- Email: [email protected]
Differential inhibition of CYP1-catalyzed regioselective hydroxylation of estradiol by berberine and its oxidative metabolites
Drug Metabolism and Pharmacokinetics 30 (2015) 374e383
Contents lists available at ScienceDirect
Drug Metabolism and Pharmacokinetics journal homepage: http://www.journals.elsevier.com/drug-metabolism-andpharmacokinetics
Regular article
Differential inhibition of CYP1-catalyzed regioselective hydroxylation of estradiol by berberine and its oxidative metabolites Yu-Ping Chang a, Chiung-Chiao Huang a, Chien-Chang Shen b, Keng-Chang Tsai c, Yune-Fang Ueng a, d, e, * a
Divisions of Basic Chinese Medicine, Taiwan, ROC Chinese Medicinal Chemistry, Taiwan, ROC c Chinese Materia Medica Development, National Research Institute of Chinese Medicine, Ministry of Health and Welfare, Taipei, Taiwan, ROC d Department and Institute of Pharmacology, National Yang-Ming University, Taipei, Taiwan, ROC e Institute of Medical Sciences, Taipei Medical University, Taipei, Taiwan, ROC b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 16 July 2015 Received in revised form 17 August 2015 Accepted 18 August 2015 Available online 28 August 2015
Berberine is a pharmacologically active alkaloid present in widely used medicinal plants, such as Coptis chinensis (Huang-Lian). The hormone estradiol is oxidized by cytochrome P450 (CYP) 1B1 to primarily form the genotoxic metabolite 4-hydroxyestradiol, whereas CYP1A1 and CYP1A2 predominantly generate 2-hydroxyestradiol. To illustrate the effect of berberine on the regioselective oxidation of estradiol, effects of berberine and its metabolites on CYP1 activities were studied. Among CYP1s, CYP1B1.1, 1.3 (L432V), and 1.4 (N453S)-catalyzed 4-hydroxylation were preferentially inhibited by berberine. Differing from the competitive inhibition of CYP1B1.1 and 1.3, N453S substitution in CYP1B1 allowed a non-competitive or mixed-type pattern. An N228T in CYP1B1 highly decreased its activity and preference to 4-hydroxylation. A reverse mutation of T223N in CYP1A2 retained its 2-hydroxylation preference, but enhanced its inhibition susceptibility to berberine. Compared with berberine, metabolites demethyleneberberine and thalifendine caused weaker inhibition of CYP1A1 and CYP1B1 activities. Unexpectedly, thalifendine was more potent than berberine in the inhibition of CYP1A2, in which case an enhanced interaction through polar hydrogen-p bond was predicted from the docking analysis. These results demonstrate that berberine preferentially inhibits the estradiol 4-hydroxylation activity of CYP1B1 variants, suggesting that 4-hydroxyestradiol-mediated toxicity might be reduced by berberine, especially in tissues/tumors highly expressing CYP1B1.
Keywords: Estradiol CYP1B1 Variants Berberine Thalifendine Demethyleneberberine
Copyright © 2015, The Japanese Society for the Study of Xenobiotics. Published by Elsevier Ltd. All rights reserved.
1. Introduction The physiologically important hormone estradiol (E2) not only plays a crucial role in maintaining normal sexual function, but also has been found to induce hormone-associated tumors in animals used in experiments [1]. E2 was oxidized by cytochrome P450 (P450, CYP) enzymes to primarily form 2- and 4-hydroxylated metabolites (Scheme 1), which could be further oxidized to 2,3- and 3,4quinones, respectively [2]. E2-3,4-quinone reacts with DNA more efficiently than the E2-2,3-quinone does [3]. In breast tissues, the level of 4-hydroxyestradiol (4-OHE2) was about 3e4 times higher than 2-hydroxyestradiol (2-OHE2) in women with breast cancer, and the level of 4-OHE2 was 4-fold higher than in women without breast cancer [4]. The ratio of 4-OHE2 to 2-OHE2 in human breast tumor
* Corresponding author. E-mail address: [email protected] (Y.-F. Ueng).
tissues was about 10-times higher than the ratio in the breast tissues of healthy women who underwent the surgery for reduction mammoplasty [5]. 4-OHE2 stimulated more transformed colony formation than 2-OHE2 in Syrian hamster embryo cells [6]. In MCF7 cells, 4-OHE2 has been reported to stimulate cell proliferation and reduce cytochrome c release [7]. In the same cell line, E2 and 4-OHE2 reduced the apoptotic effect of staurosporine, whereas 2-OHE2 exposure did not influence the staurosporine-induced apoptosis [8]. 4-OHE2 can be a risk factor for abnormal cell growth regulation related to hormone-associated cancers. P450 enzymes including CYP1A1, CYP1A2, CYP1B1, CYP2C9, and CYP3A4 are involved in the oxidation of E2 [9]. CYP1A2, CYP2C9, and CYP3A4 are primarily hepatic P450s, whereas CYP1A1 and CYP1B1 are predominantly expressed in extrahepatic tissues including the uterus and breast [10,11]. 2-OHE2 is the major product generated by most E2-oxidizing P450s, including CYP1A1 and CYP1A2. However, CYP1B1 predominantly carries out a 4-hydroxylation pathway with a specific activity of more than 2-fold greater than that of CYP1A1 [9].
http://dx.doi.org/10.1016/j.dmpk.2015.08.006 1347-4367/Copyright © 2015, The Japanese Society for the Study of Xenobiotics. Published by Elsevier Ltd. All rights reserved.
Y.-P. Chang et al. / Drug Metabolism and Pharmacokinetics 30 (2015) 374e383
375
OH
O
A
O
10
X
N
OH
N
H3CO
9
OCH3
Y
Demethyleneberberine
Berberine: X = Y = OCH3 Thalifendine: X = OH; Y = OCH3
CH3 OH
B CH3 OH
HO HO
HO
4-Hydroxyestradiol
CYP1
2
CH3
NADPH, O2 4
OH
HO
HO
2-Hydroxyestradiol Scheme 1. Structures of the oxidative metabolites of berberine and estradiol.
About 71% of Asians have CYP1B1 with homozygous L432, which has been classified as CYP1B1.1 [12]. CYP1B1.3 carries the L432V substitution. In Caucasian population, the distributions of V/ V, V/L, and L/L are roughly equal [13,14]. The difference between the incidence of hormone cancer in humans with CYP1B1 genotypes CYP1B1*1, CYP1B1*3 (4326C > G) and CYP1B1*4 (4390A > G) remains uncertain [10,12,14]. Gajjar et al. [10] reported that CYP1B1.3 had E2 4-hydroxylation activity 1.6-fold greater than that of CYP1B1.1. However, the kinetic parameters of CYP1B1.1-, CYP1B1.3 (L432V)-, and CYP1B1.4 (N453S) -catalyzed E2 2- and 4hydroxylation have been reported to be similar [13]. Although the kinetic parameters of E2 hydroxylation by CYP1B1 variants showed variations in different expression systems and assay methods [15], there is no doubt that the contribution of CYP1B1.3 and CYP1B1.4 to human E2 4-hydroxylation is as important as CYP1B1.1. Protoberberines represent a group of natural products with a core structure of dihydroisoquinolino[2,1-b]isoquinolin and exhibit a variety of pharmacological activities including anti-cancer and anti-inflammation effects [16,17]. This type of alkaloids can be found in several medicinal plants including Coptis chinensis (Huang-Lian) and Hydrastis canadensis (goldenseal) [16,17]. There are 17e19 mM berberine (Scheme 1) present in the goldenseal extract, which has been widely used as an immunostimulant in several countries including the USA [18]. Berberine suppresses cell survival and E2-induced cell proliferation in human uterine leiomyoma cells, without affecting the cell survival in an immortalized normal smooth muscle cell line [19]. Berberine treatment does not alter the serum estradiol concentration in ovariectomized rats [20]. However, the metabolite profile of estradiol in berberine-treated animals remained unclear. CYP1-catalyzed berberine oxidation generates primarily demethyleneberberine and thalifendine
(Scheme 1). Among the CYP1 members, our previous study demonstrated that berberine preferentially inhibited CYP1B1 when 7-ethoxyresorufin was used as a substrate [21]. The berberine metabolite caused a relatively weak inhibition of CYP1B1 activity. However, the inhibitory effect of berberine had substrate dependency. Activities of CYP1B1 toward alkoxy heterocyclic compounds, 7-ethoxyresorufin, 7-methoxyresorufin and 7ethoxycoumarin were strongly inhibited by berberine with IC50 values <1 mM. However, the hydroxylation activity of CYP1B1 toward the side chain-free polycyclic aromatic hydrocarbon benzo(a) pyrene was weakly inhibited with an IC50 value of 34 mM. Different from the planar structures of alkoxy heterocyclic substrates, E2 has the angular polycyclic rings and only one of the rings is aromatic (Fig. 1). E2 carries the hydrophobic methyl and the hydrophilic hydroxyl side chains. Due to the angular structure, important physiological and toxicological roles, and the regioselective hydroxylation of E2, the effects of berberine on CYP1-catalyzed E2 hydroxylation were investigated in recombinant human enzyme systems. 2. Materials and methods 2.1. Chemicals Berberine chloride, 17b-E2, glucose-6-phosphate, glucose-6phosphate dehydrogenase (from yeast), 2-OHE2, 4-OHE2, 2methoxyestradiol and nicotinamide adenine diphosphate (NADPþ) disodium salt were purchased from SigmaeAldrich Chemical Co. (St. Louis, MO, U.S.A.). Ammonium phosphate, dimethyl sulfoxide (DMSO), ethyl acetate and methanol were purchased from Merck KGaA (Darmstadt, Germany).
376
Y.-P. Chang et al. / Drug Metabolism and Pharmacokinetics 30 (2015) 374e383
Fig. 1. The formation rate ratios of 4-hydroxyestradiol to 2-hydroxyestradiol in the oxidation of estradiol by CYP1A2.1 and CYP1B1 variants. Data represent the mean ± SE of three determinations.
