Structural Identification of Two Metabolites of Catechins and Their

Occupational Health Sciences Institute, Piscataway, New Jersey 08854, and ... in the urine and plasma of human volunteers after ingestion of green tea...
0 downloads 0 Views 210KB Size
Download PDF
Chem. Res. Toxicol. 2000, 13, 177-184

177

Structural Identification of Two Metabolites of Catechins and Their Kinetics in Human Urine and Blood after Tea Ingestion Chuan Li,† Mao-Jung Lee,† Shuqun Sheng,‡ Xiaofeng Meng,† Saileta Prabhu,† Bozena Winnik,§ Baoming Huang,§ Jee Y. Chung,† Shunqi Yan,‡ Chi-Tang Ho,| and Chung S. Yang*,† Laboratory for Cancer Research, College of Pharmacy, and Department of Chemistry, Rutgers, The State University of New Jersey, Piscataway, New Jersey 08854, Environmental and Occupational Health Sciences Institute, Piscataway, New Jersey 08854, and Department of Food Science, Rutgers, The State University of New Jersey, New Brunswick, New Jersey 08901 Received October 28, 1999

Tea is a popular beverage consumed worldwide. The metabolic fate of its major constituents, catechins, however, is not well-known. In this study, two catechin metabolites were detected in the urine and plasma of human volunteers after ingestion of green tea. These metabolites were identified by LC/ESI-MS and NMR as (-)-5-(3′,4′,5′-trihydroxyphenyl)-γ-valerolactone (M4) and (-)-5-(3′,4′-dihydroxyphenyl)-γ-valerolactone (M6). The renal excretion of M4 and M6 had a 3 h lag time and peaked 7.5-13.5 h after ingestion of a single dose of green tea, while (-)-epigallocatechin (EGC) and (-)-epicatechin peaked at 2 h. M4 and M6 were two major tea metabolites with urinary cumulative excretions as high as 8-25 times the levels of EGC and (-)-epicatechin in some of our subjects, and accounted for 6-39% of the amounts of ingested EGC and (-)-epicatechin. Both the metabolites appeared to be produced by intestinal microorganisms, with EGC and (-)-epicatechin as the precursors of M4 and M6, respectively. Repeated ingestion of green tea produced a slight accumulative effect of the metabolites. They were also detected in the plasma, exhibiting kinetics similar to those of the urinary metabolites, and in the feces. Study on these metabolites may help us further understand the cancer chemopreventive actions and other beneficial effects of tea.

Introduction Tea is the fragrant brew prepared from the dried leaves of Camellia sinensis. Originating in China, tea has been cultivated and consumed for more than a millennium and is now one of the most popular beverages worldwide. Tea has attracted a great deal of attention not only for its pleasant flavor but also for its possible beneficial health effects. Studies in many animal models have demonstrated that tea and tea polyphenols can inhibit carcinogenesis (1-4). In contrast to the laboratory studies, however, the effect of tea consumption on human cancers is inconclusive (1-3, 5-7). To understand the relationship between tea consumption and cancer, a good understanding of the bioavailability and metabolism of tea components is needed. Because of their antioxidative activity (8), tea polyphenols are believed to play an important role in the anticancer property of tea. The major green tea poly* To whom correspondence should be addressed: Laboratory for Cancer Research, College of Pharmacy, Rutgers, The State University of New Jersey, 164 Frelinghuysen Rd., Piscataway, NJ 08854-8020. Phone: (732) 445-5361. Fax: (732) 445-0687. E-mail: [email protected]. † Laboratory for Cancer Research, College of Pharmacy, Rutgers, The State University of New Jersey. ‡ Department of Chemistry, Rutgers, The State University of New Jersey. § Environmental and Occupational Health Sciences Institute. | Department of Food Science, Rutgers, The State University of New Jersey.

phenols are four derivatives of catechin (flavan-3-ol), i.e., (-)-epigallocatechin (EGC),1 (-)-epicatechin, (-)-epigallocatechin-3-gallate (EGCG), and (-)-epicatechin-3-gallate (ECG) (9). These compounds account for about 25% of the dry weight of tealeaves. Black tea is made by crushing the tealeaves, allowing the polyphenol oxidasecatalyzed oxidation and condensation of catechins to take place. There are still considerable amounts of catechins left, accounting for 3-10% of the black tea solids (10). In our previous studies, we established methods for the analysis of catechins in human blood and urine (11). The bioavailability of tea catechins following oral administration of tea or tea catechins has been demonstrated in our laboratory (11-15) and others (16-18). We observed peak plasma concentrations of 148, 55, and 119 ng/mL for EGC, (-)-epicatechin, and EGCG, respectively, in human subjects between 1.5 and 2.5 h after consumption of 1.5 g of decaffeinated green tea (12). The half-life for EGCG was about 5 h, while those for EGC and (-)-epicatechin were about 3 h. Most of these plasma catechins were found in the form of glucuronide and sulfate conjugates (11). Urinary excretion of EGC and (-)-epicatechin (often in the glucuronide and sulfate-conjugated form) took place mostly within 8 h of tea ingestion, after which the suspected metabolites were excreted (10, 13). The sus1 Abbreviations: EGC, (-)-epigallocatechin; EGCG, (-)-epigallocatechin-3-gallate; ECG, (-)-epicatechin-3-gallate; ESI, electrospray ionization; CID, collision-induced dissociation; SRM, selected reaction monitoring.

