Abstract
Silybin (a flavonolignan, the main component of silymarin, an extract from the seeds of Silybum marianum) has been used to date mostly as a hepatoprotectant. However, it also has other interesting activities, e.g., anticancer and hypocholesterolemic effects. It is also known that silybin can inhibit the activities of the cytochrome P450 (P450) enzymes. In this study, a weak interaction of silybin with human microsomal CYP2E1, 2A6, 2B6, 2C19, and 2D6 (IC50 ≥ 250 μM) was found; a moderate inhibition was observed for CYP1A2 and 2C8. The most prominent inhibition effect was found with CYP3A4 and CYP2C9 (IC50 ≤ 50 μM). Using mass spectometry detection, production of O-demethylated (the main metabolite) as well as hydroxylated derivatives of silybin formed by P450 enzymes was detected. The effect of different P450 inhibitors on the formation of O-demethylated product was also studied. In particular, a relatively specific inhibitor of CYP2C8 (quercetin) markedly inhibited the formation of this metabolite. With the help of recombinant enzymes (bactosomes), it was confirmed that the CYP2C8 enzyme is responsible for the reaction leading to O-demethylated silybin.
Silybin (also known as silibinin, CAS 22888-70-6) is a flavonolignan and an active component of silymarin, an extract from Silybum marianum (milk thistle) seeds (Gazák et al., 2007). Its hepatoprotective effects have been known for hundreds of years; novel studies indicate that the molecular basis of this effect is its antioxidative and radical scavenging property (Flora et al., 1998; Gazák et al., 2007). Recent discoveries of its other activities (chemopreventive, anticancer, and neuroprotective effects) are responsible for an increasing number of articles in peer-reviewed journals devoted to this natural compound. An interference of silybin with cell cycle-regulating pathways is expected to be the mechanism underlying the majority of the effects described (Singh and Agarwal, 2006).
Although silybin (Fig. 1) is generally considered to be safe with only few adverse effects (involving mostly gastrointestinal discomfort), a possibility of drug interactions based on metabolism mediated by cytochromes P450 (P450) was recently investigated. Nifedipine oxidation, one of the major CYP3A4 activities, was found to be inhibited by silybin in the micromolar range (Beckmann-Knopp et al., 2000; Zuber et al., 2002; Sridar et al., 2004). Also, CYP2D6 and CYP2C9 as well as glucuronidation activities were found to be inhibited. These facts have led to studies on the potential influence of silybin on pharmacokinetics of a typical CYP3A4 substrate, indinavir (DiCenzo et al., 2003; Mills et al., 2005). The differences in the area under the curve of indinavir due to concomitant application of silymarin (i.e., the silybin-containing extract) were found not to be significant. The possibility of a direct interaction of silybin with P450 enzymes and its conversion to metabolites by human liver was further studied; the presence of O-demethylated and hydroxylated derivatives of silybin was detected by HPLC-electrospray ionization-ion trap mass spectrometry (Gunaratna and Zhang, 2003). However, no systematic study on the metabolism of silybin by individual P450 enzymes with an identification of a particular P450 form responsible for the formation of metabolite(s) is available.
This work is a detailed investigation of the role of individual human liver microsomal P450 enzymes in the metabolism of silybin 1) by inhibition of prototypical microsomal P450 activities by silybin, 2) by inhibition of silybin metabolism by known substrates or inhibitors of specific P450 forms, and finally 3) by confirmation of the involvement of CYP2C8 in the formation of O-demethylated product, the main metabolite of silybin.
Materials and Methods
Chemicals. Silybin was a gift from Ivax-CR a.s. (Opava, Czech Republic). Chlorzoxazone, 6-hydroxychlorzoxazone, diclofenac, 4-hydroxydiclofenac, bufuralol, 6-hydroxybufuralol, and 6β-hydroxytestosterone were supplied by Ultrafine Chemicals (Salford, UK). P450-Glo substrate for determination of CYP2C8 activities by luminescence spectrometry was the product of Promega (Madison, WI) obtained through East Port (Prague, Czech Republic). 7-Ethoxy-4-(trifluoromethyl)coumarin was supplied by Fluka (Buchs, Switzerland). All other chemicals were purchased from Sigma-Aldrich CZ (Prague, Czech Republic).