system containing 0.5 mM NADPþ, 5 mM glucose-6-phosphate and 0.5 u/ml glucose-6-phosphate dehydrogenase. The generation of 2- and 4-OHE2 was determined following the HPLC method reported by Satoh et al. [25]. E2 and hydroxyestradiols were dissolved in methanol. In the preparation of standards for quantification, known concentrations of 2- and 4-OHE2 were added to an additional incubation without glucose-6-phosphate dehydrogenase. 2-Methoxyestradiol was added as an internal control of the assay. Aliquots of the ethyl acetate extract were dried under nitrogen, redissolved in methanol, and subjected into a HPLC system equipped with a UV detector (Agilent 1100 series, Agilent Technologies, Santa Clara, USA). A pre-filter (Frit SS, 2 mm, IDEX Health & Science, Oak Harbor, WA, USA) and a C18 guard column (4.3 10 mm, ODS, IDEX Health & Science) were connected prior to the analytical C18 column (4.6 250 mm, I.D.: 5 mm, 5C18-AR-II, COSMOSIL, NA CALAI TESQUE, Inc., Kyoto, Japan). Metabolites were separated with a mobile phase consisting of methanol/50 mM ammonium phosphate, pH 2.5 (1:1, v/v) at a flow rate of 1 ml/min, at room temperature and detected by measuring the absorbance at 280 nm. 2.4. Kinetic analysis of P450 inhibition
Demethyleneberberine and thalifendine were synthesized following methods as described in a previous report [22]. The purities of synthetic metabolites were 98% based on 1H NMR and HPLC analyses. 2.2. Preparation of bacterial membranes expressing polymorphic human CYP1 enzymes Bicistronic human constructs consisting of the modified coding sequence of human CYP1A1*1, CYP1A2*1 and CYP1B1*3 followed by that of human NADPH-P450 reductase were generously provided by Dr. F. Peter Guengerich (Nashville, TN, U.S.A.). Single nucleotide point mutations were introduced into CYP1A2*1 and CYP1B1*3 by the primer-directed enzymatic amplification method following the instruction manual from the QuickChange Lightning site-directed mutagenesis kit (Stratagene, Agilent Technologies Company, La Jolla, CA, USA). Oligonucleotide primer sets used for the mutagenesis were designed with a primer containing a mutated base, and the sequences of mutated constructs were determined [21,22]. CYP1B1*4 and CYP1B1*VS constructs were prepared from a single point mutation from CYP1B1*1 and CYP1B1*3, respectively. CYP1B1NT (N228T) and CYP1A2TN (T223N) were prepared as previously described [21]. P450 constructs were transformed to Escherichia coli DH5a by electroporation (Gene pulser II, Bio-Rad, Hercules, CA, U. S. A.). P450 expression and bacterial membrane preparation were performed as described in previous report [21]. The P450 content was determined following the spectrophotometric methods [23]. The P450 contents of membranes expressing CYP1A1.1, CYP1A2.1, CYP1B1.4, CYP1B1VS (V432S453) and CYP1B1NT were 4.1, 37.1, 13.2, 10.1, 0.83, 3.52 and 0.30 nmol P450/ ml, respectively. The reduction activity of NADPH-P450 reductase (NADPH-cytochrome c reductase) was determined using cytochrome c as a substrate following the method reported by Phillips and Langdon [24]. The cytochrome c reduction activity of P450 reductase in CYP1A1.1, CYP1A2.1, CYP1B1.1, CYP1B1.3, CYP1B1.4, CYP1B1VS and CYP1B1NT were 3.77, 1.02, 2.65, 1.59, 4.55, 2.15 and 2.94 mmol/min/nmol P450, respectively. 2.3. Enzyme assays E2 hydroxylation activities were determined using 10 pmol P450/ml and 50 mM E2 in the presence of a NADPH-generating
In competitive inhibition, kinetic analysis of P450 activity was done following MichaeliseMenten kinetic property. Vmax and I are the maximal velocity and protoberberine concentration, respectively. Ki and KI are the inhibitor constants for the binding of an inhibitor to a P450 and P450-substrate complex, respectively. For noncompetitive inhibition, Ki is equivalent to KI. Values of velocities (v) at various substrate concentrations (S) were fitted by nonlinear least-squares regression without weight according to the MichaeliseMenten equation: competitive inhibition: v ¼ Vmax $ S/{S þ Km[1 þ (I/Ki)]}; noncompetitive inhibition: v ¼ Vmax $ S/[1 þ (I/Ki)][S þ Km]; mixed type of inhibition: v ¼ Vmax $ S/{S[1 þ (I/KI)] þ Km[1 þ (I/Ki)]} (Sigma Plot, Jandel Scientific, San Rafael, CA, U.S.A.). Estimates of variance and the coefficient of variation (CV, %) of the estimates of variance are presented from the analysis of individual sets of data. 2.5. Computer modeling and data analyses Models of berberine and thalifendine in a complex with CYP1A2.1 (PDB ID: 2HI4) and CYP1B1.1 (PDB ID: 3PM0) isoforms were generated as described in a previous report [21]. The docking of inhibitors to the putative active site of CYP1A1 was generated using the crystallographic structure of CYP1A1.1 (PDB ID: 4I8V) complexed with a-naphthoflavone. The concentrations of protoberberines required for 50% inhibition of catalytic activities (IC50) were calculated by curve fitting (start at 0, define end, Grafit, Erithacus Software Ltd., Staines, UK). Estimates of variance (denoted by ±) are presented from the analyses of individual sets of data. 3. Results 3.1. Inhibition of E2 hydroxylation activities of CYP1A1.1, CYP1A2.1, and CYP1B1.1 by berberine Among the wild types of CYP1 enzymes, berberine caused the least inhibitory effect on CYP1A2.1-catalyzed E2 2- and 4hydroxylation with similar IC50 values (Table 1), indicating that the berberine-mediated inhibition did not show regioselectivity. Berberine inhibited the CYP1A1.1-catalyzed 2-hydroxylation with an IC50 value of 4% of the value for CYP1A2.1 inhibition. The 4hydroxylation activity of CYP1A1.1 was too low to determine the IC50 value. The isoform most inhibited by berberine was CYP1B1.1. Berberine potently inhibited the CYP1B1.1-mediated estradiol 4-
Y.-P. Chang et al. / Drug Metabolism and Pharmacokinetics 30 (2015) 374e383
377
Table 1 Inhibition of estradiol hydroxylation activities of CYP1A1.1, CYP1A2.1, and CYP1B1 variants by berberine. P450
Estradiol hydroxylation site
Hydroxylation activity nmol/min/nmol P450
CYP1A1.1
242424242424-
8.57 0.18 6.30 0.27 1.07 2.92 0.58 1.92 1.35 3.82 0.68 2.10
CYP1A2.1 CYP1B1.1 CYP1B1.3 CYP1B1.4 CYP1B1VS
± ± ± ± ± ± ± ± ± ± ± ±
0.16 0.03 0.09 0.04 0.03 0.13 0.01 0.21 0.25 0.10 0.05 0.16
Activity ratio (4-/2-)
IC50, mM
0.02
1.59 n.d. 35.7 40.2 0.56 0.33 0.45 0.20 0.29 0.18 0.28 0.22
0.04 2.7 3.3 2.8 3.1
IC50 ratio (4-/2-)
± 0.23
e
± ± ± ± ± ± ± ± ± ±
1.1
0.9 13.0 0.04 0.02 0.08 0.01 0.07 0.02 0.04 0.03
0.6 0.4 0.6 0.8
CYP1 activities were determined using 10 pmol P450/ml and 50 mM estradiol in an estradiol hydroxylation assay. Estradiol hydroxylation activity was determined as described in the Materials and Methods. Results of activity determinations show the mean ± SEM of 3 individual experiments with duplicate determinations in each experiment. The IC50 values represent estimates and estimated variance (denoted by ±) of the curve fittings. n.d.: not determinable, the activity was too low to determine the IC50 value. The symbol “e” indicates that the ratio could not be calculated.
Table 2 The inhibition kinetic constants of berberine-mediated inhibition of estradiol hydroxylation activities of CYP1B1 variants. Variants
1B1.1 1B1.3 1B1.4 1B1.VS
4-Hydroxylation
2-Hydroxylation
Type of inhibition
Ki, nM
Competitive Competitive Noncompetitive Mixed type
20.2 10.9 167 33.5
± ± ± ±
3.5 2.3 12 3.7
KI, nM
Type of inhibition
Ki, nM
895 ± 153
Competitive Competitive Noncompetitive Competitive
25.1 14.0 250 25.3
± ± ± ±
1.8 2.1 40 2.1
Data represent the mean ± SE of inhibition constants calculated from the activity determination in the presence of three different berberine concentrations. The estimated variance was 6e27% of the parameter estimates generated from non-linear regression of each set of data.
hydroxylation with an IC50 value of 1% of the value for CYP1A2.1 inhibition. However, the IC50 value for the inhibition of 2hydroxylation activity of CYP1B1.1 was two-fold higher than that for the inhibition of 4-hydroxylation, revealing the preferential inhibition of CYP1B1.1-mediated E2 4-hydroxylation by berberine. 3.2. Inhibitory effects of berberine on E2 hydroxylation activities of CYP1B1 variants To determine the inhibitory effect of berberine on the CYP1B1 variants, bacterial membranes expressing CYP1B1.3, CYP1B1.4 and CYP1B1VS were prepared. All CYP1B1 variants showed a 4hydroxylation preference and the activity ratio of 4-hydroxylation to 2-hydroxylation was about 3 (Table 1 and Fig. 1). In the presence of increasing concentrations of berberine, the metabolite formation rate ratio of 4-OHE2 to 2-OHE2 were decreased (Fig. 1). However, in the CYP1A2.1 system, the metabolite formation rate ratio of 4-OHE2 to 2-OHE2 was not affected by berberine. In E2 hydroxylation, 4-hydroxylation activities of CYP1B1.1, CYP1B1.3 and CYP1B1.4 were potently inhibited by berberine. The 2hydroxylation activities of CYP1B1.4 and CYP1B1VS variants carrying an N453S substitution were inhibited more potently than CYP1B1.1 and CYP1B1.3, resulting in an increase in the IC50 ratio of 4-hydroxylation to 2-hydroxylation, especially in CYP1B1VS (Table 2).
concentrations in E2 2- and 4-hydroxylation catalyzed by CYP1B1 variants showed a hyperbolic property (Fig. 2). The Km values for 2hydroxylation by CYP1B1.1, CYP1B1.3, CYP1B1.4 and CYP1B1VS were 2.7 ± 0.3, 2.4 ± 0.3, 11.1 ± 2.8 and 4.8 ± 0.5 mM, respectively. The Km values for the 4-hydroxylation by CYP1B1.1, CYP1B1.3, CYP1B1.4 and CYP1B1VS were 2.9 ± 0.6, 2.4 ± 0.5, 6.0 ± 1.1 and 3.4 ± 0.5 mM, respectively. The Vmax values for 2-hydroxylation by CYP1B1.1, CYP1B1.3, CYP1B1.4 and CYP1B1VS were 0.61 ± 0.01, 0.54 ± 0.02, 2.02 ± 0.18 and 1.17 ± 0.03 nmol/min/nmol P450, respectively. The Vmax values for 4-hydroxylation were 3.26 ± 0.16, 1.79 ± 0.11, 5.37 ± 0.29, and 2.79 ± 0.10 nmol/min/nmol P450, respectively. The intrinsic clearance (Clin ¼ Vmax/Km) values of E2 through 4-hydroxylation by CYP1B1.1, CYP1B1.3, CYP1B1.4, and CYP1B1VS were 1.14, 0.75, 0.90 and 0.82 ml/min/nmol P450, respectively. Linear regression analyses of the LineweavereBurk plots revealed that berberine competitively inhibited CYP1B1.1and CYP1B1.3-catalyzed E2 2- and 4-hydroxylation (Fig. 3A and B). Nonlinear regression analysis generated Ki values as listed in Table 2. In the CYP1B1.1 and CYP1B1.3 enzyme systems, the Ki for the inhibition of 2-hydroxylation was higher than that for the inhibition of 4-hydroxylation. However, noncompetitive and mixedtyped inhibition properties were present in the inhibition of CYP1B1.4 and CYP1B1VS carrying an N453S substitution (Fig. 3C and D), suggesting the influence of N453S in the inhibition pattern of berberine.
3.3. Kinetic analysis of the inhibitory behavior of berberine on E2 hydroxylation by CYP1B1 variants
3.4. Inhibition of E2 hydroxylation activities of CYP1B1NT and CYP1A2TN by berberine
Effects of berberine on hydroxylation by CYP1B1 variants were determined to be due to the potent inhibition of CYP1B1.1catalyzed 4-hydroxylation by berberine when estradiol was used as a substrate. The plots of the velocity versus substrate
Our previous report reveals that Asn228 is a crucial amino acid residue of CYP1B1 for its oxidation activity toward 7ethoxyresorufin and its susceptibility to inhibition by berberine [21]. The corresponding amino acid in CYP1A2.1 is Thr223, which
378
Y.-P. Chang et al. / Drug Metabolism and Pharmacokinetics 30 (2015) 374e383
Fig. 2. Plots of velocity versus estradiol concentration of the inhibition of estradiol 2- and 4-hydroxylation activities of CYP1B1.1 (LN) (A), CYP1B1.3 (VN) (B), CYP1B1.4 (LS) (C), and CYP1B1VS (D) by berberine. The left and right panels show the plots of velocity versus estradiol concentration in the CYP1B1-catalyzed estradiol 2- and 4-hydroxylation, respectively. Estradiol hydroxylation velocity was determined in an assay containing 10 pmol P450/ml. The data represent the mean ± SE of three determinations.
did not allow hydrogen bond interaction with the methoxyl moiety in berberine. The substitution of T223N (CYP1A2TN) clearly enhanced the inhibition susceptibility of CYP1A2.1 to berberine. Thus, estradiol hydroxylation activities of CYP1A2TN and CYP1B1NT (N228T) were determined to examine the influence of substitution on regioselective hydroxylation by CYP1A2 and CYP1B1 and CYP1B1-preferential inhibition by berberine. Compared with the wild type enzyme, CYP1A2TN and CYP1B1NT had low E2 hydroxylation activities (Table 3). CYP1A2TN remained the preference in 2-hydroxylation toward estradiol. T223N substitution in CYP1A2.1 enhanced the inhibition susceptibility of CYP1A2.1 to berberine. CYP1B1NT catalyzed oxidation generated 2hydroxylated and 4-hydroxylated metabolites at nearly equal velocities. The activity of CYP1B1NT was too low to allow us to determine the IC50 value of berberine-mediated inhibition. To examine the detrimental changes of functional CYP1B1 protein by
N228T substitution, the activity toward another substrate 7-ethoxycoumarin was determined. 7-Ethoxycoumarin O-deethylation activity was decreased to 5% of the wild type (data not shown). The extremely low activity of CYP1B1NT toward different substrates revealed that Asn228 was not only important for the interaction with berberine but also very important for maintaining the function of CYP1B1. 3.5. Inhibition of CYP1A1.1, CYP1A2.1, and CYP1B1.1-catalyzed E2 hydroxylation by demethyleneberberine and thalifendine Demethyleneberberine and thalifendine inhibited the 2hydroxylation activity of CYP1A1 with IC50 values of 13.4 ± 0.7 and 6.6 ± 1.3 mM, respectively (Fig. 4). Demethyleneberberine and thalifendine inhibited the 2-hydroxylation activity of CYP1A2.1 with IC50 values of 92.9 ± 17.2 and 23.2 ± 2.0 mM, respectively. E2 2-
Y.-P. Chang et al. / Drug Metabolism and Pharmacokinetics 30 (2015) 374e383
379
Fig. 3. LineweavereBurk plots of the inhibition of estradiol 2- and 4-hydroxylation activities of CYP1B1.1 (LN) (A), CYP1B1.3 (VN) (B), CYP1B1.4 (LS) (C), and CYP1B1VS (D) by berberine. The left and right panels show the plots of velocity versus estradiol concentration in CYP1B1-catalyzed estradiol 2- and 4-hydroxylation, respectively. The estradiol hydroxylation velocity was determined in an assay containing 10 pmol P450/ml. The data represent the reciprocals of the means of three determinations shown in Fig. 2.