10.1021/tx9901837 CCC: $19.00 © 2000 American Chemical Society Published on Web 02/24/2000

178

Chem. Res. Toxicol., Vol. 13, No. 3, 2000

pected tea metabolites, the chromatographic peaks of which were labeled peaks 4 and 6 (13), are referred to as M4 and M6, respectively. The objective of this study is to establish the chemical identity of M4 and M6 and to determine their pharmacokinetic properties.

Li et al. Table 1. HPLC Eluent Conditions gradienta G-1

Materials and Methods Materials. Decaffeinated green tea solids (dehydrated water extract of green tea, 1 g of powder was derived from 6 g of dry leaves) were obtained from Thomas J. Lipton Inc. (Englewood Cliffs, NJ). EGC, (-)-epicatechin, EGCG, ECG, β-D-glucuronidase (EC 3.2.1.31, Sigma catalog no. G-7896, from Escherichia coli with 9 000 000 units/g of solid), sulfatase (EC 3.1.6.1, Sigma catalog no. S-9754, from Abalone entrails with 23 000 units/g of solid), methyl-d3 alcohol-d (CD3OD, 99.8 at. % D), and TMS were purchased from Sigma (St. Louis, MO). Other reagents and HPLC-grade solvents were obtained from Fisher (Pittsburgh, PA). Human Samples. Five healthy male volunteers (31-35 years old; weighing 55-80 kg) were recruited from our department. All volunteers were nonsmokers and did not drink alcohol. The protocol was approved by the Institutional Review Board for the Protection of Human Subjects in Research (IRB no. 92-034M) at Rutgers University (Piscataway, NJ). The subjects abstained from tea, tea-related products, or other herbal products for at least 2 weeks before the experiment. Each volunteer drank a cup of tea, which was prepared by infusing 1.2 g of green tea powder in 300 mL of hot water (95 °C) for 2 min, at 9:00 a.m. after breakfast. Urine samples were collected 0, 1, 3, 6, 9, 12, 15, 23, 29, 35, 39, and 47 h following tea intake. During this period, the subjects did not consume any tea beverage, but drank water ad libitum. In a second experiment, one of the subjects drank a cup of green tea (containing 1.2 g of green tea solids in 300 mL of warm water) every 12 h, 12 consecutive times. Urine samples were collected at different time points. In a third experiment, the subject was on a 3 day cephalexin (1 g/day) regimen, followed by a 3 day washout period, and then drank one cup of green tea (1.2 g of solids). Serial blood samples were drawn into a heparinized vacutainer from an indwelling venous catheter immediately before tea consumption and 0.5, 2, 4.5, 7.5, 10.5, 13.5, 19, 26, 32, 37, and 43 h later. Urine samples were also collected at different time intervals. To determine the precursors of M4 and M6, another subject drank a solution of pure EGC or (-)-epicatechin, and urine samples were collected at different time points. The urine and plasma samples were frozen at -80 °C until they could be processed for subsequent HPLC analysis. Sample Preparation for Analysis of Urinary and Plasma Tea Polyphenols. The urine samples were treated as described previously (12) for the detection and LC/MS analysis of M4 and M6. In brief, the thawed urine (300 µL) was mixed with 150 µL of 0.1 M phosphate buffer (pH 7.4), 20 µL of ascorbic acid-EDTA solution [0.4 M phosphate buffer containing 20% ascorbic acid and 0.1% EDTA disodium (pH 7.2)] as a preservative, 10 µL of β-D-glucuronidase (250 units), and 10 µL of sulfatase (1 unit). The mixture was incubated at 37 °C for 45 min. The reaction mixture was extracted consecutively with 1 and 0.75 mL of ethyl acetate. The combined ethyl acetate extract (1.50 mL) was then reduced to dryness in a Savant vacuum centrifugal evaporator (Farmingdale, NY). The residue was reconstituted in 80 µL of a 10% acetonitrile aqueous solution. The resultant solution was filtered in a 0.2 µm Nanosep MF microconcentrator (Pall Filtron, Northborough, MA), and 10 µL of the filtrant was analyzed by HPLC. The thawed plasma (200 µL) was mixed with 10 µL of the ascorbic acid-EDTA solution and 30 µL of β-D-glucuronidase (500 units) and sulfatase (1 unit). The mixture was then incubated at 37 °C for 45 min and precipitated with 400 µL of ascorbic acid-saturated acetonitrile. After centrifugation at 14000g for 5 min, 600 µL of the supernatant was treated in the vacuum centrifuge concentrator for 15 min to evaporate part of the

G-2

G-3

time (min)

flow rate (mL/min)

0 7 25 31 37 38 43 50 0 3 17 22 36 0 6 18 22 30

1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 0.200 0.200 0.200 0.200 0.200 1.50 1.50 1.50 1.50 1.50

solvent (%, v/v) A

B

96 96 84 72 67 2 2 96 90 69 67 0 90 100 91 86 0 100

4 4 16 28 33 98 98 4 10 31 33 100 10 0 9 14 100 0

gradient curveb 6 6 6 6 6 6 1 6 6 1 1 11 6 1 1

a The mobile phases were as follows: solvent A, CH CN/THF/ 3 H2O [70:5:3925, v/v, in 30 mM phosphate buffer (pH 3.35)], and solvent B, CH3CN/THF/H2O [585:125:290, v/v, in 30 mM phosphate buffer (pH 3.45)], for G-1; solvent A, MeOH/H2O (5:495, v/v), and solvent B, MeOH/H2O (350:150, v/v), for G-2; and solvent A, CH3CN/H2O/CH3COOH (20:1970:10, v/v), and solvent B, CH3CN/ H2O/CH3COOH (1400:590:10, v/v), for G-3. b The effects of gradient curves 1, 6, and 11 were immediately going to specified conditions, linear, and maintaining start conditions until the next step, respectively.

