented at few hundred nM concentrations. Furthermore, the steady-state concentration of WAR during the anticoagulant therapy is few mM (typically 1.5–5.0 mM) [24,40], suggesting that the plasma concentration of total Q can comparable or slightly lower compared to WAR. 3.3. Investigation of the displacing effect of Q and Q metabolite combinations Fig. 4. Decrease of the albumin-bound fraction of warfarin (1 mM warfarin + 3.5 mM human serum albumin) in the presence of increasing flavonoid concentrations (0.5, 1.0, 2.0, 3.0, 4.0, and 5.0 mM) in PBS (pH 7.4). [Q: quercetin, Q30 S: quercetin-30 sulfate, Q3G: quercetin-3-glucuronide, I3G: isorhamnetin-3-glucuronide, IR: isorhamnetin, and TAM: tamarixetin].
vs. WAR (Table 1 and Fig. 4). Our results indicate that each tested Q metabolite is able to compete with WAR. Furthermore, these data also highlight that Q30 S, IR, and TAM are similarly active competitors than Q. The glucuronide conjugates bind with lower affinity to HSA and they are less effective competitors than Q, however, they are still able to displace WAR from HSA. The displacement of WAR by the tested metabolites is consistent with our previous hypothesis that the binding site of the metabolites (including the conjugates) needs to be located on Sudlow’s Site I in the same way as Q and WAR [17,18]. The observation that even micromolar concentrations of Q metabolites result in a considerable displacement of WAR strongly suggests that Q metabolites are able to disrupt the albumin binding of WAR.
After the exposition/treatment with Q-containing foods or dietary supplements, Q circulates, mainly as its metabolites in the blood [16,38,41] resulting in the possibility of a combined effect. To test this, the displacing abilities of mixtures of Q and its metabolites were examined in combination. Samples containing 1 mM WAR and 3.5 mM HSA were studied in the absence and presence of a flavonoid pair (1 mM concentrations of both applied flavonoids). Based on the result of the previous experiment (see in 3.2.) the displaced fraction of WAR by 1 mM flavonoid alone was determined, and the sum of these data was calculated in order to compare the combined flavonoid effect with the expected additive effect of the two compounds. Fig. 5 demonstrates that in most of the cases the displacing abilities of flavonoid pairs showed additive or quasi-additive property. Some of the flavonoid pairs represented lower than additive effect but no dramatic differences were observed compared to the calculated and the measured data. These results indicate that in the presence of Q and its metabolites, the combined displacing effect will be close to additive making the
Fig. 5. Investigation of the potential additive property of the displacing effects of quercetin and quercetin metabolites. Displaced amounts (%) of warfarin (WAR) were quantified in the presence of 1 mM flavonoids alone (see in 3.2. and Fig. 4), then flavonoid pairs were formed. The displacing effect of the flavonoid pairs (1 + 1 mM) was added based on their separate displacing effects (calculated values). Thereafter, these data were compared with the result of the measurements where the combined effects of flavonoid pairs were determined in the presence of 1 mM WAR, 3.5 mM human serum albumin, and 1 mM of both tested flavonoids in PBS (measured values; lexc = 317 nm, lem = 379 nm). [Q: quercetin, Q30 S: quercetin-30 -sulfate, Q3G: quercetin-3-glucuronide, I3G: isorhamnetin-3-glucuronide, IR: isorhamnetin, and TAM: tamarixetin].
M. Poór et al. / Biomedicine & Pharmacotherapy 88 (2017) 574–581
strong displacement of WAR from HSA by Q and its metabolites more likely. 3.4. Ultrafiltration studies To further confirm the displacing abilities of Q and its metabolites, ultrafiltration studies were performed. Because HSA is a large protein (approximately 67.5 kDa), it is unable to pass through a filter unit with a 10 kDa molecular weight cut-off value, therefore only the free (not HSA-bound) WAR will appear in the filtrate. Because the duration of the applied ultrafiltration process is 10 min, the equilibrium in the supernatant is continuously changing. For this reason, the determined WAR concentration in the filtrate is not exactly equal with the free fraction of WAR in the sample, but the elevated concentration of WAR in the filtrate indicates the increase of its free fraction. During this experiment, 1 mM WAR and 5 mM HSA was applied in PBS. Under these circumstances, the free WAR concentration in the filtrate is approximately 50% compared to the WAR content measured when WAR was filtered alone. To investigate the effect of Q and its metabolites on WAR-HSA interaction, 7.5 and 15 mM flavonoids were applied. When samples contained only WAR and flavonoids (in the absence of HSA), Q and its metabolites did not change the concentration of WAR in the filtrate compared to the experiments when WAR was filtered alone. However, in the presence of WAR, HSA and flavonoids, most of the tested flavonoids significantly increased the WAR content of the filtrate indicating the displacement of WAR from HSA (Fig. 6). During these experiments, Q, Q30 S, IR, and TAM showed again strong displacing abilities, while Q3G and I3G caused only non-significant increase of the filtered WAR under the applied conditions. Therefore, the ultrafiltration experiments further supported the fluorescence spectroscopic measurements (see in 3.2.) again indicating that Q metabolites (except the glucuronide conjugates) are able to strongly interfere with the albumin binding of WAR. 3.5. Inhibition of CYP2C9 by Q and Q metabolites Displacement of WAR from HSA results in the increase of free WAR concentration in the circulation, therefore it is plausible to
579
hypothesize its increased uptake by liver cells as well as the elevated speed of its metabolic inactivation by CYP2C9. On the other hand, previous studies demonstrated that Q and IR can competitively inhibit CYP2C9 enzyme [28,29] suggesting that Q and/or some of its metabolites are able to disrupt the metabolic elimination of WAR. To test this idea, the inhibitory effects of Q and Q metabolites relative to WAR were tested using diclofenac, the FDA-recommended substrate of CYP2C9. Fig. 7 demonstrates the formation of 40 -hydroxydiclofenac in the absence and presence of 25 or 50 mM WAR or flavonoid after 60, 90, and 120 min incubations. Under the applied circumstances, WAR strongly decreased the CYP2C9-catalyzed 40 -hydroxydiclofenac formation in each time period, while the same concentrations of Q3G and I3G did not cause any differences. Significant inhibition of CYP2C9 was observed in the presence of higher (50 mM) Q and Q30 S concentrations; however, IR resulted in significant impact only after 90 and 120 min incubations (Fig. 7 and Fig. S2). On the other hand, the inhibitory effects of Q, Q30 S, and IR were considerably lower compared to WAR. After 60 and 90 min incubations, strong inhibition of CYP2C9 enzyme was observed by TAM resulting in the similar effect of 25 mM TAM than 25 mM WAR; however, 50 mM TAM did not lead to further decrease of 40 hydroxydiclofenac formation (Fig. 7 and Fig. S2). Furthermore, inhibitory action of TAM was negligible compared to WAR after 120 min incubation. Based on these results, Q and/or some of its metabolites (Q30 S or IR) may slightly inhibit the metabolic elimination of WAR, however, it is unlikely that their inhibitory effect on CYP2C9 enzyme results in relevant interaction in the body. It is also important to note that COMT enzyme favors the formation of IR vs. TAM therefore relatively low amount of TAM is produced from Q in the human liver [13,14], suggesting the low pharmacological relevance of the inhibitory effect of TAM on CYP2C9 enzyme. Finally, the CYP2C9 inhibition assay was also performed in the presence of both WAR (25 mM) and flavonoids (25 mM). Combined application of WAR and Q metabolites results in only slight but non-significant effect compared to WAR alone (Fig. S2). A further minor decrease of 40 -hydroxydiclofenac formation was observed in the presence of Q, Q30 S, IR, and TAM; while in the presence of the glucuronide metabolites the formation of 40 -hydroxydiclofenac
Fig. 6. Results of the ultrafiltration experiments: Effects of quercetin (Q) or its metabolites on warfarin (WAR) concentrations in the filtrate. Before ultrafiltration, samples contained 1 mM WAR, 5 mM human serum albumin (HSA) in the absence and presence of 7.5 or 15 mM flavonoid concentrations in PBS (pH 7.4). Data are expressed in percentage, where 100% means the WAR concentration in the filtrate when the samples contained neither HSA nor flavonoids (*p < 0.05; **p < 0.01). [Q30 S: quercetin-30 sulfate, Q3G: quercetin-3-glucuronide, I3G: isorhamnetin-3-glucuronide, IR: isorhamnetin, and TAM: tamarixetin].