Table 3 The activities and berberine-mediated inhibition of estradiol hydroxylation by CYP1A2TN and CYP1B1NT. P450 mutants
Estradiol hydroxylation
Hydroxylation activity nmol/min/nmol P450
CYP1A2TN
2-hydroxylation 4-hydroxylation 2-hydroxylation 4-hydroxylation
1.63 0.20 0.15 0.14
CYP1B1NT
± ± ± ±
0.09 0.01 0.01 0.03
Activity ratio (4-/2-)
IC50, mM
0.13
2.6 ± 0.2 n.d. n.d. n.d.
0.9
Estradiol hydroxylation activity was determined as described in Table 1. Data from the activity represent the means ± SE of 3 determinations. n.d.: not determinable, the activity was too low to determine the IC50 value.
hydroxylation activities of CYP1A1.1 and CYP1A2.1 were less inhibited by demethyleneberberine than by berberine. However, in CYP1A2-catalyzed 2-hydroxylation, activity was more sensitive to the inhibition by thalifendine than by berberine. Demethyleneberberine and thalifendine inhibited CYP1B1-catalyzed 2-
hydroxylation with IC50 values of 19.2 ± 2.5 and 13.7 ± 1.2 mM, respectively. Demethyleneberberine and thalifendine inhibited CYP1B1-catalyzed 4-hydroxylation with IC50 values of 9.9 ± 3.5 and 4.1 ± 0.6 mM, respectively. The IC50 values for the inhibition of CYP1B1-catalyzed E2 hydroxylation by demethyleneberberine and
380
Y.-P. Chang et al. / Drug Metabolism and Pharmacokinetics 30 (2015) 374e383
thalifendine may form polar hydrogen-p bond interaction [26] with the phenyl moiety of Phe256 of CYP1A2.1 (Fig. 5B). Unlike CYP1A2.1, the corresponding amino acid residues were Leu254 in CYP1A1.1 and Leu264 in CYP1B1.1, in which the potential polar hydrogen-p bond interaction was unfavorable (Fig. 5C and D). Using PLP1 scoring, the scores for the docking of thalifendine and berberine to CYP1A2.1 were 27.49 and 10.08, respectively. 4. Discussion
Fig. 4. Inhibition of CYP1A1.1, CYP1A2.1, and CYP1B1.1-catalyzed estradiol 2- and 4hydroxylation using demethyleneberberine (B), thalifendine ( ), and berberine (-). Berberine chloride, demethyleneberberine, and thalifendine were dissolved in dimethyl sulfoxide and a same volume of dimethyl sulfoxide was added to the control assay. The data represent the mean ± SE of three determinations.
▵
thalifendine were 2e34 times greater than the values for the inhibition by berberine. Compared with the inhibition by berberine, demethyleneberberine and thalifendine generally caused fewer inhibitory effects on the E2 hydroxylation activity of CYP1 isoforms but the CYP1A2.1 activity was inhibited by thalifendine with a greater potency. 3.6. Docking of berberine and thalifendine into CYP1A1.1, CYP1A2.1, and CYP1B1.1 Since the inhibitory effect of thalifendine was more potent than that of berberine on E2 hydroxylation by CYP1A2.1, but not for CYP1A1.1 and CYP1B1.1, computer modeling of the binding of thalifendine into CYP1 enzymes was performed. The binding of berberine in the putative active site of CYP1A2.1 was superimposed with the binding of E2 in an orientation favorable for the 2hydroxylation pathway (Fig. 5A). The 10-hydroxyl moiety of
The non-planar polycyclic steroid, E2, plays a crucial role in the modulation of not only the development of sexual organs, but also the pathological events unrelated to the sexual function, such as tumorigenesis and blood pressure regulation [1,27]. An elevated level of CYP1B1 has been reported in a variety of tumors, especially hormone-related cancers including breast and endometrial cancers [10,28,29]. Microsomal E2 4-hydroxylation activity of breast tumor tissues was about 4-fold greater than that from normal tissues [30]. Selective inhibition of CYP1B1-catalyzed E2 hydroxylation, especially preferential inhibition of a 4-hydroxylation pathway, may prove beneficial in the chemoprotection against estrogenassociated cancers. A non-selective CYP1A2 and CYP1B1 inhibitor a-naphthoflavone reduced the breast tumor incidence in female August Copenhagen Irish (ACI) rats with long term exposure to E2, suggesting a potential chemopreventive effect of a-naphthoflavone [31]. A synthetic 2,30 ,4,50 -tetramethoxystilbene has been reported to preferentially inhibit CYP1B1-catalyzed 7-ethoxyresorufin Odeethylation and E2 hydroxylation activities [32]. In E2 4hydroxylation, berberine inhibited CYP1B1 activity with an IC50 value within the range of IC50 values for the inhibition by 2,30 ,4,50 tetramethoxystilbene (IC50: 0.09 mM in the recombinant enzyme system expressed in bacterial membranes and 0.39 mM in a reconstituted system of purified enzyme) [32]. Although CYP1B1catalyzed E2 hydroxylation generates mainly the carcinogenic metabolite 4-OHE2 [30], more evidence reveals that extrahepatic CYP1B1 may play an important role in the regulation of physiological response, such as blood pressure [33]. Angiotensin IIinduced hypertension was more pronounced in male Cyp1b1þ/þ mice than in Cyp1b1/ mice. CYP1B1-catalyzed testosterone hydroxylation has been suggested to be involved in the hypertensive effect in Cyp1b1þ/þ mice. A potent CYP1B1 inhibitor 2,30 ,4,50 tetramethoxy stilbene ameliorates the hypertension status and renal injury caused by angiotensin II-treatment in male rats and mice. However, 4-OHE2 but not 2-OHE2 enhanced angiotensin IIinduced hypertension in female Cyp1b1þ/þ mice [27]. Berberine ameliorated the severity of hypertension and hypertensioninduced renal damage in male spontaneously hypertensive rats [34]. However, the effect of berberine in female rats has not yet been reported. Meta-analysis of 27 randomized controlled clinical trials, in which male and female patients were included, showed that berberine may be beneficial for the hypotensive effects [35]. The decreased E2 4-hydroxylation by berberine suggests its potential chemopreventive effect against estrogen-related cancers. Berberine may become a lead candidate for further development of agents for the protection against 4-OHE2-associated toxicity. On the other hand, the influence of hormone homeostasis on blood pressure affected by berberine should be noted. Different from a similar O-deethylation product, resorufin generated from CYP1A1/2- and CYP1B1-catalyzed oxidation toward 7-ethoxyresorufin and 7-methoxyresorufin, docking results suggested that E2 preferentially binds to CYP1A1/2 and CYP1B1 in a distinct orientation to generate primarily the 2-hydroxylation and 4-hydroxylation metabolite, respectively [36]. Several amino acid residues have been suggested or proved to be important for the regioselective metabolism of E2 by CYP1 enzymes. Nishida et al.
Y.-P. Chang et al. / Drug Metabolism and Pharmacokinetics 30 (2015) 374e383
381
Fig. 5. The docking of estradiol, berberine and thalifendine to the putative active sites of CYP1A1.1, CYP1A2.1 and CYP1B1.1. Panel A shows the binding of estradiol (blue) and berberine (yellow thin sticks) superimposed within the active site of CYP1A2.1 based on molecular modeling [21,36]. Panels B, C and D show the binding of thalifendine (yellow thick sticks) to CYP1A2.1, CYP1A1.1 and CYP1B1.1, respectively. The side chains of amino acid residues of CYP1B1.1 interacting with berberine are shown in light gray. The nitrogen and oxygen atoms are shown in dark blue and red, respectively. The heme prosthetic group of P450 is shown in magenta and iron is shown as white.
