Inhibition of Cytochrome P450 2C8-Mediated Drug Metabolism by the Flavonoid Diosmetin

Drug Metab. Pharmacokinet. 26 (6): 559568 (2011). Copyright © 2011 by the Japanese Society for the Study of Xenobiotics (JSSX) Regular Article Inhibition of Cytochrome P450 2C8-mediated Drug Metabolism by the Flavonoid Diosmetin

Luigi QUINTIERI1,PietroPALATINI1,StefanoMORO2 and Maura FLOREANI1,* 1Department of Pharmacology and Anaesthesiology, University of Padova, Italy 2Molecular Modeling Section (MMS), Department of Pharmaceutical Sciences, University of Padova, Italy

Full text of this paper is available at http://www.jstage.jst.go.jp/browse/dmpk

Summary: The aim of this study was to assess the effects of diosmetin and hesperetin, two ﬂavonoids present in various medicinal products, on CYP2C8 activity of human liver microsomes using paclitaxel oxidation to 6¡-hydroxy-paclitaxel as a probe reaction. Diosmetin and hesperetin inhibited 6¡-hydroxy- paclitaxel production in a concentration-dependent manner, diosmetin being about 16-fold more potent

than hesperetin (mean IC50 values 4.25 « 0.02 and 68.5 « 3.3 µM for diosmetin and hesperetin, respectively). Due to the low inhibitory potency of hesperetin, we characterized the mechanism of diosmetin-induced inhibition only. This ﬂavonoid proved to be a reversible, dead-end, full inhibitor of

CYP2C8, its mean inhibition constant (Ki)being3.13« 0.11 µM. Kinetic analysis showed that diosmetin caused mixed-type inhibition, since it signiﬁcantly decreased the Vmax (maximum velocity) and increased the Km value (substrate concentration yielding 50% of Vmax) of the reaction. The results of kinetic analyses were consistent with those of molecular docking simulation, which showed that the putative binding site of diosmetin coincided with the CYP2C8 substrate binding site. The demonstration that diosmetin inhibits CYP2C8 at concentrations similar to those observed after in vivo administration (in the low micromolar range) is of potential clinical relevance, since it may cause pharmacokinetic interactions with co- administered drugs metabolized by this CYP.

Keywords: CYP2C8 inhibition; diosmetin; paclitaxel; drugdrug interactions, molecular docking simulations

present in pharmaceutical preparations. For example, Introduction diosmin, the 7-rutinoside ¤rhamnosyl-D-glucoside¥ of the Flavonoids are polyphenolic compounds ¤produced as flavone diosmetin ¤3$,5,7-trihydroxy-4-methoxyflavone¥, secondary plant metabolites¥ widely ingested with the diet alone or in association ¤9:1, w/w¥ with hesperidin, the by humans. Their beneficial effects against oxidative stress 7-rutinoside of the flavanone hesperetin ¤3$,5,7-trihydroxy- and related pathologies1®3¥ have been mainly ascribed to their 4-methoxyflavanone¥¤Fig. 1¥, is widely used for the antioxidant activity,4¥ although flavonoids appear to exert management of chronic venous insufficiency, venous ulcers more complex biological effects, including inhibition of and hemorrhoids.9¥ However, the safe use of herbs or tumor proliferation, modulation of several enzymes and medications containing flavonoids should take adequate interaction with different signal transduction pathways.5¥ account of their potential for interactions with co- Moreover, they seem to exhibit cancer-preventive proper- administered drugs, since various clinical studies have ties, possibly due to inhibition of cytochromes P450 ¤CYP¥ demonstrated that some flavonoids modify the oral 1A1, CYP1A2 and CYP1B1, the CYP forms principally disposition kinetics of numerous drugs.10®15¥ For example, involved in the metabolism of pro-carcinogens.6®8¥ For these Rajnarayana et al. showed that oral treatment of healthy reasons, the use of dietary supplements and/or herbal volunteers with diosmin increases significantly the bioavail- medications containing flavonoids has consistently increased ability of metronidazole,12¥ diclofenac14¥ and chlorzoxa- in recent years. In many countries, some flavonoids are also zone,15¥ and hypothesized that diosmin-induced inhibition

Received; May 30, 2011, Accepted; July 14, 2011 J-STAGE Advance Published Date: July 26, 2011, doi:10.2133/dmpk.DMPK-11-RG-048 *To whom correspondence should be addressed: Prof. Maura FLOREANI, Department of Pharmacology and Anaesthesiology, University of Padova, Largo Meneghetti 2, 35131 Padova, Italy. Tel. +39 049 827 5088, Fax. +39 049 827 5093, E-mail: m.ﬂ[email protected] This work was supported by grants from the University of Padova, Italy. The molecular modeling work coordinated by S. M. was carried out also with ﬁnancial support from the Italian Ministry of Education, University and Research (MIUR), Rome, Italy.

559 560 Luigi QUINTIERI, et al.