Microsomes and Recombinant Enzymes. Pooled human liver microsomes were purchased from Advancell (Barcelona, Spain). Microsomes were obtained in accordance with ethical regulations of the country of origin (Spain). They were from 10 donors (5 males and 5 females) with a protein content of 38.4 mg/ml; CYP1A2, CYP2A6, CYP2B6, CYP2C9, CYP2E1, and CYP3A4 enzyme activities are accessible at the Advancell website (http:/www.advancell.net, batch reference no. 102091201). Bactosomes (bacterial membrane fractions from Escherichia coli) containing recombinant human cytochromes P450 enzymes (CYP1A2, 2C8, 2C9, 3A4, 2D6, 2B6, and 2E1) coexpressed with human NADPH-cytochrome P450 reductase, were purchased from Cypex (Dundee, UK).
Structure of silybin.
Effect of silybin on specific activities of cytochromes P450. Concentrations of silybin in reaction mixture were 0, 10, 50, 100, 150, 200, and 400 μM. Experiments were performed in duplicate with results expressed as means; as a rule, the data obtained did not differ more than 5%.
P450 Activities. Individual P450 activities were measured according to established protocols. The following microsomal P450 activities were tested: CYP3A4, testosterone 6β-hydroxylation (Guengerich et al.,1986); CYP2C9, diclofenac 4′-hydroxylation (Crespi et al., 1998a); CYP2E1, chlorzoxazone 6-hydroxylation (Lucas et al., 1996); CYP1A2, 7-ethoxyresorufin O-deethylation (Chang and Waxman, 1998); CYP2D6, bufuralol 1′-hydroxylation (Crespi et al., 1998b); CYP2A6, coumarin 7-hydroxylation (Waxman and Chang, 1998); CYP2B6, 7-ethoxy-4-(trifluoromethyl)coumarin O-deethylation (Donato et al., 2004); CYP2C19, S-mephenytoin 4′-hydroxylation (http://www.cypex.co.uk/intro.htm, Cypex 2C19 QC assays); and CYP2C8, luciferinmethyl ester demethylation (Promega Technical Bulletin no.325, http://www.promega.com). For determination of metabolites formed from specific substrates, an HPLC system (Class VP; Shimadzu, Kyoto, Japan) with UV (6β-hydroxytestosterone, 6-hydroxychlorzoxazone, 4′-hydroxydiclofenac, and 4′-hydroxymephenytoin) or with fluorescence detection (1′-hydroxybufuralol) was used. A Tecan GENios absorbance/fluorescence/luminescence reader (Tecan, Vienna, Austria) was used for detection of other metabolites [7-ethoxyresorufin, 7-hydroxycoumarin, and 7-ethoxy-4-(trifluoromethyl)coumarin, luciferin].
Inhibition of P450 Enzymes by Silybin in Microsomal Fractions. First, for each enzyme assay the Km and Vmax values were determined to get the substrate concentration suitable for inhibition experiments. The Km values corresponded well to known literature data (e.g., Walsky and Obach, 2004). Substrate concentrations were used near the Km (Table 1). Inhibition experiments were performed with six concentration levels of silybin (10, 50, 100, 150, 200, and 400 μM); the stock solution was 25 mM in 60% (v/v) dimethyl sulfoxide, except for measurement of activities of CYP2E1 and 2C19 for which the stock solution contained 17.3 mM silybin in acetonitrile. Experimental conditions were the same as for determination of individual P450 activities; preincubation of reaction mixtures with inhibitor (silybin) for 30 min was kept in all determinations. Inhibition of individual P450 activities was in all cases evaluated by plotting respective remaining activity against the inhibitor concentration. When an inhibition of a particular P450 activity was pronounced, the Ki values were determined as averages from Dixon plots with three substrate concentrations used (corresponding to 0.5 Km, Km, and 2Km). Parameters of the enzyme kinetics (Km and Vmax) as well the IC50 and intercept values (for Ki determination) were obtained using the Sigma Plot 8.0.2 scientific graphing software (SPSS, Chicago, IL).
Enzyme activities of microsomal P450 enzymes
Inhibition by silybin expressed as IC50 or Ki ± S.E.M.