580
M. Poór et al. / Biomedicine & Pharmacotherapy 88 (2017) 574–581
was practically unchanged. These results suggests again the poor inhibitory effect of Q metabolites on CYP2C9 enzyme compared to WAR. 4. Conclusions In this study, the pharmacokinetic interactions of Q and its main metabolites with WAR were investigated. Our results show that Q metabolites bind to HSA with high affinity, and we also demonstrated that methylated and sulfate metabolites have similar or even higher displacing abilities than the parent compound Q. Glucuronide conjugates showed lower affinities towards HSA and their abilities to compete with WAR to bind were also lower compared to Q. On the other hand, the fact that glucuronide metabolites of Q are also able to displace WAR and the close to additive behavior of their displacing effect in the presence of more than one metabolites suggest that frequent exposure to high doses of Q (e.g., through the consumption of dietary supplements containing large amounts of Q) can result in significant interaction with WAR. Because the strong displacement of WAR from HSA can lead to bleeding, it can cause serious consequences. However, our results also indicate that the tested flavonoids have considerably lower effect on CYP2C9-catalysed hydroxylation than WAR, suggesting that this is not, in fact, an additional concern. For these listed reasons, we need to consider the potentially hazardous consequences of the consumption of high levels of quercetin-containing dietary supplements in WARtreated patients. Acknowledgements The present scientific contribution is dedicated to the 650th anniversary of the foundation of the University of Pécs, Hungary. PAK and PWN acknowledge the support of the Biotechnology and Biological Sciences Research Council (UK) through an Institute Strategic Programme Grant (‘Food and Health’; Grant No. BB/ J004545/1) to the Institute of Food Research. This work was supported by the GINOP-2.3.2-15-2016-00049 grant. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.biopha.2017. 01.092. References
Fig. 7. Formation of 40 -hydroxydiclofenac from diclofenac by CYP2C9 enzyme in the absence and presence of warfarin (WAR) or flavonoids (see details in 2.4.) after 60, 90, and 120 min incubations (**p < 0.01) [Q: quercetin, Q30 S: quercetin-30 -sulfate, Q3G: quercetin-3-glucuronide, I3G: isorhamnetin-3-glucuronide, IR: isorhamnetin, and TAM: tamarixetin].
[1] B.H. Havsteen, The biochemistry and medical significance of the flavonoids, Pharmacol. Ther. 96 (2002) 67–202. [2] M. Ø. Andersen, K.R. Markham, Flavonoids: Chemistry, Biochemistry and Applications, CRC, Press, Taylor & Francis Group, Boca Raton, FL, US, 2006. [3] C. Amato, Advantage of a micronized flavonoidic fraction (Daflon 500 mg) in comparison with a nonmicronized diosmin, Angiology 45 (1994) 531–536. [4] S.H. Thilakarathna, H.P. Rupasinghe, Flavonoid bioavailability and attempts for bioavailability enhancement, Nutrients 5 (2013) 3367–3387. [5] C. Dufour, O. Dangles, Flavonoid-serum albumin complexation: determination of binding constants and binding sites by fluorescence spectroscopy, Biochim. Biophys. Acta 1721 (2005) 164–173. [6] J. Xiao, Y. Zhao, H. Wang, Y. Yuan, F. Yang, C. Zhang, K. Yamamoto, Noncovalent interaction of dietary polyphenols with common human plasma proteins, J. Agric. Food Chem. 59 (2011) 10747–10754. [7] S. Pal, C. Saha, A review on structure-affinity relationship of dietary flavonoids with serum albumins, J. Biomol. Struct. Dyn. 32 (2014) 1132–1147. [8] J. Peterson, J. Dwyer, Flavonoids Dietary occurrence and biochemical activity, Nutr. Res. 18 (1998) 1995–2018. [9] G.S. Kelly, Quercetin. Monograph, Altern. Med. Rev. 16 (2011) 172–194. [10] J.A. Conquer, G. Maiani, E. Azzini, A. Raguzzini, B.J. Holub, Supplementation with quercetin markedly increases plasma quercetin concentration without effect on selected risk factors for heart disease in healthy subjects, J. Nutr. 128 (1998) 593–597.
M. Poór et al. / Biomedicine & Pharmacotherapy 88 (2017) 574–581 [11] S.A. Heinz, D.A. Henson, D.C. Nieman, M.D. Austin, F. Jin, A 12-week supplementation with quercetin does not affect natural killer cell activity, granulocyte oxidative burst activity or granulocyte phagocytosis in female human subjects, Br. J. Nutr. 104 (2010) 849–857. [12] A.W. Boots, G.R. Haenen, A. Bast, Health effects of quercetin: from antioxidant to nutraceutical, Eur. J. Pharmacol. 585 (2008) 325–337. [13] P.T. Männistö, S. Kaakkola, Catechol-O-methyltransferase (COMT): biochemistry, molecular biology, pharmacology, and clinical efficacy of the new selective COMT inhibitors, Pharmacol. Rev. 51 (1999) 593–628. [14] C. De Santi, A. Pietrabissa, F. Mosca, G.M. Pacifici, Methylation of quercetin and fisetin flavonoids widely distributed in edible vegetables, fruits and wine, by human liver, Int. J. Clin. Pharmacol. Ther. 40 (2002) 207–212. [15] M. Poór, Z. Zrínyi, T. KÅszegi, Structure related effects of flavonoid aglycones on cell cycle progression of HepG2 cells: metabolic activation of fisetin and quercetin by catechol-O-methyltransferase (COMT), Biomed. Pharmacother. 83 (2016) 998–1005. [16] W. Mullen, C.A. Edwards, A. Crozier, Absorption, excretion and metabolite profiling of methyl-, glucuronyl-, glucosyl- and sulpho-conjugates of quercetin in human plasma and urine after ingestion of onions, Br. J. Nutr. 96 (2006) 107– 116. [17] M. Poór, S. Kunsági-Máté, T. Bencsik, J. Petrik, S. Vladimir-Kneževi c, T. KÅszegi, Flavonoid aglycones can compete with Ochratoxin A for human serum albumin: a new possible mode of action, Int. J. Biol. Macromol. 51 (2012) 279– 283. [18] M. Poór, Y. Li, S. Kunsági-Máté, J. Petrik, S. Vladimir-Kneževi c, T. KÅszegi, Molecular displacement of warfarin from human serum albumin by flavonoid aglycones, J. Lumin. 142 (2013) 122–127. [19] O. Dangles, C. Dufour, C. Manach, C. Morand, C. Remesy, Binding of flavonoids to plasma proteins, Methods Enzymol. 335 (2001) 319–333. [20] P.A. Tideman, R. Tirimacco, A. St John, G.W. Roberts, How to manage warfarin therapy, Aust. Prescr. 38 (2015) 44–48. [21] A.M. Holbrook, J.A. Pereira, R. Labiris, H. McDonald, J.D. Douketis, M. Crowther, P.S. Wells, Systematic overview of warfarin and its drug and food interactions, Arch. Intern. Med. 165 (2005) 1095–1106. [22] P.M. Leite, M.A. Martins, R.O. Castilho, Review on mechanisms and interactions in concomitant use of herbs and warfarin therapy, Biomed. Pharmacother. 83 (2016) 14–21. [23] N.H. Holford, Clinical pharmacokinetics and pharmacodynamics of warfarin. Understanding the dose-effect relationship, Clin. Pharmacokinet. 11 (1986) 483–504. [24] E. Chan, A.J. McLachlan, M. Pegg, A.D. MacKay, R.B. Cole, M. Rowland, Disposition of warfarin enantiomers and metabolites in patients during multiple dosing with rac-warfarin, Br. J. Clin. Pharmacol. 37 (1994) 563–569. [25] L. Di Bari, S. Ripoli, S. Pradhan, P. Salvadori, Interactions between quercetin and warfarin for albumin binding: a new eye on food/drug interference, Chirality 22 (2010) 593–596. [26] D.R. Flora, A.E. Rettie, R.C. Brundage, T.S. Tracy, CYP2C9 genotype-dependent warfarin pharmacokinetics: impact of CYP2C9 genotype on R- and S-Warfarin