Fig. 1. Basic structures and numbering system of flavones, flavanones and flavonols The figure also shows the substitutions for diosmetin (3′,5,7-trihydroxy-4-methoxyflavone) and hesperetin (3′,5,7-trihydroxy-4-methoxyflavanone), and for some flavonols which also inhibit CYP2C8.24) Moreover, the structure of diosmin and hesperidin, natural glycosides of diosmetin and hesperetin, respectively, are reported. of the CYP enzymes catalyzing the biotransformation of terize the mechanism of CYP2C8 inhibition caused by these drugs was responsible for the observed pharmacoki- diosmetin, we performed both enzyme kinetic studies and netic interactions. We have recently demonstrated that molecular docking simulations, since these two techniques diosmetin, the absorbable aglycone of diosmin,16¥ to which provide useful complementary information on the mecha- orally administered diosmin is converted by rhamnosidases nisms of drug-enzyme interactions. of Enterobacteriaceae,17¥ is a potent in vitro inhibitor of human ¥ ¥ Materials and Methods CYP3A4/518 and CYP2C9,19 the CYP forms responsible for the metabolism of metronidazole and diclofenac, Human liver microsomes ¤HLMs¥ and reagents: respectively,12,14¥ thereby confirming the hypothesis of Pooled mixed-gender HLMs ¤obtained from 25 female and Rajnarayana et al.12,14¥ 27 male donors¥ were provided by Xenotech LLC ¤Lenexa, Since CYP2C9 shares more than 80% amino acid KS, USA¥. The microsomal fractions were stored in aliquots sequence identity with CYP2C8,20¥ a CYP form expressed at %80ôC until use. at relatively high levels ¤6®7% of total CYP content¥ in Diosmetin and hesperetin, purchased from Chromadex human liver21¥ and which plays a major role in the ¤Irvine, CA, USA¥, were dissolved in N,N-dimethylform- 22¥ ¤ ¥ ¤ ¥ fl metabolism of several therapeutically important drugs, amide DMF /H2O 50:50, v/v . Solutions of the avonoid the main aim of the present study was to assess whether were prepared daily and kept in ice until use. The final diosmetin is also an effective inhibitor of human CYP2C8. solvent concentration in the incubation medium was 0.5%. Because in some pharmaceutical preparations the flavone Paclitaxel was a kind gift from Indena ¤Milan, Italy¥, diosmin is associated with the flavanone hesperidin, which is whereas its 6Æ-hydroxy metabolite ¤6Æ-OH paclitaxel¥ was also likely to be hydrolyzed by intestinal microflora to its purchased from SPI-Bio ¤Montigny le Bretonneux, France¥. aglycone hesperetin,17¥ a further aim of our study was to test Paclitaxel and 6Æ-OH paclitaxel were dissolved in aceto- ¤ ¥ ¤ ¥ the effect of hesperetin on CYP2C8 activity. Although this nitrile CH3CN /H2O 50:50, v/v .CH3CN was chosen as enzyme has been shown to be inhibited by some flavonols, the solvent for paclitaxel since, at the final concentration such as quercetin, morin and fisetin ¤see Fig. 1¥,23,24¥ to the used in our incubation mixtures ¤1%¥, it had been shown best of our knowledge, no data are available in the literature to produce a negligible effect on CYP2C8-mediated concerning the effects of flavones and/or flavanones on paclitaxel hydroxylation.27¥ NADPH and DMF were CYP2C8-catalyzed reactions. In this study, we used the purchased from Sigma-Aldrich Italy ¤Milan, Italy¥. Potassium ¤ ¥ antitumor agent paclitaxel as a probe substrate for CYP2C8 dihydrogen phosphate KH2PO4 ,CH3CN and methanol activity of human liver microsomes ¤HLMs¥ for two main ¤all of HPLC grade¥ were from Carlo Erba Reagenti reasons: ¤1¥ 6Æ-hydroxylation of paclitaxel is a validated ¤Milan, Italy¥. Ultrapure water was obtained by means of a marker reaction of CYP2C8 activity,22,23¥ since CYP2C8 is Pure-Lab Option Q ¤Elga Lab-Water, High Wycombe, UK¥ the only CYP form catalyzing this reaction in humans;25¥ ¤2¥ apparatus. a possible inhibitory effect of diosmetin and/or hesperetin on Determination of 6Æ-OH-paclitaxel formation this metabolic reaction would be of great clinical relevance, rate by HLMs: Incubation of paclitaxel with HLMs was since formation of the inactive metabolite 6Æ-hydroxy- carried out in conditions yielding linear reaction rates with paclitaxel ¤6Æ-OH-paclitaxel¥ accounts for more than 60% of respect to microsomal protein concentration and incubation paclitaxel metabolic disposition in humans.25,26¥ To charac- time at all paclitaxel concentrations used for our assays.

HLMs ¤final protein concentration: 0.5 mg/ml¥ were n © 6¥ in the presence of heat-inactivated microsomes ¤final incubated in a mixture ¤total volume: 200 µl¥ containing concentration: 0.5 mg/ml¥ processed exactly as the samples ¤ ¥ 100 mM KH2PO4 pH 7.4 , 0.5 mM NADPH, and increas- obtained from kinetic experiments. The calibration curves ing concentrations ¤n © 9¥ of paclitaxel ¤from 3 to 40 µM¥. were linear in the above-reported concentration range Moreover, in some experiments the incubation medium ¤r2 ; 0.99¥, the lowest value of the range representing the contained 0.5% ¤final concentration¥ of the flavonoid solvent limit of quantification of the assay. The inter- and intra-assay DMF. After a 3-min thermal equilibration of the incubation coefficients of variation ¤CV¥ for 6Æ-OH-paclitaxel determi- mixtures at 37ôC, the reaction was started by the addition of nation ¤n © 5¥ at 0.04 and 0.4 nmol/0.2 ml were both less paclitaxel and was allowed to proceed for 10 min in a than 2%. shaking water bath at 37ôC in aerobic conditions before Data analysis: Initial velocity ¤v¥ data for 6Æ-OH- being stopped by addition of 50 µl of ice-cold CH3CN. After paclitaxel formation catalyzed by HLMs were evaluated by denatured proteins were removed by centrifugation for graphical analysis with the Eadie-Hofstee plot ¤v vs v/ªS«¥ and 10 min at 20000 g, aliquots of the supernatants ¤100 µl¥ were by best-fitting procedures using GraphPad Prism software, analyzed by high-performance liquid chromatography version 5.03 ¤GraphPad Software Inc., San Diego, CA, ¤HPLC¥ with UV detection, as described below. USA¥. The F test was used to discriminate between different Evaluation of the effects of diosmetin and kinetic models ¤one- or two-site hyperbolic Michaelis- hesperetin on 6Æ-OH-paclitaxel production by Menten model, or Michaelis-Menten kinetics with substrate ¥ HLMs: To determine the effects of diosmetin and inhibition . The estimated kinetic parameters were: Vmax, Æ ¤fi hesperetin on 6 -OH-paclitaxel production, HLMs nal maximum velocity of reaction; Km, substrate concentra- ¥ protein concentration 0.5 mg/ml were incubated in the tion yielding 50% of Vmax; and CLint, intrinsic metabolic absence or presence of increasing concentrations of clearance, calculated as Vmax/Km. ¤ ¥ ¤ ¤ diosmetin from 0.5 to 25 µM or hesperetin from 5 to IC50 values inhibitor concentrations yielding 50% of 200 µM¥ in the above-specified incubation mixture contain- inhibition¥ were obtained graphically from each experiment. ¤ ¥ Æ ing 10 µM paclitaxel. Control samples contained 0.5% The maximal inhibition imax of paclitaxel conversion to 6 - DMF. Incubations were carried out at 37ôC as specified OH-paclitaxel caused by diosmetin was evaluated by means above and were stopped after 10 min by addition of 50 µl of the graphical method described by Palatini.28¥ In this plot, of ice-cold CH3CN. The samples were then processed as if the experimental concentrations of the inhibitor are marked described above. on a negative horizontal axis and the corresponding fractional In order to characterize the kinetic mechanism of inhibition ¤i¥ values are marked on a vertical axis, the straight inhibition of CYP2C8-mediated metabolism of paclitaxel lines drawn through each pair of points intersect at a common Æ by diosmetin, 6 -OH-paclitaxel production catalyzed by point whose ordinal coordinate represents the imax value. HLMs ¤final protein concentration 0.5 mg/ml¥ was deter- The kinetic mechanism of diosmetin-induced inhibition of mined in the above-specified incubation mixture containing 6Æ-OH-paclitaxel formation was inferred from its effect on ¤ fi increasing concentrations of paclitaxel ranging from 3 to Vmax and Km values obtained by tting initial velocity data to 40 µM¥ in the absence or presence of three concentrations of the one-site Michaelis-Menten model. The Dixon plot ¤1/v diosmetin ¤1, 3 and 5 µM¥. vs ªI«¥ was used for graphical presentation of the inhibitory HPLC analysis: Quantitative evaluation of 6Æ-OH- effects of diosmetin on 6Æ-OH-paclitaxel formation. paclitaxel in the supernatants of incubation mixtures was Since preliminary dilution experiments showed that carried out using a Hewlett-Packard series 1100 HPLC diosmetin caused reversible inhibition of CYP2C8 activity system equipped with degasser, quaternary pump, auto- ¤results not shown¥, we calculated the dissociation constant ¤ ¤ ¥ sampler and multiple-wavelength detector Agilent, of the inhibitor-enzyme complex Ki . To this purpose, formerly Hewlett-Packard GMBH, Germany¥; chromato- untransformed kinetic data were fitted to noncompetitive graphic data were collected and integrated by means of and mixed-type ¤competitive-noncompetitive¥ inhibition Hewlett-Packard ChemStation software ¤version A.06.03¥. models by nonlinear regression analysis ¤GraphPad Prism, Chromatographic conditions were: column, Symmetry-C8 version 5.03¥ using the appropriate equations.29¥ The F test ¤4.6 ' 250 mm, 5 µm, Agilent Technologies Inc., Palo Alto, was again used to discriminate between the two different ¥ ¤ ¥ CA, USA ; mobile phase, methanol/H2O 65:35, v/v ; run kinetic models. time, 20 min; flow rate, 0.9 ml/min; injection volume, Statistical analyses were performed using GraphPad 100 µl; column temperature, 40ôC; and detection, UV Prism 5.03 software. All values are expressed as arithmetic absorbance at 230 nm. Under the above conditions, means + S.E.M. Comparison of experimental data obtained retention times of 6Æ-OH-paclitaxel and paclitaxel were in the presence of increasing concentrations of the inhibitor 11.1 and 13.4 min, respectively. Quantitative determination with data obtained in the absence of the inhibitor ¤control of 6Æ-OH-paclitaxel was carried out by means of standard values¥ was made by one-way analysis of variance ¤ANOVA¥ calibration curves obtained with increasing concentrations of followed by Dunnettös post-hoc test. P g 0.05 was authentic 6Æ-OH-paclitaxel ¤from 0.04 to 0.4 nmol/0.2 ml, considered statistically significant.