acetonitrile and then extracted twice with ethyl acetate. The combined ethyl acetate solution (1.5 mL) was evaporated to dryness. The residue was redissolved in a 10% acetonitrile aqueous solution and filtered in the Nanosep MF microconcentrator before HPLC analysis. Detection of M4 and M6 by HPLC Coupled with UV and EC Detectors. The ethyl acetate-extracted urine sample (50 µL) was injected with a Waters 717 autosampler (Milford, MA) onto the HPLC system equipped with a NBS C18 column (150 mm × 4.6 mm i.d., Niko Bioscience, Tokyo, Japan). The composition of the mobile phase is given in Table 1. The gradient (G-1) was applied by two Waters 510 pumps, and the eluent was monitored by a Waters 440 absorbance detector with wavelength settings at 254 and 280 nm and a tandem ESA 5600 coulochem electrode array detector (Chelmsford, MA) set at operating potentials of -90, -10, 70, and 150 mV. The fractions containing M4 and M6 were collected from the outlet of the UV detector for subsequent LC/MS analysis. LC/MS Analysis of M4 and M6. LC/MS analysis was performed using a Waters 2690 LC separation module coupled to a Waters 996 photodiode array UV detector and a tandem Finnigan MAT LCQ mass detector (San Jose, CA) fitted with an electrospray ionization (ESI) interface. The LC flow, at 200 µL/min, was introduced into the ESI interface following detection by UV absorption from 200 to 400 nm. The chromatography of the M4 or M6 fraction was performed on a 5 µm Supelcosil LC-18 column (150 mm × 2.1 mm i.d., Supelco, Bellefonte, PA) at 30 °C. The mobile phase and gradient are shown in Table 1 (G-2). The voltage on the ESI interface was maintained at approximately -4 kV. High-purity (99% pure) nitrogen gas was supplied to the ESI source as the sheath gas at a flow rate of 90 arb and the auxiliary gas at 5 arb, respectively. The heated capillary temperature and voltage were maintained at 220 °C and -30 V, respectively. The tube lens offset voltage was set at -20 V. The LCQ MS detector was operated in four scan mode/ scan power combinations, that is, “full scan MS” for the determination of the molecular weights of M4 and M6, “zoom scan MS” for obtaining information about the charge state of M4 and M6 ions, “full scan MS/MS” for the analysis of fragment ions of M4 and M6 through collision-induced dissociation (CID) to obtain more structural information, and “selected reaction

Tea Catechin Metabolites monitoring” (SRM) for the quantitative analysis of urinary M4, M6, and their parent catechins. The precursor-to-product ion transitions of m/z 223 f 179 for M4, m/z 207 f 163 for M6, m/z 305 f 261 for EGC, and m/z 289 f 245 for (-)-epicatechin were used for SRM with a relative collision energy setting of 18%. The ESI ion source was set in negative ion polarity mode for acquiring all mass spectrometry data. Semipreparative Isolation of M4 and M6 from Human Urine. The urine samples exhibiting high chromatographic peaks of M4 and M6 were used for the isolation. Thawed urine (100 mL) was mixed with 50 mL of 1 M phosphate buffer (pH 6.8), 1.4 g of ascorbic acid, and β-D-glucuronidase (5000 units). The mixture was incubated at 37 °C for 45 min. The treated urine was extracted with 150 mL and then 100 mL of EtOAc/MeOH (9:1, v/v), and the combined organic phase was evaporated to dryness under reduced pressure at 35 °C. The residue was dissolved in 2 mL of a 10% acetonitrile aqueous solution and subjected to a Whatman syringe filter with a pore size of 0.2 µm. The resultant filtrate (200 µL) was repeatedly introduced into a 5 µm Supelcosil C-18 HPLC column (150 mm × 4.6 mm i.d.) maintained at 30 °C. A gradient elution program (G-3 in Table 1) was used, and the flow rate was 1.50 mL/min. M4 and M6 fractions with retention times of 12.0-12.6 and 18.0-18.6 min, respectively, were collected from the outlet of the UV detector. The fractions containing M4 and M6 were extracted into ethyl acetate and then evaporated to dryness in the vacuum centrifuge concentrator at ambient temperature. The resultant residues were reconstituted in a 10% acetonitrile aqueous solution and rechromatographed on another 5 µm Supelcosil C-18 column (150 mm × 4.6 mm i.d.) at 30 °C. In isocratic pump mode, THF/H2O/CH3COOH (10:975:5, v/v) and THF/H2O/CH3COOH (48:946:6, v/v) mixtures were used as mobile phases for the purification of M4 and M6, respectively. The purified M4 and M6 fractions were lyophilized, and their purity was further checked by HPLC using gradient G-3 (Table 1). 1H NMR Analysis of Purified M4 and M6. NMR spectra were recorded on either a Varian Inova 600 spectrometer or a Varian Unity 400 spectrometer equipped with a triple-resonance probe (1H, 13C, and 15N). Two samples for each of the compounds (M4 and M6) were prepared by dissolving ∼2 mg of M4 or M6 in 0.25 mL of CD3OD and CD3OH, respectively. The samples were then transferred to Shigemi CD3OD-matching tubes (Tokyo, Japan). The NMR spectra of M4 and M6 in CD3OD were recorded at 25 °C, while those in CD3OH were recorded at 10 °C to slow exchange between hydroxyl protons of M4 or M6 and CD3OH. Chemical shifts are referenced to TMS at 0.000 ppm. Solvent suppression was achieved by the WATERGATE sequence (18).

