- Email: [email protected]
Simultaneous detection of green tea catechins and gallic acid in human serum after ingestion of green tea tablets using ion-pair high-performance liquid chromatography with electrochemical detection
Journal of Chromatography B, 945–946 (2014) 147–153
Contents lists available at ScienceDirect
Journal of Chromatography B journal homepage: www.elsevier.com/locate/chromb
Simultaneous detection of green tea catechins and gallic acid in human serum after ingestion of green tea tablets using ion-pair high-performance liquid chromatography with electrochemical detection Keiko Narumi a , Jun-Ichiro Sonoda a,∗ , Keita Shiotani a , Michihiro Shigeru a , Masayuki Shibata a , Akio Kawachi a , Erisa Tomishige a , Keizo Sato b , Toshiro Motoya a a First Department of Clinical Pharmacy, School of Pharmaceutical Sciences, Kyushu University of Health & Welfare, 1714-1 Yoshino-cho, Nobeoka, Miyazaki 882-8508, Japan b Department of Clinical Biochemistry, School of Pharmaceutical Sciences, Kyushu University of Health & Welfare, 1714-1 Yoshino-cho, Nobeoka, Miyazaki 882-8508, Japan
a r t i c l e
i n f o
Article history: Received 28 June 2013 Accepted 4 November 2013 Available online 1 December 2013 Keywords: Green tea catechins Gallic acid HPLC Ion-pair reagent Electrochemical detection Pharmacokinetic study
a b s t r a c t We developed an analytical method for the simultaneous determination of tea catechins and gallic acid (GA) in human serum using ion-pair high-performance liquid chromatography (HPLC) with electrochemical detection. GA was measured to estimate the amount of gallate moiety produced by degradation of gallated catechins ((−)-epicatechin-3-gallate, ECG; (−)-epigallocatechin-3-gallate, EGCG). Ethyl gallate was adopted as an internal standard to correct for the extraction efficiency. To maximize extraction efficiency, a hydrophobic polytetrafluoroethylene (PTFE) filter was selected for pre-treatment prior to separation. HPLC separation was performed using a C18 reversed-phase column with a gradient mobile phase of phosphate buffer (pH 2.5) containing tetrahexylammonium hydrogensulfate as an ion-pair reagent. Using this method, (−)-epicatechin (EC), (−)-epigallocatechin (EGC), ECG, EGCG, ethyl gallate, and GA were detected as single peaks. The resolution values for target analytes were 4.0–13.0 and the mean values of the absolute recoveries of catechins and GA were 77.3–93.9%. The detection limits for catechins and GA in serum were 0.4–3.1 ng/mL. The serum catechin levels of eight healthy volunteers after ingestion of a single dose of green tea tablets were measured using this method. The concentration of total catechins (free + conjugated forms) in serum peaked 60 min after ingestion. From these results, this method is thought to enable the simultaneous quantification of GA, the hydrolysis product of gallated catechins, and target catechins, and to be sufficiently sensitive for pharmacokinetic studies of catechins following oral administration of green tea. © 2013 Elsevier B.V. All rights reserved.
1. Introduction Tea, including green, black, and oolong varieties, is made from the leaves of the unique tea plant (Camellia sinensis), and has been the most widely consumed drink in the world for thousands of years. Among the different types of tea, green (Japanese) tea has been reported to have beneficial health effects, such as reducing the risk of cancer and cardiovascular diseases [1–4]. The active constituents of green tea are thought to be polyphenols, commonly known as tea catechins. The major tea catechins are (−)-epicatechin
∗ Corresponding author at: First Department of Clinical Pharmacy, School of Pharmaceutical Sciences, Kyushu University of Health & Welfare, 1714-1 Yoshino-cho, Nobeoka, Miyazaki 882-8508, Japan. Tel.: +81 982 23 5529; fax: +81 982 23 5536. E-mail address: [email protected] (J.-I. Sonoda). 1570-0232/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jchromb.2013.11.007
(EC), (−)-epigallocatechin (EGC), (−)-epicatechin-3-gallate (ECG), and (−)-epigallocatechin-3-gallate (EGCG) (Fig. 1). One cup of regular green tea contains several hundred milligrams of catechins. After ingestion, the serum concentration of catechins reaches a maximum between several tens and several hundreds of ng/mL. The reasons for the poor transfer of catechins to the bloodstream and the wide variation in pharmacokinetics of catechins after green tea ingestion are not clearly understood [5]. In order to assess the kinetics and the beneficial health effects of catechins, a sensitive, reproducible, and straightforward method is needed to determine serum catechins. Generally, tea catechins are analyzed using high-performance liquid chromatography (HPLC) with ultraviolet (UV) [6], chemiluminescence (CL) [7], or electrochemical detection (ECD); also commonly used are capillary electrophoresis (CE) [8] and gas chromatography with mass spectrometry (GC–MS) [9]. These methods
148
K. Narumi et al. / J. Chromatogr. B 945–946 (2014) 147–153
from Nacalai Tesque (Kyoto, Japan). Ethyl gallate was purchased from Tokyo Chemical Industry (Tokyo, Japan). EC, EGC, EGCG, L-ascorbic acid, -glucuronidase (G8420), sulfatase (Type VIII: S9754), and tetrahexylammonium hydrogensulfate (THA) were purchased from Sigma–Aldrich (St. Louis, MO, USA). Other chemicals and solvents used were of HPLC grade. Decaffeinated green tea tablets were provided by Sathuma Green Tea International, Inc. (Kagoshima, Japan). 2.2. Apparatus
Fig. 1. Structures of the four target catechins.
have been used for determining catechins in tea [8,10], rat plasma [11], human plasma [6,7,12,13], and urine [10]. Although GC–MS can be used for the determination of catechins with high sensitivity [9], the required sample pre-treatment is too demanding for routine analysis. However, HPLC is easier to use and more widely available. Previous work has shown that HPLC-ECD is more sensitive for the determination of catechins [14] than HPLC-UV [15]. Moreover, ECD is well suited to the detection of polyphenolic compounds like catechins because these analytes can be selectively oxidized by the appropriate electric potential to flow current. The magnitude of this current is related to the analyte concentration. Therefore, HPLC-ECD is an excellent candidate technique for the selective and highly sensitive detection of catechins in human serum. After the ingestion of green tea, a large proportion of catechins are present as glucuronate or sulfate conjugates in the bloodstream. For the quantification of total catechins in serum, samples were treated with -glucuronidase and sulfatase to hydrolyze the conjugates to free form catechins. Type X-A -glucuronidase is typically used to hydrolyze glucuronic acid conjugates [14,16]. It has been reported that type H-2 -glucuronidase, which contains -glucuronidases other than sulfatase, effectively hydrolyzed conjugates in plasma samples but concomitantly cleaved the ester bonds of EGCG and ECG [6]. As such, these enzymes converted gallated catechins to their non-gallated forms (EGC and EC). In this case, total catechins are expressed as EGCG + EGC and ECG + EC, which unfortunately makes it difficult to elucidate the exact concentrations of each individual catechin because some EGC and EC are derived from EGCG and ECG via treatment with type H-2 glucuronidase. Therefore, gallic acid (GA), derived from the gallated catechins, must be measured simultaneously to elucidate whether or not cleavage occurs. In this study, we developed a sensitive and practical method for catechin determination consisting of digestion of conjugates followed by HPLC-ECD. Using this method, we evaluated the timedependent changes in the concentrations of tea catechins in serum after the ingestion of a single dose of green tea tablets (equivalent to one cup of typical green tea) in human subjects. 2. Materials and methods 2.1. Chemicals and reagents GA, ECG, acetonitrile, phosphoric acid, and di-sodium dihydrogen ethylenediamine-tetraacetate dihydrate (EDTA) were obtained
A Shimadzu (Kyoto, Japan) HPLC system consisting of a DGU14AM degasser, LC-10AD VP pumps, SIL-10AD autosampler, and SCL-10A system controller was used. The polyphenolic compounds were detected using a Coulochem III coulometric ECD system with a guard cell (Model 5020) and analytical cell (Model 5011; ESA, MA, USA). HPLC separations were performed on a C18 reversed-phase column (Mightysil RP-18 GP aqua, 250 mm × 4.6 mm, i.d., particle size, 5 m; Kanto Chemical, Tokyo, Japan). Data acquisition and analysis were performed with Chromeleon version 6.40 software (Dionex, CA, USA). 2.3. HPLC conditions Phosphate buffer (50 mM) containing 0.05 mM EDTA and 5 mM THA as an ion-pair reagent was adjusted to pH 2.5 with phosphoric acid. Mobile phase A was a deaerated mixture of 99:1 (v/v) phosphate buffer (pH 2.5):acetonitrile. Mobile phase B was a deaerated mixture of 77:23 (v/v) phosphate buffer (pH 2.5):acetonitrile. The gradient program was as follows: 0–6.00 min, 100% A; 6.01–30.00 min, B 0–42%; 30.01–35.00 min, B 42–100%; 35.01–37.00 min, 100% B; 37.01–39.00 min, B 100–95%; 39.01–41.00 min, B 95%; 41.01–46.00 min, B 95–80%; 46.01–57.00 min, B 80%; and 57.01–85.0 min, 100% A. The eluent was monitored by coulometric ECD (filter mode) with the guard cell potential at 250 mV and analytical cell potentials for electrode 1 and 2 at −50 and 200 mV, respectively. The flow rate was kept at 1.0 mL/min from 0 to 57.00 min and then increased at 1.5 mL/min from 57.01 to 80.00 min to shorten the conditioning time. The injected volume was 10 L. The column was housed in a temperature-regulated compartment maintained at 40 ◦ C, and the autosampler was maintained at 4 ◦ C. 2.4. Preparation of standard and serum samples Standard stock solutions of EC, EGC, ECG, EGCG, GA, and ethyl gallate as an internal standard were prepared at a concentration of 100 g/mL in 0.2 M phosphate buffer (pH 3.6) with 2% ascorbic acid/0.01% EDTA. -Glucuronidase and sulfatase were prepared by dissolving in 75 mM phosphate buffer (pH 6.8) to concentrations of 250 U/10 L and 25 U/10 L, respectively, and were stored in small aliquots at −30 ◦ C until use. An aliquot of 200 L of the serum sample was mixed with 20 L of ascorbate-EDTA solution (0.2 M phosphate buffer containing 20% ascorbic acid and 0.1% EDTA, pH 3.6), 20 L of 0.2 M phosphate buffer (pH 7.3), and 20 L of a mixture of 10 L of -glucuronidase (250 U) and 10 L of sulfatase (25 U). The mixture was incubated at 37 ◦ C for 45 min to allow for enzymatic hydrolysis. After incubation, 10 L of ethyl gallate (500 ng/mL) was added as an internal standard. The free catechins in solution were extracted twice by vortexing for 5 min with 2 mL of ethyl acetate. The ethyl acetate was pooled and evaporated to dryness using a nitrogen stream. The dried residue was redissolved in 100 L of mobile phase A. The solution was centrifuged at 2600 g for 5 min, and the supernatant was filtered through a 0.45 m membrane filter (Minisart SRP4). A