Evaluation of the metabolic stability of diosmetin gradient was g0.1 kcal/mol/!. GB/SA approximation36¥ in the presence of HLMs: Diosmetin ¤final concen- has been used to model the electrostatic contribution to tration: 25 µM¥ was incubated for 20 min at 37ôCina the free energy of solvation in a continuum solvent model. ¤fi ¥ medium nal volume 200 µl containing 0.1 M KH2PO4 The interaction energy values were calculated as the energy ¤pH 7.4¥, 0.5 mM NADPH and 0.5 mg/ml HLMs. The of the complex minus the energy of the ligand minus the ¼ © ®¤ ¦ ¥ reaction was started by the addition of NADPH and stopped energy of CYP2C8: Einter E¤complex¥ E¤L¥ E¤CYP2C8¥ . by means of 50 µl ice-cold acetonitrile. Proteins were Results removed by centrifugation for 10 min at 20000 g, and aliquots ¤100 µl¥ of the clear supernatants were analyzed by Formation of 6Æ-OH-paclitaxel from paclitaxel HPLC with UV detection. Separation of the flavonoid by HLMs: Paclitaxel conversion to 6Æ-OH-paclitaxel by contained in the supernatant of the incubation mixtures was HLMs was preliminarily evaluated over a substrate carried out using the Hewlett-Packard series 1100 HPLC concentration range from 3 to 40 µM in the absence and system described above. Chromatographic conditions were: presence of DMF ¤0.5% final concentration¥, the solvent column, Eclipse XDB-C8 ¤4.6 ' 150 mm, 5 µm, Agilent used to solubilize diosmetin. Untransformed initial velocity Technologies Inc., Palo Alto, CA, USA¥ protected by a data were best-fitted by the one-site Michaelis-Menten 4.6 ' 12.5 mm pre-column of the same stationary phase; model, both in the absence and presence of 0.5% DMF mobile phase, 20 mM ammonium acetate ¤pH 4.5¥/meth- ¤Fig. 2¥. Graphical analysis of the data according to Eadie- anol ¤70:30 v/v, solvent A¥ and acetonitrile ¤solvent B¥; Hofstee37¥ yielded linear plots ¤insets of Fig. 2¥, confirming elution program, isocratic elution with 100% solvent A for the involvement of a single enzyme in the formation of 6Æ- 5 min, linear gradient elution from 0% to 15% solvent B in OH-paclitaxel by HLMs.38¥ The addition of 0.5% DMF to 15 min, isocratic elution with 15% solvent B for a further the incubation medium did not modify significantly the fl + 11 min; postrun 5 min with 100% solvent A; ow rate, kinetic parameters of the reaction: Vmax was 442 19 and 1.2 ml/min; injection volume, 100 µl; column temperature, 401 + 55 pmol/mg/min in the absence and presence of ô + 30 C; detection, UV absorbance at 254 nm. In the above DMF, respectively, and Km was 24.2 1.4 and conditions, the retention time of authentic diosmetin was 21.8 + 2.9 µM, respectively ¤P h 0.05 for both parame- ¥ 22.5 min. ters . As a consequence, CLint was virtually unaltered Molecular docking simulations: ¤18.3 + 0.3 and 18.6 + 2.3 µl/mg/min in the absence and Target structures presence of DMF, respectively¥. Human CYP2C8 complexed with montelukast was Effect of increasing concentrations of diosmetin retrieved from the Protein Data Bank ¤PDB code: and hesperetin on 6Æ-OH-paclitaxel formation by 2NNI¥30¥ and processed in order to remove the ligand and HLMs: The effect of increasing concentrations of dio- water molecules. Hydrogen atoms were added using smetin and hesperetin was evaluated at a paclitaxel standard geometries to the protein structure with the concentration of 10 µM. Both diosmetin and hesperetin Molecular Operation Environment ¤MOE, version 2010.11¥ decreased the rate of CYP2C8-mediated biotransformation program.31¥ To minimize contacts between hydrogen atoms, of paclitaxel in a concentration-dependent manner, but they the structures were subjected to Amber9432¥ force field exhibited significantly different inhibitory potencies ¤Fig. 3¥, + + minimization until the rms of the conjugate gradient was mean IC50 values being 4.25 0.02 and 68.5 3.3 µM for g0.1 kcal/mol/!, keeping the heavy atoms fixed at their diosmetin and hesperetin, respectively. crystallographic positions. Characterization of the mechanism of inhibition Molecular docking protocol of 6Æ-OH-paclitaxel formation by diosmetin: Be- All docked ligands were built using the Builder module of cause the above results indicate that hesperetin is about MOE.31¥ Ligands were docked into the putative heme 16 times less potent than diosmetin as an inhibitor of binding sites using flexible MOE-Dock methodology.31¥ The CYP2C8-mediated paclitaxel metabolism, and because the purpose of MOE-Dock is to search for favorable binding hesperetin concentrations that can be reached in vivo39,40¥ are fi fl con gurations between a small, exible ligand and a rigid considerably lower than its IC50 value, we characterized the macromolecular target. Searching is conducted within a mechanism of inhibition by diosmetin only. user-specified 3D docking box, using the Tabù search The effect of diosmetin at concentrations exceeding 25 µM protocol33¥ and MMFF94 force field.34¥ Charges for ligands could not be assessed because the amount of 6Æ-OH- were imported from the MOPAC ¤Molecular Orbital paclitaxel recovered in the incubation mixture was below PACkage¥ program35¥ output files. MOE-Dock performs a the quantification limit. Therefore, in order to ascertain user-specified number of independent docking runs ¤55 in whether diosmetin is a full or partial inhibitor, we analyzed the presented case¥ and writes the resulting conformations experimental data by means of the plot described by and their energies to a molecular database file. The resulting Palatini,28¥ which allows a straightforward determination of ligand-CYP2C8 complexes were subjected to MMFF94 all- maximal fractional inhibition values even when the effect atom energy minimization until the rms of the conjugate of saturating inhibitor concentrations cannot be tested.