Chem. Res. Toxicol., Vol. 13, No. 3, 2000 179

Figure 1. HPLC electrochemical detection of M4, M6, EGC, and (-)-epicatechin in human urine. The urine samples were digested with β-D-glucuronidase and sulfatase before sample extraction for HPLC: trace 1, green tea extract ingested; trace 2, urine sample collected before ingestion of green tea; trace 3, urine sample collected 3-6 h after ingestion of green tea; trace 4, urine sample collected after oral administration of pure EGC; and trace 5, urine sample collected after oral administration of pure (-)-epicatechin.

279

Results Detection and Confirmation of M4 and M6 as the Metabolites of Catechins. After ingestion of green tea, M4 and M6 were initially detected in human urine with the electrochemical detector in our HPLC system. The M4 peak was prominent in the channels set at -90 and -10 mV, whereas the M6 peak was prominently detected at operating potentials of -10 and 70 mV. As shown in Figure 1, neither M4 nor M6 was detected in the green tea preparation, nor in the blank urine samples collected from the subjects who had not consumed tea. After oral administration of pure EGC, we detected M4 and a small amount of M6 in the human urine, whereas administration of pure (-)-epicatechin resulted in only M6 in the urine (Figure 1). This observation suggests that EGC and (-)-epicatechin are precursors of M4 and M6, respectively. The similarity of the UV spectra of M4 and EGC, as well as that of M6 and (-)-epicatechin, is consistent with such a precursor-metabolite relationship (Figure 2).

Figure 2. UV spectra of M4, EGC, M6, and (-)-epicatechin. The spectra were recorded on the HPLC system coupled to a photodiode array UV detector.

Chemical Identification of M4 and M6. We obtained information about the molecular weights of M4 and M6 from the LC/ESI-MS analysis. Electrospray ionization is one of the softest ionization techniques. As depicted in Figure 3, the deprotonated molecular ion peaks at m/z 223 for M4 and at m/z 207 for M6 indicated that their molecular weights are 224 and 208, respectively. The difference of the molecular weight between M4 and M6 is 16 mass units, which is the same as that between their precursors EGC and (-)-epicatechin, suggesting that M4 may have one more phenolic hydroxyl group than M6. The product ion mass spectra

180

Chem. Res. Toxicol., Vol. 13, No. 3, 2000

Li et al.

Figure 3. HPLC/PDA-UV/ESI-MS analysis of urinary M4 and M6. For panel 1, M4 and M6 in a treated urine sample following green tea ingestion were separated under the chromatographic conditions denoted G-1 in Table 1. For panels 2 and 3, the fractions containing M4 and M6 were collected separately and rechromatographed under the conditions denoted G-2 in Table 1 for photodiode array analysis. For panels 4 and 5, the ESI-MS chromatograms and spectra of M4 and M6, respectively, were obtained in tandem in the full scan MS mode. For panels 6 and 7, in the full scan MS/MS mode deprotonated M4 and M6 ions were selected and then dissociated through CID to produce their fragment ions, respectively. Table 2. 1H NMR (600 MHz, CD3OD) Spectral Dataa of Metabolites M4 and M6 proton

M4

M6

H-3b H-3a H-2b H-2a H-5b H-5a H-4 H-6′ H-2′ H-5′ OH-3′ b OH-4′ OH-5′

1.958 dddd (7.3, 8.8, 9.5, 16.7) 2.241 dddd (4.8, 6.8, 9.7, 16.7) 2.344 ddd (4.8, 9.5, 12.7) 2.484 ddd (8.8, 9.7, 12.7) 2.720 dd (6.1, 14.0) 2.817 dd (6.1, 14.0) 4.711 dddd (6.1, 6.1, 6.8, 7.3) 6.245 s 6.245 s 8.857 s 8.857 s 8.056 s

1.957 dddd (7.3, 8.8, 9.5, 12.8) 2.240 dddd (4.9, 6.9, 9.8, 12.8) 2.332 ddd (4.9, 9.6, 17.7) 2.482 ddd (8.8, 9.8, 17.7) 2.792 dd (6.1, 13.9) 2.875 dd (6.1, 13.9) 4.725 dddd (6.1, 6.1, 6.9, 7.3) 6.562 dd (8.0, 2.0) 6.685 d (2.0) 6.695 d (8.0) 8.821 s 8.821 s -

a Chemical shifts are expressed in parts per million downfield from the signal for TMS in CD3OD or CD3OH, and coupling constants in hertz are in parentheses. b The hydroxyl regions of the 1H NMR spectra of M4 and M6 were recorded at 10 °C in CD OH. 3