K. Narumi et al. / J. Chromatogr. B 945–946 (2014) 147–153
149
10 L aliquot was injected onto the HPLC. Blank extracts of serum were also prepared and analyzed concurrently.
and gave written informed consent to participate in this trial before the first treatment.
2.5. Selection of pre-treatment filter
2.9. Pharmacokinetic analysis
Blank serum spiked with standard analytes were extracted and filtered by six different filters. The filtrates were analyzed by HPLC as described above, using triplicate determination. The recovery was calculated by comparing the peak heights of known amounts of analytes spiked into blank serum and extracted with those obtained from the same amount of analytes spiked directly into post-extraction solvent. The tested filters were Ultrafree MC (Millipore, Tokyo, Japan), Nanosep MF (Pall Co., Port Washington, NY, US), Minisart NY15, Vivaspin 500, Minisart GF, and Minisart SRP4 (Sartorius Stedim, Tokyo, Japan).
The pharmacokinetic parameters of four target catechins from eight subjects were estimated using a one-compartment model. The maximum serum concentration (Cmax ) value and the corresponding time (Tmax ) were obtained directly from the individual serum catechin concentration–time course data. The area under the serum concentration–time curve (AUC) was calculated using the trapezoidal rule. The elimination rate constant (ke ) was obtained by linear regression analysis of the semi-logarithmic serum concentration–time curve and the elimination half-life (t1/2 ) was calculated from the ratio of 0.693/ke . The absorption rate constant (ka ) was calculated using the feathering method. The apparent volume of distribution (V/F) was calculated from AUC(0–∞) data according to V/F = (D/AUC(0–∞) )/ke . Numerical values are expressed as mean ± SD.
2.6. Recovery (extraction efficiency) The extraction efficiency of each analyte (four catechins and GA) was determined by triplicate determination in serum at three different concentration levels of 25, 100, and 400 ng/mL for EC, ECG, and GA; 50, 200, and 800 ng/mL for EGC and EGCG. The recovery was calculated by comparing the peak heights of known amounts of analytes spiked into blank serum and extracted with those obtained from the same amount of analytes spiked directly into post-extraction solvent. 2.7. Accuracy and precision Intra- and inter-assay precisions were assessed at three different concentrations using the coefficient of variation (CV): % CV =
standard deviation × 100 mean concentration
(1)
Intra-assay precision and accuracy were calculated using triplicate determinations for each concentration of the spiked serum sample during a single analytical run. Inter-assay precision and accuracy were calculated using the triplicate determinations of each concentration performed over 3 days. 2.8. Pharmacokinetic study Eight healthy volunteers (6 men and 2 women), between 23 and 59 years of age, participated in the study. The subjects did not ingest tea or tea-related beverages, food, or supplements for at least two days prior to the experiment. On an empty stomach, the volunteers ingested three tablets of decaffeinated green tea extract containing 16.7 mg of EC, 44.9 mg of EGC, 11.1 mg of ECG, and 42.9 mg of EGCG; GA was not a component of the tablet. Afterwards, no other beverages were ingested except water (the volume was not controlled). Blood samples from the subjects were collected in blood collection tubes before ingestion and at 0.5, 1.0, 1.5, 2, 3, 4, 6, and 8 h after ingestion of the green tea tablets. After collection, the blood samples were allowed to settle for approximately 30 min and centrifuged at 1200 g for 15 min to separate the serum supernatant. All samples were stored at −30 ◦ C until analysis. We determined the abundance of free and conjugated forms of catechin. -Glucuronidase and sulfatase were used before the extraction procedure to obtain the combined (total) amounts of free and conjugated forms of catechins. To determine the amount of free form catechins, a similar extraction procedure without the enzyme digestion was carried out. The study protocol and the procedure for informed consent were approved by the ethics committee of Kyushu University of Health & Welfare. The volunteer received written and verbal information
3. Results 3.1. Analysis of catechins and GA by HPLC First, the four target catechins, GA, and ethyl gallate were separated and quantified using the presented method. Since GA is a hydrophilic compound that elutes in high-polarity solvent along with components of the serum and enzyme solution, it was difficult to attain a single GA peak. Therefore, we adopted ion-pair chromatography using THA to prolong the retention time of GA and more effectively separate it from matrix components. The coulochem III detector consists of an analytical cell with two electrodes and a guard cell. In the analytical cell, one electrode is used as the filter electrode (E1) to eliminate possible electrochemical interference and the second electrode is used as the analytical electrode (E2). A guard cell was placed immediately before the injector to eliminate background noise derived from the solvent. To examine the reactivity of tea catechins to E2, the potential of E1 was fixed at 0 mV, and that of E2 was charged to 50–500 mV. Typically, the potential of the guard cell is set 50 mV higher than that of E2. EC, ECG, EGCG, and GA were efficiently oxidized with an applied potential of 200 mV (data not shown). Although the highest oxidative reaction for EGC was observed at potential of over 400 mV, EGC had a relatively stable response at a potential of 200 mV. Therefore, the potentials of E2 and the guard cell were set at 200 mV and 250 mV, respectively. To examine the reactivity of the catechins to E1, the potential of E2 was fixed at 200 mV, and that of E1 was charged to −50 mV to −200 mV. When the oxidative potential of E2 for catechins was set at 200 mV, the E1 potential less than −50 mV obtains an almost maximal response to E2. According to these findings, the applied potentials of E1, E2, and the guard cell were optimal at −50 mV, 200 mV, and 250 mV, respectively. Fig. 2 shows a typical chromatogram of (A) a standard mixture containing the four catechins and GA, (B) human serum blank, and (C) human serum after ingestion of green tea tablets. By comparing the chromatograph of the catechin standard with that of the serum blank, we determined that all the analysis compounds were separated and obtained as single peaks. The retention times for GA, EGC, EC, EGCG, ethyl gallate, and ECG were 18.4, 32.5, 39.0, 46.1, 47.4, and 56.1 min and their resolution values were 4.1, 12.0, 7.2, 4.0, 13.0, and 7.6, respectively. 3.2. Selection of pre-treatment filter Sample filtration as part of pre-treatment is a simple, economical practice that extends the life of consumable HPLC parts,
150
K. Narumi et al. / J. Chromatogr. B 945–946 (2014) 147–153 Table 1 Recovery of catechins and GA, and detection limit. Recovery
EC
EGC
ECG
EGCG
GA Fig. 2. HPLC chromatograms of (A) tea catechins and GA standards, (B) human serum blank, and (C) tea catechins and GA in human serum after ingestion of three green tea tablets. The serum was digested with -glucuronidase and sulfatase, extracted, and analyzed as described in Section 2.4.