[27]
[28]
[29]
[30] [31]
[32]
[33] [34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
581
and their oxidative metabolites, J. Clin. Pharmacol. (2016), doi:http://dx.doi. org/10.1002/jcph.813 in press. T.Y. Chan, Adverse interactions between warfarin and nonsteroidal antiinflammatory drugs: mechanisms, clinical significance, and avoidance, Ann. Pharmacother. 29 (1995) 1274–1283. D. Si, Y. Wang, Y.H. Zhou, Y. Guo, J. Wang, H. Zhou, Z.S. Li, J.P. Fawcett, Mechanism of CYP2C9 inhibition by flavones and flavonols, Drug Metab. Dispos. 37 (2009) 629–634. Y. Kimura, H. Ito, R. Ohnishi, T. Hatano, Inhibitory effects of polyphenols on human cytochrome P450 3A4 and 2C9 activity, Food Chem. Toxicol. 48 (2010) 429–435. P.W. Needs, P.A. Kroon, Convenient syntheses of metabolically important quercetin glucuronides and sulfates, Tetrahedron 62 (2006) 6862–6868. M. Poór, Y. Li, S. Kunsági-Máté, Z. Varga, A. Hunyadi, B. Dankó, F.R. Chang, Y.C. Wu, T. KÅszegi, Protoapigenone derivatives: albumin binding properties and effects on HepG2 cells, J. Photochem. Photobiol. B 124 (2013) 20–26. Y. Li, Z. Czibulya, M. Poór, S. Lecomte, L. Kiss, E. Harte, T. KÅszegi, S. KunságiMáté, Thermodynamic study of the effects of ethanol on the interaction of ochratoxin A with human serum albumin, J. Lumin. 148 (2014) 18–25. M. Poór, B. Lemli, M. Bálint, C. Hetényi, N. Sali, T. KÅszegi, S. Kunsági-Máté, Interaction of citrinin with human serum albumin, Toxins 7 (2015) 5155–5166. M. Poór, Y. Li, G. Matisz, L. Kiss, S. Kunsági-Máté, T. KÅszegi, Quantitation of species differences in albumin-ligand interactions for bovine, human and rat serum albumins using fluorescence spectroscopy: a test case with some Sudlow’s site I ligands, J. Lumin. 145 (2014) 767–773. K.M. Janisch, G. Williamson, P. Needs, G.W. Plumb, Properties of quercetin conjugates: modulation of LDL oxidation and binding to human serum albumin, Free Radic. Res. 38 (2004) 877–884. F. Zsila, Z. Bikádi, M. Simonyi, Probing the binding of the flavonoid quercetin to human serum albumin by circular dichroism, electronic absorption spectroscopy and molecular modelling methods, Biochem. Pharmacol. 65 (2003) 447–456. S. Bi, L. Ding, Y. Tian, D. Song, X. Zhou, X. Liu, H. Zhang, Investigation of the interaction between flavonoids and human serum albumin, J. Mol. Struct. 703 (2004) 37–45. L. Wang, M.E. Morris, Liquid chromatography-tandem mass spectroscopy assay for quercetin and conjugated quercetin metabolites in human plasma and urine, J. Chromatogr. 821 (2005) 194–201. F. Jin, D.C. Nieman, R.A. Shanely, A.M. Knab, M.D. Austin, W. Sha, The variable plasma quercetin response to 12-week quercetin supplementation in humans, Eur. J. Clin. Nutr. 64 (2010) 692–697. S. Sun, M. Wang, L. Su, J. Li, H. Li, D. Gu, Study on warfarin plasma concentration and its correlation with international normalized ratio, J. Pharm. Biomed. Anal. 42 (2006) 218–222. P.A. Kroon, M.N. Clifford, A. Crozier, A.J. Day, J.L. Donovan, C. Manach, G. Williamson, How should we assess the effects of exposure to dietary polyphenols in vitro? Am. J. Clin. Nutr. 80 (2004) 15–21.
Available online at
ScienceDirect www.sciencedirect.com
Original article
Interaction of quercetin and its metabolites with warfarin: Displacement of warfarin from serum albumin and inhibition of CYP2C9 enzyme Miklós Poóra,* , Gabriella Bodaa , Paul W. Needsb , Paul A. Kroonb , Beáta Lemlic,d , Tímea Bencsike a
Department of Pharmacology, University of Pécs, Faculty of Pharmacy, Szigeti út 12, Pécs, H-7624, Hungary Institute of Food Research, Norwich Research Park, Norwich, NR4 7UA, UK c Department of General and Physical Chemistry, University of Pécs, Ifjúság útja 6, Pécs, H-7624, Hungary d János Szentágothai Research Center, Ifjúság útja 20, Pécs, H-7624, Hungary e Institute of Pharmacognosy, University of Pécs, Faculty of Pharmacy, Rókus utca 2, Pécs, H-7624, Hungary b
A R T I C L E I N F O
Article history: Received 7 November 2016 Received in revised form 14 January 2017 Accepted 15 January 2017 Keywords: Warfarin Quercetin metabolites Human serum albumin Food-drug interaction CYP2C9
A B S T R A C T
Flavonoids are ubiquitous molecules in nature with manifold pharmacological effects. Flavonoids interact with several proteins, and thus potentially interfere with the pharmacokinetics of various drugs. Though much is known about the protein binding characteristics of flavonoid aglycones, the behaviour of their metabolites, which are extensively formed in the human body has received little attention. In this study, the interactions of the flavonoid aglycone quercetin and its main metabolites with the albumin binding of the oral anticoagulant warfarin were investigated by fluorescence spectroscopy and ultrafiltration. Furthermore, the inhibitory effects of these flavonoids on CYP2C9 enzyme were tested because the metabolic elimination of warfarin is catalysed principally by this enzyme. Herein, we demonstrate that each tested flavonoid metabolite can bind to human serum albumin (HSA) with high affinity, some with similar or even higher affinity than quercetin itself. Quercetin metabolites are able to strongly displace warfarin from HSA suggesting that high quercetin doses can strongly interfere with warfarin therapy. On the other hand, tested flavonoids showed no or weaker inhibition of CYP2C9 compared to warfarin, making it very unlikely that quercetin or its metabolites can significantly inhibit the CYP2C9-mediated inactivation of warfarin. © 2017 Elsevier Masson SAS. All rights reserved.