(Figure 3), obtained in the full scan MS/MS mode, exhibiting fragment ions [(M - 44) - H]- at m/z 179 for M4 and at m/z 163 for M6 also suggested that the two metabolites may have similar chemical structures except for the number of OH groups. Further structural determination of M6 was carried out using its 1H NMR data (Table 2), which could almost be superimposed on the reported data for 5-(3′,4′-dihydroxyphenyl)-γ-valerolactone (19). Most of the signals in the NMR spectrum of M4 were similar to those of M6 (see Table 2). However, integration of the area under the singlet at 6.25 ppm suggests that M4 possesses two equivalent aromatic protons. In addition, the appearance of the OH proton

signals at 8.06 and 8.82 ppm in the spectrum for M4 suggests that it has one more phenolic hydroxyl group than M6 (Table 2). EGC was the precursor of M4, which also exhibited two aromatic OH proton signals at 8.08 and 8.86 ppm. Thus, M4 was identified as 5-(3′,4′,5′-trihydroxyphenyl)-γ-valerolactone. The structures of M4 and M6 are shown in Figure 4. The carbon at position 4 in M4 or M6 is a chiral center. Although we cannot give the absolute configurations of their chiral centers, both urinary M4 and M6 obtained in our experiment were optically active, the [R]D(methanol) values of which were -6.23° (c ) 1.12 × 10-3 g/mL) and -10.00° (c ) 1.80 × 10-3 g/mL), respectively. The optical rotations were measured with a Perkin-Elmer 241 polarimeter (Norwalk, CT) at 23 °C. Kinetics and Quantification of Renal Excretion of M4, M6, and Their Precursors. The renal excretion rate-time curves of EGC and (-)-epicatechin following ingestion of a single dose of green tea were similar to their plasma level-time curves observed previously (13). As shown in Figure 5, the excretion rates of EGC and (-)-epicatechin quickly reached peak values at 2 h in all five subjects and then declined to negligible rates after 6-9 h. The renal excretion rate-time curve for EGC or (-)-epicatechin was unimodal. This did not give us evidence to support the hydrolysis of EGCG to EGC or ECG to (-)-epicatechin by intestinal flora in the colon, although in vitro experiments have shown that the biotransformation is possible (20). The renal excretion rate-time profiles of M4 and M6 were, however, quite different. There was approximately a 3 h lag time for the

Tea Catechin Metabolites

Chem. Res. Toxicol., Vol. 13, No. 3, 2000 181

Figure 4. Proposed pathway for conversion of EGC and (-)-epicatechin to M4 and M6, respectively.

Figure 5. Renal excretion rate-time profiles of M4, M6, EGC, and (-)-epicatechin following ingestion of a cup of green tea (1.2 g) by five human subjects.

excretion of M4 or M6. The maximum renal excretion rates of M4 and M6 occurred between 7.5 and 13.5 h later than those of EGC and (-)-epicatechin (Figure 5). In subjects A-C, after the major excretion peak there was a smaller later peak for M6. The second peak might arise from the conversion of M4 or its precursors to additional M6. In Subjects D and E, a secondary excretion peak of M6 also seemed to occur at 39-47 h, even though the levels were low. A comparison between the cumulative amount excreted in urine and the amount administered from tea showed that, after ingestion of a single dose of green tea, only 0.7-2.5% of the administered EGC and only 0.7-3.7% of the administered (-)-epicatechin were excreted in urine of the five subjects. As shown in Table 3, the cumulative amount of excreted M4 was 25, 13, and 10 times greater than that of EGC excreted in subjects A-C, respectively, and accounted for approximately 34, 32, and 8% of the administered EGC from green tea, respectively. Similarly, the cumulative amount of excreted M6 was 15, 10, and 8 times greater than that of (-)-epicatechin and accounted for 36, 39, and 6% of the administered (-)-

epicatechin from green tea in subjects A-C, respectively. In subjects D and E, the cumulative amounts of excreted M4 and M6 were comparable to the levels of EGC and (-)-epicatechin, respectively, and accounted for 0.6-4.2% of the ingested EGC and (-)-epicatechin. On the basis of the renal excretion of M4 and M6, there appeared to be two classes of metabolizer among the five human subjects, i.e., three rapid metabolizers and two poor metabolizers. There was no statistically significant correlation in the cumulative excretion between M4 and EGC or between M6 and (-)-epicatechin. Also, the intersubject differences in M4 and M6 excretion were much larger than those in the excretion of EGC and (-)epicatechin (Table 3), suggesting that most of the formation of M4 and M6 did not follow the absorption of their precursors. Excretion of M4 and M6 during and after Repeated Ingestion of Green Tea. Renal excretion ratetime curves for M4 and M6 in subject A, as well as their precursors EGC and (-)-epicatechin, during and after repeated consumption of tea are shown in Figure 6. The renal excretion rates of M4 and M6 were higher than the excretion rates of EGC and (-)-epicatechin, respectively. Since tea was ingested every 12 h, at which time the metabolites were not cleared in the urine, a cumulative effect was observed for M4. However, the excretion peaks and troughs did not strictly follow an additive model, which suggests adaptive responses. Consistent with this idea is the fact that the ratio of renal excretion between M4 and EGC or between M6 and (-)-epicatechin, as well as the percentages of ingested EGC and (-)-epicatechin excreted as M4 and M6, was lower than those observed following ingestion of a single dose of green tea. Plasma Levels and Urinary Excretion of M4 and M6. The kinetics of the urinary excretion of M4 and M6 suggested the involvement of intestinal microflora in their formation reactions. Therefore, an experiment was conducted with subject A after administration of the antibiotic cephalexin for 3 days. As shown in Figure 7,

182

Chem. Res. Toxicol., Vol. 13, No. 3, 2000

Li et al.