decreases system wear and tear, and preserves the integrity of the HPLC system. Particularly vulnerable was the coulometric ECD system, which contains porous graphite electrodes susceptible to clogging by particulate matter. Initially, we used a hydrophilic PTFE membrane filter, shown in Fig. 3 (filter A, Ultrafree MC, hydrophilic PTFE membrane), for pre-treatment, but it appeared that the recoveries of ECG and EGCG using the filter were only 20% and 23%, respectively. To minimize filtration loss, six types of filter were evaluated. Filters B (Nanosep MF, hydrophilic polypropylene membrane) and C (Minisart NY15, nylon membrane) removed nearly all the catechins and ethyl gallate. Filter D (Vivaspin 500, polyethersulfone membrane) removed the gallated catechins completely. Filters E (Minisart GF, glass fiber membrane) appeared that the recoveries of analytes were similar to those of Filter A. Filter F (Minisart SRP4, hydrophobic PTFE membrane) provided optimal recovery among the six filters tested, so it was selected for sample pre-treatment. 3.3. Determination of extraction efficiency, recovery, and detection limit Analytical methods for determining low concentrations of catechins in biological matrices ideally include quantitative extraction
Detection limit (ng/mL)
Concentration (ng/mL)
Mean (%)
SD (%)
n
400 100 25 800 200 50 400 100 25 800 200 50 400 100 25
85.1 85.0 77.7 81.4 82.3 77.3 82.3 82.7 80.0 86.0 92.3 84.6 93.9 91.0 89.6
1.3 2.7 1.1 0.8 2.8 0.2 2.0 1.3 2.7 1.9 0.7 2.1 0.5 1.1 0.2
3 3 3 3 3 3 3 3 3 3 3 3 3 3 3
1.4
3.1
2.6
2.3
0.4
followed by sensitive and specific detection. To determine the recovery during sample preparation, 10 L of serial amounts of catechins (5, 20, and 80 ng for EC, ECG, and GA; 10, 40, and 160 ng for EGC and EGCG) were added to 190 L of control serum, and the sample was incubated, extracted, and analyzed. The recovery of the analytes was determined by comparing the recovery of known amounts of standard added to post-extraction solvent and to blank serum during the sample preparation. The overall recoveries of catechins and GA were 77.3–93.9% (Table 1). Using the experimental conditions described above, the detection limits (exceeding a signal-to-noise ratio of 3) of catechins and GA were 0.4–3.1 ng/mL (Table 1).
3.4. Linearity When different amounts of standard catechins and GA were added to serum samples, regression analysis of the peak height versus concentration exhibited linearity over the range of 10–1000 ng/mL in serum with correlation coefficient values R > 0.995.
Fig. 3. Comparison of catechin recovery rates using six different filter types: Filter A (Ultrafree MC), hydrophilic PTFE; Filter B (Nanosep MF), hydrophilic polypropylene; Filter C (Minisart NY15), nylon; Filter D (Vivaspin 500), polyethersulfone; Filter E (Minisart GF), glass fiber; Filter F (Minisart SRP4), hydrophobic PTFE.
K. Narumi et al. / J. Chromatogr. B 945–946 (2014) 147–153
151
Table 2 Intra-day and inter-day reproducibility of human serum samples. Intra-day assay Amount (ng/mL) EC
EGC
ECG
EGCG
GA
400.2 99.1 25.2 800.3 198.6 52.5 400.4 97.7 25.9 800.2 199.2 51.2 400.5 97.7 26.4
± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
5.6 1.4 0.6 11.3 1.9 0.6 10.7 1.7 0.9 20.8 5.8 1.6 5.6 0.9 0.1
Inter-day assay (3 days) CV (%)
n
Amount (ng/mL)
1.4 1.4 2.2 1.4 1.0 1.1 2.7 1.7 3.4 2.6 2.9 3.0 1.4 1.0 0.4
3 3 3 3 3 3 3 3 3 3 3 3 3 3 3
399.8 100.7 25.4 799.4 202.7 52.9 399.8 100.3 25.8 799.2 203.6 53.1 400.0 99.9 25.9
± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
9.9 2.5 1.1 16.6 5.3 1.5 10.5 2.8 1.2 20.1 4.9 2.1 8.9 2.0 0.7
CV (%)
n
2.5 2.4 4.5 2.1 2.6 2.9 2.6 2.8 4.8 2.5 2.4 3.9 2.2 2.0 2.8
9 9 9 9 9 9 9 9 9 9 9 9 9 9 9
The data are expressed as means ± SD and coefficient of variation. CV: coefficient of variation.
3.5. Accuracy and precision The intra- and inter-day reproducibilities of our method are shown in Table 2. The intra-day precision was in the range of 0.4–3.4%. The range of the inter-day precision was 2.0–4.8%. 3.6. Application to pharmacokinetic study Our method was successfully used within the context of a human pharmacokinetic study, readily allowing drug quantitation up to 8 h following oral ingestion of three green tea tablets. The mean serum total catechins (free + conjugated form) concentration–time curves are illustrated in Fig. 4A and pharmacokinetic parameters are summarized in Table 3. The concentration of target total catechins in serum reached its maximum at 60 min after ingestion. The highest concentration was shown for EGC, followed by EGCG, EC, and ECG. GA was not found under these conditions (Fig. 4A). The free form levels in serum were determined and plotted with the total catechin levels (Fig. 4B). When comparing the AUC(0–8 h) of total catechins with that of free form catechins, EC and EGC were mostly in the conjugated form in the serum (86.1% and 83.2%), whereas ECG and EGCG were mostly in the free form (68.0% and 64.7%). 4. Discussion The understanding of catechin pharmacokinetics after green tea ingestion in human subjects provides critical information to elucidate their physiological activities that give rise to their beneficial health effects, including reduction in the risk of cancer and cardiovascular disease as well as antioxidant and anti-inflammatory activities. Most catechins exist as glucuronized or sulfate conjugates in vivo, since their hydroxyl groups are an easy target for metabolic reactions, meaning that only a small portion of catechins are present in their free forms. To measure total catechins in serum, conjugated forms of catechins need to be hydrolyzed enzymatically. In general, -glucuronidase and sulfatase are typically used to convert conjugated catechins into the free form [6]. However, this treatment is simultaneously able to convert gallated catechins to their non-gallated form. The galloyl ester of GA, called a depside bond, is comparatively easy to cleave. It has been reported that galloyl ester is hydrolyzed by enzymes such as -glucuronidase and sulfatase as well as other enzymes (like tannase) produced by intestinal bacteria [17], decomposing gallated catechins into
Fig. 4. Mean serum concentration of the four target catechins versus time plot after ingestion of green tea tablets. (A) Total catechins of four catechins and GA, (B) the total catechins (a solid line) and free form catechins (a dotted line) for each catechin. For total catechins, the serum samples were digested with -glucuronidase and sulfatase and analyzed as described in Section 2. For free form catechins, the serum samples were analyzed without enzymatic digestion. Catechins: () EGC; () EC; () EGCG; () ECG; () GA.
non-gallated catechins and GA. This structural instability of gallated catechins makes precise quantitation of total catechins in biological matrices difficult. The nonspecific degradation reaction entails the hydrolysis of an ester bond in gallated catechins with the production of GA. Therefore, to evaluate the extent of degradation of the gallated catechins during the enzymatic hydrolysis process, we developed a new HPLC method for the simultaneous quantitation of GA and four target catechins. Conventional reversed-phase HPLC separation [16] did not enable us to isolate the GA peak from water-soluble impurities in the serum and enzyme solution, as the hydrophilic GA and these impurities overlapped with each other within a relatively small retention time. In order to overcome this issue and detect GA as a single peak, ion-pair chromatography using THA as an ion-pair agent was adopted. In reversed-phase chromatography, ionic analytes have weak interactions with the stationary phase because ionic compounds are surrounded by water molecules and therefore bear an electrical charge. By adding an ion with the opposite charge of the target analytes to the eluent, the low-polarity ionic analytes form an ion pair with the counter-ion to neutralize the charge, increasing their interaction with the stationary phase. Since GA contains a carboxyl group, it forms an ion pair with THA, significantly extending the retention time. The resolution values for the target analytes were 4.0–13.0, confirming complete separation of the target peaks. Thus, separation of GA from
152
K. Narumi et al. / J. Chromatogr. B 945–946 (2014) 147–153
Table 3 Pharmacokinetic parameters of catechins after administration of green tea tablets. Dosea (mg) EC EGC ECG EGCG a
16.7 44.9 11.1 42.9
± ± ± ±
0.2 0.4 0.1 0.5
Cmax (ng/mL) 34.7 60.7 20.9 42.8
± ± ± ±
24.8 41.8 8.2 8.2
Tmax (h) 1.4 1.3 1.2 0.9
± ± ± ±
0.3 0.3 0.4 0.2
AUC(0–8 h) (ng h/mL) 91.8 193.2 52.1 174.0
± ± ± ±
76.5 208.0 40.0 78.0
AUC(0–∞) (ng h/mL) 155.0 372.2 75.9 435.1
± ± ± ±
161.3 562.9 88.7 255.5
ke (h−1 ) 0.6 0.5 0.3 0.2
± ± ± ±
0.4 0.3 0.1 0.1
t1/2 (h) 1.5 1.9 2.5 4.4
± ± ± ±
0.9 1.0 0.6 1.6
ka (h−1 ) 2.2 2.7 3.6 6.3
± ± ± ±
1.3 1.4 0.9 2.2
Vd /F (L) 654.2 1058.7 797.5 1728.7
± ± ± ±
528.4 613.2 274.8 744.4
The contents of green tea extract tablets (3T) were measured by HPLC after extracted with 50% acetonitrile aqueous solution and expressed as mean ± SD, n = 3.