1. Introduction Flavonoids are ubiquitous molecules in the nature. They are able to interact with many proteins (e.g., enzymes, transporters) in the human organism leading to their manifold biochemical and pharmacological effects in the body [1,2]. Flavonoids occur in several foods, drinks, herbal products, dietary supplements as well as some medications [2,3]. After oral consumption of flavonoidcontaining foods or products, flavonoid aglycones undergo extensive presystemic elimination, resulting in the relatively low oral bioavailability of these compounds [4]. The rest of the parent compound and/or its metabolites can then reach the systemic circulation. Therefore the plasma concentration of the orally consumed or administered flavonoid aglycones is relatively
* Corresponding author. E-mail address: [email protected] (M. Poór). http://dx.doi.org/10.1016/j.biopha.2017.01.092 0753-3322/© 2017 Elsevier Masson SAS. All rights reserved.
low. Most of the flavonoids can bind to serum albumin with high affinity resulting in their strong albumin binding property [5–7]. Quercetin (Q) is one of the most commonly occurring flavonoid in the nature, it is contained by many fruits, vegetables, and grains [8,9]. Therefore, Q is part of the normal diet; furthermore, there are several dietary supplements containing high doses of Q (even 500–1000 mg in a tablet). During the normal diet, human plasma concentrations of Q and its metabolites are in the nanomolar range [9], while continuous supplementation with high doses of Q (500–1000 mg) can result in few micromolar levels in the circulation [10,11]. Like other flavonoid aglycones, Q is also highly metabolised [9,12]. Methylation of Q by catechol-O-methyltransferase (COMT) leads to the formation of isorhamnetin (IR; 30 -Omethylquercetin) and tamarixetin (TAM; 40 -O-methylquercetin); as COMT prefers more the 30 -O-methylation vs. 40 -O-methylation, considerably higher amounts of IR are formed [13–15]. Furthermore, sulfate and glucuronide conjugates of Q also appear [9,12];
M. Poór et al. / Biomedicine & Pharmacotherapy 88 (2017) 574–581
the dominant circulating metabolites of Q are quercetin-30 -sulfate (Q30 S), quercetin-3-glucuronide (Q3G), and isorhamnetin-3-glucuronide (I3G) [16]. Q binds to human serum albumin (HSA) with high affinity [6]. Recent in vitro studies demonstrate that Q is able to effectively displace strongly albumin-bound molecules from HSA causing the disruption of their albumin binding [17,18]. Previous investigations also suggest that some Q metabolites can bind to HSA with lower, similar or even higher affinity compared to the parent compound [19]; however, displacing abilities of the most important circulating Q metabolites (Q30 S, Q3G, and I3G) have not been reported. Warfarin (WAR) is an orally-administered anticoagulant drug which is commonly used to prevent thrombosis and thromboembolism [20]. Because WAR is a drug with a narrow therapeutic window, the interaction of WAR with other drugs or dietary supplements can result in thrombus formation or bleedings in WAR-treated patients [21,22]. Therefore, the pharmacokinetic interactions of WAR are of high pharmacological as well as toxicological importance. More than 99% of WAR in the human circulation is albumin-bound form [23,24]. Thus, the displacement of only a few percent of WAR from HSA can dramatically increase the free concentration of WAR in the circulation. Previous in vitro studies demonstrated the strong displacing ability of flavonoids vs. WAR [18,25]. The main process responsible for the elimination of WAR is its biotransformation by CYP2C9 enzyme [26]. If the free concentration of WAR increases in the blood, and consequently in the liver cells, it is plausible to hypothesize that the metabolic inactivation of WAR by CYP2C9 becomes elevated as well, which can partly compensate the increased pharmacological effects of WAR. On the other hand, it was previously reported that the displacement of WAR from HSA by other drugs can cause bleeding [21,27]. These observations suggest that the strong displacement of WAR from albumin can result in serious consequences despite of the compensatory effects in the body. Furthermore, Q and IR are able to competitively inhibit the CYP2C9 enzyme [28,29], and the inhibition of CYP2C9-mediated metabolic inactivation of WAR by flavonoids can further aggravate the toxic consequences of the increased free (not HSA-bound) concentration of WAR. In this study, we examined the albumin binding abilities of Q and its metabolites (Q30 S, Q3G, I3G, IR, and TAM; see in Fig. 1) as well as the potential inhibitory effects of these compounds on CYP2C9 enzyme. After the binding constants of flavonoid-HSA complexes were quantified by fluorescence quenching method, displacing abilities of Q and Q metabolites vs. WAR were demonstrated using fluorescence spectroscopic and ultrafiltration
techniques. The effects of flavonoids on CYP2C9-catalysed hydroxylation of diclofenac were evaluated, and compared to the influence of the same amounts of WAR. 2. Materials and methods 2.1. Reagents All reagents were of spectroscopic or analytical grade. Quercetin (Q) was purchased from Fluka. Isorhamnetin (IR) and tamarixetin (TAM) were purchased from Extrasynthese. Quercetin-30 -sulfate (Q30 S), quercetin-3-glucuronide (Q3G), and isorhamnetin-3-glucuronide (I3G) were synthetized as described previously [30]. Warfarin (WAR), human serum albumin (HSA) and CypExpressTM 2C9 (Cytochrome P450 human) kit were purchased from Sigma. Diclofenac (free acid) and 40 -hydroxydiclofenac were purchased from Carbosynth. 2.2. Fluorescence spectroscopic measurements Fluorescence measurements were carried out employing Hitachi F-4500 fluorescence spectrophotometer. All measurements were performed at 25 C. In order to mimic extracellular physiological conditions, albumin-ligand interactions and competitive interaction of flavonoids with WAR were investigated in phosphate buffered saline (PBS; pH 7.4). Binding constants of Q and its metabolites with HSA were quantified with fluorescence quenching experiments. Fluorescence emission spectra of 2 mM HSA in the absence and presence of increasing flavonoid concentrations (0.5, 1.0, 2.0, 3.0, 4.0, 5.0, and 6.0 mM) were recorded. Assuming 1:1 stoichiometry, binding constants were calculated by non-linear fitting with Hyperquad2006 (Protonic Software) using 295 nm and 340 nm as excitation and emission wavelengths, respectively [31,32]: ðIHG I0 Þ I ¼ I0 þ 2 ½H0 0 1 @½H0 þ ½G0 þ K
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi1 1 2 ½H0 þ ½G0 þ 4 ½H0 ½G0 A K
ð1Þ
where I and I0 denote the fluorescence emission intensity of HSA with and without flavonoids, respectively; IHG is the fluorescence emission intensity of pure flavonoid-HSA complex (calculated by the Hyperquad2006); K denotes the binding constant (with the unit of dm3/mol); while [H]0 and [G]0 are the total concentrations of HSA and flavonoids, respectively. Evaluation of spectroscopic data with the graphical application of Stern-Volmer equation was also performed: I0 ¼ 1 þ K SV ½Q I
Fig. 1. Chemical structure of quercetin and its metabolites.