Table 3. Cumulative Amounts of M4, M6, EGC, and (-)-Epicatechin Excreted in Human Urine subject single dose (1.2 g of green tea A B C D E mean ( SD (CV%)

M4 (µmol)

M6 (µmol)

EGC (µmol)

(-)-epicatechin (µmol)

90.25 86.97 20.32 1.58 2.05 40.23 ( 44.82 (111.41%)

39.19 43.18 6.66 4.64 2.86 19.31 ( 20.07 (103.94%)

3.68 6.73 1.96 5.07 2.82 4.05 ( 1.89 (46.67%)

2.57 4.13 0.81 2.58 1.08 2.23 ( 1.34 (60.09%)

524.34

153.41

56.54

17.03

solidsa)

multiple doses (twelve 1.2 g portions of green tea solids) A a

The administered amounts of EGC and (-)-epicatechin from 1.2 g of green tea solids were 268.85 and 110.34 µmol, respectively. The molecular weights of M4, M6, EGC, and (-)-epicatechin are 224, 208, 306, and 290, respectively.

Figure 6. Renal excretion rate-time profiles of M4, M6, EGC, and (-)-epicatechin during and after repeated consumption of green tea by a human subject. The subject consumed 1.2 g of green tea every 12 h, 12 consecutive times.

Figure 7. Renal excretion rate-time profiles (left) and plasma level-time profiles (right) of M4, EGC, M6, and (-)-epicatechin following ingestion of a cup of green tea (1.2 g) by one human subject. The subject had orally received 3 day cephalexin treatment and then drank one dose of green tea following a 3 day washout period. The time for the urinary data is the midpoint time of urine collection in hours, while the time for the plasma data is the time of blood collection in hours.

after ingestion of a dose of green tea, the rates of renal excretion of M4 and M6 in subject A were much lower than those shown in Figure 5. The corresponding plasma level-time curves of M4, M6, EGC, and (-)-epicatechin are also shown in Figure 7. The plasma levels of M4 and M6 were lower than the levels of EGC and (-)-epicatechin, respectively. The profiles of plasma levels were similar to the corresponding profiles of the rates of urinary excretion (Figure 7). To obtain additional information about the plasma levels of M4 and M6 as well as their relationship with urinary excretion, a set of samples collected for detailed pharmacokinetic studies of catechins (the results will be published elsewhere) were analyzed for M4 and M6. In two subjects (S1 and S2), peak plasma levels of M4 and M6 were observed 5-12 h after tea ingestion (Figure 8). These levels were lower than the peak plasma levels of EGC and (-)-epicatechin which appeared between 1 and 2 h. The areas under the curves of plasma EGC and (-)-epicatechin were 2.4-13.3 times larger than those of M4 and M6. In the same two subjects, M4 and M6 were present in the urine samples collected at 3-8 h

Figure 8. Two male human subjects (S1 and S2) ingested a single dose of decaffeinated green tea (20 mg/kg of body weight). The blood samples were collected at 0, 0.25, 0.5, 1, 2, 3, 5, 8, 12, and 24 h, while the urine samples were collected at 0, 3, 8, and 24 h.

(3.70-6.82 µg/mL) at levels that were 2 orders of magnitude higher than their peak plasma values. For the samples collected at 8-24 h, M4 and M6 concentrations (0.63 µg/mL) were lower than those in the 3-8 h samples for S2; but for S1, the levels (5.26 and 7.42 µg/mL, respectively) were higher. The cumulative urinary excretion levels of M4 and M6 are shown in Figure 8. S2, who exhibited plasma level peaks of M4 and M6 at 5 h, exhibited most M4 and M6 excretion by 8 h, whereas the delayed plasma level peaks in S1 were associated with delayed renal excretion. The total amounts of M4 and M6 excreted in 24 h account for 3.8-25.5% of the ingested EGC and (-)-epicatechin.

Discussion Knowledge of the bioavailability and biotransformation of tea components is important for understanding the biological effects of tea. In this study, we identified two catechin metabolites in the human urine following oral administration of green tea. Large intersubject differences in the pattern and quantity of their excretion were observed. Quantitative analysis indicated that the cumulative amounts of M4 and M6 excreted in urine were 0.3-25 times greater than those of EGC and (-)epicatechin in the subjects following ingestion of a single dose of green tea (Table 3), and accounted for 0.6-39% of the amounts of ingested EGC and (-)-epicatechin. In addition, the cumulative amount of renal excretion of M4 or M6 was more than 9-fold greater than that of EGC or (-)-epicatechin in subject A during and after repeated consumption of green tea. M4 and M6 were also detected in the plasma, the peak levels of which were lower than those of EGC and (-)-epicatechin, and exhibited kinetics similar to those of their urinary metabolites. Both M4 and M6 were also detected in human urine following ingestion of black tea, but the levels were less than onetenth of those produced by green tea due to the lower EGC and (-)-epicatechin levels in the black tea (data not shown). We also rechecked our previous chromatographic