impurity peaks was achieved, making simultaneous determination of the four target catechins and GA possible. In our previous procedure, tetrahydrofuran (THF) and acetonitrile were used an organic constitution of mobile phase B to separate catechin peaks from blank peaks. However, when THF was contained in mobile phase B, the conditioning time after analysis was prolonged, presumably because the miscibility between mobile phases A and B was decreased. Therefore, we removed THF from the mobile phase and optimized the separation conditions by changing the gradient program. After this modification, conditioning time was shortened by 2/3, and the sensitivity and reproducibility of this method were improved. The concentration of catechins in plasma after oral ingestion of tea is on the order of a few hundred ng/mL [5,9,14,18], and does not necessarily scale to the amount ingested. Since the electrodes of the ECD (coulochem III) are made from very porous graphite, electroactive compounds like catechins in the eluent flow although the electrodes are completely oxidized (reduced), which maximizes the sensitivity of the analysis. We attempted to modify the extraction procedure to improve the recovery rate to increase the sensitivity of this method further. The evaluation of extraction efficiency was carried out in various solvents. The highest extraction efficiency among the seven extraction solvents tested was obtained from ethyl acetate (data not shown). Furthermore, extraction efficiency was improved by doubling the volume of solvent used. Another factor influencing recovery rate was the filter used in the pre-treatment procedure to protect the column and graphite electrodes of the ECD system from insoluble particulates in the sample. When filtered with a pre-wet PTFE membrane (prior to hydrophilic treatment), the recovery rate of gallated catechins decreased by half. Catechins are water-soluble because of their polyphenolic structure, and gallated catechins contain even more hydroxyl groups than non-gallated ones, causing them to adsorb strongly to the polar membrane. On the other hand, there was almost no adsorption of catechins to the PTFE membrane with no hydrophilic processing. By improving the recovery rate through optimization of the extraction procedure, particularly in selecting the best pretreatment filter and extraction solvent, we attained higher sensitivity than that of our previous method. The accuracy, reproducibility, and sensitivity of our improved method were sufficient to measure the target catechins in serum and GA concentrations, and are adequate for use to elucidate the pharmacokinetics of catechins in humans. Eight healthy volunteers participated in the study. The serum levels of catechins after the ingestion of green tea tablets were measured using this method, and the changes in the levels of the four catechins were tracked over time. Three green tea extract tablets taken by the subjects contained 16.7 mg of EC, 44.9 mg of EGC, 11.1 mg of ECG, and 42.9 mg of EGCG, and the mean Cmax (n = 8) reached 34.7, 60.6, 20.9, and 42.8 ng/mL, respectively (Table 3). These Cmax values indicated that the amount of catechin absorption was relatively low. One of the gallated catechins, EGCG, remained in serum 8 h after intake of the tablets, meaning that EGCG had a longer half-life than non-gallated catechins (Table 3), probably because gallated catechins remained in the free form in the bloodstream (Fig. 4A).
No GA was detected in any of the samples. In our preliminary experiments, samples that were spiked with the same amount of standard catechins in serum, treated with the deconjugation enzyme, yielded a smaller amount of gallated catechin and a larger amount of non-gallated catechin than expected. However, no GA peak was observed during the analysis of serum catechins (data not shown). These results suggest that enzymatic decomposition of gallated catechins was prevented by an unknown component in serum. It has been reported that gallated catechins, specifically EGCG, preferentially bind with albumin in blood and may be avoided by metabolic conjugation [19–21]. On the other hand, it is hypothesized that binding to protein in the blood protects the vulnerable gallate functional group from undergoing enzymatic hydrolysis, leading to the absence of GA from the human serum sample tested. Indeed, we demonstrated that the abundance of the free form was greater for gallated catechins (ECG, EGCG) than for non-gallated catechins (EC, EGC) after deproteinization by ethyl acetate (Fig. 4B). This may reflect that gallated catechins bind with higher affinity to blood protein than non-gallated ones. However, there are only a few reports on the protein binding rate, pharmacokinetics, and predominant metabolic form of catechins in blood, and more detailed research is need to confirm this hypothesis. 5. Conclusions We developed an ion-pair HPLC-ECD method for the simultaneous quantitation of GA and four target catechins. Target catechins were enzymatically deconjugated to measure unconjugated free catechins. GA was measured to estimate the amount of gallate moiety produced by degradation of gallated catechins. The accuracy, reproducibility, and sensitivity of our improved method were sufficient to measure the target catechins in serum and GA concentrations. Catechin concentrations in human serum after green tea tablet ingestion were determined using this assay. The four target catechins were successfully detected, and no GA was found under these conditions. This method is applicable to assess the kinetics of catechins after tea consumption and is considered to be useful to clarify the beneficial health effects of catechins. Acknowledgements We sincerely thank Dr. Shunro Sonoda, President CEO of Satsuma Green Tea International, Inc. (Kagoshima, Japan) for providing decaffeinated green tea tablets. References [1] [2] [3] [4] [5] [6] [7] [8]
C.S. Yang, Z.Y. Wang, J. Natl. Cancer Inst. 85 (1993) 1038. I.E. Dreosti, M.J. Wargovich, C.S. Yang, Crit. Rev. Food Sci. Nutr. 37 (1997) 761. P.C. Hollman, L.B. Tijburg, C.S. Yang, Crit. Rev. Food Sci. Nutr. 37 (1997) 719. R. Mohan, S. Katiyar, S. Khan, M. Zaim, H. Mukhtar, R. Agarwal, Int. J. Oncol. 8 (1996) 1079. M.J. Lee, Z.Y. Wang, H. Li, L. Chen, Y. Sun, S. Gobbo, D.A. Balentine, C.S. Yang, Cancer Epidemiol. Biomarkers Prev. 4 (1995) 393. K. Umegaki, A. Sugisawa, K. Yamada, M. Higuchi, J. Nutr. Sci. Vitaminol. 47 (2001) 402. K. Nakagawa, T. Miyazawa, Anal. Biochem. 248 (1997) 41. B.L. Lee, C.N. Ong, J. Chromatogr. A 881 (2000) 439.
K. Narumi et al. / J. Chromatogr. B 945–946 (2014) 147–153 [9] G.J. Soleas, J. Yan, D.M. Goldberg, J. Chromatogr. B: Biomed. Sci. Appl. 757 (2001) 161. [10] B. Yang, K. Arai, F. Kusu, Anal. Biochem. 283 (2000) 77. [11] T. Unno, T. Takeo, Biosci. Biotechnol. Biochem. 59 (1995) 1558. [12] S. Carando, P.L. Teissedre, J.C. Cabanis, J. Chromatogr. B: Biomed. Sci. Appl. 707 (1998) 195. [13] T. Unno, K. Kondo, H. Itakura, T. Takeo, Biosci. Biotechnol. Biochem. 60 (1996) 2066. [14] A. Kotani, N. Miyashita, F. Kusu, J. Chromatogr. B 788 (2003) 269. [15] T. Fu, J. Liang, G. Han, L. Lv, N. Li, J. Chromatogr. B 875 (2008) 363. [16] M.J. Lee, S. Prebhu, X. Meng, C. Li, C.S. Yang, Anal. Biochem. 279 (2000) 164.
153
[17] Y.-H. Hong, E.Y. Jung, K.-S. Shin, T.Y. Kim, K.-W. Yu, U.J. Chang, H.J. Suh, Appl. Biochem. Biotechnol. 166 (2012) 165. [18] H.-H.S. Chow, Y. Cai, D.S. Alberts, I. Hakim, R. Dorr, F. Shahi, J.A. Crowell, C.S. Yang, Y. Hara, Cancer Epidemiol. Biomarkers Prev. 10 (2001) 53. [19] T. Ishii, T. Ichikawa, K. Minoda, K. Kusaka, S. Ito, Y. Suzuki, M. Akagawa, K. Mochizuki, T. Goda, T. Nakayama, Biosci. Biotechnol. Biochem. 75 (2011) 100. [20] K. Minoda, T. Ichikawa, T. Katsumata, K. Onobori, T. Mori, Y. Suzuki, T. Ishii, T. Nakayama, J. Nutr. Sci. Vitaminol. 56 (2010) 331. [21] M.J. Bae, T. Ishii, K. Minoda, Y. Kawada, T. Ichikawa, T. Mori, M. Kamihira, T. Nakayama, Mol. Nutr. Food Res. 53 (2009) 709.