Inhibition of Metabolite Formation by Carbon Monoxide. The incubation was performed in 0.05 M potassium phosphate buffer (pH 7.4) containing an NADPH-generating system (0.5 mM NADP, 3.7 mM citric acid, and 0.5 unit/ml isocitric acid dehydrogenase), 5 mM MgCl2, and 250 pmol of microsomal CYP, in a total volume of 1 ml. After 10 min of preincubation at 37°C, carbon monoxide (Linde Technoplyn, Prague, CZ) was applied by gentle bubbling through the reaction mixture for 1 min. Then silybin was added as 20 μl of a 5 mM stock solution in 60% (v/v) dimethyl sulfoxide giving the final concentration of 100 μM, and the reaction mixture was incubated for 30 min in a shaking water bath. The reaction was stopped by the addition of 2 ml of ethyl acetate. The next procedure was performed in accordance with the method of Gunaratna and Zhang (2003). After vigorous stirring and centrifugation (5000g for 10 min), 1 ml of supernatant was evaporated under a stream of nitrogen, and samples were dissolved in 100 μl of the mobile phase (acetonitrile, methanol, and 0.1% formic acid, 27:5:68) and transferred into vials of the autosampler. Formation of metabolites was followed by high-performance liquid chromatography according to the original work (Gunaratna and Zhang, 2003).
Inhibition of Microsomal Silybin Metabolism by Specific Inhibitors of P450 Forms. To examine the effects of different P450 inhibitors on the metabolism of silybin in human microsomes, inhibition of individual P450 activities was studied. Furafylline (22.5 μM) was used as a specific inhibitor of CYP1A2, sulfaphenazole (3 μM) was used for CYP2C9, triacetyloleandomycin (3 μM) was used for CYP3A4, diethyldithiocarbamate (75 μM) was used for CYP2E1, quinidine (0.3 μM) was used for CYP2D6, 8-methoxypsoralen (1 μM) was used for CYP2A6, 7-pentoxyresorufin (2 μM) was used for CYP2B6, S-mephenytoin (160 μM) was used for CYP2C19, and an inhibitor of CYP2C8, quercetin (100 μM), was also added to the reaction mixture (Baldwin et al., 1995; Newton et al., 1995; Dierks et al., 2001; Goldstein et al., 1994.). Every inhibitor used was preincubated with the reaction mixture (see the preceding paragraph) containing human liver microsomes for 30 min, and then the silybin (50 μM) was added.
Identification of P450 Enzyme Involved in Metabolite Formation in Bactosomes. The metabolism of silybin by CYP1A2, CYP2B6, CYP2C8, CYP2C9, CYP3A4, CYP2D6, and CYP2E1 was determined using bactosomes containing individual CYP enzymes. The experiments were done according to the supplier's recommendation (Cypex). The incubation was performed in 0.1 M Tris-HCl buffer (pH 7.4) containing 5 pmol of P450 enzyme (together with P450 reductase), 5 mM MgCl2, an NADPH-generating system (0.5 mM NADP, 3.7 mM citric acid, and 0.5 unit/ml isocitric acid dehydrogenase), and 25 or 50 μM silybin. The reaction was terminated by the addition of 1 ml of ethyl acetate. Samples were centrifuged at 5000g for 10 min to take out the precipitated protein. The supernatant was evaporated under a stream of nitrogen.
Identification of Silybin Metabolites by μLC/MS. Residue obtained after incubation with microsomes or with bactosomes was dissolved in acetonitrile (50 μl) and further diluted by 200 μl of mobile phase A (see below). μLC/MS analyses were performed using a CapLC XE system (Waters, Milford, MA), a C18 column Gemini (150 mm × 300 μm i.d.; Phenomenex, Torrance, CA) at a mobile phase flow rate of 6 μl/min. The elution was realized at the isocratic condition with 65% of mobile phase A (5.7 mM acetic acid + 5% acetonitrile) and 35% of mobile phase B (acetonitrile); 1 μl of sample solution was injected using an autosampler. Accurate mass measurement and MS/MS experiments were performed to confirm the identity of metabolites on a quadrupole time of flight mass spectrometer (Waters Micromass Q-Tof Premier Mass Spectrometer). Optimized parameters of electrospray were capillary voltage -2.8 kV (negative mode), sampling cone 45 V, source temperature 80°C, desolvation temperature 180°C, cone gas flow 50 liters/h, and desolvation gas flow 400 liters/h. A collision energy ramp in the range of 5 to 30 eV was used for fragmentation experiments. Data were obtained in a single V mode.
