Article pubs.acs.org/jpr
Rapid and Sustained Systemic Circulation of Conjugated Gut Microbial Catabolites after Single-Dose Black Tea Extract Consumption John van Duynhoven,*,†,‡,⊥,▽ Justin J. J. van der Hooft,*,§,⊥,▽ Ferdinand A. van Dorsten,†,⊥ Sonja Peters,†,⊥ Martin Foltz,† Victoria Gomez-Roldan,∥,#,⊥ Jacques Vervoort,§,⊥ Ric C. H. de Vos,∥,#,⊥,△ and Doris M. Jacobs†,⊥,△ †
Unilever Discover Vlaardingen, Olivier van Noortlaan 120, 3133 AT Vlaardingen, The Netherlands Laboratory of Biophysics and Wageningen NMR Centre, Wageningen University, 6700 ET Wageningen, The Netherlands § Laboratory of Biochemistry, Wageningen University, 6700 ET Wageningen, The Netherlands ∥ Plant Research International, Droevendaalsesteeg 1, 6708 PB Wageningen, The Netherlands ⊥ Netherlands Metabolomics Centre, Einsteinweg 55, 2333 CL Leiden, The Netherlands # Centre for Biosystems Genomics, Droevendaalsesteeg 1, 6708 PB Wageningen, The Netherlands ‡
S Supporting Information *
ABSTRACT: Gut microbial catabolites of black tea polyphenols (BTPs) have been proposed to exert beneficial cardiovascular bioactivity. This hypothesis is difficult to verify because the conjugation patterns and pharmacokinetics of these catabolites are largely unknown. The objective of our study was to identify, quantify, and assess the pharmacokinetics of conjugated BTP metabolites in plasma of healthy humans by means of an a priori untargeted LC−MS-based metabolomics approach. In a randomized, open, placebocontrolled, crossover study, 12 healthy men consumed a single bolus of black tea extract (BTE) or a placebo. The relative and, in several cases, absolute concentrations of a wide range of metabolites were determined using U(H)PLC-LTQ-Orbitrap-FTMS. Following BTE consumption, a kinetic response in plasma was observed for 59 BTP metabolites, 11 of these in a quantitative manner. Conjugated and unconjugated catechins appeared in plasma without delay, at 2−4 h, followed by a range of microbial catabolites. Interindividual variation in response was greater for gut microbial catabolites than for directly absorbed BTPs. The rapid and sustained circulation of conjugated catabolites suggests that these compounds may be particularly relevant to proposed health benefits of BTE. Their presence and effects may depend on individual variation in catabolic capacity of the gut microbiota. KEYWORDS: black tea, polyphenols, gut microbial conversion, metabolite identification, metabolite quantification, catabolites
■
INTRODUCTION
black tea polyphenols (BTPs) as the potential source of bioactives.15 It has been strongly argued that in vitro bioactivity testing is only meaningful when the actual circulating species are considered at relevant physiological concentrations.16 Currently, our insights into the bioavailability of BTPs are, however, limited. A major obstacle is the complex metabolic fate of BTPs upon ingestion. Only 20−30% of BTPs are lowmolecular-weight flavonoids, while the major part consists of oligomeric polyphenols such as theaflavins and thearubigins.17 Because of their high molecular weights, their absorption in the small intestine is low.18 Most BTPs persist to the colon, where they undergo extensive catabolism by the resident gut
Epidemiological studies have clearly shown that sustained consumption of black tea is associated with a lower risk of cardiovascular diseases1−4 including stroke.5,6 Recently, randomized controlled intervention studies on subjects with mildly elevated blood pressure convincingly demonstrated beneficial effects of black tea on surrogate cardiovascular end points such as BP7,8 and rate of blood pressure variation.9 Other intervention studies have shown acute and chronic effects of black tea on endothelial-dependent vasodilation, which may enhance a healthy blood flow.10 So far, no conclusive evidence has been found for systemic anti-inflammatory or antioxidant effects of black tea.11−13 Instead, local and specific molecular actions at the vascular endothelium have been proposed as underlying mechanism for beneficial effects of tea,14 including © 2014 American Chemical Society
Received: February 12, 2014 Published: March 27, 2014 2668