575
ð2Þ
where I0 and I are fluorescence emission intensities of HSA in the absence and presence of flavonoids, respectively. KSV is the SternVolmer quenching constant while [Q] is the concentration of the quencher. To investigate the displacement of WAR from HSA by flavonoids, our previously described method was applied [31,33]. The complex formation of WAR with HSA results in the strong increase of its fluorescence (lexc = 317 nm, lem = 379 nm), therefore the complex formation as well as the displacement of WAR from HSA can be precisely followed. The increasing concentrations of HSA result in the gradual elevation of the fluorescence signal of WAR at 379 nm (Fig. S1) thus the concentration of the free and HSA-bound WAR can be calculated, as described previously [18]. In the presence of 1 mM WAR and 3.5 mM HSA, approximately 70% of the WAR molecules are presented in albumin-bound form. To achieve a
576
M. Poór et al. / Biomedicine & Pharmacotherapy 88 (2017) 574–581
further significant increase of the bound fraction of WAR, relatively high amounts of HSA are necessary (Fig. S1), which would make our model less sensitive regarding the displacement. To examine the displacing ability of Q and its metabolites, increasing flavonoid concentrations (0.5, 1.0, 2.0, 3.0, 4.0, and 5.0 mM) were added to 1 mM WAR and 3.5 mM HSA. 2.3. Ultrafiltration Ultrafiltration studies were performed as described elsewhere [33]. After ultrafiltration of 2.5 mL samples (7500 g, 10 min, 25 C) with Amicon Ultra-4 centrifugal filter units (10 kDa molecular weight cut-off value; from Merck Millipore), the concentration of WAR in the filtrate was determined by fluorescence spectroscopy using a calibration curve (0.1–1.0 mM). Emission intensities of the filtrates were measured at 389 nm (lexc = 309 nm). Under the applied circumstances, no interferences were observed between the spectra of WAR and the examined flavonoids. The constant recovery of WAR was certified regardless of its applied concentration (0.2–1.0 mM) [15]. 2.4. CYP2C9 assay In order to investigate a potential inhibitory action of Q and/or its metabolites on CYP2C9 enzyme, CypExpressTM 2C9 kit was used. Following the guide of this assay kit, the FDA-recommended substrate of CYP2C9 diclofenac was applied. The reaction of diclofenac with CYP2C9 results in the formation of its 40 -hydroxy derivative. Samples contained 8 mg/mL CypExpressTM 2C9 as well as 15 mM diclofenac in 50 mM potassium phosphate buffer (pH 7.5). Under these conditions, approximately 90 percent (89.3 3.0%) of diclofenac was converted to 40 -hydroxydiclofenac during 2-h incubation. The complete volume of incubates was 200 mL. The inhibitory effects of 25 and 50 mM concentrations of WAR and flavonoids were tested. Each sample (including the control) contained the same amount of DMSO (approximately 2 v/v%) which was the solvent of diclofenac, WAR, and flavonoids. Incubations were started by adding the enzyme. Samples were incubated in a Thermomixer (Eppendorf) for 120 min at 30 C and 700 rpm. The incubation was stopped by 100 mL HPLC eluent (see below), after which samples were vortexed and the enzyme suspension was centrifuged for 10 min at room temperature and 14,000g. Then the supernatant was carefully removed and the
concentrations of 40 -hydroxydiclofenac as well as diclofenac were quantified by HPLC. In order to quantify diclofenac and 40 -hydroxydiclofenac, the recommended HPLC method of CypExpressTM 2C9 kit was used with some modifications. Integrated HPLC system (Thermo Scientific Dionex UltiMate 3000) was equipped with a pump (LPG-3400SD), an autosampler column compartment (ACC-3000; with integrated autosampler and column oven), and a diode array detector (DAD-3000). Separation was performed on a LiChroCART 125-4 RP-8 (5 mm) analytical column with a guard column (Teknokroma TR-C-160K1) using isocratic mobile phase containing methanol (Promochem), water (Promochem), and acetic acid (69.5:30:0.5 v/v%) as a mobile phase with 1 mL/min flow rate. The column temperature was 35 C, peak areas were monitored at 275 nm, and the injected volume was 20 mL. Data were recorded and evaluated using Dionex Chromeleon 7 software. 2.5. Statistical analyses Mean data SEM values were shown on figures. A One-Way ANOVA test (IBM SPSS Statistics, Version 21) was applied for statistical analyses, where the level of significance was set at a minimum of p < 0.05 with a maximum of p < 0.01. 3. Results and discussion 3.1. Interaction of quercetin metabolites with human serum albumin In order to investigate the binding ability of quercetin metabolites to HSA, fluorescence quenching method was applied. During these experiments, increasing flavonoid concentrations (0–6 mM) were added to standard amount of HSA (2 mM). Under the applied circumstances (lexc = 295 nm, lem = 340 nm), the presence of each tested compound resulted in significant reduction of the fluorescence signal of HSA (Figs. 2 and 3). Fluorescence and absorption spectra of flavonoids were also recorded, no spectroscopic interferences were observed. Q30 S, IR, and TAM caused similar while the presence of glucuronide metabolites of quercetin (Q3G and I3G) led to lower fluorescence quenching compared to Q (Fig. 2). Based on these data, the Stern-Volmer quenching constants (KSV) as well as the binding constants (K) of flavonoidHSA complexes were calculated by applying the Stern-Volmer equation and Hyperquad2006 program package, respectively (see details in 2.2.). Both methods suggested 1:1 stoichiometry
Fig. 2. Fluorescence quenching of emission intensities of 2 mM human serum albumin (lexc = 295 nm, lem = 340 nm) in the presence of increasing quercetin (Q; left), quercetin-30 -sulfate (Q30 S; middle), and quercetin-3-glucuronide (Q3G; right) concentrations (0–6 mM) in PBS (pH 7.4).
M. Poór et al. / Biomedicine & Pharmacotherapy 88 (2017) 574–581
577
Fig. 3. Fluorescence emission intensities (lexc = 295 nm, lem = 340 nm) of human serum albumin (HSA; 2 mM) in the presence of increasing flavonoid concentrations (0.5, 1.0, 2.0, 3.0, 4.0, 5.0, and 6.0 mM) in PBS (pH 7.4; left); and Stern-Volmer plots of the same flavonoid-HSA interactions (right). [Q: quercetin, Q30 S: quercetin-30 -sulfate, Q3G: quercetin-3-glucuronide, I3G: isorhamnetin-3-glucuronide, IR: isorhamnetin, and TAM: tamarixetin].
during the complex formation (Fig. 3). The values and tendencies of KSV and K values were in a good agreement and suggest that each metabolite of Q forms stable complex with HSA; however, the stabilities of these flavonoid-HSA complexes show some differences (Table 1). Q30 S and IR bind with slightly higher while TAM binds slightly lower affinity to HSA when compared to Q. Furthermore, the binding constants of the glucuronide metabolites (Q3G and I3G) were considerably lower than the other compounds. On the other hand, it is important to note that despite having a lower affinity for HSA than Q, the stabilities of the Q3G-HSA and I3G-HSA complexes (logK = 4.5–4.7) were still high enough for significant interaction with HSA. The logK value of the WAR-HSA complex is approximately 5.3 [34] demonstrating that Q, Q30 S, IR, and TAM bind with similar affinity while Q3G and I3G bind with lower affinity to HSA than WAR. Interaction of some of these Q metabolites with HSA has been investigated by fluorescence spectroscopy before, and the described results are in a good agreement with our observation that affinity order of the metabolites is Q30 S > Q > Q3G > I3G [35]. The significantly lower binding affinity of glucuronides toward HSA can be explained by the followings. The benzopyrone moiety of Q is located within the binding pocket of subdomain IIA, and the nitrogen atom of Lys195 forms intermolecular H-bond or ionic interaction with the 3-hydroxyl group of Q [36]. The presence of the bulky polar substituent in position 3 may decrease the binding affinity of glucuronides because of steric reasons; furthermore, the absence of the interaction of 3-hydroxyl moiety with Lys195 can result in the lower complex stabilities as well. Complex formations of Q glycosides with HSA follow a similar fashion, binding constants of rutin-HSA and hyperoside-HSA complexes are significantly lower compared to the Q-HSA complex resulted from the presence of the bulky polar rutinose and galactose substituents in position 3, respectively [37]. Previous studies suggest that most of the flavonoid aglycones that form stable complexes with HSA occupy the binding site called Sudlow’s Site I (or Drug Binding Site I; on subdomain IIA) [5–7]. Because the single tryptophan residue (Trp-214) of HSA is located in this part of the protein, binding of flavonoids to the HSA
molecule results in the strong quenching of its fluorescence [6]. For this reason the interaction of flavonoids with HSA can be precisely followed using the fluorescence quenching method. Q metabolites significantly decreased the fluorescence of HSA at 340 nm (Figs. 2 and 3), therefore it is plausible to hypothesize that the binding site of each tested metabolite is located on Sudlow’s Site I. 3.2. Displacement of warfarin from albumin by quercetin metabolites Because most of the Q metabolites (except the glucuronides) bind with similar affinity to HSA as Q, it is very likely that Q metabolites are also able to displace the Sudlow’s Site I ligand warfarin from HSA in the same way as Q and other flavonoids [18]. To test this hypothesis, our previously described model was applied [33]. Since WAR shows a much stronger fluorescence signal in HSA-bound form than free WAR, the interaction of WAR with HSA as well as the displacement of HSA-bound WAR results in significant changes to the fluorescence spectrum of WAR [18]. Using these principles, increasing flavonoid concentrations (0–5 mM) were added to standard WAR (1 mM) and HSA (3.5 mM), and fluorescence emission spectra of WAR were recorded (lexc = 317 nm). Applying our previously described method [18], the concentration of HSA-bound WAR was quantified based on the emission intensities of WAR (lem = 379 nm). Our results demonstrate that the presence of each tested metabolite resulted in a significant decrease of the fluorescence signal of WAR suggesting that all Q metabolites are able to displace WAR from HSA (Fig. 4). Flavonoids and HSA alone did not give a fluorescence signal at the wavelengths used under the applied conditions. Furthermore, to exclude the potential involvement of an internal filter effect, the absorption spectra of the flavonoids were also recorded. Only negligible absorbance values of Q and Q metabolites were observed. The strongest displacement effects were caused by IR and Q30 S, followed by Q and TAM, while considerably lower but significant impacts resulted from Q3G and I3G treatment. These data are in a very good agreement with our previous results (see in 3.1.) because the competitors which showed higher binding constants with HSA demonstrated stronger displacing abilities
Table 1 Stabilities of flavonoid-HSA complexes: Stern-Volmer quenching constants (KSV) and binding constants (K) (the unit of KSV and K is dm3/mol).