[37] successfully showed that the formation rate ratio of 4-OHE2/2OHE2 was shifted from 5 to 0.5 by substituting Val395Leu. This demonstrated that Val395 was important for CYP1B1 in a 4hydroxylation pathway. Based on the scoring results of active site docking, Asn265 and Ala133 (Fig. 4) have also been suggested to be important for E2 recognition by CYP1B1 [36]. In our previous study, when we used 7-ethoxyresorufin and caffeine as substrates, CYP1A2.1 was resistant to inhibition by berberine (IC50 > 60 mM) [21]. Although the IC50 value was still high and may not occur at the pharmacological concentration of berberine in patients, current results show that CYP1A2.1-catalyzed E2 2-hydroxylation activity was more sensitive than its activities toward 7-ethoxyresorufin and caffeine to the inhibition by berberine. Previously, our findings revealed that the substitution of T223N in CYP1A2.1 increased the sensitivity of CYP1A2.1 to inhibition by berberine. In this report, results revealed that the substitution of T223N in CYP1A2.1 not only enhanced the inhibition sensitivity of CYP1A2.1 to berberine, but also resulted in the increase of the formation rate ratio of 4-OHE2/2OHE2. In E2 hydroxylation, the substitution of Asn228 with Thr in CYP1B1 dramatically reduced activity by 70%e90% and the production of 2-OHE2 and 4-OHE2 by CYP1B1NT became equal. Like the E2 hydroxylation activity, activities of CYP1B1NT toward 7ethoxyresorufin [21] and 7-ethoxycoumarin (data not shown) were less than 30% of CYP1B1.1 activities. This marked decrease in CYP1B1 activity revealed that Asn228 is not only important for the berberine-mediated CYP1B1 inhibition, but very important for its catalytic function and the 4-hydroxylation preference of CYP1B1. Consistent with the summary results reported by Shimada et al. [15], determination of the formation rate ratios of 4-OHE2 to 2OHE2 by CYP1B1 variants and mutants indicated that the amino acid substitutions Leu432Val and Asn453Ser did not affect the
preference of CYP1B1 in the 4-hydroxylation pathway. However, the substitution of N453S tended to increase the inhibitory effect of berberine on the E2 2-hydroxylation activity of CYP1B1 carrying either Leu432 or Val432. These results were unexpected since Asn453 is located on the opposite side of the substrate binding pocket of CYP1B1. The Km values for the 4-hydroxylation using CYP1B1.1, CYP1B1.3, CYP1B1.4 and CYP1B1VS were 3.5e14, 3.8e40, 7e11 and 3.7e17 mM, respectively [13,15]. The Vmax values for the 4hydroxylation using CYP1B1.1, CYP1B1.3, CYP1B1.4, and CYP1B1VS were 0.91e3.7, 1.45e3.3, 1.5e3.6, and 1.8e4.5 nmol/min/nmol P450, respectively [13,15]. Using 10 pmol P450/ml in the assay, the kinetic analysis generated parameters within the range of reported values (Supplementary Table 1). Any differences can be attributed to not only the distinct recombinant enzyme systems as discussed in the report of Shimada et al. [15], but also the P450 concentrations in the assay. Results from the inhibition kinetic analysis showed that the berberine inhibited CYP1B1.1 and CYP1B1.3 in a competitive manner, whereas CYP1B1.4 and CYP1B1VS carrying an N453S substitution could be inhibited by berberine in a non-competitive or mixed type manner. An N453S substitution may change the intra-protein interactions, leading to the differential inhibition behavior of berberine. Consistent with reports that CYP1B1 had a lower formation rate difference (3-fold) in 4-OHE2 and 2-OHE2 than CYP1A1/2 (>25fold) did, the docking results indicated that the oxidation pocket of CYP1B1 was rounder than for CYP1A1/2 [36]. Orientation of E2 in the CYP1B1 putative active site may be more flexible than in CYP1A1/2. The IC50 value for berberine-mediated inhibition of the 4-hydroxylation pathway was about half that of the 2hydroxylation pathway. The presence of berberine resulted in a concentration-dependent decrease in the metabolite formation
382
Y.-P. Chang et al. / Drug Metabolism and Pharmacokinetics 30 (2015) 374e383
rate ratio of 4-OHE2 to 2-OHE2. The activity of CYP1B1 through two hydroxylation pathways was almost equal in the presence of 3e5 mM berberine. Our previous report revealed that berberine was poorly oxidized by CYP1B1, generating mainly demethyleneberberine and thalifendine [22]. Using 7-ethoxyresorufin as the substrate, the inhibitory effect on CYP1A2.1 might be too weak (IC50 > 60 mM) to differentiate the inhibition by berberine or thalifendine. However, using E2 as a substrate, our findings indicated that thalifendine caused CYP1A2.1 inhibition with a potency greater than that of berberine. The inhibition of CYP1A1-and CYP1B1-mediated E2 hydroxylations by demethyleneberberine and thalifendine was not as potent as the inhibition by berberine. Docking results suggested that the hydroxyl group at C10 may present polar hydrogen-p interaction with the phenyl moiety of the Phe256 of CYP1A2.1. However, the corresponding amino acid residues Leu254 in CYP1A1 and Leu264 in CYP1B1 did not allow this interaction. In summary, berberine preferentially inhibited E2 4hydroxylation catalyzed by CYP1B1 variants. However, the inhibition of CYP1B1 variants carrying N453S showed an inhibition kinetic pattern different from the wild type. Docking of thalifendine to the putative active site of CYP1 isoforms suggested that a polar hydrogen-p bond might form between the 10-hydroxyl group of thalifendine and Phe256 in CYP1A2.1, leading to a stronger inhibition by thalifendine than by berberine, even though the inhibition of CYP1B1 variants by berberine was stronger. Berberine may provide benefit in the reduction of 4-OHE2-induced toxicity, especially in tissues or tumors expressing high levels of CYP1B1. The benefit of berberine in the prevention of 4-OHE2-mediated toxicity needs further investigation in vivo. On the other hand, berberine has been reported to inhibit human liver microsomal CYP2D6 and CYP3A4 activities, but the IC50 values (45 and 400 mM, respectively) were relatively high [18]. Repeated administration of 300 mg berberine t.i.d. for 2 weeks has been reported to increase the AUC012 and AUC0-∞ values of CYP3A4 substrate midazolam by ~40% in healthy participants [38]. Thus, potential drug-interaction should be considered in the dosing regimen of berberine for chemoprotection against the toxicity of 4-OHE2. Acknowledgments We truly appreciate Dr. Guengerich for kindly providing us with the P450 constructs. This work was mainly supported by Grant MOST104-2320-B-077-001 from Ministry of Science and Technology (MOST), Taipei and partially supported by Grant NSC102-2923B-077-001-MY3 from National Science Council (now MOST), Taipei, and the National Research Institute of Chinese Medicine, Taipei. Appendix A. Supplementary data Supplementary data related to this article can be found online at http://dx.doi.org/10.1016/j.dmpk.2015.08.006. References [1] Roy D, Cai Q, Felty Q, Narayan S. Estrogen-induced generation of reactive oxygen and nitrogen species, gene damage, and estrogen-dependent cancers. J Toxicol Environ Health B 2007;10:235e57. [2] Tsuchiya Y, Nakajima M, Yokoi T. Cytochrome P450-mediated metabolism of estrogens and its regulation in human. Cancer Lett 2005;227:115e24. [3] Zahid M, Kohli E, Saeed M, Rogan E, Cavalieri E. The greater reactivity of estradiol-3,4-quinone vs estradiol-2,3-quinone with DNA in the formation of depurinating adducts: implications for tumor-initiating activity. Chem Res Toxicol 2006;19:164e72. [4] Rogan EG, Badawi AF, Devanesan PD, Meza JL, Edney JA, West WW, et al. Relative imbalance in estrogen metabolism and conjugation in breast tissue of women with carcinoma: potential biomarkers of susceptibility to cancer. Carcinogenesis 2003;24:697e702.
[5] Castagnetta L, Granata OM, Cocciadiferro L, Saetta A, Polito L, Bronte G, et al. Sex steroids, carcinogenesis, and cancer progression. Ann N Y Acad Sci 2004;1024:233e46. [6] Tsutsui T, Tamura Y, Yagi E, Barrett JC. Involvement of genotoxic effects in the initiation of estrogen-induced cellular transformation: studies using Syrian Hamster embryo cells treated with 17b-estradiol and eight of its metabolites. Int J Cancer 2000;86:8e14. [7] Seeger H, Wallwiener D, Kraemer E, Mueck AO. Comparison of possible carcinogenic estradiol metabolites: effects on proliferation, apoptosis and metastasis of human breast cancer cells. Maturitas 2006;54:72e7. [8] Gregoraszuk E, Ptak A. Involvement of caspase-9 but not caspase-8 in the antiapoptotic effects of estradiol and 4-OH-estradiol in MCF-7 human breast cancer cells. Endocr Regul 2011;45:3e8. [9] Lee AJ, Cai MX, Thomas PE, Conney AH, Zhu BT. Characterization of the oxidative metabolites of 17b-estradiol and estrone formed by 15 selectively expressed human cytochrome P450 isoforms. Endocrinology 2003;144: 3382e98. [10] Gajjar K, Martin-Hirsch PL, Martin FL. CYP1B1 and hormone-induced cancer. Cancer Lett 2012;324:13e30. [11] Singh MN, Stringfellow HF, Taylor SE, Ashton KM, Ahmad M, Abdo KR, et al. Elevated expression of CYP1A1 and g-SYNUCLEIN in human ectopic (ovarian) endometriosis compared with eutopic endometrium. Mol Hum Reprod 2008;14:655e63. [12] Paracchini V, Raimondi S, Gram IT, Kang D, Kocabas NA, Kristensen VN, et al. Meta- and pooled analyses of the cytochrome P-450 1B1 Val432Leu polymorphism and breast cancer: a HuGE-GSEC review. Am J Epidemiol 2006;165: 115e25. [13] Aklillu E, Oscarson M, Hidestrand M, Leidvik B, Otter C, Ingelman-Sunberg M. Functional analysis of six different polymorphic CYP1B1 enzyme variants found in an Ethiopian population. Mol Pharmacol 2002;61:586e94. [14] Yao L, Fang F, Wu Q, Zhong Y, Yu L. No association between CYP1B1 Val432Leu polymorphism and breast cancer risk: a meta-analysis involving 40, 303 subjects. Breast Cancer Res Treat 2010;122:237e42. [15] Shimada T, Watanabe J, Inoue K, Guengerich FP, Gillam EMJ. Specificity of 17bestradiol and benzo(a)pyrene oxidation by polymorphic human cytochrome P4501B1 variants substituted at residues 48, 119 and 432. Xenobiotica 2001;31:163e76. [16] Kulkarni SK, Dhir A. Berberine: a plant alkaloid with therapeutic potential for central nervous system disorders. Phytochem 2010;24:317e24. [17] Grycova L, Dostal J, Marek R. Quaternary protoberberine alkaloids. Phytochem 2007;68:150e7. [18] Chatterjee P, Franklin MR. Human cytochrome P450 inhibition and metabolicintermediate complex formation by goldenseal extract and its methylenedioxyphenyl components. Drug Metab Dispos 2003;31:1391e7. [19] Wu HL, Chuang TY, Al-Hendy A, Diamond MP, Azziz R, Chen YH. Berberine inhibits the proliferation of human uterine leiomyoma cells. Fertil Steril 2015;103:1098e106. [20] Li H, Miyahara T, Tezuka Y, Namba T, Suzuki T, Dowaki R, et al. The effect of Kampo formulae on bone resorption in vitro and in vivo. II detailed study of berberine. Biol Pharm Bull 1999;22:391e6. [21] Lo SN, Chang YP, Tsai KC, Chang CY, Wu TS, Ueng YF. Inhibition of CYP1 by berberine, palmatine, and jatrorrhizine: selectivity, kinetic characterization, and molecular modeling. Toxicol Appl Pharmacol 2013;272: 671e80. [22] Lo SN, Shen CC, Chang CY, Tsai KC, Huang CC, Wu TS, et al. The effect of oxidation on berberine-mediated CYP1 inhibition: oxidative behavior and metabolite-mediated inhibition. Drug Metab Dispos 2015;43:1100e7. [23] Omura T, Sato R. The carbon monoxide-binding pigment of liver microsomes. I. Evidence for its hemoprotein nature. J Biol Chem 1964;239:2370e8. [24] Phillips AH, Langdon RG. Hepatic triphosphopyridine nucleotide-cytochrome c reductase: isolation, characterization, and kinetic studies. J Biol Chem 1962;237:2652e60. [25] Satoh T, Fujita KI, Munakata H, Itoh S, Nakamura K, Kamataki T, et al. Studies on the interactions between drugs and estrogen: analytical method for prediction system of gynecomastia induced by drugs on the inhibitory metabolism of estradiol using Escherichia coli coexpressing human CYP3A4 with human NADPH-cytochrome P450 reductase. Anal Biochem 2000;286: 179e86. [26] Du QS, Wang QY, Du LQ, Chen D, Huang RB. Theoretical study on the polar hydrogen-p (Hp-p) interactions between protein side chains. Chem Cent J 2013;7:92. [27] Jennings BL, George W, Pingili AK, Khan NS, Estes AM, Fang XR, et al. Estrogen metabolism by cytochrome P450 1B1 modulates the hypertensive effect of angiotensin II in female mice. Hypertension 2014;64:134e40. [28] McKay JA, Melvin WT, Ah-See AK, Ewen SWB, Greenlee WF, Marcus CB, et al. Expression of cytochrome P450 CYP1B1 in breast cancer. FEBS Lett 1995;374: 270e2. [29] Shayalova EG, Klein-Szanto AJP, Devarajan K, Cukieman E, Clapper M. Estrogen and cytochrome P450 1B1 contribute to both early- and late-stage head and neck carcinogenesis. Cancer Prev Res 2011;4:107e15. [30] Liehr JG, Ricci MJ. 4-Hydroxylation of estrogens as marker of human mammary tumors. Proc Natl Acad Sci 1996;93:3294e6. [31] Mense SM, Singh B, Remotti F, Liu X, Bhat HK. Vitamin C and a-naphthoflavone prevent estrogen-induced mammary tumors and decrease oxidative stress in female ACI rats. Carcinogenesis 2009;30:1202e8.
Y.-P. Chang et al. / Drug Metabolism and Pharmacokinetics 30 (2015) 374e383 [32] Chun YJ, Kim S, Kim D, Lee SK, Guengerich FP. A new selective and potent inhibitor of human cytochrome P450 1B1 and its application to antimutagenesis. Cancer Res 2001;61:8164e70. [33] Jennings BL, Sahan-Firat S, Estes AM, Das K, Farjana N, Fang XR, et al. Cytochrome P450 1B1 contributes to angiotensin II-induced hypertension and associated pathophysiology. Hypertension 2010;56:667e74. [34] Guo Z, Sun H, Zhang H, Zhang Y. Anti-hypertensive and renoprotective effects of berberine in spontaneously hypertensive rats. Clin Exp Hypertens 2015;37: 332e9.
383
[35] Lan J, Zhao Y, Dong F, Yan Z, Zheng W, Fan J, et al. Meta-analysis of the effect and safety of berberine in the treatment of type 2 diabetes mellitus, hyperlipemia and hypertension. J Ethnopharmacol 2015;161:69e81. [36] Itoh T, Takemura H, Shimoi K, Yamamoto KA. 3D model of CYP1B1 explains the dominant 4-hydroxylation of estradiol. J Chem Inf Model 2010;50:1173e8. [37] Nishida CR, Everett S, Ortiz de Montellano P. Specificity determinants of CYP1B1 estradiol hydroxylation. Mol Pharmacol 2013;84:451e8. [38] Guo Y, Chen Y, Tan ZR, Klassen CD, Zhou HH. Repeated administration of berberine inhibits cytochromes P450 in humans. Eur J Clin Pharmacol 2012;68:213e7.