Contents lists available at ScienceDirect
Journal of Chromatography B journal homepage: www.elsevier.com/locate/chromb
Simultaneous detection of green tea catechins and gallic acid in human serum after ingestion of green tea tablets using ion-pair high-performance liquid chromatography with electrochemical detection Keiko Narumi a , Jun-Ichiro Sonoda a,∗ , Keita Shiotani a , Michihiro Shigeru a , Masayuki Shibata a , Akio Kawachi a , Erisa Tomishige a , Keizo Sato b , Toshiro Motoya a a First Department of Clinical Pharmacy, School of Pharmaceutical Sciences, Kyushu University of Health & Welfare, 1714-1 Yoshino-cho, Nobeoka, Miyazaki 882-8508, Japan b Department of Clinical Biochemistry, School of Pharmaceutical Sciences, Kyushu University of Health & Welfare, 1714-1 Yoshino-cho, Nobeoka, Miyazaki 882-8508, Japan
a r t i c l e
i n f o
Article history: Received 28 June 2013 Accepted 4 November 2013 Available online 1 December 2013 Keywords: Green tea catechins Gallic acid HPLC Ion-pair reagent Electrochemical detection Pharmacokinetic study
a b s t r a c t We developed an analytical method for the simultaneous determination of tea catechins and gallic acid (GA) in human serum using ion-pair high-performance liquid chromatography (HPLC) with electrochemical detection. GA was measured to estimate the amount of gallate moiety produced by degradation of gallated catechins ((−)-epicatechin-3-gallate, ECG; (−)-epigallocatechin-3-gallate, EGCG). Ethyl gallate was adopted as an internal standard to correct for the extraction efficiency. To maximize extraction efficiency, a hydrophobic polytetrafluoroethylene (PTFE) filter was selected for pre-treatment prior to separation. HPLC separation was performed using a C18 reversed-phase column with a gradient mobile phase of phosphate buffer (pH 2.5) containing tetrahexylammonium hydrogensulfate as an ion-pair reagent. Using this method, (−)-epicatechin (EC), (−)-epigallocatechin (EGC), ECG, EGCG, ethyl gallate, and GA were detected as single peaks. The resolution values for target analytes were 4.0–13.0 and the mean values of the absolute recoveries of catechins and GA were 77.3–93.9%. The detection limits for catechins and GA in serum were 0.4–3.1 ng/mL. The serum catechin levels of eight healthy volunteers after ingestion of a single dose of green tea tablets were measured using this method. The concentration of total catechins (free + conjugated forms) in serum peaked 60 min after ingestion. From these results, this method is thought to enable the simultaneous quantification of GA, the hydrolysis product of gallated catechins, and target catechins, and to be sufficiently sensitive for pharmacokinetic studies of catechins following oral administration of green tea. © 2013 Elsevier B.V. All rights reserved.
1. Introduction Tea, including green, black, and oolong varieties, is made from the leaves of the unique tea plant (Camellia sinensis), and has been the most widely consumed drink in the world for thousands of years. Among the different types of tea, green (Japanese) tea has been reported to have beneficial health effects, such as reducing the risk of cancer and cardiovascular diseases [1–4]. The active constituents of green tea are thought to be polyphenols, commonly known as tea catechins. The major tea catechins are (−)-epicatechin
∗ Corresponding author at: First Department of Clinical Pharmacy, School of Pharmaceutical Sciences, Kyushu University of Health & Welfare, 1714-1 Yoshino-cho, Nobeoka, Miyazaki 882-8508, Japan. Tel.: +81 982 23 5529; fax: +81 982 23 5536. E-mail address: [email protected] (J.-I. Sonoda). 1570-0232/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jchromb.2013.11.007
(EC), (−)-epigallocatechin (EGC), (−)-epicatechin-3-gallate (ECG), and (−)-epigallocatechin-3-gallate (EGCG) (Fig. 1). One cup of regular green tea contains several hundred milligrams of catechins. After ingestion, the serum concentration of catechins reaches a maximum between several tens and several hundreds of ng/mL. The reasons for the poor transfer of catechins to the bloodstream and the wide variation in pharmacokinetics of catechins after green tea ingestion are not clearly understood [5]. In order to assess the kinetics and the beneficial health effects of catechins, a sensitive, reproducible, and straightforward method is needed to determine serum catechins. Generally, tea catechins are analyzed using high-performance liquid chromatography (HPLC) with ultraviolet (UV) [6], chemiluminescence (CL) [7], or electrochemical detection (ECD); also commonly used are capillary electrophoresis (CE) [8] and gas chromatography with mass spectrometry (GC–MS) [9]. These methods
148
K. Narumi et al. / J. Chromatogr. B 945–946 (2014) 147–153
from Nacalai Tesque (Kyoto, Japan). Ethyl gallate was purchased from Tokyo Chemical Industry (Tokyo, Japan). EC, EGC, EGCG, L-ascorbic acid, -glucuronidase (G8420), sulfatase (Type VIII: S9754), and tetrahexylammonium hydrogensulfate (THA) were purchased from Sigma–Aldrich (St. Louis, MO, USA). Other chemicals and solvents used were of HPLC grade. Decaffeinated green tea tablets were provided by Sathuma Green Tea International, Inc. (Kagoshima, Japan). 2.2. Apparatus
Fig. 1. Structures of the four target catechins.
have been used for determining catechins in tea [8,10], rat plasma [11], human plasma [6,7,12,13], and urine [10]. Although GC–MS can be used for the determination of catechins with high sensitivity [9], the required sample pre-treatment is too demanding for routine analysis. However, HPLC is easier to use and more widely available. Previous work has shown that HPLC-ECD is more sensitive for the determination of catechins [14] than HPLC-UV [15]. Moreover, ECD is well suited to the detection of polyphenolic compounds like catechins because these analytes can be selectively oxidized by the appropriate electric potential to flow current. The magnitude of this current is related to the analyte concentration. Therefore, HPLC-ECD is an excellent candidate technique for the selective and highly sensitive detection of catechins in human serum. After the ingestion of green tea, a large proportion of catechins are present as glucuronate or sulfate conjugates in the bloodstream. For the quantification of total catechins in serum, samples were treated with -glucuronidase and sulfatase to hydrolyze the conjugates to free form catechins. Type X-A -glucuronidase is typically used to hydrolyze glucuronic acid conjugates [14,16]. It has been reported that type H-2 -glucuronidase, which contains -glucuronidases other than sulfatase, effectively hydrolyzed conjugates in plasma samples but concomitantly cleaved the ester bonds of EGCG and ECG [6]. As such, these enzymes converted gallated catechins to their non-gallated forms (EGC and EC). In this case, total catechins are expressed as EGCG + EGC and ECG + EC, which unfortunately makes it difficult to elucidate the exact concentrations of each individual catechin because some EGC and EC are derived from EGCG and ECG via treatment with type H-2 glucuronidase. Therefore, gallic acid (GA), derived from the gallated catechins, must be measured simultaneously to elucidate whether or not cleavage occurs. In this study, we developed a sensitive and practical method for catechin determination consisting of digestion of conjugates followed by HPLC-ECD. Using this method, we evaluated the timedependent changes in the concentrations of tea catechins in serum after the ingestion of a single dose of green tea tablets (equivalent to one cup of typical green tea) in human subjects. 2. Materials and methods 2.1. Chemicals and reagents GA, ECG, acetonitrile, phosphoric acid, and di-sodium dihydrogen ethylenediamine-tetraacetate dihydrate (EDTA) were obtained
A Shimadzu (Kyoto, Japan) HPLC system consisting of a DGU14AM degasser, LC-10AD VP pumps, SIL-10AD autosampler, and SCL-10A system controller was used. The polyphenolic compounds were detected using a Coulochem III coulometric ECD system with a guard cell (Model 5020) and analytical cell (Model 5011; ESA, MA, USA). HPLC separations were performed on a C18 reversed-phase column (Mightysil RP-18 GP aqua, 250 mm × 4.6 mm, i.d., particle size, 5 m; Kanto Chemical, Tokyo, Japan). Data acquisition and analysis were performed with Chromeleon version 6.40 software (Dionex, CA, USA). 2.3. HPLC conditions Phosphate buffer (50 mM) containing 0.05 mM EDTA and 5 mM THA as an ion-pair reagent was adjusted to pH 2.5 with phosphoric acid. Mobile phase A was a deaerated mixture of 99:1 (v/v) phosphate buffer (pH 2.5):acetonitrile. Mobile phase B was a deaerated mixture of 77:23 (v/v) phosphate buffer (pH 2.5):acetonitrile. The gradient program was as follows: 0–6.00 min, 100% A; 6.01–30.00 min, B 0–42%; 30.01–35.00 min, B 42–100%; 35.01–37.00 min, 100% B; 37.01–39.00 min, B 100–95%; 39.01–41.00 min, B 95%; 41.01–46.00 min, B 95–80%; 46.01–57.00 min, B 80%; and 57.01–85.0 min, 100% A. The eluent was monitored by coulometric ECD (filter mode) with the guard cell potential at 250 mV and analytical cell potentials for electrode 1 and 2 at −50 and 200 mV, respectively. The flow rate was kept at 1.0 mL/min from 0 to 57.00 min and then increased at 1.5 mL/min from 57.01 to 80.00 min to shorten the conditioning time. The injected volume was 10 L. The column was housed in a temperature-regulated compartment maintained at 40 ◦ C, and the autosampler was maintained at 4 ◦ C. 2.4. Preparation of standard and serum samples Standard stock solutions of EC, EGC, ECG, EGCG, GA, and ethyl gallate as an internal standard were prepared at a concentration of 100 g/mL in 0.2 M phosphate buffer (pH 3.6) with 2% ascorbic acid/0.01% EDTA. -Glucuronidase and sulfatase were prepared by dissolving in 75 mM phosphate buffer (pH 6.8) to concentrations of 250 U/10 L and 25 U/10 L, respectively, and were stored in small aliquots at −30 ◦ C until use. An aliquot of 200 L of the serum sample was mixed with 20 L of ascorbate-EDTA solution (0.2 M phosphate buffer containing 20% ascorbic acid and 0.1% EDTA, pH 3.6), 20 L of 0.2 M phosphate buffer (pH 7.3), and 20 L of a mixture of 10 L of -glucuronidase (250 U) and 10 L of sulfatase (25 U). The mixture was incubated at 37 ◦ C for 45 min to allow for enzymatic hydrolysis. After incubation, 10 L of ethyl gallate (500 ng/mL) was added as an internal standard. The free catechins in solution were extracted twice by vortexing for 5 min with 2 mL of ethyl acetate. The ethyl acetate was pooled and evaporated to dryness using a nitrogen stream. The dried residue was redissolved in 100 L of mobile phase A. The solution was centrifuged at 2600 g for 5 min, and the supernatant was filtered through a 0.45 m membrane filter (Minisart SRP4). A