Fig. 2. Michaelis-Menten kinetics of 6¡-OH-paclitaxel formation Panel A by HLMs evaluated in the absence ( ) and presence Fig. 3. Effect of increasing concentrations of diosmetin ( ) and Panel B ( ) of 0.5% DMF in the incubation medium hesperetin ( )on6¡-OH-paclitaxel formation by HLMs In the insets, Eadie-Hofstee plots are shown. The data are Formation of 6¡-OH-paclitaxel was evaluated at 10 µM paclitaxel. + means S.E.M. of the results from 5 independent experiments. Under these conditions, the mean value of 6¡-OH-paclitaxel produc- TheS.E.M.isnotshownwherethesizeofdatapointsislargerthan tion, determined in the absence of flavonoids, was 98.5 + 13.8 pmol/ the S.E.M. bar. mg protein/min. Results are means + S.E.M. of at least 6 experiments. S.E.M. values are not shown where the size of data points is larger than the S.E.M. bar. Statistical analyses of the data were Figure 4 clearly shows that diosmetin behaved as a full performed by means of Prism 5.03 software, using ANOVA followed inhibitor of 6Æ-OH-paclitaxel formation by HLMs, since the by Dunnett’s post-hoc test to evaluate the effect of increasing fl vs g lines intersected at a common point whose ordinal concentrations of each avonoid . its control value (*P 0.05; ¤ ¥ **P g 0.01). coordinate representing imax is equal to 1. In order to characterize the mechanism of CYP2C8 inhibition by diosmetin, the kinetics of 6Æ-OH-paclitaxel the negative x-axis may also indicate pure competitive formation was examined in the absence and presence of three inhibition,37¥ so we analyzed our data by means of a secondary concentrations of diosmetin ¤1, 3 and 5 µM¥. The data plot in which the slopes of the straight lines are replotted reported in Figure 5 ¤Panel A¥ clearly show that the vs. the reciprocal of substrate concentration.37¥ The resulting presence of diosmetin did not modify the kinetic pattern of line indicates competitive or mixed-type inhibition depending 6Æ-OH-paclitaxel production by HLMs, which followed a on whether it goes through the origin or intersects the ordinal one-site Michaelis-Menten model. The results of non-linear axis at a value greater than zero, respectively.37¥ As clearly regression analysis ¤Table 1¥ indicate that diosmetin decreas- shown in Panel C of Figure 5, the y-intercept of the straight fi ed the Vmax value and increased the Km value of the reaction, line obtained from this replotting procedure was signi cantly although statistically significant modifications were observed different from zero ¤0.41 ' 10%3; 95% Confidence Interval: ' %3¥ fi only at diosmetin concentrations of 3 and 5 µM. CLint 0.25 to 0.56 10 , con rming that diosmetin behaved as a decreased significantly ¤P g 0.01¥ at all diosmetin concen- mixed-type inhibitor. trations. These results indicate that diosmetin behaved as a Preliminary experiments showed that the inhibition mixed-type inhibitor, which is consistent with the Dixon exerted by diosmetin did not increase with the incubation plot reported in Panel B of Figure 5, since the straight lines time and could be reversed upon dilution ¤results not intersected at a common point over the negative x-axis. shown¥, indicating that diosmetin caused a reversible However, in this type of plot an intersection of the lines over inhibition of CYP2C8 activity. Therefore, we calculated

Table 1. Kinetic parameters for 6¡-OH-paclitaxel formation by HLMs in the absence and presence of increasing concentrations of diosmetin

Diosmetin Vmax Km CLint ¤µM¥ ¤pmol/mg/min¥ ¤µM¥ ¤µl/mg/min¥ 0 475 + 7 18.7 + 0.3 25.3 + 0.1 1 430 + 28 20.1 + 1.6 21.7 + 0.4** 3 336 + 12** 25.8 + 1.4* 13.1 + 0.3** 5 246 + 15** 27.8 + 2.4** 9.1 + 0.22** Results are means + S.E.M. from 6 independent experiments. Statistical analysis of the data was performed by means of ANOVA followed by Dunnettös post-hoc test. g g vs. Fig. 4. Determination of maximal inhibition (imax) of diosmetin *P 0.05; **P 0.01 control value. on 6¡-OH-paclitaxel formation by means of the plot described by Palatini30) (i) indicates fractional inhibition. Formation of 6¡-OH-paclitaxel was evaluated at 10 µM paclitaxel. Data represent fractional inhibition values (i) obtained from a typical experiment.

A B C

Fig. 5. Effect of diosmetin on the kinetics of 6¡-OH-paclitaxel formation by HLMs (A) untransformed initial velocity data of 6¡-OH-paclitaxel formation in the absence ( )orpresenceof1µM( ), 3 µM ( )and5µM( ) diosmetin were ﬁtted to the one-site Michaelis-Menten model. (B) Dixon plot, obtained by linear regression analysis of the data, for the effect of diosmetin on 6¡-OH-paclitaxel formation, evaluated at 5 ( ), 7.5 ( ), 15 ( )and30µM( ) paclitaxel. (C) secondary plot obtained according to Segel37) by replotting the slopes of the Dixon plots (of Panel B) vs. the reciprocal of substrate concentrations. Data are means + S.E.M. from 6 independent experiments.