Tea Catechin Metabolites

data of 24 h urine samples collected from 938 tea drinkers in a case-control study for gastric cancer in China. Pooling all his or her urine collected in a day made the urine sample of each subject. The urinary concentrations of M4 and/or M6 in 23% of the subjects were at least 3 times greater than the levels of the parent EGC and/or (-)-epicatechin. Using LC/MS and NMR, we identified urinary tea metabolites M4 and M6 as (-)-5-(3′,4′,5′-trihydroxyphenyl)-γ-valerolactone and (-)-5-(3′,4′-dihydroxyphenyl)-γvalerolactone, respectively. The details of the metabolic pathway for the EGC f M4 or (-)-epicatechin f M6 reaction, however, are not clear. Meselhy et al. (20) found (-)-1-(3′,4′-dihydroxyphenyl)-3-(2′′,4′′,6′′-trihydroxyphenyl)propan-2-ol (labeled compound 5 in ref 20) after anaerobic incubation of (-)-epicatechin with human intestinal bacteria. Furthermore, reincubation of this compound gave (-)-5-(3′,4′-dihydroxyphenyl)-γ-valerolactone (labeled compound 7 in ref 20). Thus, they proposed a possible metabolic pathway for the (-)epicatechin f compound 7 reaction. On the basis of these results, we propose that M4 and M6 are formed by the pathways depicted in Figure 5. EGC and (-)epicatechin first undergo reductive cleavage to give 1-(3′,4′,5′-trihydroxyphenyl)-3-(2′′,4′′,6′′-trihydroxyphenyl)propan-2-ol and 1-(3′,4′-dihydroxyphenyl)-3-(2′′,4′′,6′′trihydroxyphenyl)propan-2-ol, which are further lactonized to give M4 and M6, respectively. Examples of ring fission have also been reported for other flavonoid derivatives, including flavonols, flavones, and flavanones (21). The large intestine, harboring a huge population of anaerobic bacteria, is likely to be the main site for the formation of M4 and M6. Other types of flavonoid ring cleavage are also known to be caused by the colonic microorganisms. In a preliminary study, we found neither M4 nor M6 when EGC or (-)-epicatechin was incubated with human liver microsomes or homogenate (data not shown). The fact that renal excretion of M4 or M6 had a 3 h lag time and peaked at 7.5-13.5 h [in comparison to those of EGC or (-)-epicatechin at 2 h] following ingestion of a single dose of green tea also suggests that M4 and M6 are formed in the colon, absorbed, and then excreted (predominantly in conjugated forms) in urine. Such pharmacokinetic features were also observed for the colonic metabolites of other herbal compounds after oral administration of aqueous herbal extracts. Two examples are the delayed renal excretion of davidigenin, a colonic metabolite of (R/S)-liquiritigenins, and the bimodal renal excretion of (R/S)-liquiritigenins, baicalein, wogonin, and oroxylin A following ingestion of the herbal extract TJ-9 by humans (22, 23). Das and Griffiths observed that the formation of the ring fission products from (+)-catechin in guinea pig was arrested when the antibacterial compounds aureomycin and phthaloylsulfathiazole were administered (24). In the study presented here, we also observed lower levels of M4 and M6 in an individual after administration of an antibacterial agent. Both M4 and M6 were found in human fecal samples (11 and 10 µg/g wet weight, respectively) collected 0-9 h following ingestion of green tea, and the fecal levels of EGC, (-)epicatechin, EGCG, and ECG were 18, 12, 54, and 22 µg/ g, respectively (unpublished results). In subjects A-E, the total cumulative molar amount of EGC and M4 excreted in urine was 2-35% of the ingested molar amount of EGC from green tea in our subjects, while the total amount of (-)-epicatechin and M6 excreted was

Chem. Res. Toxicol., Vol. 13, No. 3, 2000 183

4-43% of the ingested amount of (-)-epicatechin. In vitro anaerobic incubation of (-)-epicatechin with human intestinal bacteria gave nine metabolites (20). Also, two monomethylated (-)-epicatechin derivatives were identified as 3′-O-methyl-(-)-epicatechin and 4′-O-methyl(-)-epicatechin, which were isolated from rat urine after oral administration of pure (-)-epicatechin (25). Other major catechin metabolites, besides valerolactones, may also be formed after tea consumption. In the study presented here, we have established the chemical identity of two major tea catechin metabolites, i.e., (-)-5-(3′,4′,5′-trihydroxyphenyl)-γ-valerolactone (M4) and (-)-5-(3′,4′-dihydroxyphenyl)-γ-valerolactone (M6), in human urine and blood. These metabolites appeared to be formed by the intestinal flora in the human colon and then absorbed. On the basis of the structures, we predict antioxidative activities for M4 and M6. The possible cancer prevention and other health effects of M4 and M6 need to be investigated. Study of these metabolites may help us better understand the biological effects of tea.

Acknowledgment. We thank all volunteers for their participation in the tea experiments. Dr. Sandra Mohr from Environmental and Occupational Health Sciences Institute kindly conducted the blood sampling. The LC/ MS analysis was conducted in the analytical center (directed by Dr. Brian Buckley) at Environmental and Occupational Health Sciences Institute. The optical rotation of M4 and M6 was measured in Dr. Edmond J. LaVoie’s laboratory at Rutgers University. This study was supported by NIH Grant CA 56673.