K. Narumi et al. / J. Chromatogr. B 945–946 (2014) 147–153
149
10 L aliquot was injected onto the HPLC. Blank extracts of serum were also prepared and analyzed concurrently.
and gave written informed consent to participate in this trial before the first treatment.
2.5. Selection of pre-treatment filter
2.9. Pharmacokinetic analysis
Blank serum spiked with standard analytes were extracted and filtered by six different filters. The filtrates were analyzed by HPLC as described above, using triplicate determination. The recovery was calculated by comparing the peak heights of known amounts of analytes spiked into blank serum and extracted with those obtained from the same amount of analytes spiked directly into post-extraction solvent. The tested filters were Ultrafree MC (Millipore, Tokyo, Japan), Nanosep MF (Pall Co., Port Washington, NY, US), Minisart NY15, Vivaspin 500, Minisart GF, and Minisart SRP4 (Sartorius Stedim, Tokyo, Japan).
The pharmacokinetic parameters of four target catechins from eight subjects were estimated using a one-compartment model. The maximum serum concentration (Cmax ) value and the corresponding time (Tmax ) were obtained directly from the individual serum catechin concentration–time course data. The area under the serum concentration–time curve (AUC) was calculated using the trapezoidal rule. The elimination rate constant (ke ) was obtained by linear regression analysis of the semi-logarithmic serum concentration–time curve and the elimination half-life (t1/2 ) was calculated from the ratio of 0.693/ke . The absorption rate constant (ka ) was calculated using the feathering method. The apparent volume of distribution (V/F) was calculated from AUC(0–∞) data according to V/F = (D/AUC(0–∞) )/ke . Numerical values are expressed as mean ± SD.
2.6. Recovery (extraction efficiency) The extraction efficiency of each analyte (four catechins and GA) was determined by triplicate determination in serum at three different concentration levels of 25, 100, and 400 ng/mL for EC, ECG, and GA; 50, 200, and 800 ng/mL for EGC and EGCG. The recovery was calculated by comparing the peak heights of known amounts of analytes spiked into blank serum and extracted with those obtained from the same amount of analytes spiked directly into post-extraction solvent. 2.7. Accuracy and precision Intra- and inter-assay precisions were assessed at three different concentrations using the coefficient of variation (CV): % CV =
standard deviation × 100 mean concentration
(1)
Intra-assay precision and accuracy were calculated using triplicate determinations for each concentration of the spiked serum sample during a single analytical run. Inter-assay precision and accuracy were calculated using the triplicate determinations of each concentration performed over 3 days. 2.8. Pharmacokinetic study Eight healthy volunteers (6 men and 2 women), between 23 and 59 years of age, participated in the study. The subjects did not ingest tea or tea-related beverages, food, or supplements for at least two days prior to the experiment. On an empty stomach, the volunteers ingested three tablets of decaffeinated green tea extract containing 16.7 mg of EC, 44.9 mg of EGC, 11.1 mg of ECG, and 42.9 mg of EGCG; GA was not a component of the tablet. Afterwards, no other beverages were ingested except water (the volume was not controlled). Blood samples from the subjects were collected in blood collection tubes before ingestion and at 0.5, 1.0, 1.5, 2, 3, 4, 6, and 8 h after ingestion of the green tea tablets. After collection, the blood samples were allowed to settle for approximately 30 min and centrifuged at 1200 g for 15 min to separate the serum supernatant. All samples were stored at −30 ◦ C until analysis. We determined the abundance of free and conjugated forms of catechin. -Glucuronidase and sulfatase were used before the extraction procedure to obtain the combined (total) amounts of free and conjugated forms of catechins. To determine the amount of free form catechins, a similar extraction procedure without the enzyme digestion was carried out. The study protocol and the procedure for informed consent were approved by the ethics committee of Kyushu University of Health & Welfare. The volunteer received written and verbal information
3. Results 3.1. Analysis of catechins and GA by HPLC First, the four target catechins, GA, and ethyl gallate were separated and quantified using the presented method. Since GA is a hydrophilic compound that elutes in high-polarity solvent along with components of the serum and enzyme solution, it was difficult to attain a single GA peak. Therefore, we adopted ion-pair chromatography using THA to prolong the retention time of GA and more effectively separate it from matrix components. The coulochem III detector consists of an analytical cell with two electrodes and a guard cell. In the analytical cell, one electrode is used as the filter electrode (E1) to eliminate possible electrochemical interference and the second electrode is used as the analytical electrode (E2). A guard cell was placed immediately before the injector to eliminate background noise derived from the solvent. To examine the reactivity of tea catechins to E2, the potential of E1 was fixed at 0 mV, and that of E2 was charged to 50–500 mV. Typically, the potential of the guard cell is set 50 mV higher than that of E2. EC, ECG, EGCG, and GA were efficiently oxidized with an applied potential of 200 mV (data not shown). Although the highest oxidative reaction for EGC was observed at potential of over 400 mV, EGC had a relatively stable response at a potential of 200 mV. Therefore, the potentials of E2 and the guard cell were set at 200 mV and 250 mV, respectively. To examine the reactivity of the catechins to E1, the potential of E2 was fixed at 200 mV, and that of E1 was charged to −50 mV to −200 mV. When the oxidative potential of E2 for catechins was set at 200 mV, the E1 potential less than −50 mV obtains an almost maximal response to E2. According to these findings, the applied potentials of E1, E2, and the guard cell were optimal at −50 mV, 200 mV, and 250 mV, respectively. Fig. 2 shows a typical chromatogram of (A) a standard mixture containing the four catechins and GA, (B) human serum blank, and (C) human serum after ingestion of green tea tablets. By comparing the chromatograph of the catechin standard with that of the serum blank, we determined that all the analysis compounds were separated and obtained as single peaks. The retention times for GA, EGC, EC, EGCG, ethyl gallate, and ECG were 18.4, 32.5, 39.0, 46.1, 47.4, and 56.1 min and their resolution values were 4.1, 12.0, 7.2, 4.0, 13.0, and 7.6, respectively. 3.2. Selection of pre-treatment filter Sample filtration as part of pre-treatment is a simple, economical practice that extends the life of consumable HPLC parts,
150
K. Narumi et al. / J. Chromatogr. B 945–946 (2014) 147–153 Table 1 Recovery of catechins and GA, and detection limit. Recovery
EC
EGC
ECG
EGCG
GA Fig. 2. HPLC chromatograms of (A) tea catechins and GA standards, (B) human serum blank, and (C) tea catechins and GA in human serum after ingestion of three green tea tablets. The serum was digested with -glucuronidase and sulfatase, extracted, and analyzed as described in Section 2.4.