Results
Inhibition of Specific Activities in Human Liver Microsomes by Silybin. The results of in vitro inhibition of nine CYP enzymes by silybin in microsomal fraction are given in Fig. 2. Silybin displayed a weak or no interaction with CYP2E1, 2A6, 2B6, 2C19, and 2D6 (IC50 ≥ 250 μM); a moderate inhibition was observed for CYP1A2 and CYP2C8. The most prominent inhibition effect was found with CYP3A4 and CYP2C9 (IC50 = 49.8 and 34.1 μM) (Table 1).
Identification of Metabolites by μLC/MS Analysis. Incubation of silybin with human liver microsomes confirmed the formation of metabolites of silybin (Gunaratna and Zhang, 2003). The major metabolite was identified by μLC/MS analysis as O-demethylated silybin; the minor ones were silybin mono- and dihydroxy derivatives. Figure 3 shows chromatograms for selected m/z values (silybin m/z = 481; O-demethylated metabolite m/z = 467; monohydroxy m/z = 497 and dihydroxy m/z = 513) and fragmentation spectra of quasimolecular ions of the metabolites mentioned. It is evident that modification of a molecule of silybin occurs in two different positions for the monohydroxy metabolite (two peaks for m/z = 497 with retention times of 4.49 and 5.31 min, respectively). It has to be mentioned that chromatographic peaks in retention times of silybin (6.7-6.8 min) correspond to products of modification of silybin in the ion source. This fact was verified by an analysis of standard (unmodified) silybin, and it underlines the necessity of chromatographic separation. Identity of metabolites was verified by interpretation of MS/MS spectra and by accurate mass measurements. Experimental masses of metabolites were in good agreement with the calculated ones. Their relative errors were 0.6 ppm or less.
Effect of Carbon Monoxide and of the Specific P450 Inhibitors on the Formation of Metabolites of Sylibin. To prove the role of cytochromes P450 in the metabolism of silybin, inhibition of this reaction by carbon monoxide was examined. Results indicated a considerable reduction (by 80%) of the main metabolite level in the reaction mixture, suggesting a role for a P450-mediated mechanism of metabolite formation (Fig. 4).
To find which P450 form is responsible for the formation of silybin metabolite(s), the effect of different inhibitors of particular P450 activities was studied. An overview of inhibitors used together with their effect on the formation of O-demethylated silybin is shown in Fig. 5. Among the inhibitors used, quercetin (a relatively specific inhibitor of CYP2C8) (Walsky et al., 2005) appeared to be the most potent inhibitor, causing 80% inhibition of O-demethylated silybin formation.
Incubation of Silybin with Recombinant Enzymes (Bactosomes). Although the inhibition experiments described in the preceding paragraph indicated a possible involvement of several P450 forms in an interaction with silybin and, hence, a possibility of their ability to form silybin metabolite(s), experiments using bacterial membrane fractions from E. coli containing recombinant human cytochrome P450 enzymes (CYP1A2, 2C8, 2C9, 3A4, 2D6, 2B6, and 2E1) coexpressed with human NADPH-cytochrome P450 reductase (bactosomes) confirmed (Fig. 6) a significant role only for CYP2C8 (and a minor contribution of CYP3A4, results not shown) in the formation of the main metabolite of silybin, i.e., of its O-demethylated derivative by HPLC/MS.
Discussion
The results of the inhibition of prototypic activities of individual microsomal CYP enzymes by silybin revealed that there are at least four P450 activities influenced by the presence of this compound, namely, those of CYP1A2, 2C8, 2C9, and 3A4 (Fig. 2). In the earlier literature, data showing a significant inhibition of CYP3A4 and CYP2C9 were published (Beckmann-Knopp et al., 2000; Zuber et al., 2002; Sridar et al., 2004). The character of inhibition was also studied using standard Dixon plots (results not shown), which indicated the presence of a noncompetitive inhibition of CYP2C9 activity and of a competitive mechanism of inhibition of CYP3A4 activity. When the inhibition was more pronounced, IC50 and Ki values were also determined (Table 1). The results obtained here with the CYP3A4 and CYP2C9 are in agreement with the previous ones; on the other hand, they do not confirm the results of a certain degree of inhibition obtained also with CYP2E1 and CYP2D6 (Beckmann-Knopp et al., 2000; Zuber et al., 2002). Taken together, the results are indicative of an interaction between silybin and at least two microsomal P450 enzymes.
Silybin was shown recently to yield several metabolites when incubated with human liver microsomes (Gunaratna and Zhang, 2003). The main metabolite was found to be the O-demethylated product; the mono- and dihydroxy silybins were the minor metabolites. In this work, the formation of these metabolites in human liver microsomes has been confirmed (Fig. 3); as the next step, we were interested in the identification of the particular P450 form involved in the formation of the main metabolite.