¤ ¥ fi the inhibition constant Ki for diosmetin by tting experimental conditions, diosmetin was not metabolized by untransformed initial velocity data to both a mixed and a CYP enzymes. noncompetitive inhibition model. The F test, used to Molecular docking simulation: Recently, Schoch discriminate between the two types of inhibition, indicated and co-workers30¥ determined the X-ray crystal structures that our experimental data fitted better to a mixed-type of CYP2C8 in complex with montelukast ¤2.8 !, PDB inhibition model, confirming the indications obtained from code: 2NNI¥, troglitazone ¤2.7 !, 2VN0¥, felodipine ¤2.3 !, Table 1 and Figure 5. Nonlinear regression analysis of the 2NNJ¥, or two molecules of 9-cis-retinoic acid ¤2.6 !, data, using the equation describing mixed-type inhibition, 2NNH¥. Normally, CYP2C8 exhibits two substrate-binding + gave a mean Ki value of 3.13 0.11 µM. channels that exit the active site cavity on either side of the Evaluation of the metabolic stability of diosmetin helix B-C loop. The two cavities merge and terminate just in the presence of HLMs: To explore the possibility beyond the heme iron, where the reactive iron-oxo that diosmetin may be a substrate for CYP enzymes, some intermediate is generated during catalysis. CYP2C8 exhibits experiments were performed to assess its metabolic stability a relatively large substrate-binding cavity ¤ca 1440 !3¥ in the presence of HLMs and NADPH. The HPLC profile of comparable to that of CYP3A4 ¤ca 1390 !3¥ and slightly authentic diosmetin proved to be identical to that obtained smaller than that of CYP2C9 ¤ca 1670 !3¥.41¥ Therefore, after analysis of the incubation mixtures containing diosmetin CYP2C8 can accommodate several large substrates and plus HLMs and NADPH ¤chromatograms not shown¥. inhibitors, such as paclitaxel ¤MW: 853.9¥ and montelukast Moreover, the chromatographic peak area of diosmetin ¤MW: 586.2¥, respectively. Since in the present study was virtually identical after the flavonoid was incubated in paclitaxel was used as the prototypical substrate of CYP2C8 the presence or absence of HLMs, indicating that, in our and its 6Æ-hydroxylation was utilized as the marker reaction

the 3$ and/or the 4$ position of quercetin and diosmetin, or by the carbonyl group present in the flavonol or flavone moiety of quercetin or diosmetin, respectively. These stabilizing interactions may guarantee an adequate perma- nence time of these two tiny ligands inside the large catalytic site of CYP2C8. Looking at the most energetically stable conformations ¤shown in Fig. 6¥, both flavonoids were oriented with the hydroxyl-phenyl moieties at the 2-position ¤see Fig. 1¥ exposed to the solvent in the large active site cavity. Moreover, a 180ô rotation about the axis perpendic- ular to the heme plane is observed in the orientation of the flavonol moiety of quercetin with respect to the flavone moiety of diosmetin ¤Fig. 6 C and B, respectively¥. In this specific binding pose, the 7-OH group of quercetin is close to the heme group at a coordination distance to iron. By Fig. 6. Most favorable binding poses of paclitaxel (A), diosmetin contrast, in the flipped orientation of diosmetin, the 5-OH (B) and quercetin (C) in the substrate binding site of CYP2C8, group is in proximity distance to the heme-iron. However, it obtained by molecular docking simulations is worth underlining that both orientations were collected for either flavonoid, indicating that a stabilizing flip-driven mechanics can be operating during the recognition process of of CYP2C8 activity, a molecular docking study was carried these flavonoids. out using the MOE-Dock program31¥ to better understand Discussion the hypothetical structure of the CYP2C8-paclitaxel complex. Figure 6¤A¥ depicts the energetically more stable In a recent study we demonstrated that diosmetin and binding mode of paclitaxel in the CYP2C8 heme active site. hesperetin, the absorbable aglycones of the flavonoids Interestingly, the paclitaxel molecule complemented the size diosmin and hesperidin, respectively,16,17¥ inhibit in a and shape of the active-site cavity, and no major changes in concentration-dependent manner the drug-metabolizing the tertiary structure were requested compared with the activity of human CYP2C9.19¥ Since CYP2C9 shares more montelukast-bound structure 2NNI. In accordance with the than 80% amino acid sequence identity with CYP2C8,20¥ the above-mentioned experimental evidence, the C6 position of present study was specifically designed to evaluate in vitro the tetracyclic 17-carbon ¤heptadecane¥ skeleton of paclitaxel the possible inhibitory effects of diosmetin and hesperetin on was positioned close to the heme at a distance of 6.2 ! from CYP2C8-mediated hepatic drug metabolism, a biochemical the catalytic iron, this conformation being compatible with effect of potential clinical relevance since these flavonoids are its 6Æ-hydroxylation. present in widely used pharmaceutical preparations and since To corroborate the experimental evidence that diosmetin CYP2C8 metabolizes various important drugs, including acts as a mixed competitive-noncompetitive inhibitor of amiodarone, cerivastatin, loperamide, paclitaxel, pioglita- CYP2C8 activity, molecular docking studies were carried zone, rosiglitazone and repaglinide.22¥ To characterize the out with the flavonoid. Docking simulation was also mechanism of the inhibitory effect, we used HLMs rather performed in comparison with quercetin, due to its well- than cDNA-expressed CYP2C8 for two main reasons: ¤1¥ documented inhibitory effect on CYP2C8 activity.23,24¥ the inhibitory potency of a compound may be overestimated Docking simulations showed that both quercetin and by the use of certain recombinant enzymes, which confounds diosmetin can be easily accommodated in the large catalytic the prediction of potential drug®drug interactions.42¥ For binding cleft, adopting several energetically feasible con- example, it has been demonstrated that, in comparison with fi formations. This fact may be explained by considering the HLMs, Supersomes exhibit signi cantly lower Ki values huge difference between the volume of the CYP2C8 ligand- for nine inhibitors assessed towards CYP2C9-mediated binding cavity and the corresponding occupancy volumes of ¤S¥-flurbiprofen metabolism;42¥ ¤2¥ HLMs úappear the most diosmetin and quercetin. However, docking simulations applicable in vitro system ¤for the purpose of scaling¥, suggest that most conformations of both quercetin and since the various CYPs are present in proportion to their diosmetin are stabilized by a network of hydrophobic in vivo representation.û 43¥ Although diosmetin can interact interactions including Phe205, Leu208, Ile213, Ile476 with other CYPs, such as CYP1A,6¥ CYP2C919¥ and and Val477. The simulations also suggest that in some CYP3A4/5,18¥ this approach is methodologically correct, conformations a hydrogen bonding interaction can take since CYP2C8 is the only CYP form catalyzing paclitaxel 6- place either with the back bond of Pro211 or with the side Æ-hydroxylation in humans.25¥ chain of Thr301. In these cases, the above-mentioned H- We first evaluated, under rigorously controlled initial-rate bond interactions are mediated by the phenolic group¤s¥ at conditions, the kinetics of CYP2C8-catalyzed paclitaxel