decreases system wear and tear, and preserves the integrity of the HPLC system. Particularly vulnerable was the coulometric ECD system, which contains porous graphite electrodes susceptible to clogging by particulate matter. Initially, we used a hydrophilic PTFE membrane filter, shown in Fig. 3 (filter A, Ultrafree MC, hydrophilic PTFE membrane), for pre-treatment, but it appeared that the recoveries of ECG and EGCG using the filter were only 20% and 23%, respectively. To minimize filtration loss, six types of filter were evaluated. Filters B (Nanosep MF, hydrophilic polypropylene membrane) and C (Minisart NY15, nylon membrane) removed nearly all the catechins and ethyl gallate. Filter D (Vivaspin 500, polyethersulfone membrane) removed the gallated catechins completely. Filters E (Minisart GF, glass fiber membrane) appeared that the recoveries of analytes were similar to those of Filter A. Filter F (Minisart SRP4, hydrophobic PTFE membrane) provided optimal recovery among the six filters tested, so it was selected for sample pre-treatment. 3.3. Determination of extraction efficiency, recovery, and detection limit Analytical methods for determining low concentrations of catechins in biological matrices ideally include quantitative extraction
Detection limit (ng/mL)
Concentration (ng/mL)
Mean (%)
SD (%)
n
400 100 25 800 200 50 400 100 25 800 200 50 400 100 25
85.1 85.0 77.7 81.4 82.3 77.3 82.3 82.7 80.0 86.0 92.3 84.6 93.9 91.0 89.6
1.3 2.7 1.1 0.8 2.8 0.2 2.0 1.3 2.7 1.9 0.7 2.1 0.5 1.1 0.2
3 3 3 3 3 3 3 3 3 3 3 3 3 3 3
1.4
3.1
2.6
2.3
0.4
followed by sensitive and specific detection. To determine the recovery during sample preparation, 10 L of serial amounts of catechins (5, 20, and 80 ng for EC, ECG, and GA; 10, 40, and 160 ng for EGC and EGCG) were added to 190 L of control serum, and the sample was incubated, extracted, and analyzed. The recovery of the analytes was determined by comparing the recovery of known amounts of standard added to post-extraction solvent and to blank serum during the sample preparation. The overall recoveries of catechins and GA were 77.3–93.9% (Table 1). Using the experimental conditions described above, the detection limits (exceeding a signal-to-noise ratio of 3) of catechins and GA were 0.4–3.1 ng/mL (Table 1).
3.4. Linearity When different amounts of standard catechins and GA were added to serum samples, regression analysis of the peak height versus concentration exhibited linearity over the range of 10–1000 ng/mL in serum with correlation coefficient values R > 0.995.
Fig. 3. Comparison of catechin recovery rates using six different filter types: Filter A (Ultrafree MC), hydrophilic PTFE; Filter B (Nanosep MF), hydrophilic polypropylene; Filter C (Minisart NY15), nylon; Filter D (Vivaspin 500), polyethersulfone; Filter E (Minisart GF), glass fiber; Filter F (Minisart SRP4), hydrophobic PTFE.
K. Narumi et al. / J. Chromatogr. B 945–946 (2014) 147–153
151
Table 2 Intra-day and inter-day reproducibility of human serum samples. Intra-day assay Amount (ng/mL) EC
EGC
ECG
EGCG
GA
400.2 99.1 25.2 800.3 198.6 52.5 400.4 97.7 25.9 800.2 199.2 51.2 400.5 97.7 26.4
± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
5.6 1.4 0.6 11.3 1.9 0.6 10.7 1.7 0.9 20.8 5.8 1.6 5.6 0.9 0.1
Inter-day assay (3 days) CV (%)
n
Amount (ng/mL)
1.4 1.4 2.2 1.4 1.0 1.1 2.7 1.7 3.4 2.6 2.9 3.0 1.4 1.0 0.4
3 3 3 3 3 3 3 3 3 3 3 3 3 3 3
399.8 100.7 25.4 799.4 202.7 52.9 399.8 100.3 25.8 799.2 203.6 53.1 400.0 99.9 25.9
± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
9.9 2.5 1.1 16.6 5.3 1.5 10.5 2.8 1.2 20.1 4.9 2.1 8.9 2.0 0.7
CV (%)
n
2.5 2.4 4.5 2.1 2.6 2.9 2.6 2.8 4.8 2.5 2.4 3.9 2.2 2.0 2.8
9 9 9 9 9 9 9 9 9 9 9 9 9 9 9
The data are expressed as means ± SD and coefficient of variation. CV: coefficient of variation.
3.5. Accuracy and precision The intra- and inter-day reproducibilities of our method are shown in Table 2. The intra-day precision was in the range of 0.4–3.4%. The range of the inter-day precision was 2.0–4.8%. 3.6. Application to pharmacokinetic study Our method was successfully used within the context of a human pharmacokinetic study, readily allowing drug quantitation up to 8 h following oral ingestion of three green tea tablets. The mean serum total catechins (free + conjugated form) concentration–time curves are illustrated in Fig. 4A and pharmacokinetic parameters are summarized in Table 3. The concentration of target total catechins in serum reached its maximum at 60 min after ingestion. The highest concentration was shown for EGC, followed by EGCG, EC, and ECG. GA was not found under these conditions (Fig. 4A). The free form levels in serum were determined and plotted with the total catechin levels (Fig. 4B). When comparing the AUC(0–8 h) of total catechins with that of free form catechins, EC and EGC were mostly in the conjugated form in the serum (86.1% and 83.2%), whereas ECG and EGCG were mostly in the free form (68.0% and 64.7%). 4. Discussion The understanding of catechin pharmacokinetics after green tea ingestion in human subjects provides critical information to elucidate their physiological activities that give rise to their beneficial health effects, including reduction in the risk of cancer and cardiovascular disease as well as antioxidant and anti-inflammatory activities. Most catechins exist as glucuronized or sulfate conjugates in vivo, since their hydroxyl groups are an easy target for metabolic reactions, meaning that only a small portion of catechins are present in their free forms. To measure total catechins in serum, conjugated forms of catechins need to be hydrolyzed enzymatically. In general, -glucuronidase and sulfatase are typically used to convert conjugated catechins into the free form [6]. However, this treatment is simultaneously able to convert gallated catechins to their non-gallated form. The galloyl ester of GA, called a depside bond, is comparatively easy to cleave. It has been reported that galloyl ester is hydrolyzed by enzymes such as -glucuronidase and sulfatase as well as other enzymes (like tannase) produced by intestinal bacteria [17], decomposing gallated catechins into
Fig. 4. Mean serum concentration of the four target catechins versus time plot after ingestion of green tea tablets. (A) Total catechins of four catechins and GA, (B) the total catechins (a solid line) and free form catechins (a dotted line) for each catechin. For total catechins, the serum samples were digested with -glucuronidase and sulfatase and analyzed as described in Section 2. For free form catechins, the serum samples were analyzed without enzymatic digestion. Catechins: () EGC; () EC; () EGCG; () ECG; () GA.
non-gallated catechins and GA. This structural instability of gallated catechins makes precise quantitation of total catechins in biological matrices difficult. The nonspecific degradation reaction entails the hydrolysis of an ester bond in gallated catechins with the production of GA. Therefore, to evaluate the extent of degradation of the gallated catechins during the enzymatic hydrolysis process, we developed a new HPLC method for the simultaneous quantitation of GA and four target catechins. Conventional reversed-phase HPLC separation [16] did not enable us to isolate the GA peak from water-soluble impurities in the serum and enzyme solution, as the hydrophilic GA and these impurities overlapped with each other within a relatively small retention time. In order to overcome this issue and detect GA as a single peak, ion-pair chromatography using THA as an ion-pair agent was adopted. In reversed-phase chromatography, ionic analytes have weak interactions with the stationary phase because ionic compounds are surrounded by water molecules and therefore bear an electrical charge. By adding an ion with the opposite charge of the target analytes to the eluent, the low-polarity ionic analytes form an ion pair with the counter-ion to neutralize the charge, increasing their interaction with the stationary phase. Since GA contains a carboxyl group, it forms an ion pair with THA, significantly extending the retention time. The resolution values for the target analytes were 4.0–13.0, confirming complete separation of the target peaks. Thus, separation of GA from
152
K. Narumi et al. / J. Chromatogr. B 945–946 (2014) 147–153
Table 3 Pharmacokinetic parameters of catechins after administration of green tea tablets. Dosea (mg) EC EGC ECG EGCG a
16.7 44.9 11.1 42.9
± ± ± ±
0.2 0.4 0.1 0.5
Cmax (ng/mL) 34.7 60.7 20.9 42.8
± ± ± ±
24.8 41.8 8.2 8.2
Tmax (h) 1.4 1.3 1.2 0.9
± ± ± ±
0.3 0.3 0.4 0.2
AUC(0–8 h) (ng h/mL) 91.8 193.2 52.1 174.0
± ± ± ±
76.5 208.0 40.0 78.0
AUC(0–∞) (ng h/mL) 155.0 372.2 75.9 435.1
± ± ± ±
161.3 562.9 88.7 255.5
ke (h−1 ) 0.6 0.5 0.3 0.2
± ± ± ±
0.4 0.3 0.1 0.1
t1/2 (h) 1.5 1.9 2.5 4.4
± ± ± ±
0.9 1.0 0.6 1.6
ka (h−1 ) 2.2 2.7 3.6 6.3
± ± ± ±
1.3 1.4 0.9 2.2
Vd /F (L) 654.2 1058.7 797.5 1728.7
± ± ± ±
528.4 613.2 274.8 744.4
The contents of green tea extract tablets (3T) were measured by HPLC after extracted with 50% acetonitrile aqueous solution and expressed as mean ± SD, n = 3.