log KSV log K
QUE
Q30 S
Q3G
I3G
IR
TAM
5.24 0.03 5.37 0.00
5.35 0.02 5.49 0.01
4.72 0.03 4.73 0.00
4.54 0.08 4.55 0.00
5.27 0.04 5.43 0.01
5.10 0.01 5.19 0.00
578
M. Poór et al. / Biomedicine & Pharmacotherapy 88 (2017) 574–581
After oral consumption, the molar amount of Q itself gives only a few percent of the plasma concentration [38]. Approximately 50, 25, and 10% of the circulating Q is presented by Q30 S, Q3G, and I3G, respectively [16], indicating that Q30 S is the dominant metabolite of Q in the human plasma. Considering the observation that continuous administration of 1000 mg Q daily can result in even 2 mM plasma concentration of total Q (including Q as well as its metabolites) [11,39], it is plausible to hypothesize that the plasma concentration of Q30 S may exceed the 1 mM, while Q3G and I3G likely presented at few hundred nM concentrations. Furthermore, the steady-state concentration of WAR during the anticoagulant therapy is few mM (typically 1.5–5.0 mM) [24,40], suggesting that the plasma concentration of total Q can comparable or slightly lower compared to WAR. 3.3. Investigation of the displacing effect of Q and Q metabolite combinations Fig. 4. Decrease of the albumin-bound fraction of warfarin (1 mM warfarin + 3.5 mM human serum albumin) in the presence of increasing flavonoid concentrations (0.5, 1.0, 2.0, 3.0, 4.0, and 5.0 mM) in PBS (pH 7.4). [Q: quercetin, Q30 S: quercetin-30 sulfate, Q3G: quercetin-3-glucuronide, I3G: isorhamnetin-3-glucuronide, IR: isorhamnetin, and TAM: tamarixetin].
vs. WAR (Table 1 and Fig. 4). Our results indicate that each tested Q metabolite is able to compete with WAR. Furthermore, these data also highlight that Q30 S, IR, and TAM are similarly active competitors than Q. The glucuronide conjugates bind with lower affinity to HSA and they are less effective competitors than Q, however, they are still able to displace WAR from HSA. The displacement of WAR by the tested metabolites is consistent with our previous hypothesis that the binding site of the metabolites (including the conjugates) needs to be located on Sudlow’s Site I in the same way as Q and WAR [17,18]. The observation that even micromolar concentrations of Q metabolites result in a considerable displacement of WAR strongly suggests that Q metabolites are able to disrupt the albumin binding of WAR.
After the exposition/treatment with Q-containing foods or dietary supplements, Q circulates, mainly as its metabolites in the blood [16,38,41] resulting in the possibility of a combined effect. To test this, the displacing abilities of mixtures of Q and its metabolites were examined in combination. Samples containing 1 mM WAR and 3.5 mM HSA were studied in the absence and presence of a flavonoid pair (1 mM concentrations of both applied flavonoids). Based on the result of the previous experiment (see in 3.2.) the displaced fraction of WAR by 1 mM flavonoid alone was determined, and the sum of these data was calculated in order to compare the combined flavonoid effect with the expected additive effect of the two compounds. Fig. 5 demonstrates that in most of the cases the displacing abilities of flavonoid pairs showed additive or quasi-additive property. Some of the flavonoid pairs represented lower than additive effect but no dramatic differences were observed compared to the calculated and the measured data. These results indicate that in the presence of Q and its metabolites, the combined displacing effect will be close to additive making the
Fig. 5. Investigation of the potential additive property of the displacing effects of quercetin and quercetin metabolites. Displaced amounts (%) of warfarin (WAR) were quantified in the presence of 1 mM flavonoids alone (see in 3.2. and Fig. 4), then flavonoid pairs were formed. The displacing effect of the flavonoid pairs (1 + 1 mM) was added based on their separate displacing effects (calculated values). Thereafter, these data were compared with the result of the measurements where the combined effects of flavonoid pairs were determined in the presence of 1 mM WAR, 3.5 mM human serum albumin, and 1 mM of both tested flavonoids in PBS (measured values; lexc = 317 nm, lem = 379 nm). [Q: quercetin, Q30 S: quercetin-30 -sulfate, Q3G: quercetin-3-glucuronide, I3G: isorhamnetin-3-glucuronide, IR: isorhamnetin, and TAM: tamarixetin].