Copyright © 2011 by the Japanese Society for the Study of Xenobiotics (JSSX) 566 Luigi QUINTIERI, et al. metabolism by HLMs. Our results confirm previous 2. As previously observed for CYP2C9,19¥ diosmetin is a observations24®27,44,45¥ that untransformed initial velocity full inhibitor of CYP2C8, whereas it was shown to data of 6Æ-OH-paclitaxel formation are best-fitted by using cause only partial inhibition of CYP3A4/5.18¥ More- the one-site Michaelis-Menten equation, indicating the over, it exhibits similar inhibitory potency towards Æ ¤ © + + involvement of a single enzyme in the 6 -hydroxylation of CYP2C8 and CYP2C9 Ki 3.13 0.11 and 1.71 paclitaxel. The values of the kinetic parameters are also 0.58 µM,19¥ respectively¥, whereas it is a much less ¤ © + ¥ 18¥ similar to those previously reported by various au- effective inhibitor of CYP3A4 Ki 37.0 8.1 µM . thors.25,27,44,45¥ An additional observation of methodological This is likely explained by the high degree of amino interest that emerged from our study is that DMF, the acid sequence identity of the two proteins belonging to solvent used to dissolve diosmetin and hesperetin, does not the CYP2C subfamily.20¥ significantly modify the kinetic parameters of the reaction. 3. Diosmetin is not metabolized upon incubation with Since we are not aware of any report investigating the effect HLMs, indicating that it acts as a dead-end inhibitor, of DMF on CYP2C8 activity, our results provide the first not as an alternative substrate for CYP2C8. Some information about the suitability of 0.5% DMF to solubilize in vitro studies have demonstrated a possible oxidative non-hydrophilic compounds that are to be added to the demethylation, by microsomal preparations isolated incubation medium for assay of CYP2C8 activity. from different animal species, of those flavonoids in The main findings concerning the inhibition of 6Æ- which the methoxy group is at the 4$-position of the hydroxylation of paclitaxel by diosmetin and hesperetin B-ring,39,47,48¥ as in the case of diosmetin ¤Fig. 1¥. may be summarized as follows: However, literature data clearly indicate that the 1. The formation rate of 6Æ-OH-paclitaxel, the major extent of CYP-mediated oxidation of flavonoids is metabolite of paclitaxel in humans,25,26¥ is decreased in negligible ¤g6%¥ compared to that of conjugation a concentration-dependent manner by both diosmetin reactions.39,49¥ Moreover, oxidative demethylation of and hesperetin, but diosmetin is about 16 times more diosmetin has been observed only in the presence of potent than hesperetin as an inhibitor of CYP2C8- recombinant human CYP1A1 and CYP1A2 but not in catalyzed paclitaxel metabolism. This difference in the presence of CYP2C8.50¥ Thus, previous observa- inhibitory potency between diosmetin and hesperetin tions and our data warrant that the interpretation of with respect to CYP2C8 activity resembles the rank our kinetic results, based on the theory describing order of potency of the two flavonoids as inhibitors of dead-end inhibition, is correct.37¥ CYP2C9 activity, since diosmetin is about 12-times 4. Kinetic analysis shows that diosmetin behaves as more potent than hesperetin as an inhibitor of a mixed ¤competitive-noncompetitive¥-type inhibitor CYP2C9-mediated diclofenac metabolism.19¥ Because of the 6Æ-hydroxylation of paclitaxel. Interestingly, the only structural difference between diosmetin and mixed-type inhibition of CYP2C8 activity has been hesperetin is the absence of the C2-C3 double bond in observed by Václavíková et al.24¥ for the flavonols the latter flavonoid ¤see Fig. 1¥, this functionality fisetin, morin and quercetin. Our kinetic results are seems to play a critical role for the inhibition of both supported by molecular docking simulations showing CYP2C8 and CYP2C9 activity. These results are in that diosmetin accommodates in the large cavity of the accordance with the findings of Václavíková et al.,24¥ CYP2C8 catalytic site with the 5-OH group of its who reported that the flavonols morin, quercetin and chromone moiety close to the heme group. This fisetin ¤in which the C2-C3 double bond is present¥ conformation is compatible with a competition toward inhibited significantly CYP2C8-mediated paclitaxel paclitaxel for binding to the active site of CYP2C8 and metabolism, whereas the flavanone naringenin, which explains the competitive component of the mixed-type lacks this double bond, is a poor inhibitor. This inhibition caused by diosmetin. Similar conclusions can structural requirement for the inhibition of CYP2C8 be drawn for quercetin, which accommodates in the activity by flavonoids also resembles that for the cavity of CYP2C8, although with a different preferred inhibition of CYP1A activity,6,46¥ since flavones orientation. An intrinsic limitation of our docking containing the C2-C3 double bond are more potent simulations is the uncertainty regarding how many CYP1A inhibitors than flavanones lacking this specific water molecules fill the distal empty portion of the functionality.6¥ ligand binding cavity. Lack of this information can Considering the low bioavailability of orally adminis- drastically condition the conformational sampling of tered flavonoids, whose peak plasma concentrations are the ligands and also the calculation of the potential in the low micromolar range,39,40¥ hesperetin concen- energy values of the binding process. ¤ + ¥ trations in the range of its IC50 value 68.5 3.3 µM 5. The Ki value for diosmetin is similar to those reported are unlikely to be reached in vivo. Therefore, in the for fisetin, morin and quercetin by Václavíková et al.24¥ 23¥ present study we characterized the inhibitory effect of and for quercetin by Bun et al. This Ki value of diosmetin only. diosmetin makes its pharmacokinetic interactions with