References (1) Yang, C. S., and Wang, Z.-Y. (1993) Tea and cancer. J. Natl. Cancer Inst. 85, 1038-1049. (2) Yang, C. S., Chen, L., Lee, M.-J., and Landau, J. M. (1996) Effects of tea on carcinogenesis in animal models and human. In Dietary Phytochemicals in Cancer Prevention and Treatment (Butrum, R. R., Ed.) pp 51-61, Plenum Press, New York. (3) Wang, Z.-Y., Chen, L., Lee, M.-J., and Yang, C. S. (1996) Tea and cancer prevention. In Hypernutritious Food (Finley, J. W., Armstrong, D. J., Nagy, S., and Robinson, S. F., Eds.) pp 239260, Agscience, Inc., Auburndale, FL. (4) Dreosti, I. E., Wargovich, M. J., and Yang, C. S. (1997) Inhibition of carcinogenesis by tea: the evidence from experimental studies. Crit. Rev. Food Sci. Nutr. 37, 761-770. (5) Blot, W. J., Chow, W. H., and McLaughlin, J. K. (1996) Tea and cancer: a review of the epidemiologic evidence. Eur. J. Cancer Prev. 5, 425-438. (6) Kohlmeier, L., Weterings, K., Steck, S., and Kok, F. J. (1997) Tea and cancer prevention: an evaluation of the epidemiologic literature. Nutr. Cancer 27, 1-3. (7) Blot, W. J., McLaughlin, J. K., and Chow, H. W. (1997) Cancer rates among drinkers of black tea. Crit. Rev. Food Sci. Nutr. 37, 739-760. (8) Wiseman, S. A., Balentine, D. A., and Frei, B. (1997) Antioxidants in tea. Crit. Rev. Food Sci. Nutr. 37, 705-718. (9) Balentine, D. A., Wiseman, S. A., and Bouwens, L. C. M. (1997) The chemistry of tea flavonoids. Crit. Rev. Food Sci. Nutr. 37, 693-704. (10) Graham, H. N. (1992) Green tea composition, and polyphenol chemistry. Prev. Med. 21, 334-350. (11) Lee, M.-J., Wang, Z.-Y., Li, H., Chen, L., Sun, Y., Gobbo, S., Balentine, D. A., and Yang, C. S. (1995) Analysis of plasma and urinary tea polyphenols in human subjects. Cancer Epidemiol., Biomarkers Prev. 4, 393-399. (12) Chen, L., Lee, M.-J., Li, H., and Yang, C. S. (1997) Absorption, distribution, and elimination of tea polyphenols in rats. Drug Metab. Dispos. 25, 1045-1050. (13) Yang, C. S., Chen, L., Lee, M.-J., Balentine, D. A., Kuo, M. C., and Schantz, S. P. (1998) Blood and urine level of tea catechins after ingestion of different amounts of green tea by human volunteers. Cancer Epidemiol., Biomarkers Prev. 7, 351-354. (14) Yang, C. S., Lee, M.-J., and Chen, L. (1999) Human salivary tea catechin levels and catechin esterase activities: implication in human cancer prevention studies. Cancer Epidemiol., Biomarkers Prev. 8, 83-89.

184

Chem. Res. Toxicol., Vol. 13, No. 3, 2000

(15) Van het Hof, K. H., Wiseman, S. A., Yang, C. S., and Tijburg, L. B. M. (1999) Plasma and lipoprotein levels of tea catechins following repeated tea consumption. Proc. Soc. Exp. Biol. Med. 220, 203-209. (16) Okushio, K., Matsumoto, N., Suzuki, M., Nanjo, F., and Hara, Y. (1995) Absorption of (-)-epigallocatechin gallate into rat portal vein. Biol. Pharm. Bull. 18, 190-191. (17) Okushio, K., Matsumoto, N., Kohri, T., Suzuki, M., Nanjo, F., and Hara, Y. (1996) Absorption of tea catechins into rat portal vein. Biol. Pharm. Bull. 19, 326-329. (18) Unno, T., Kondo, K., Itakura, H., and Takeo, T. (1996) Analysis of (-)-epigallocatechin gallate in human serum obtained after ingestion of green tea. Biosci., Biotechnol., Biochem. 60, 20662068. (19) Piotto, M., Saudek, V., and Sklenar, V. (1992) Gradient-tailored excitation for single-quantum NMR spectroscopy of aqueous solutions. J. Biomol. NMR 2, 661-665. (20) Meselhy, M. R., Nakamura, N., and Hattori, M. (1997) Biotransformation of (-)-epicatechin 3-O-gallate by human intestinal bacteria. Chem. Pharm. Bull. 45, 888-893.

Li et al. (21) Hollman, P. C. H., and Katan, M. B. (1998) Absorption, metabolism, and bioavailability of flavonoids. In Flavonoids in Health and Disease (Rice-Evans, C. A., and Packer, L., Eds.) pp 483522, Marcel Dekker, New York, Basel, and Hong Kong. (22) Li, C. (1998) Chemistry and pharmacokinetics of Shosaiko-to, a herbal medicine. Ph.D. Thesis, pp 59-60, Tokyo University of Pharmacy and Life Science, 1432-1 Horinouchi, Hachioji, Tokyo 192-0392, Japan. (23) Li, C., Homma, M., and Oka, K. (1998) Characteristics of delayed excretion of flavonoids in human urine after administration of Shosaiko-to, a herbal medicine. Biol. Pharm. Bull. 21, 1251-1257. (24) Das, N. P., and Griffiths, L. A. (1968) Studies on flavonoid metabolism, metabolism of (+)-catechin in the guinea pig. Biochem. J. 110, 449-456. (25) Okushio, K., Suzuki, M., Matsumoto, N., Nanjo, F., and Hara, Y. (1999) Identification of (-)-epicatechin metabolites and their metabolic fate in the rat. Drug Metab. Dispos. 27, 309-316.

TX9901837