Contents lists available at ScienceDirect
Drug Metabolism and Pharmacokinetics journal homepage: http://www.journals.elsevier.com/drug-metabolism-andpharmacokinetics
Regular article
Differential inhibition of CYP1-catalyzed regioselective hydroxylation of estradiol by berberine and its oxidative metabolites Yu-Ping Chang a, Chiung-Chiao Huang a, Chien-Chang Shen b, Keng-Chang Tsai c, Yune-Fang Ueng a, d, e, * a
Divisions of Basic Chinese Medicine, Taiwan, ROC Chinese Medicinal Chemistry, Taiwan, ROC c Chinese Materia Medica Development, National Research Institute of Chinese Medicine, Ministry of Health and Welfare, Taipei, Taiwan, ROC d Department and Institute of Pharmacology, National Yang-Ming University, Taipei, Taiwan, ROC e Institute of Medical Sciences, Taipei Medical University, Taipei, Taiwan, ROC b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 16 July 2015 Received in revised form 17 August 2015 Accepted 18 August 2015 Available online 28 August 2015
Berberine is a pharmacologically active alkaloid present in widely used medicinal plants, such as Coptis chinensis (Huang-Lian). The hormone estradiol is oxidized by cytochrome P450 (CYP) 1B1 to primarily form the genotoxic metabolite 4-hydroxyestradiol, whereas CYP1A1 and CYP1A2 predominantly generate 2-hydroxyestradiol. To illustrate the effect of berberine on the regioselective oxidation of estradiol, effects of berberine and its metabolites on CYP1 activities were studied. Among CYP1s, CYP1B1.1, 1.3 (L432V), and 1.4 (N453S)-catalyzed 4-hydroxylation were preferentially inhibited by berberine. Differing from the competitive inhibition of CYP1B1.1 and 1.3, N453S substitution in CYP1B1 allowed a non-competitive or mixed-type pattern. An N228T in CYP1B1 highly decreased its activity and preference to 4-hydroxylation. A reverse mutation of T223N in CYP1A2 retained its 2-hydroxylation preference, but enhanced its inhibition susceptibility to berberine. Compared with berberine, metabolites demethyleneberberine and thalifendine caused weaker inhibition of CYP1A1 and CYP1B1 activities. Unexpectedly, thalifendine was more potent than berberine in the inhibition of CYP1A2, in which case an enhanced interaction through polar hydrogen-p bond was predicted from the docking analysis. These results demonstrate that berberine preferentially inhibits the estradiol 4-hydroxylation activity of CYP1B1 variants, suggesting that 4-hydroxyestradiol-mediated toxicity might be reduced by berberine, especially in tissues/tumors highly expressing CYP1B1.
Keywords: Estradiol CYP1B1 Variants Berberine Thalifendine Demethyleneberberine
Copyright © 2015, The Japanese Society for the Study of Xenobiotics. Published by Elsevier Ltd. All rights reserved.
1. Introduction The physiologically important hormone estradiol (E2) not only plays a crucial role in maintaining normal sexual function, but also has been found to induce hormone-associated tumors in animals used in experiments [1]. E2 was oxidized by cytochrome P450 (P450, CYP) enzymes to primarily form 2- and 4-hydroxylated metabolites (Scheme 1), which could be further oxidized to 2,3- and 3,4quinones, respectively [2]. E2-3,4-quinone reacts with DNA more efficiently than the E2-2,3-quinone does [3]. In breast tissues, the level of 4-hydroxyestradiol (4-OHE2) was about 3e4 times higher than 2-hydroxyestradiol (2-OHE2) in women with breast cancer, and the level of 4-OHE2 was 4-fold higher than in women without breast cancer [4]. The ratio of 4-OHE2 to 2-OHE2 in human breast tumor
* Corresponding author. E-mail address: [email protected] (Y.-F. Ueng).
tissues was about 10-times higher than the ratio in the breast tissues of healthy women who underwent the surgery for reduction mammoplasty [5]. 4-OHE2 stimulated more transformed colony formation than 2-OHE2 in Syrian hamster embryo cells [6]. In MCF7 cells, 4-OHE2 has been reported to stimulate cell proliferation and reduce cytochrome c release [7]. In the same cell line, E2 and 4-OHE2 reduced the apoptotic effect of staurosporine, whereas 2-OHE2 exposure did not influence the staurosporine-induced apoptosis [8]. 4-OHE2 can be a risk factor for abnormal cell growth regulation related to hormone-associated cancers. P450 enzymes including CYP1A1, CYP1A2, CYP1B1, CYP2C9, and CYP3A4 are involved in the oxidation of E2 [9]. CYP1A2, CYP2C9, and CYP3A4 are primarily hepatic P450s, whereas CYP1A1 and CYP1B1 are predominantly expressed in extrahepatic tissues including the uterus and breast [10,11]. 2-OHE2 is the major product generated by most E2-oxidizing P450s, including CYP1A1 and CYP1A2. However, CYP1B1 predominantly carries out a 4-hydroxylation pathway with a specific activity of more than 2-fold greater than that of CYP1A1 [9].
http://dx.doi.org/10.1016/j.dmpk.2015.08.006 1347-4367/Copyright © 2015, The Japanese Society for the Study of Xenobiotics. Published by Elsevier Ltd. All rights reserved.
Y.-P. Chang et al. / Drug Metabolism and Pharmacokinetics 30 (2015) 374e383
375
OH
O
A
O
10
X
N
OH
N
H3CO
9
OCH3
Y
Demethyleneberberine
Berberine: X = Y = OCH3 Thalifendine: X = OH; Y = OCH3
CH3 OH
B CH3 OH
HO HO
HO
4-Hydroxyestradiol
CYP1
2
CH3
NADPH, O2 4
OH
HO
HO
2-Hydroxyestradiol Scheme 1. Structures of the oxidative metabolites of berberine and estradiol.
About 71% of Asians have CYP1B1 with homozygous L432, which has been classified as CYP1B1.1 [12]. CYP1B1.3 carries the L432V substitution. In Caucasian population, the distributions of V/ V, V/L, and L/L are roughly equal [13,14]. The difference between the incidence of hormone cancer in humans with CYP1B1 genotypes CYP1B1*1, CYP1B1*3 (4326C > G) and CYP1B1*4 (4390A > G) remains uncertain [10,12,14]. Gajjar et al. [10] reported that CYP1B1.3 had E2 4-hydroxylation activity 1.6-fold greater than that of CYP1B1.1. However, the kinetic parameters of CYP1B1.1-, CYP1B1.3 (L432V)-, and CYP1B1.4 (N453S) -catalyzed E2 2- and 4hydroxylation have been reported to be similar [13]. Although the kinetic parameters of E2 hydroxylation by CYP1B1 variants showed variations in different expression systems and assay methods [15], there is no doubt that the contribution of CYP1B1.3 and CYP1B1.4 to human E2 4-hydroxylation is as important as CYP1B1.1. Protoberberines represent a group of natural products with a core structure of dihydroisoquinolino[2,1-b]isoquinolin and exhibit a variety of pharmacological activities including anti-cancer and anti-inflammation effects [16,17]. This type of alkaloids can be found in several medicinal plants including Coptis chinensis (Huang-Lian) and Hydrastis canadensis (goldenseal) [16,17]. There are 17e19 mM berberine (Scheme 1) present in the goldenseal extract, which has been widely used as an immunostimulant in several countries including the USA [18]. Berberine suppresses cell survival and E2-induced cell proliferation in human uterine leiomyoma cells, without affecting the cell survival in an immortalized normal smooth muscle cell line [19]. Berberine treatment does not alter the serum estradiol concentration in ovariectomized rats [20]. However, the metabolite profile of estradiol in berberine-treated animals remained unclear. CYP1-catalyzed berberine oxidation generates primarily demethyleneberberine and thalifendine
(Scheme 1). Among the CYP1 members, our previous study demonstrated that berberine preferentially inhibited CYP1B1 when 7-ethoxyresorufin was used as a substrate [21]. The berberine metabolite caused a relatively weak inhibition of CYP1B1 activity. However, the inhibitory effect of berberine had substrate dependency. Activities of CYP1B1 toward alkoxy heterocyclic compounds, 7-ethoxyresorufin, 7-methoxyresorufin and 7ethoxycoumarin were strongly inhibited by berberine with IC50 values <1 mM. However, the hydroxylation activity of CYP1B1 toward the side chain-free polycyclic aromatic hydrocarbon benzo(a) pyrene was weakly inhibited with an IC50 value of 34 mM. Different from the planar structures of alkoxy heterocyclic substrates, E2 has the angular polycyclic rings and only one of the rings is aromatic (Fig. 1). E2 carries the hydrophobic methyl and the hydrophilic hydroxyl side chains. Due to the angular structure, important physiological and toxicological roles, and the regioselective hydroxylation of E2, the effects of berberine on CYP1-catalyzed E2 hydroxylation were investigated in recombinant human enzyme systems. 2. Materials and methods 2.1. Chemicals Berberine chloride, 17b-E2, glucose-6-phosphate, glucose-6phosphate dehydrogenase (from yeast), 2-OHE2, 4-OHE2, 2methoxyestradiol and nicotinamide adenine diphosphate (NADPþ) disodium salt were purchased from SigmaeAldrich Chemical Co. (St. Louis, MO, U.S.A.). Ammonium phosphate, dimethyl sulfoxide (DMSO), ethyl acetate and methanol were purchased from Merck KGaA (Darmstadt, Germany).
376
Y.-P. Chang et al. / Drug Metabolism and Pharmacokinetics 30 (2015) 374e383
Fig. 1. The formation rate ratios of 4-hydroxyestradiol to 2-hydroxyestradiol in the oxidation of estradiol by CYP1A2.1 and CYP1B1 variants. Data represent the mean ± SE of three determinations.