M. Poór et al. / Biomedicine & Pharmacotherapy 88 (2017) 574–581
strong displacement of WAR from HSA by Q and its metabolites more likely. 3.4. Ultrafiltration studies To further confirm the displacing abilities of Q and its metabolites, ultrafiltration studies were performed. Because HSA is a large protein (approximately 67.5 kDa), it is unable to pass through a filter unit with a 10 kDa molecular weight cut-off value, therefore only the free (not HSA-bound) WAR will appear in the filtrate. Because the duration of the applied ultrafiltration process is 10 min, the equilibrium in the supernatant is continuously changing. For this reason, the determined WAR concentration in the filtrate is not exactly equal with the free fraction of WAR in the sample, but the elevated concentration of WAR in the filtrate indicates the increase of its free fraction. During this experiment, 1 mM WAR and 5 mM HSA was applied in PBS. Under these circumstances, the free WAR concentration in the filtrate is approximately 50% compared to the WAR content measured when WAR was filtered alone. To investigate the effect of Q and its metabolites on WAR-HSA interaction, 7.5 and 15 mM flavonoids were applied. When samples contained only WAR and flavonoids (in the absence of HSA), Q and its metabolites did not change the concentration of WAR in the filtrate compared to the experiments when WAR was filtered alone. However, in the presence of WAR, HSA and flavonoids, most of the tested flavonoids significantly increased the WAR content of the filtrate indicating the displacement of WAR from HSA (Fig. 6). During these experiments, Q, Q30 S, IR, and TAM showed again strong displacing abilities, while Q3G and I3G caused only non-significant increase of the filtered WAR under the applied conditions. Therefore, the ultrafiltration experiments further supported the fluorescence spectroscopic measurements (see in 3.2.) again indicating that Q metabolites (except the glucuronide conjugates) are able to strongly interfere with the albumin binding of WAR. 3.5. Inhibition of CYP2C9 by Q and Q metabolites Displacement of WAR from HSA results in the increase of free WAR concentration in the circulation, therefore it is plausible to
579
hypothesize its increased uptake by liver cells as well as the elevated speed of its metabolic inactivation by CYP2C9. On the other hand, previous studies demonstrated that Q and IR can competitively inhibit CYP2C9 enzyme [28,29] suggesting that Q and/or some of its metabolites are able to disrupt the metabolic elimination of WAR. To test this idea, the inhibitory effects of Q and Q metabolites relative to WAR were tested using diclofenac, the FDA-recommended substrate of CYP2C9. Fig. 7 demonstrates the formation of 40 -hydroxydiclofenac in the absence and presence of 25 or 50 mM WAR or flavonoid after 60, 90, and 120 min incubations. Under the applied circumstances, WAR strongly decreased the CYP2C9-catalyzed 40 -hydroxydiclofenac formation in each time period, while the same concentrations of Q3G and I3G did not cause any differences. Significant inhibition of CYP2C9 was observed in the presence of higher (50 mM) Q and Q30 S concentrations; however, IR resulted in significant impact only after 90 and 120 min incubations (Fig. 7 and Fig. S2). On the other hand, the inhibitory effects of Q, Q30 S, and IR were considerably lower compared to WAR. After 60 and 90 min incubations, strong inhibition of CYP2C9 enzyme was observed by TAM resulting in the similar effect of 25 mM TAM than 25 mM WAR; however, 50 mM TAM did not lead to further decrease of 40 hydroxydiclofenac formation (Fig. 7 and Fig. S2). Furthermore, inhibitory action of TAM was negligible compared to WAR after 120 min incubation. Based on these results, Q and/or some of its metabolites (Q30 S or IR) may slightly inhibit the metabolic elimination of WAR, however, it is unlikely that their inhibitory effect on CYP2C9 enzyme results in relevant interaction in the body. It is also important to note that COMT enzyme favors the formation of IR vs. TAM therefore relatively low amount of TAM is produced from Q in the human liver [13,14], suggesting the low pharmacological relevance of the inhibitory effect of TAM on CYP2C9 enzyme. Finally, the CYP2C9 inhibition assay was also performed in the presence of both WAR (25 mM) and flavonoids (25 mM). Combined application of WAR and Q metabolites results in only slight but non-significant effect compared to WAR alone (Fig. S2). A further minor decrease of 40 -hydroxydiclofenac formation was observed in the presence of Q, Q30 S, IR, and TAM; while in the presence of the glucuronide metabolites the formation of 40 -hydroxydiclofenac
Fig. 6. Results of the ultrafiltration experiments: Effects of quercetin (Q) or its metabolites on warfarin (WAR) concentrations in the filtrate. Before ultrafiltration, samples contained 1 mM WAR, 5 mM human serum albumin (HSA) in the absence and presence of 7.5 or 15 mM flavonoid concentrations in PBS (pH 7.4). Data are expressed in percentage, where 100% means the WAR concentration in the filtrate when the samples contained neither HSA nor flavonoids (*p < 0.05; **p < 0.01). [Q30 S: quercetin-30 sulfate, Q3G: quercetin-3-glucuronide, I3G: isorhamnetin-3-glucuronide, IR: isorhamnetin, and TAM: tamarixetin].
580
M. Poór et al. / Biomedicine & Pharmacotherapy 88 (2017) 574–581
was practically unchanged. These results suggests again the poor inhibitory effect of Q metabolites on CYP2C9 enzyme compared to WAR. 4. Conclusions In this study, the pharmacokinetic interactions of Q and its main metabolites with WAR were investigated. Our results show that Q metabolites bind to HSA with high affinity, and we also demonstrated that methylated and sulfate metabolites have similar or even higher displacing abilities than the parent compound Q. Glucuronide conjugates showed lower affinities towards HSA and their abilities to compete with WAR to bind were also lower compared to Q. On the other hand, the fact that glucuronide metabolites of Q are also able to displace WAR and the close to additive behavior of their displacing effect in the presence of more than one metabolites suggest that frequent exposure to high doses of Q (e.g., through the consumption of dietary supplements containing large amounts of Q) can result in significant interaction with WAR. Because the strong displacement of WAR from HSA can lead to bleeding, it can cause serious consequences. However, our results also indicate that the tested flavonoids have considerably lower effect on CYP2C9-catalysed hydroxylation than WAR, suggesting that this is not, in fact, an additional concern. For these listed reasons, we need to consider the potentially hazardous consequences of the consumption of high levels of quercetin-containing dietary supplements in WARtreated patients. Acknowledgements The present scientific contribution is dedicated to the 650th anniversary of the foundation of the University of Pécs, Hungary. PAK and PWN acknowledge the support of the Biotechnology and Biological Sciences Research Council (UK) through an Institute Strategic Programme Grant (‘Food and Health’; Grant No. BB/ J004545/1) to the Institute of Food Research. This work was supported by the GINOP-2.3.2-15-2016-00049 grant. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.biopha.2017. 01.092. References
Fig. 7. Formation of 40 -hydroxydiclofenac from diclofenac by CYP2C9 enzyme in the absence and presence of warfarin (WAR) or flavonoids (see details in 2.4.) after 60, 90, and 120 min incubations (**p < 0.01) [Q: quercetin, Q30 S: quercetin-30 -sulfate, Q3G: quercetin-3-glucuronide, I3G: isorhamnetin-3-glucuronide, IR: isorhamnetin, and TAM: tamarixetin].
[1] B.H. Havsteen, The biochemistry and medical significance of the flavonoids, Pharmacol. Ther. 96 (2002) 67–202. [2] M. Ø. Andersen, K.R. Markham, Flavonoids: Chemistry, Biochemistry and Applications, CRC, Press, Taylor & Francis Group, Boca Raton, FL, US, 2006. [3] C. Amato, Advantage of a micronized flavonoidic fraction (Daflon 500 mg) in comparison with a nonmicronized diosmin, Angiology 45 (1994) 531–536. [4] S.H. Thilakarathna, H.P. Rupasinghe, Flavonoid bioavailability and attempts for bioavailability enhancement, Nutrients 5 (2013) 3367–3387. [5] C. Dufour, O. Dangles, Flavonoid-serum albumin complexation: determination of binding constants and binding sites by fluorescence spectroscopy, Biochim. Biophys. Acta 1721 (2005) 164–173. [6] J. Xiao, Y. Zhao, H. Wang, Y. Yuan, F. Yang, C. Zhang, K. Yamamoto, Noncovalent interaction of dietary polyphenols with common human plasma proteins, J. Agric. Food Chem. 59 (2011) 10747–10754. [7] S. Pal, C. Saha, A review on structure-affinity relationship of dietary flavonoids with serum albumins, J. Biomol. Struct. Dyn. 32 (2014) 1132–1147. [8] J. Peterson, J. Dwyer, Flavonoids Dietary occurrence and biochemical activity, Nutr. Res. 18 (1998) 1995–2018. [9] G.S. Kelly, Quercetin. Monograph, Altern. Med. Rev. 16 (2011) 172–194. [10] J.A. Conquer, G. Maiani, E. Azzini, A. Raguzzini, B.J. Holub, Supplementation with quercetin markedly increases plasma quercetin concentration without effect on selected risk factors for heart disease in healthy subjects, J. Nutr. 128 (1998) 593–597.