impurity peaks was achieved, making simultaneous determination of the four target catechins and GA possible. In our previous procedure, tetrahydrofuran (THF) and acetonitrile were used an organic constitution of mobile phase B to separate catechin peaks from blank peaks. However, when THF was contained in mobile phase B, the conditioning time after analysis was prolonged, presumably because the miscibility between mobile phases A and B was decreased. Therefore, we removed THF from the mobile phase and optimized the separation conditions by changing the gradient program. After this modification, conditioning time was shortened by 2/3, and the sensitivity and reproducibility of this method were improved. The concentration of catechins in plasma after oral ingestion of tea is on the order of a few hundred ng/mL [5,9,14,18], and does not necessarily scale to the amount ingested. Since the electrodes of the ECD (coulochem III) are made from very porous graphite, electroactive compounds like catechins in the eluent flow although the electrodes are completely oxidized (reduced), which maximizes the sensitivity of the analysis. We attempted to modify the extraction procedure to improve the recovery rate to increase the sensitivity of this method further. The evaluation of extraction efficiency was carried out in various solvents. The highest extraction efficiency among the seven extraction solvents tested was obtained from ethyl acetate (data not shown). Furthermore, extraction efficiency was improved by doubling the volume of solvent used. Another factor influencing recovery rate was the filter used in the pre-treatment procedure to protect the column and graphite electrodes of the ECD system from insoluble particulates in the sample. When filtered with a pre-wet PTFE membrane (prior to hydrophilic treatment), the recovery rate of gallated catechins decreased by half. Catechins are water-soluble because of their polyphenolic structure, and gallated catechins contain even more hydroxyl groups than non-gallated ones, causing them to adsorb strongly to the polar membrane. On the other hand, there was almost no adsorption of catechins to the PTFE membrane with no hydrophilic processing. By improving the recovery rate through optimization of the extraction procedure, particularly in selecting the best pretreatment filter and extraction solvent, we attained higher sensitivity than that of our previous method. The accuracy, reproducibility, and sensitivity of our improved method were sufficient to measure the target catechins in serum and GA concentrations, and are adequate for use to elucidate the pharmacokinetics of catechins in humans. Eight healthy volunteers participated in the study. The serum levels of catechins after the ingestion of green tea tablets were measured using this method, and the changes in the levels of the four catechins were tracked over time. Three green tea extract tablets taken by the subjects contained 16.7 mg of EC, 44.9 mg of EGC, 11.1 mg of ECG, and 42.9 mg of EGCG, and the mean Cmax (n = 8) reached 34.7, 60.6, 20.9, and 42.8 ng/mL, respectively (Table 3). These Cmax values indicated that the amount of catechin absorption was relatively low. One of the gallated catechins, EGCG, remained in serum 8 h after intake of the tablets, meaning that EGCG had a longer half-life than non-gallated catechins (Table 3), probably because gallated catechins remained in the free form in the bloodstream (Fig. 4A).
No GA was detected in any of the samples. In our preliminary experiments, samples that were spiked with the same amount of standard catechins in serum, treated with the deconjugation enzyme, yielded a smaller amount of gallated catechin and a larger amount of non-gallated catechin than expected. However, no GA peak was observed during the analysis of serum catechins (data not shown). These results suggest that enzymatic decomposition of gallated catechins was prevented by an unknown component in serum. It has been reported that gallated catechins, specifically EGCG, preferentially bind with albumin in blood and may be avoided by metabolic conjugation [19–21]. On the other hand, it is hypothesized that binding to protein in the blood protects the vulnerable gallate functional group from undergoing enzymatic hydrolysis, leading to the absence of GA from the human serum sample tested. Indeed, we demonstrated that the abundance of the free form was greater for gallated catechins (ECG, EGCG) than for non-gallated catechins (EC, EGC) after deproteinization by ethyl acetate (Fig. 4B). This may reflect that gallated catechins bind with higher affinity to blood protein than non-gallated ones. However, there are only a few reports on the protein binding rate, pharmacokinetics, and predominant metabolic form of catechins in blood, and more detailed research is need to confirm this hypothesis. 5. Conclusions We developed an ion-pair HPLC-ECD method for the simultaneous quantitation of GA and four target catechins. Target catechins were enzymatically deconjugated to measure unconjugated free catechins. GA was measured to estimate the amount of gallate moiety produced by degradation of gallated catechins. The accuracy, reproducibility, and sensitivity of our improved method were sufficient to measure the target catechins in serum and GA concentrations. Catechin concentrations in human serum after green tea tablet ingestion were determined using this assay. The four target catechins were successfully detected, and no GA was found under these conditions. This method is applicable to assess the kinetics of catechins after tea consumption and is considered to be useful to clarify the beneficial health effects of catechins. Acknowledgements We sincerely thank Dr. Shunro Sonoda, President CEO of Satsuma Green Tea International, Inc. (Kagoshima, Japan) for providing decaffeinated green tea tablets. References [1] [2] [3] [4] [5] [6] [7] [8]
C.S. Yang, Z.Y. Wang, J. Natl. Cancer Inst. 85 (1993) 1038. I.E. Dreosti, M.J. Wargovich, C.S. Yang, Crit. Rev. Food Sci. Nutr. 37 (1997) 761. P.C. Hollman, L.B. Tijburg, C.S. Yang, Crit. Rev. Food Sci. Nutr. 37 (1997) 719. R. Mohan, S. Katiyar, S. Khan, M. Zaim, H. Mukhtar, R. Agarwal, Int. J. Oncol. 8 (1996) 1079. M.J. Lee, Z.Y. Wang, H. Li, L. Chen, Y. Sun, S. Gobbo, D.A. Balentine, C.S. Yang, Cancer Epidemiol. Biomarkers Prev. 4 (1995) 393. K. Umegaki, A. Sugisawa, K. Yamada, M. Higuchi, J. Nutr. Sci. Vitaminol. 47 (2001) 402. K. Nakagawa, T. Miyazawa, Anal. Biochem. 248 (1997) 41. B.L. Lee, C.N. Ong, J. Chromatogr. A 881 (2000) 439.
K. Narumi et al. / J. Chromatogr. B 945–946 (2014) 147–153 [9] G.J. Soleas, J. Yan, D.M. Goldberg, J. Chromatogr. B: Biomed. Sci. Appl. 757 (2001) 161. [10] B. Yang, K. Arai, F. Kusu, Anal. Biochem. 283 (2000) 77. [11] T. Unno, T. Takeo, Biosci. Biotechnol. Biochem. 59 (1995) 1558. [12] S. Carando, P.L. Teissedre, J.C. Cabanis, J. Chromatogr. B: Biomed. Sci. Appl. 707 (1998) 195. [13] T. Unno, K. Kondo, H. Itakura, T. Takeo, Biosci. Biotechnol. Biochem. 60 (1996) 2066. [14] A. Kotani, N. Miyashita, F. Kusu, J. Chromatogr. B 788 (2003) 269. [15] T. Fu, J. Liang, G. Han, L. Lv, N. Li, J. Chromatogr. B 875 (2008) 363. [16] M.J. Lee, S. Prebhu, X. Meng, C. Li, C.S. Yang, Anal. Biochem. 279 (2000) 164.
153
[17] Y.-H. Hong, E.Y. Jung, K.-S. Shin, T.Y. Kim, K.-W. Yu, U.J. Chang, H.J. Suh, Appl. Biochem. Biotechnol. 166 (2012) 165. [18] H.-H.S. Chow, Y. Cai, D.S. Alberts, I. Hakim, R. Dorr, F. Shahi, J.A. Crowell, C.S. Yang, Y. Hara, Cancer Epidemiol. Biomarkers Prev. 10 (2001) 53. [19] T. Ishii, T. Ichikawa, K. Minoda, K. Kusaka, S. Ito, Y. Suzuki, M. Akagawa, K. Mochizuki, T. Goda, T. Nakayama, Biosci. Biotechnol. Biochem. 75 (2011) 100. [20] K. Minoda, T. Ichikawa, T. Katsumata, K. Onobori, T. Mori, Y. Suzuki, T. Ishii, T. Nakayama, J. Nutr. Sci. Vitaminol. 56 (2010) 331. [21] M.J. Bae, T. Ishii, K. Minoda, Y. Kawada, T. Ichikawa, T. Mori, M. Kamihira, T. Nakayama, Mol. Nutr. Food Res. 53 (2009) 709.