system containing 0.5 mM NADPþ, 5 mM glucose-6-phosphate and 0.5 u/ml glucose-6-phosphate dehydrogenase. The generation of 2- and 4-OHE2 was determined following the HPLC method reported by Satoh et al. [25]. E2 and hydroxyestradiols were dissolved in methanol. In the preparation of standards for quantification, known concentrations of 2- and 4-OHE2 were added to an additional incubation without glucose-6-phosphate dehydrogenase. 2-Methoxyestradiol was added as an internal control of the assay. Aliquots of the ethyl acetate extract were dried under nitrogen, redissolved in methanol, and subjected into a HPLC system equipped with a UV detector (Agilent 1100 series, Agilent Technologies, Santa Clara, USA). A pre-filter (Frit SS, 2 mm, IDEX Health & Science, Oak Harbor, WA, USA) and a C18 guard column (4.3 10 mm, ODS, IDEX Health & Science) were connected prior to the analytical C18 column (4.6 250 mm, I.D.: 5 mm, 5C18-AR-II, COSMOSIL, NA CALAI TESQUE, Inc., Kyoto, Japan). Metabolites were separated with a mobile phase consisting of methanol/50 mM ammonium phosphate, pH 2.5 (1:1, v/v) at a flow rate of 1 ml/min, at room temperature and detected by measuring the absorbance at 280 nm. 2.4. Kinetic analysis of P450 inhibition
Demethyleneberberine and thalifendine were synthesized following methods as described in a previous report [22]. The purities of synthetic metabolites were 98% based on 1H NMR and HPLC analyses. 2.2. Preparation of bacterial membranes expressing polymorphic human CYP1 enzymes Bicistronic human constructs consisting of the modified coding sequence of human CYP1A1*1, CYP1A2*1 and CYP1B1*3 followed by that of human NADPH-P450 reductase were generously provided by Dr. F. Peter Guengerich (Nashville, TN, U.S.A.). Single nucleotide point mutations were introduced into CYP1A2*1 and CYP1B1*3 by the primer-directed enzymatic amplification method following the instruction manual from the QuickChange Lightning site-directed mutagenesis kit (Stratagene, Agilent Technologies Company, La Jolla, CA, USA). Oligonucleotide primer sets used for the mutagenesis were designed with a primer containing a mutated base, and the sequences of mutated constructs were determined [21,22]. CYP1B1*4 and CYP1B1*VS constructs were prepared from a single point mutation from CYP1B1*1 and CYP1B1*3, respectively. CYP1B1NT (N228T) and CYP1A2TN (T223N) were prepared as previously described [21]. P450 constructs were transformed to Escherichia coli DH5a by electroporation (Gene pulser II, Bio-Rad, Hercules, CA, U. S. A.). P450 expression and bacterial membrane preparation were performed as described in previous report [21]. The P450 content was determined following the spectrophotometric methods [23]. The P450 contents of membranes expressing CYP1A1.1, CYP1A2.1, CYP1B1.4, CYP1B1VS (V432S453) and CYP1B1NT were 4.1, 37.1, 13.2, 10.1, 0.83, 3.52 and 0.30 nmol P450/ ml, respectively. The reduction activity of NADPH-P450 reductase (NADPH-cytochrome c reductase) was determined using cytochrome c as a substrate following the method reported by Phillips and Langdon [24]. The cytochrome c reduction activity of P450 reductase in CYP1A1.1, CYP1A2.1, CYP1B1.1, CYP1B1.3, CYP1B1.4, CYP1B1VS and CYP1B1NT were 3.77, 1.02, 2.65, 1.59, 4.55, 2.15 and 2.94 mmol/min/nmol P450, respectively. 2.3. Enzyme assays E2 hydroxylation activities were determined using 10 pmol P450/ml and 50 mM E2 in the presence of a NADPH-generating
In competitive inhibition, kinetic analysis of P450 activity was done following MichaeliseMenten kinetic property. Vmax and I are the maximal velocity and protoberberine concentration, respectively. Ki and KI are the inhibitor constants for the binding of an inhibitor to a P450 and P450-substrate complex, respectively. For noncompetitive inhibition, Ki is equivalent to KI. Values of velocities (v) at various substrate concentrations (S) were fitted by nonlinear least-squares regression without weight according to the MichaeliseMenten equation: competitive inhibition: v ¼ Vmax $ S/{S þ Km[1 þ (I/Ki)]}; noncompetitive inhibition: v ¼ Vmax $ S/[1 þ (I/Ki)][S þ Km]; mixed type of inhibition: v ¼ Vmax $ S/{S[1 þ (I/KI)] þ Km[1 þ (I/Ki)]} (Sigma Plot, Jandel Scientific, San Rafael, CA, U.S.A.). Estimates of variance and the coefficient of variation (CV, %) of the estimates of variance are presented from the analysis of individual sets of data. 2.5. Computer modeling and data analyses Models of berberine and thalifendine in a complex with CYP1A2.1 (PDB ID: 2HI4) and CYP1B1.1 (PDB ID: 3PM0) isoforms were generated as described in a previous report [21]. The docking of inhibitors to the putative active site of CYP1A1 was generated using the crystallographic structure of CYP1A1.1 (PDB ID: 4I8V) complexed with a-naphthoflavone. The concentrations of protoberberines required for 50% inhibition of catalytic activities (IC50) were calculated by curve fitting (start at 0, define end, Grafit, Erithacus Software Ltd., Staines, UK). Estimates of variance (denoted by ±) are presented from the analyses of individual sets of data. 3. Results 3.1. Inhibition of E2 hydroxylation activities of CYP1A1.1, CYP1A2.1, and CYP1B1.1 by berberine Among the wild types of CYP1 enzymes, berberine caused the least inhibitory effect on CYP1A2.1-catalyzed E2 2- and 4hydroxylation with similar IC50 values (Table 1), indicating that the berberine-mediated inhibition did not show regioselectivity. Berberine inhibited the CYP1A1.1-catalyzed 2-hydroxylation with an IC50 value of 4% of the value for CYP1A2.1 inhibition. The 4hydroxylation activity of CYP1A1.1 was too low to determine the IC50 value. The isoform most inhibited by berberine was CYP1B1.1. Berberine potently inhibited the CYP1B1.1-mediated estradiol 4-
Y.-P. Chang et al. / Drug Metabolism and Pharmacokinetics 30 (2015) 374e383
377
Table 1 Inhibition of estradiol hydroxylation activities of CYP1A1.1, CYP1A2.1, and CYP1B1 variants by berberine. P450
Estradiol hydroxylation site
Hydroxylation activity nmol/min/nmol P450
CYP1A1.1
242424242424-
8.57 0.18 6.30 0.27 1.07 2.92 0.58 1.92 1.35 3.82 0.68 2.10
CYP1A2.1 CYP1B1.1 CYP1B1.3 CYP1B1.4 CYP1B1VS
± ± ± ± ± ± ± ± ± ± ± ±
0.16 0.03 0.09 0.04 0.03 0.13 0.01 0.21 0.25 0.10 0.05 0.16
Activity ratio (4-/2-)
IC50, mM
0.02
1.59 n.d. 35.7 40.2 0.56 0.33 0.45 0.20 0.29 0.18 0.28 0.22
0.04 2.7 3.3 2.8 3.1
IC50 ratio (4-/2-)
± 0.23
e
± ± ± ± ± ± ± ± ± ±
1.1
0.9 13.0 0.04 0.02 0.08 0.01 0.07 0.02 0.04 0.03
0.6 0.4 0.6 0.8
CYP1 activities were determined using 10 pmol P450/ml and 50 mM estradiol in an estradiol hydroxylation assay. Estradiol hydroxylation activity was determined as described in the Materials and Methods. Results of activity determinations show the mean ± SEM of 3 individual experiments with duplicate determinations in each experiment. The IC50 values represent estimates and estimated variance (denoted by ±) of the curve fittings. n.d.: not determinable, the activity was too low to determine the IC50 value. The symbol “e” indicates that the ratio could not be calculated.
Table 2 The inhibition kinetic constants of berberine-mediated inhibition of estradiol hydroxylation activities of CYP1B1 variants. Variants
1B1.1 1B1.3 1B1.4 1B1.VS
4-Hydroxylation
2-Hydroxylation
Type of inhibition
Ki, nM
Competitive Competitive Noncompetitive Mixed type
20.2 10.9 167 33.5
± ± ± ±
3.5 2.3 12 3.7
KI, nM
Type of inhibition
Ki, nM
895 ± 153
Competitive Competitive Noncompetitive Competitive
25.1 14.0 250 25.3
± ± ± ±
1.8 2.1 40 2.1
Data represent the mean ± SE of inhibition constants calculated from the activity determination in the presence of three different berberine concentrations. The estimated variance was 6e27% of the parameter estimates generated from non-linear regression of each set of data.
hydroxylation with an IC50 value of 1% of the value for CYP1A2.1 inhibition. However, the IC50 value for the inhibition of 2hydroxylation activity of CYP1B1.1 was two-fold higher than that for the inhibition of 4-hydroxylation, revealing the preferential inhibition of CYP1B1.1-mediated E2 4-hydroxylation by berberine. 3.2. Inhibitory effects of berberine on E2 hydroxylation activities of CYP1B1 variants To determine the inhibitory effect of berberine on the CYP1B1 variants, bacterial membranes expressing CYP1B1.3, CYP1B1.4 and CYP1B1VS were prepared. All CYP1B1 variants showed a 4hydroxylation preference and the activity ratio of 4-hydroxylation to 2-hydroxylation was about 3 (Table 1 and Fig. 1). In the presence of increasing concentrations of berberine, the metabolite formation rate ratio of 4-OHE2 to 2-OHE2 were decreased (Fig. 1). However, in the CYP1A2.1 system, the metabolite formation rate ratio of 4-OHE2 to 2-OHE2 was not affected by berberine. In E2 hydroxylation, 4-hydroxylation activities of CYP1B1.1, CYP1B1.3 and CYP1B1.4 were potently inhibited by berberine. The 2hydroxylation activities of CYP1B1.4 and CYP1B1VS variants carrying an N453S substitution were inhibited more potently than CYP1B1.1 and CYP1B1.3, resulting in an increase in the IC50 ratio of 4-hydroxylation to 2-hydroxylation, especially in CYP1B1VS (Table 2).
concentrations in E2 2- and 4-hydroxylation catalyzed by CYP1B1 variants showed a hyperbolic property (Fig. 2). The Km values for 2hydroxylation by CYP1B1.1, CYP1B1.3, CYP1B1.4 and CYP1B1VS were 2.7 ± 0.3, 2.4 ± 0.3, 11.1 ± 2.8 and 4.8 ± 0.5 mM, respectively. The Km values for the 4-hydroxylation by CYP1B1.1, CYP1B1.3, CYP1B1.4 and CYP1B1VS were 2.9 ± 0.6, 2.4 ± 0.5, 6.0 ± 1.1 and 3.4 ± 0.5 mM, respectively. The Vmax values for 2-hydroxylation by CYP1B1.1, CYP1B1.3, CYP1B1.4 and CYP1B1VS were 0.61 ± 0.01, 0.54 ± 0.02, 2.02 ± 0.18 and 1.17 ± 0.03 nmol/min/nmol P450, respectively. The Vmax values for 4-hydroxylation were 3.26 ± 0.16, 1.79 ± 0.11, 5.37 ± 0.29, and 2.79 ± 0.10 nmol/min/nmol P450, respectively. The intrinsic clearance (Clin ¼ Vmax/Km) values of E2 through 4-hydroxylation by CYP1B1.1, CYP1B1.3, CYP1B1.4, and CYP1B1VS were 1.14, 0.75, 0.90 and 0.82 ml/min/nmol P450, respectively. Linear regression analyses of the LineweavereBurk plots revealed that berberine competitively inhibited CYP1B1.1and CYP1B1.3-catalyzed E2 2- and 4-hydroxylation (Fig. 3A and B). Nonlinear regression analysis generated Ki values as listed in Table 2. In the CYP1B1.1 and CYP1B1.3 enzyme systems, the Ki for the inhibition of 2-hydroxylation was higher than that for the inhibition of 4-hydroxylation. However, noncompetitive and mixedtyped inhibition properties were present in the inhibition of CYP1B1.4 and CYP1B1VS carrying an N453S substitution (Fig. 3C and D), suggesting the influence of N453S in the inhibition pattern of berberine.
3.3. Kinetic analysis of the inhibitory behavior of berberine on E2 hydroxylation by CYP1B1 variants
3.4. Inhibition of E2 hydroxylation activities of CYP1B1NT and CYP1A2TN by berberine
Effects of berberine on hydroxylation by CYP1B1 variants were determined to be due to the potent inhibition of CYP1B1.1catalyzed 4-hydroxylation by berberine when estradiol was used as a substrate. The plots of the velocity versus substrate
Our previous report reveals that Asn228 is a crucial amino acid residue of CYP1B1 for its oxidation activity toward 7ethoxyresorufin and its susceptibility to inhibition by berberine [21]. The corresponding amino acid in CYP1A2.1 is Thr223, which
378
Y.-P. Chang et al. / Drug Metabolism and Pharmacokinetics 30 (2015) 374e383
Fig. 2. Plots of velocity versus estradiol concentration of the inhibition of estradiol 2- and 4-hydroxylation activities of CYP1B1.1 (LN) (A), CYP1B1.3 (VN) (B), CYP1B1.4 (LS) (C), and CYP1B1VS (D) by berberine. The left and right panels show the plots of velocity versus estradiol concentration in the CYP1B1-catalyzed estradiol 2- and 4-hydroxylation, respectively. Estradiol hydroxylation velocity was determined in an assay containing 10 pmol P450/ml. The data represent the mean ± SE of three determinations.
did not allow hydrogen bond interaction with the methoxyl moiety in berberine. The substitution of T223N (CYP1A2TN) clearly enhanced the inhibition susceptibility of CYP1A2.1 to berberine. Thus, estradiol hydroxylation activities of CYP1A2TN and CYP1B1NT (N228T) were determined to examine the influence of substitution on regioselective hydroxylation by CYP1A2 and CYP1B1 and CYP1B1-preferential inhibition by berberine. Compared with the wild type enzyme, CYP1A2TN and CYP1B1NT had low E2 hydroxylation activities (Table 3). CYP1A2TN remained the preference in 2-hydroxylation toward estradiol. T223N substitution in CYP1A2.1 enhanced the inhibition susceptibility of CYP1A2.1 to berberine. CYP1B1NT catalyzed oxidation generated 2hydroxylated and 4-hydroxylated metabolites at nearly equal velocities. The activity of CYP1B1NT was too low to allow us to determine the IC50 value of berberine-mediated inhibition. To examine the detrimental changes of functional CYP1B1 protein by
N228T substitution, the activity toward another substrate 7-ethoxycoumarin was determined. 7-Ethoxycoumarin O-deethylation activity was decreased to 5% of the wild type (data not shown). The extremely low activity of CYP1B1NT toward different substrates revealed that Asn228 was not only important for the interaction with berberine but also very important for maintaining the function of CYP1B1. 3.5. Inhibition of CYP1A1.1, CYP1A2.1, and CYP1B1.1-catalyzed E2 hydroxylation by demethyleneberberine and thalifendine Demethyleneberberine and thalifendine inhibited the 2hydroxylation activity of CYP1A1 with IC50 values of 13.4 ± 0.7 and 6.6 ± 1.3 mM, respectively (Fig. 4). Demethyleneberberine and thalifendine inhibited the 2-hydroxylation activity of CYP1A2.1 with IC50 values of 92.9 ± 17.2 and 23.2 ± 2.0 mM, respectively. E2 2-
Y.-P. Chang et al. / Drug Metabolism and Pharmacokinetics 30 (2015) 374e383
379
Fig. 3. LineweavereBurk plots of the inhibition of estradiol 2- and 4-hydroxylation activities of CYP1B1.1 (LN) (A), CYP1B1.3 (VN) (B), CYP1B1.4 (LS) (C), and CYP1B1VS (D) by berberine. The left and right panels show the plots of velocity versus estradiol concentration in CYP1B1-catalyzed estradiol 2- and 4-hydroxylation, respectively. The estradiol hydroxylation velocity was determined in an assay containing 10 pmol P450/ml. The data represent the reciprocals of the means of three determinations shown in Fig. 2.