Metabolites of silybin identified by μLC/MS analysis. The major metabolite was identified as O-demethylated silybin; the minor ones were silybin mono and dihydroxy derivatives.
HPLC analysis of silybin metabolites generated by human hepatic microsomes before and after treatment with carbon monoxide. M1 to M3 are metabolites of silybin. mAU, milli-absorbance units.
Effect of selective inhibitors of P450 activities on the rate of formation of demethylated silybin by human hepatic microsomes. Furafylline, sulfaphenazole, triacetyloleandomycin (TAO), diethyldithiocarbamate (DEDC), quinidine, 8-methoxypsoralen, 7-pentoxyresorufin, S-mephenytoin, and quercetin were used to inhibit the respective P450 activities. Experiments were performed in duplicate with results expressed as means; as a rule, the data obtained did not differ more than 5%.
Carbon monoxide is known to bind strongly to the heme iron of all cytochromes P450, yielding a complex that is unable to bind molecular oxygen and perform the catalytic reaction (Cooper et al., 1977). The results (Fig. 4) have shown a clear inhibition of the formation of the O-demethylated silybin; also, the levels of hydroxylated metabolites were diminished. The inhibitors used in the literature to examine the function of a particular P450 enzyme have been subsequently used to find which form of P450 is responsible for the formation of the main silybin metabolite.
HPLC analysis of silybin metabolized by a CYP2C8-bactosomal extract. M1 is the O-demethylated silybin. A blank sample was obtained with the reaction stopped by an immediate addition of ethyl acetate. The results obtained with the other P450-containing bactosomes were similar to those with the blank. mAU, milli-absorbance units.
Among the inhibitors used, quercetin, an inhibitor of CYP2C8 (Walsky et al., 2005), has been found to inhibit the O-demethylation of silybin to the greatest extent. The minor inhibiting effects seen with sulfaphenazole and triacetyloleandomycin (inhibitors of CYP2C9 and CYP3A4) were in line with findings on the inhibition of specific P450 activities by silybin in microsomes (Fig. 2) as well as with the earlier results of Beckmann-Knopp et al. (2000) and of Sridar et al. (2004). The use of bactosomes, expressing a single P450 form, has confirmed conclusively that the CYP2C8 form is catalyzing the O-demethylation of silybin. Bactosomal preparation with CYP3A4 has also been able to produce the demethylated metabolite but to a much lesser extent. No formation of metabolites has been found with bactosomes expressing CYP2C9 or other P450 enzymes. The inhibition of silybin metabolism by sulfaphenazole as well as the inhibition of CYP2C9 activity by silybin (Figs. 2 and 3) can apparently be explained by a strong structural similarity of the CYP2C8 and CYP2C9 forms (77.8%, P450 database at http://cpd.ibmh.msk.su), which is reflected also in a broad overlap of substrates and inhibitors of both forms; however, only CYP2C8 is able to metabolize silybin.
The results obtained here both with the metabolism of silybin as well as with the inhibition of P450 activities do not seem to constitute a rational base for the clinical importance of silybin-drug interactions. The percentage inhibition of the metabolic pathway may be expressed (in an ideal case of competitive inhibition) (Segel, 1975; Boxenbaum, 1999) as % inhibition = 100 CI/(CI + Ki) where CI is the inhibitor concentration and the Ki has the known meaning as the inhibition constant. As the silybin concentration in the systemic circulation does exceed 1 μM (Weyhenmeyer et al., 1992; van Erp et al., 2005), the degree of inhibition as estimated from the in vitro experiment should be expected to be on the order of several percentage points. Silybin, hence, should be assumed to be a relatively safe drug whose metabolism does not interfere significantly with major P450-catalyzed routes of drug biotransformation.
Footnotes
-
The authors gratefully acknowledge the financial support from the Czech Ministry of Education (Grant MSM 6198959216).
-
doi:10.1124/dmd.107.016410.
-
ABBREVIATIONS: P450, cytochrome P450; HPLC, high performance liquid chromatography; μLC/MS, micro liquid chromatography coupled with mass spectrometry; MS/MS, tandem mass spectrometry.
- Received May 3, 2007.
- Accepted July 30, 2007.
- The American Society for Pharmacology and Experimental Therapeutics