drugs metabolized by CYP2C8 rather likely in vivo, 4¥ Rice-Evans, C. A., Miller, N. J. and Paganga, G.: Structure® fl since plasma concentrations in the low micromolar antioxidant activity relationships of avonoids and phenolic acids. ¤ + ¥ Free Radic. Biol. Med., 20: 933®956 ¤1996¥. range Cmax 1.40 0.31 µM have been observed after 5¥ DöArchivio, M., Filesi, C., Varì, R., Scazzocchio, B. and Masella, single-dose administration ¤10 mg/kg¥ of its natural Int. J. ¥ R.: Bioavailability of polyphenols: Status and controversies. rutinoside diosmin.16 Considering the high recom- Mol. Sci., 11: 1321®1342 ¤2010¥. mended dosages ¤up to six 450 mg-doses a day¥ at 6¥ Doostdar, H., Burke, M. D. and Mayer, R. T.: Bioflavonoids: selective substrates and inhibitors for cytochrome P450 CYP1A and which diosmin is prescribed and the long in vivo ® ¤ ¥ ¤ ¥ 16¥ CYP1B1. Toxicology, 144:31 38 2000 . elimination half-life of diosmetin 31 h , a consid- 7¥ Moon, Y. J., Wang, X. and Morris, M. E.: Dietary flavonoids: erable accumulation is to be expected upon repeated effects on xenobiotic and carcinogen metabolism. Toxicol. In Vitro, doses and, consequently, plasma concentrations sig- 20: 187®210 ¤2006¥. nificantly higher than those observed after single 8¥ Androutsopoulos, V. P., Papakyriakou, A., Vourloumis, D., fl administration. Plasma concentrations in the nano- Tsatsakis, A. M. and Spandidos, D. A.: Dietary avonoids in cancer therapy and prevention: substrates and inhibitors of cytochrome molar range ¤median 48, range 2®132 nM¥ have been ® ¤ ¥ ¤ P450 CYP1 enzymes. Pharmacol. Ther., 126:9 20 2010 . observed after repeated administration 500 mg 3 9¥ Lyseng-Williamson, K. A. and Perry, C. M.: Micronised purified times daily for 7 days¥ of the structurally allied flavonoid fraction: a review of its use in chronic venous flavonoid quercetin.51¥ This may explain why no insufficiency, venous ulcers and haemorrhoids. Drugs, 63:71®100 ¤ ¥ pharmacokinetic interaction with the CYP2C8 sub- 2003 . 10¥ Bailey, D. G. and Dresser, G. K.: Interactions between grapefruit strate rosiglitazone has been observed in healthy juice and cardiovascular drugs. Am. J. Cardiovasc. Drugs, 4: 281®297 volunteers after repeated administration ¤500 mg/day ¤ ¥ ¥ 2004 . for 3 weeks¥ of quercetin.52 This large difference in the 11¥ Peng, W. X., Li, H. D. and Zhou, H. H.: Effect of daidzein on observed plasma concentrations of diosmetin and CYP1A2 activity and pharmacokinetics of theophylline in healthy ® ¤ ¥ quercetin is most likely because, unlike diosmetin, volunteers. Eur. J. Clin. Pharmacol., 59: 237 241 2003 . ¤ ¥ 12¥ Rajnarayana, K., Reddy, M. S. and Krishna, D. R.: Diosmin which is only administered as its rutinoside diosmin , pretreatment affects bioavailability of metronidazole. Eur. J. Clin. quercetin was administered as an aglycone, in the cited Pharmacol. ® ¤ ¥ ¥ , 58: 803 807 2003 . pharmacokinetic studies.51,52 Monoglucosides and 13¥ Choi, J. S., Choi, B. C. and Choi, K. E.: Effect of quercetin on the aglycones of flavonoids are taken up by enterocytes pharmacokinetics of oral cyclosporine. Am. J. Health Syst. Pharm., ® ¤ ¥ where they are extensively metabolized by conjugation 61: 2406 2409 2004 . ¥ 14¥ Rajnarayana, K., Venkatesham, A. and Krishna, D. R.: Bioavail- with glucuronic acid and/or sulphate.5,39,53 As a ¤ ability of diclofenac sodium after pretreatment with diosmin in rutinoside i.e., conjugated with a disaccharide; see healthy volunteers. Drug Metabol. Drug Interact., 22: 165®174 Introduction¥, diosmin is not absorbed through the ¤2007¥. mucosa of the small intestine, but reaches the colon 15¥ Rajnarayana, K., Venkatesham, A., Nagulu, M., Srinivas, M. and fl and cecum unmodified, where it is absorbed after Krishna, D. R.: In uence of diosmin pretreatment on the pharmacokinetics of chlorzoxazone in healthy male volunteers. hydrolysis to its aglycone diosmetin by rhamnosidases Drug Metabol. Drug Interact., 23: 311®322 ¤2008¥. 16,17¥ of Enterobacteriaceae. In these distal regions of the 16¥ Cova, D., De Angelis, L., Giavarini, F., Palladini, G. and Perego, intestine, the activity of phase II drug-metabolizing R.: Pharmacokinetics and metabolism of oral diosmin in healthy enzymes is negligible.54¥ volunteers. Int. J. Clin. Pharmacol. Ther. Toxicol., 30:29®33 ¤1992¥. ¥ In conclusion, this study provides the first demonstration 17 Serra, H., Mendes, T., Bronze, M. R. and Simplicio, A. L.: Prediction of intestinal absorption and metabolism of pharmaco- that diosmetin, the absorbable aglycone of diosmin, is a logically active flavones and flavanones. Bioorg. Med. Chem., 16: potent inhibitor of CYP2C8 at concentrations similar to 4009®4018 ¤2008¥. those reached in vivo after administration of diosmin- 18¥ Quintieri, L., Palatini, P., Nassi, A., Ruzza, P. and Floreani, M.: containing pharmaceutical preparations.16¥ This finding may Flavonoids diosmetin and luteolin inhibit midazolam metabolism by have important clinical implications, since CYP2C8 metab- human liver microsomes and recombinant CYP 3A4 and CYP3A5 enzymes. Biochem. Pharmacol., 75: 1426®1437 ¤2008¥. olizes various important drugs. 19¥ Quintieri, L., Bortolozzo, S., Stragliotto, S., Moro, S., Pavanetto, M., Nassi, A., Palatini, P. and Floreani, M.: Flavonoids diosmetin References and hesperetin are potent inhibitors of cytochrome P450 2C9- mediated drug metabolism in vitro. Drug Metab. Pharmacokinet., 25: 1¥ Middleton, E., Jr., Kandaswami, C. and Theoharides, T. C.: The 466®476 ¤2010¥. effects of plant flavonoids on mammalian cells: implications for 20¥ Ridderström, M., Zamora, I., Fjellström, O. and Andersson, inflammation, heart disease, and cancer. Pharmacol. Rev., 52: 673® T. B.: Analysis of selective regions in the active sites of human 751 ¤2000¥. cytochromes P450, 2C8, 2C9, 2C18, and 2C19 homology models 2¥ Havsteen, B. H.: The biochemistry and medical significance of the using GRID/CPCA. J. Med. Chem., 44: 4072®4081 ¤2001¥. flavonoids. Pharmacol. Ther., 96:67®202 ¤2002¥. 21¥ Rendic, S. and Carlo, F. J.: Human cytochrome P450 enzymes: a 3¥ Benavente-García, O. and Castillo, J.: Update on uses and status report summarizing their reactions, substrates, inducers, and properties of citrus flavonoids: new findings in anticancer, inhibitors. Drug Metab. Rev., 29: 413®580 ¤1997¥. cardiovascular, and anti-inflammatory activity. J. Agric. Food Chem., 22¥ Totah, R. A. and Rettie, A. E.: Cytochrome P450 2C8: substrates, 56: 6185®6205 ¤2008¥. inhibitors, pharmacogenetics, and clinical relevance. Clin. Pharmacol.