Table 3 The activities and berberine-mediated inhibition of estradiol hydroxylation by CYP1A2TN and CYP1B1NT. P450 mutants
Estradiol hydroxylation
Hydroxylation activity nmol/min/nmol P450
CYP1A2TN
2-hydroxylation 4-hydroxylation 2-hydroxylation 4-hydroxylation
1.63 0.20 0.15 0.14
CYP1B1NT
± ± ± ±
0.09 0.01 0.01 0.03
Activity ratio (4-/2-)
IC50, mM
0.13
2.6 ± 0.2 n.d. n.d. n.d.
0.9
Estradiol hydroxylation activity was determined as described in Table 1. Data from the activity represent the means ± SE of 3 determinations. n.d.: not determinable, the activity was too low to determine the IC50 value.
hydroxylation activities of CYP1A1.1 and CYP1A2.1 were less inhibited by demethyleneberberine than by berberine. However, in CYP1A2-catalyzed 2-hydroxylation, activity was more sensitive to the inhibition by thalifendine than by berberine. Demethyleneberberine and thalifendine inhibited CYP1B1-catalyzed 2-
hydroxylation with IC50 values of 19.2 ± 2.5 and 13.7 ± 1.2 mM, respectively. Demethyleneberberine and thalifendine inhibited CYP1B1-catalyzed 4-hydroxylation with IC50 values of 9.9 ± 3.5 and 4.1 ± 0.6 mM, respectively. The IC50 values for the inhibition of CYP1B1-catalyzed E2 hydroxylation by demethyleneberberine and
380
Y.-P. Chang et al. / Drug Metabolism and Pharmacokinetics 30 (2015) 374e383
thalifendine may form polar hydrogen-p bond interaction [26] with the phenyl moiety of Phe256 of CYP1A2.1 (Fig. 5B). Unlike CYP1A2.1, the corresponding amino acid residues were Leu254 in CYP1A1.1 and Leu264 in CYP1B1.1, in which the potential polar hydrogen-p bond interaction was unfavorable (Fig. 5C and D). Using PLP1 scoring, the scores for the docking of thalifendine and berberine to CYP1A2.1 were 27.49 and 10.08, respectively. 4. Discussion
Fig. 4. Inhibition of CYP1A1.1, CYP1A2.1, and CYP1B1.1-catalyzed estradiol 2- and 4hydroxylation using demethyleneberberine (B), thalifendine ( ), and berberine (-). Berberine chloride, demethyleneberberine, and thalifendine were dissolved in dimethyl sulfoxide and a same volume of dimethyl sulfoxide was added to the control assay. The data represent the mean ± SE of three determinations.
▵
thalifendine were 2e34 times greater than the values for the inhibition by berberine. Compared with the inhibition by berberine, demethyleneberberine and thalifendine generally caused fewer inhibitory effects on the E2 hydroxylation activity of CYP1 isoforms but the CYP1A2.1 activity was inhibited by thalifendine with a greater potency. 3.6. Docking of berberine and thalifendine into CYP1A1.1, CYP1A2.1, and CYP1B1.1 Since the inhibitory effect of thalifendine was more potent than that of berberine on E2 hydroxylation by CYP1A2.1, but not for CYP1A1.1 and CYP1B1.1, computer modeling of the binding of thalifendine into CYP1 enzymes was performed. The binding of berberine in the putative active site of CYP1A2.1 was superimposed with the binding of E2 in an orientation favorable for the 2hydroxylation pathway (Fig. 5A). The 10-hydroxyl moiety of
The non-planar polycyclic steroid, E2, plays a crucial role in the modulation of not only the development of sexual organs, but also the pathological events unrelated to the sexual function, such as tumorigenesis and blood pressure regulation [1,27]. An elevated level of CYP1B1 has been reported in a variety of tumors, especially hormone-related cancers including breast and endometrial cancers [10,28,29]. Microsomal E2 4-hydroxylation activity of breast tumor tissues was about 4-fold greater than that from normal tissues [30]. Selective inhibition of CYP1B1-catalyzed E2 hydroxylation, especially preferential inhibition of a 4-hydroxylation pathway, may prove beneficial in the chemoprotection against estrogenassociated cancers. A non-selective CYP1A2 and CYP1B1 inhibitor a-naphthoflavone reduced the breast tumor incidence in female August Copenhagen Irish (ACI) rats with long term exposure to E2, suggesting a potential chemopreventive effect of a-naphthoflavone [31]. A synthetic 2,30 ,4,50 -tetramethoxystilbene has been reported to preferentially inhibit CYP1B1-catalyzed 7-ethoxyresorufin Odeethylation and E2 hydroxylation activities [32]. In E2 4hydroxylation, berberine inhibited CYP1B1 activity with an IC50 value within the range of IC50 values for the inhibition by 2,30 ,4,50 tetramethoxystilbene (IC50: 0.09 mM in the recombinant enzyme system expressed in bacterial membranes and 0.39 mM in a reconstituted system of purified enzyme) [32]. Although CYP1B1catalyzed E2 hydroxylation generates mainly the carcinogenic metabolite 4-OHE2 [30], more evidence reveals that extrahepatic CYP1B1 may play an important role in the regulation of physiological response, such as blood pressure [33]. Angiotensin IIinduced hypertension was more pronounced in male Cyp1b1þ/þ mice than in Cyp1b1/ mice. CYP1B1-catalyzed testosterone hydroxylation has been suggested to be involved in the hypertensive effect in Cyp1b1þ/þ mice. A potent CYP1B1 inhibitor 2,30 ,4,50 tetramethoxy stilbene ameliorates the hypertension status and renal injury caused by angiotensin II-treatment in male rats and mice. However, 4-OHE2 but not 2-OHE2 enhanced angiotensin IIinduced hypertension in female Cyp1b1þ/þ mice [27]. Berberine ameliorated the severity of hypertension and hypertensioninduced renal damage in male spontaneously hypertensive rats [34]. However, the effect of berberine in female rats has not yet been reported. Meta-analysis of 27 randomized controlled clinical trials, in which male and female patients were included, showed that berberine may be beneficial for the hypotensive effects [35]. The decreased E2 4-hydroxylation by berberine suggests its potential chemopreventive effect against estrogen-related cancers. Berberine may become a lead candidate for further development of agents for the protection against 4-OHE2-associated toxicity. On the other hand, the influence of hormone homeostasis on blood pressure affected by berberine should be noted. Different from a similar O-deethylation product, resorufin generated from CYP1A1/2- and CYP1B1-catalyzed oxidation toward 7-ethoxyresorufin and 7-methoxyresorufin, docking results suggested that E2 preferentially binds to CYP1A1/2 and CYP1B1 in a distinct orientation to generate primarily the 2-hydroxylation and 4-hydroxylation metabolite, respectively [36]. Several amino acid residues have been suggested or proved to be important for the regioselective metabolism of E2 by CYP1 enzymes. Nishida et al.
Y.-P. Chang et al. / Drug Metabolism and Pharmacokinetics 30 (2015) 374e383
381
Fig. 5. The docking of estradiol, berberine and thalifendine to the putative active sites of CYP1A1.1, CYP1A2.1 and CYP1B1.1. Panel A shows the binding of estradiol (blue) and berberine (yellow thin sticks) superimposed within the active site of CYP1A2.1 based on molecular modeling [21,36]. Panels B, C and D show the binding of thalifendine (yellow thick sticks) to CYP1A2.1, CYP1A1.1 and CYP1B1.1, respectively. The side chains of amino acid residues of CYP1B1.1 interacting with berberine are shown in light gray. The nitrogen and oxygen atoms are shown in dark blue and red, respectively. The heme prosthetic group of P450 is shown in magenta and iron is shown as white.