M. Poór et al. / Biomedicine & Pharmacotherapy 88 (2017) 574–581 [11] S.A. Heinz, D.A. Henson, D.C. Nieman, M.D. Austin, F. Jin, A 12-week supplementation with quercetin does not affect natural killer cell activity, granulocyte oxidative burst activity or granulocyte phagocytosis in female human subjects, Br. J. Nutr. 104 (2010) 849–857. [12] A.W. Boots, G.R. Haenen, A. Bast, Health effects of quercetin: from antioxidant to nutraceutical, Eur. J. Pharmacol. 585 (2008) 325–337. [13] P.T. Männistö, S. Kaakkola, Catechol-O-methyltransferase (COMT): biochemistry, molecular biology, pharmacology, and clinical efficacy of the new selective COMT inhibitors, Pharmacol. Rev. 51 (1999) 593–628. [14] C. De Santi, A. Pietrabissa, F. Mosca, G.M. Pacifici, Methylation of quercetin and fisetin flavonoids widely distributed in edible vegetables, fruits and wine, by human liver, Int. J. Clin. Pharmacol. Ther. 40 (2002) 207–212. [15] M. Poór, Z. Zrínyi, T. KÅszegi, Structure related effects of flavonoid aglycones on cell cycle progression of HepG2 cells: metabolic activation of fisetin and quercetin by catechol-O-methyltransferase (COMT), Biomed. Pharmacother. 83 (2016) 998–1005. [16] W. Mullen, C.A. Edwards, A. Crozier, Absorption, excretion and metabolite profiling of methyl-, glucuronyl-, glucosyl- and sulpho-conjugates of quercetin in human plasma and urine after ingestion of onions, Br. J. Nutr. 96 (2006) 107– 116. [17] M. Poór, S. Kunsági-Máté, T. Bencsik, J. Petrik, S. Vladimir-Kneževi c, T. KÅszegi, Flavonoid aglycones can compete with Ochratoxin A for human serum albumin: a new possible mode of action, Int. J. Biol. Macromol. 51 (2012) 279– 283. [18] M. Poór, Y. Li, S. Kunsági-Máté, J. Petrik, S. Vladimir-Kneževi c, T. KÅszegi, Molecular displacement of warfarin from human serum albumin by flavonoid aglycones, J. Lumin. 142 (2013) 122–127. [19] O. Dangles, C. Dufour, C. Manach, C. Morand, C. Remesy, Binding of flavonoids to plasma proteins, Methods Enzymol. 335 (2001) 319–333. [20] P.A. Tideman, R. Tirimacco, A. St John, G.W. Roberts, How to manage warfarin therapy, Aust. Prescr. 38 (2015) 44–48. [21] A.M. Holbrook, J.A. Pereira, R. Labiris, H. McDonald, J.D. Douketis, M. Crowther, P.S. Wells, Systematic overview of warfarin and its drug and food interactions, Arch. Intern. Med. 165 (2005) 1095–1106. [22] P.M. Leite, M.A. Martins, R.O. Castilho, Review on mechanisms and interactions in concomitant use of herbs and warfarin therapy, Biomed. Pharmacother. 83 (2016) 14–21. [23] N.H. Holford, Clinical pharmacokinetics and pharmacodynamics of warfarin. Understanding the dose-effect relationship, Clin. Pharmacokinet. 11 (1986) 483–504. [24] E. Chan, A.J. McLachlan, M. Pegg, A.D. MacKay, R.B. Cole, M. Rowland, Disposition of warfarin enantiomers and metabolites in patients during multiple dosing with rac-warfarin, Br. J. Clin. Pharmacol. 37 (1994) 563–569. [25] L. Di Bari, S. Ripoli, S. Pradhan, P. Salvadori, Interactions between quercetin and warfarin for albumin binding: a new eye on food/drug interference, Chirality 22 (2010) 593–596. [26] D.R. Flora, A.E. Rettie, R.C. Brundage, T.S. Tracy, CYP2C9 genotype-dependent warfarin pharmacokinetics: impact of CYP2C9 genotype on R- and S-Warfarin
[27]
[28]
[29]
[30] [31]
[32]
[33] [34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
581
and their oxidative metabolites, J. Clin. Pharmacol. (2016), doi:http://dx.doi. org/10.1002/jcph.813 in press. T.Y. Chan, Adverse interactions between warfarin and nonsteroidal antiinflammatory drugs: mechanisms, clinical significance, and avoidance, Ann. Pharmacother. 29 (1995) 1274–1283. D. Si, Y. Wang, Y.H. Zhou, Y. Guo, J. Wang, H. Zhou, Z.S. Li, J.P. Fawcett, Mechanism of CYP2C9 inhibition by flavones and flavonols, Drug Metab. Dispos. 37 (2009) 629–634. Y. Kimura, H. Ito, R. Ohnishi, T. Hatano, Inhibitory effects of polyphenols on human cytochrome P450 3A4 and 2C9 activity, Food Chem. Toxicol. 48 (2010) 429–435. P.W. Needs, P.A. Kroon, Convenient syntheses of metabolically important quercetin glucuronides and sulfates, Tetrahedron 62 (2006) 6862–6868. M. Poór, Y. Li, S. Kunsági-Máté, Z. Varga, A. Hunyadi, B. Dankó, F.R. Chang, Y.C. Wu, T. KÅszegi, Protoapigenone derivatives: albumin binding properties and effects on HepG2 cells, J. Photochem. Photobiol. B 124 (2013) 20–26. Y. Li, Z. Czibulya, M. Poór, S. Lecomte, L. Kiss, E. Harte, T. KÅszegi, S. KunságiMáté, Thermodynamic study of the effects of ethanol on the interaction of ochratoxin A with human serum albumin, J. Lumin. 148 (2014) 18–25. M. Poór, B. Lemli, M. Bálint, C. Hetényi, N. Sali, T. KÅszegi, S. Kunsági-Máté, Interaction of citrinin with human serum albumin, Toxins 7 (2015) 5155–5166. M. Poór, Y. Li, G. Matisz, L. Kiss, S. Kunsági-Máté, T. KÅszegi, Quantitation of species differences in albumin-ligand interactions for bovine, human and rat serum albumins using fluorescence spectroscopy: a test case with some Sudlow’s site I ligands, J. Lumin. 145 (2014) 767–773. K.M. Janisch, G. Williamson, P. Needs, G.W. Plumb, Properties of quercetin conjugates: modulation of LDL oxidation and binding to human serum albumin, Free Radic. Res. 38 (2004) 877–884. F. Zsila, Z. Bikádi, M. Simonyi, Probing the binding of the flavonoid quercetin to human serum albumin by circular dichroism, electronic absorption spectroscopy and molecular modelling methods, Biochem. Pharmacol. 65 (2003) 447–456. S. Bi, L. Ding, Y. Tian, D. Song, X. Zhou, X. Liu, H. Zhang, Investigation of the interaction between flavonoids and human serum albumin, J. Mol. Struct. 703 (2004) 37–45. L. Wang, M.E. Morris, Liquid chromatography-tandem mass spectroscopy assay for quercetin and conjugated quercetin metabolites in human plasma and urine, J. Chromatogr. 821 (2005) 194–201. F. Jin, D.C. Nieman, R.A. Shanely, A.M. Knab, M.D. Austin, W. Sha, The variable plasma quercetin response to 12-week quercetin supplementation in humans, Eur. J. Clin. Nutr. 64 (2010) 692–697. S. Sun, M. Wang, L. Su, J. Li, H. Li, D. Gu, Study on warfarin plasma concentration and its correlation with international normalized ratio, J. Pharm. Biomed. Anal. 42 (2006) 218–222. P.A. Kroon, M.N. Clifford, A. Crozier, A.J. Day, J.L. Donovan, C. Manach, G. Williamson, How should we assess the effects of exposure to dietary polyphenols in vitro? Am. J. Clin. Nutr. 80 (2004) 15–21.