Ther., 77: 341®352 ¤2005¥. ¤2001¥. 23¥ Bun, S. S., Ciccolini, J., Bu, H., Aubert, C. and Catalin, J.: Drug 39¥ Cermak, R. and Wolffram, S.: The potential of flavonoids to interactions of paclitaxel metabolism in human liver microsomes. influence drug metabolism and pharmacokinetics by local gastro- J. Chemother., 15: 266®274 ¤2003¥. intestinal mechanisms. Curr. Drug Metab., 7: 729®744 ¤2006¥. 24¥ Václavíková, R., Horský, S., Šimek, P. and Gut, I.: Paclitaxel 40¥ Manach, C., Williamson, G., Morand, C., Scalbert, A. and metabolism in rat and human liver microsomes is inhibited by Rémésy, C.: Bioavailability and bioefficacy of polyphenols in phenolic antioxidants. Naunyn Schmiedebergs Arch. Pharmacol., 368: humans. I. Review of 97 bioavailability studies. Am. J. Clin. Nutr., 200®209 ¤2003¥. 81¤1 Suppl.¥: 230S®242S ¤2005¥. 25¥ Rahman, A., Korzekwa, K. R., Grogan, J., Gonzalez, F. J. and 41¥ Lai, X. S., Yang, L. P., Li, X. T., Liu, J. P., Zhou, Z. W. and Harris, J. W.: Selective biotransformation of taxol to 6Æ- Zhou, S. F.: Human CYP2C8: structure, substrate specificity, hydroxytaxol by human cytochrome P450 2C8. Cancer Res., 54: inhibitor selectivity, inducers and polymorphisms. Curr. Drug 5543®5546 ¤1994¥. Metab., 10: 1009®1047 ¤2009¥. 26¥ Cresteil, T., Monsarrat, B., Dubois, J., Sonnier, M., Alvinerie, P. 42¥ Kumar, V., Rock, D. A., Warren, C. J., Tracy, T. S. and and Gueritte, F.: Regioselective metabolism of taxoids by human Wahlstrom, J. L.: Enzyme source effects on CYP2C9 kinetics and CYP3A4 and 2C8: structure-activity relationship. Drug Metab. inhibition. Drug Metab. Dispos., 34: 1903®1908 ¤2006¥. Dispos., 30: 438®445 ¤2002¥. 43¥ von Moltke, L. L., Greenblatt, D. J., Schmider, J., Wright, C. E., 27¥ Vuppugalla, R., Chang, S. Y., Zhang, H., Marathe, P. H. and Harmatz, J. S. and Shader, R. I.: In vitro approaches to predicting Rodrigues, D. A.: Effect of commonly used organic solvents on the drug interactions in vivo. Biochem. Pharmacol., 55: 113®122 ¤1998¥. kinetics of cytochrome P450 2B6- and 2C8-dependent activity in 44¥ Harris, J. W., Katki, A., Anderson, L. W., Chmurny, G. N., human liver microsomes. Drug Metab. Dispos., 35: 1990®1995 Paukstelis, J. V. and Collins, J. M.: Isolation, structural ¤2007¥. determination, and biological activity of 6 alpha-hydroxytaxol, 28¥ Palatini, P.: A simple method for discriminating between full and the principal human metabolite of Taxol. J. Med. Chem., 37: 706® partial inhibitors. Biochem. Int., 16: 359®368 ¤1988¥. 709 ¤1994¥. 29¥ Copeland, R. A.: Enzymes. A Practical Introduction to Structure, 45¥ Ong, C. E., Coulter, S., Birkett, D. J., Bhasker, C. R. and Miners, Mechanism and Data Analysis, 2nd Edition. New York, John Wiley J. O.: The xenobiotic inhibitor profile of cytochrome P4502C8. Br. and Sons, 2000, pp. 397. J. Clin. Pharmacol., 50: 573®580 ¤2000¥. 30¥ Schoch, G. A., Yano, J. K., Sansen, S., Dansette, P. M., Stout, 46¥ Moon, J. Y., Lee, D. W. and Park, K. H.: Inhibition of 7- C. D. and Johnson, E. F.: Determinants of cytochrome P450 2C8 ethoxycoumarin O-deethylase activity in rat liver microsomes by substrate binding: structures of complexes with montelukast, naturally occurring flavonoids: structure-activity relationships. troglitazone, felodipine, and 9-cis-retinoic acid. J. Biol. Chem., 283: Xenobiotica, 28: 117®125 ¤1998¥. 17227®17237 ¤2008¥. 47¥ Nielsen, S. E., Breinholt, V., Justesen, U., Cornett, C. and 31¥ MOE ¤Molecular Operating Environment¥, version 2010.11; Dragsted, L. O.: In vitro biotransformation of flavonoids by rat software available from Chemical Computing Group Inc. ¤1010 liver microsomes. Xenobiotica, 28: 389®401 ¤1998¥. Sherbrooke Street West, Suite 910, Montreal, Quebec, Canada 48¥ Breinholt, V. M., Offord, E. A., Brouwer, C., Nielsen, S. E., H3A 2R7¥; http://www.chemcomp.com ¤last accessed date: April Brøsen, K. and Friedberg, T.: In vitro investigation of cytochrome 28, 2011¥. P450-mediated metabolism of dietary flavonoids. Food Chem. 32¥ Cornell, W. D., Cieplak, P., Bayly, C. I., Gould, I. R., Merz, Toxicol., 40: 609®616 ¤2002¥. K. M., Ferguson, D. M., Spellmeyer, D. C., Fox, T., Caldwell, 49¥ Otake, Y., Hsieh, F. and Walle, T.: Glucuronidation versus J. W. and Kollman, P. A.: A second generation force field for the oxidation of the flavonoid galangin by human liver microsomes and simulation of proteins, nucleic acids and organic molecules. J. Am. hepatocytes. Drug Metab. Dispos., 30: 576®581 ¤2002¥. Chem. Soc., 117: 5179®5197 ¤1995¥. 50¥ Androutsopoulos, V., Wilsher, N., Arroo, R. R. J. and Potter, 33¥ Baxter, C. A., Murray, C. W., Clark, D. E., Westhead, D. R. and G. A.: Bioactivation of the phytoestrogen diosmetin by CYP1 Eldridge, M. D.: Flexible docking using Tabù search and an cytochromes P450. Cancer Lett., 274:54®60 ¤2009¥. empirical estimate of binding affinity. Proteins: Struct., Funct., Bioinf., 51¥ Moon, Y. J., Wang, L., DiCenzo, R. and Morris, M. E.: Quercetin 33: 367®382 ¤1998¥. pharmacokinetics in humans. Biopharm. Drug Dispos., 29: 205®217 34¥ Halgren, T.: Merck Molecular Force Field. I. Basis, form, scope, ¤2008¥. parameterization, and performance of MMFF94. J. Comput. Chem., 52¥ Kim, K. A., Park, P. W., Kim, H. K., Ha, J. M. and Park, J. Y.: 17: 490®519 ¤1996¥. Effect of quercetin on the pharmacokinetics of rosiglitazone, a 35¥ MOPAC 7 ¤Molecular Orbital PACkage¥ Stewart, J. J. P.; Fujitsu CYP2C8 substrate, in healthy subjects. J. Clin. Pharmacol., 45: 941® Limited: Tokyo, Japan, 1993. http://openmopac.net ¤last accessed: 946 ¤2005¥. April 28, 2011¥. 53¥ Hollman, P. C., Bijsman, M. N., van Gameren, Y., Cnossen, 36¥ Wojciechowski, M. and Lesyng, B.: Generalized Born model: E. P., de Vries, J. H. and Katan, M. B.: The sugar moiety is a Analysis, refinement, and applications to proteins. J. Phys. Chem. B, major determinant of the absorption of dietary flavonoid glycosides 108: 18368®18376 ¤2004¥. in man. Free Radic. Res., 31: 569®573 ¤1999¥. 37¥ Segel, J. H.: Enzyme Kinetics, John Wiley & Sons, 1975, p. 957. 54¥ Peters, W. H., Kock, L., Nagengast, F. M. and Kremers, P. G.: 38¥ Venkatakrishnan, K., Von Moltke, L. L. and Greenblatt, D. J.: Biotransformation enzymes in human intestine: critical low levels in Human drug metabolism and the cytochromes P450: application the colon? Gut, 32: 408®412 ¤1991¥. and relevance of in vitro models. J. Clin. Pharmacol., 41: 1149®1179