[37] successfully showed that the formation rate ratio of 4-OHE2/2OHE2 was shifted from 5 to 0.5 by substituting Val395Leu. This demonstrated that Val395 was important for CYP1B1 in a 4hydroxylation pathway. Based on the scoring results of active site docking, Asn265 and Ala133 (Fig. 4) have also been suggested to be important for E2 recognition by CYP1B1 [36]. In our previous study, when we used 7-ethoxyresorufin and caffeine as substrates, CYP1A2.1 was resistant to inhibition by berberine (IC50 > 60 mM) [21]. Although the IC50 value was still high and may not occur at the pharmacological concentration of berberine in patients, current results show that CYP1A2.1-catalyzed E2 2-hydroxylation activity was more sensitive than its activities toward 7-ethoxyresorufin and caffeine to the inhibition by berberine. Previously, our findings revealed that the substitution of T223N in CYP1A2.1 increased the sensitivity of CYP1A2.1 to inhibition by berberine. In this report, results revealed that the substitution of T223N in CYP1A2.1 not only enhanced the inhibition sensitivity of CYP1A2.1 to berberine, but also resulted in the increase of the formation rate ratio of 4-OHE2/2OHE2. In E2 hydroxylation, the substitution of Asn228 with Thr in CYP1B1 dramatically reduced activity by 70%e90% and the production of 2-OHE2 and 4-OHE2 by CYP1B1NT became equal. Like the E2 hydroxylation activity, activities of CYP1B1NT toward 7ethoxyresorufin [21] and 7-ethoxycoumarin (data not shown) were less than 30% of CYP1B1.1 activities. This marked decrease in CYP1B1 activity revealed that Asn228 is not only important for the berberine-mediated CYP1B1 inhibition, but very important for its catalytic function and the 4-hydroxylation preference of CYP1B1. Consistent with the summary results reported by Shimada et al. [15], determination of the formation rate ratios of 4-OHE2 to 2OHE2 by CYP1B1 variants and mutants indicated that the amino acid substitutions Leu432Val and Asn453Ser did not affect the
preference of CYP1B1 in the 4-hydroxylation pathway. However, the substitution of N453S tended to increase the inhibitory effect of berberine on the E2 2-hydroxylation activity of CYP1B1 carrying either Leu432 or Val432. These results were unexpected since Asn453 is located on the opposite side of the substrate binding pocket of CYP1B1. The Km values for the 4-hydroxylation using CYP1B1.1, CYP1B1.3, CYP1B1.4 and CYP1B1VS were 3.5e14, 3.8e40, 7e11 and 3.7e17 mM, respectively [13,15]. The Vmax values for the 4hydroxylation using CYP1B1.1, CYP1B1.3, CYP1B1.4, and CYP1B1VS were 0.91e3.7, 1.45e3.3, 1.5e3.6, and 1.8e4.5 nmol/min/nmol P450, respectively [13,15]. Using 10 pmol P450/ml in the assay, the kinetic analysis generated parameters within the range of reported values (Supplementary Table 1). Any differences can be attributed to not only the distinct recombinant enzyme systems as discussed in the report of Shimada et al. [15], but also the P450 concentrations in the assay. Results from the inhibition kinetic analysis showed that the berberine inhibited CYP1B1.1 and CYP1B1.3 in a competitive manner, whereas CYP1B1.4 and CYP1B1VS carrying an N453S substitution could be inhibited by berberine in a non-competitive or mixed type manner. An N453S substitution may change the intra-protein interactions, leading to the differential inhibition behavior of berberine. Consistent with reports that CYP1B1 had a lower formation rate difference (3-fold) in 4-OHE2 and 2-OHE2 than CYP1A1/2 (>25fold) did, the docking results indicated that the oxidation pocket of CYP1B1 was rounder than for CYP1A1/2 [36]. Orientation of E2 in the CYP1B1 putative active site may be more flexible than in CYP1A1/2. The IC50 value for berberine-mediated inhibition of the 4-hydroxylation pathway was about half that of the 2hydroxylation pathway. The presence of berberine resulted in a concentration-dependent decrease in the metabolite formation
382
Y.-P. Chang et al. / Drug Metabolism and Pharmacokinetics 30 (2015) 374e383
rate ratio of 4-OHE2 to 2-OHE2. The activity of CYP1B1 through two hydroxylation pathways was almost equal in the presence of 3e5 mM berberine. Our previous report revealed that berberine was poorly oxidized by CYP1B1, generating mainly demethyleneberberine and thalifendine [22]. Using 7-ethoxyresorufin as the substrate, the inhibitory effect on CYP1A2.1 might be too weak (IC50 > 60 mM) to differentiate the inhibition by berberine or thalifendine. However, using E2 as a substrate, our findings indicated that thalifendine caused CYP1A2.1 inhibition with a potency greater than that of berberine. The inhibition of CYP1A1-and CYP1B1-mediated E2 hydroxylations by demethyleneberberine and thalifendine was not as potent as the inhibition by berberine. Docking results suggested that the hydroxyl group at C10 may present polar hydrogen-p interaction with the phenyl moiety of the Phe256 of CYP1A2.1. However, the corresponding amino acid residues Leu254 in CYP1A1 and Leu264 in CYP1B1 did not allow this interaction. In summary, berberine preferentially inhibited E2 4hydroxylation catalyzed by CYP1B1 variants. However, the inhibition of CYP1B1 variants carrying N453S showed an inhibition kinetic pattern different from the wild type. Docking of thalifendine to the putative active site of CYP1 isoforms suggested that a polar hydrogen-p bond might form between the 10-hydroxyl group of thalifendine and Phe256 in CYP1A2.1, leading to a stronger inhibition by thalifendine than by berberine, even though the inhibition of CYP1B1 variants by berberine was stronger. Berberine may provide benefit in the reduction of 4-OHE2-induced toxicity, especially in tissues or tumors expressing high levels of CYP1B1. The benefit of berberine in the prevention of 4-OHE2-mediated toxicity needs further investigation in vivo. On the other hand, berberine has been reported to inhibit human liver microsomal CYP2D6 and CYP3A4 activities, but the IC50 values (45 and 400 mM, respectively) were relatively high [18]. Repeated administration of 300 mg berberine t.i.d. for 2 weeks has been reported to increase the AUC012 and AUC0-∞ values of CYP3A4 substrate midazolam by ~40% in healthy participants [38]. Thus, potential drug-interaction should be considered in the dosing regimen of berberine for chemoprotection against the toxicity of 4-OHE2. Acknowledgments We truly appreciate Dr. Guengerich for kindly providing us with the P450 constructs. This work was mainly supported by Grant MOST104-2320-B-077-001 from Ministry of Science and Technology (MOST), Taipei and partially supported by Grant NSC102-2923B-077-001-MY3 from National Science Council (now MOST), Taipei, and the National Research Institute of Chinese Medicine, Taipei. Appendix A. Supplementary data Supplementary data related to this article can be found online at http://dx.doi.org/10.1016/j.dmpk.2015.08.006. References [1] Roy D, Cai Q, Felty Q, Narayan S. Estrogen-induced generation of reactive oxygen and nitrogen species, gene damage, and estrogen-dependent cancers. J Toxicol Environ Health B 2007;10:235e57. [2] Tsuchiya Y, Nakajima M, Yokoi T. Cytochrome P450-mediated metabolism of estrogens and its regulation in human. Cancer Lett 2005;227:115e24. [3] Zahid M, Kohli E, Saeed M, Rogan E, Cavalieri E. The greater reactivity of estradiol-3,4-quinone vs estradiol-2,3-quinone with DNA in the formation of depurinating adducts: implications for tumor-initiating activity. Chem Res Toxicol 2006;19:164e72. [4] Rogan EG, Badawi AF, Devanesan PD, Meza JL, Edney JA, West WW, et al. Relative imbalance in estrogen metabolism and conjugation in breast tissue of women with carcinoma: potential biomarkers of susceptibility to cancer. Carcinogenesis 2003;24:697e702.
[5] Castagnetta L, Granata OM, Cocciadiferro L, Saetta A, Polito L, Bronte G, et al. Sex steroids, carcinogenesis, and cancer progression. Ann N Y Acad Sci 2004;1024:233e46. [6] Tsutsui T, Tamura Y, Yagi E, Barrett JC. Involvement of genotoxic effects in the initiation of estrogen-induced cellular transformation: studies using Syrian Hamster embryo cells treated with 17b-estradiol and eight of its metabolites. Int J Cancer 2000;86:8e14. [7] Seeger H, Wallwiener D, Kraemer E, Mueck AO. Comparison of possible carcinogenic estradiol metabolites: effects on proliferation, apoptosis and metastasis of human breast cancer cells. Maturitas 2006;54:72e7. [8] Gregoraszuk E, Ptak A. Involvement of caspase-9 but not caspase-8 in the antiapoptotic effects of estradiol and 4-OH-estradiol in MCF-7 human breast cancer cells. Endocr Regul 2011;45:3e8. [9] Lee AJ, Cai MX, Thomas PE, Conney AH, Zhu BT. Characterization of the oxidative metabolites of 17b-estradiol and estrone formed by 15 selectively expressed human cytochrome P450 isoforms. Endocrinology 2003;144: 3382e98. [10] Gajjar K, Martin-Hirsch PL, Martin FL. CYP1B1 and hormone-induced cancer. Cancer Lett 2012;324:13e30. [11] Singh MN, Stringfellow HF, Taylor SE, Ashton KM, Ahmad M, Abdo KR, et al. Elevated expression of CYP1A1 and g-SYNUCLEIN in human ectopic (ovarian) endometriosis compared with eutopic endometrium. Mol Hum Reprod 2008;14:655e63. [12] Paracchini V, Raimondi S, Gram IT, Kang D, Kocabas NA, Kristensen VN, et al. Meta- and pooled analyses of the cytochrome P-450 1B1 Val432Leu polymorphism and breast cancer: a HuGE-GSEC review. Am J Epidemiol 2006;165: 115e25. [13] Aklillu E, Oscarson M, Hidestrand M, Leidvik B, Otter C, Ingelman-Sunberg M. Functional analysis of six different polymorphic CYP1B1 enzyme variants found in an Ethiopian population. Mol Pharmacol 2002;61:586e94. [14] Yao L, Fang F, Wu Q, Zhong Y, Yu L. No association between CYP1B1 Val432Leu polymorphism and breast cancer risk: a meta-analysis involving 40, 303 subjects. Breast Cancer Res Treat 2010;122:237e42. [15] Shimada T, Watanabe J, Inoue K, Guengerich FP, Gillam EMJ. Specificity of 17bestradiol and benzo(a)pyrene oxidation by polymorphic human cytochrome P4501B1 variants substituted at residues 48, 119 and 432. Xenobiotica 2001;31:163e76. [16] Kulkarni SK, Dhir A. Berberine: a plant alkaloid with therapeutic potential for central nervous system disorders. Phytochem 2010;24:317e24. [17] Grycova L, Dostal J, Marek R. Quaternary protoberberine alkaloids. Phytochem 2007;68:150e7. [18] Chatterjee P, Franklin MR. Human cytochrome P450 inhibition and metabolicintermediate complex formation by goldenseal extract and its methylenedioxyphenyl components. Drug Metab Dispos 2003;31:1391e7. [19] Wu HL, Chuang TY, Al-Hendy A, Diamond MP, Azziz R, Chen YH. Berberine inhibits the proliferation of human uterine leiomyoma cells. Fertil Steril 2015;103:1098e106. [20] Li H, Miyahara T, Tezuka Y, Namba T, Suzuki T, Dowaki R, et al. The effect of Kampo formulae on bone resorption in vitro and in vivo. II detailed study of berberine. Biol Pharm Bull 1999;22:391e6. [21] Lo SN, Chang YP, Tsai KC, Chang CY, Wu TS, Ueng YF. Inhibition of CYP1 by berberine, palmatine, and jatrorrhizine: selectivity, kinetic characterization, and molecular modeling. Toxicol Appl Pharmacol 2013;272: 671e80. [22] Lo SN, Shen CC, Chang CY, Tsai KC, Huang CC, Wu TS, et al. The effect of oxidation on berberine-mediated CYP1 inhibition: oxidative behavior and metabolite-mediated inhibition. Drug Metab Dispos 2015;43:1100e7. [23] Omura T, Sato R. The carbon monoxide-binding pigment of liver microsomes. I. Evidence for its hemoprotein nature. J Biol Chem 1964;239:2370e8. [24] Phillips AH, Langdon RG. Hepatic triphosphopyridine nucleotide-cytochrome c reductase: isolation, characterization, and kinetic studies. J Biol Chem 1962;237:2652e60. [25] Satoh T, Fujita KI, Munakata H, Itoh S, Nakamura K, Kamataki T, et al. Studies on the interactions between drugs and estrogen: analytical method for prediction system of gynecomastia induced by drugs on the inhibitory metabolism of estradiol using Escherichia coli coexpressing human CYP3A4 with human NADPH-cytochrome P450 reductase. Anal Biochem 2000;286: 179e86. [26] Du QS, Wang QY, Du LQ, Chen D, Huang RB. Theoretical study on the polar hydrogen-p (Hp-p) interactions between protein side chains. Chem Cent J 2013;7:92. [27] Jennings BL, George W, Pingili AK, Khan NS, Estes AM, Fang XR, et al. Estrogen metabolism by cytochrome P450 1B1 modulates the hypertensive effect of angiotensin II in female mice. Hypertension 2014;64:134e40. [28] McKay JA, Melvin WT, Ah-See AK, Ewen SWB, Greenlee WF, Marcus CB, et al. Expression of cytochrome P450 CYP1B1 in breast cancer. FEBS Lett 1995;374: 270e2. [29] Shayalova EG, Klein-Szanto AJP, Devarajan K, Cukieman E, Clapper M. Estrogen and cytochrome P450 1B1 contribute to both early- and late-stage head and neck carcinogenesis. Cancer Prev Res 2011;4:107e15. [30] Liehr JG, Ricci MJ. 4-Hydroxylation of estrogens as marker of human mammary tumors. Proc Natl Acad Sci 1996;93:3294e6. [31] Mense SM, Singh B, Remotti F, Liu X, Bhat HK. Vitamin C and a-naphthoflavone prevent estrogen-induced mammary tumors and decrease oxidative stress in female ACI rats. Carcinogenesis 2009;30:1202e8.
Y.-P. Chang et al. / Drug Metabolism and Pharmacokinetics 30 (2015) 374e383 [32] Chun YJ, Kim S, Kim D, Lee SK, Guengerich FP. A new selective and potent inhibitor of human cytochrome P450 1B1 and its application to antimutagenesis. Cancer Res 2001;61:8164e70. [33] Jennings BL, Sahan-Firat S, Estes AM, Das K, Farjana N, Fang XR, et al. Cytochrome P450 1B1 contributes to angiotensin II-induced hypertension and associated pathophysiology. Hypertension 2010;56:667e74. [34] Guo Z, Sun H, Zhang H, Zhang Y. Anti-hypertensive and renoprotective effects of berberine in spontaneously hypertensive rats. Clin Exp Hypertens 2015;37: 332e9.
383
[35] Lan J, Zhao Y, Dong F, Yan Z, Zheng W, Fan J, et al. Meta-analysis of the effect and safety of berberine in the treatment of type 2 diabetes mellitus, hyperlipemia and hypertension. J Ethnopharmacol 2015;161:69e81. [36] Itoh T, Takemura H, Shimoi K, Yamamoto KA. 3D model of CYP1B1 explains the dominant 4-hydroxylation of estradiol. J Chem Inf Model 2010;50:1173e8. [37] Nishida CR, Everett S, Ortiz de Montellano P. Specificity determinants of CYP1B1 estradiol hydroxylation. Mol Pharmacol 2013;84:451e8. [38] Guo Y, Chen Y, Tan ZR, Klassen CD, Zhou HH. Repeated administration of berberine inhibits cytochromes P450 in humans. Eur J Clin Pharmacol 2012;68:213e7.