dx.doi.org/10.1021/pr5001253 | J. Proteome Res. 2014, 13, 2668−2678
Journal of Proteome Research
Article
microbiota.15,19,20 A large number of BTP catabolites have already been identified in in vitro colon models,21 yet it is unclear to what extent and in which conjugated form they appear in systemic circulation. Both metabolites absorbed via the small intestine, and microbial catabolites can undergo extensive first-pass metabolism through phase-II conjugation enzymes. Although it is well-accepted that the metabolic fate of BTP involves aforementioned mechanisms, phase-II conjugation patterns and kinetics in plasma are largely unknown.22 This is mainly due to the lack of analytical capability to identify and quantify the wide range of conjugated BTP metabolites in plasma.23 Current analytical approaches involve enzymatic deconjugation of phase-II metabolites,18,24,25 allowing for straightforward quantification while discarding information on conjugation patterns. A few studies have assessed conjugated (green) tea metabolites in plasma and urine,26,27 but so far reliable absolute quantification could not be achieved and absolute configurations of conjugated forms could not be provided.22,28 In the work presented here, we aimed at the identification and quantification of as many conjugated BTP metabolites as possible in human plasma using an essentially untargeted Ion trap−Orbitrap FTMS approach. BTP phase-II metabolites and catabolites identified in and purified from urine after tea intake, as reported in our previous study,29 were used to facilitate annotation and absolute quantification of BTPderived compounds in blood plasma samples. The subjects consumed an equivalent of 4−6 cups of black tea in the form of a dissolved black tea extract (BTE) or hot water as a placebo. The kinetic curves of a diverse set of BTP metabolites will be discussed in the context of their putative role in mediating acute and chronic cardiovascular effects. Interindividual variation in circulation in plasma is assessed to estimate whether gut metabotypes30 may play a role in mediating cardiovascular benefits of BTP.
■
Lyon, France) and was registered at www.clinicaltrials.gov under number NCT01533857. The study was carried out at Eurofins-Optimed (Gières, France). Study Design
The volunteers were subjected to an exploratory, placebocontrolled, crossover, single-blinded, single-dose study. The study consisted of two treatments (placebo and active), which were separated by a 2 week washout period. A 12 subject crossover design was generated such that an equal number of subjects tested placebo-active and active-placebo. A restriction on this was the need to split subjects into two groups, each to be processed on different days. Within each group of six, the randomization allocated three active-placebo and three placeboactive sequences. Within each group of six, the approach allocated three AB and three BA sequences at random. For each treatment, subjects were placed in a metabolic ward during a run-in of 2 days before each treatment and during the 30 h treatment period. During the run-in and treatment phases, subjects received a polyphenol-poor diet. In the morning of each intervention day, the fasted subjects received one treatment. Because of the exploratory nature of the study, a sample of 12 male subjects was chosen. Diet and Study Products
One day before and during the 1 day intervention the volunteers were on a low polyphenol diet with a relatively high intake of dairy, meat, tubers, and cereals and low intake of coffee, tea, fruits, and vegetables. The diet comprised breakfast, lunch, and dinner (Table S1 in the Supporting Information). A low polyphenol lunch and dinner was provided, respectively, 4 and 12 h after the start of treatments. The placebo consisted of a hard gelatin capsule size 0 (Capsugel, Bornem, Belgium) filled with inert filler (microcrystalline cellulose; Ph Eur, Avicel PH102, FMC Biopolymer; Philadelphia, PA) served with 250 mL of hot water. As active treatment, one hard gelatin capsule size zero filled with three isotopic labeled phenolic acids and inert filler was administered next to 2650 mg of Brook Bond red label (BBRL, Unilever Synthite Industries, Kerala, India) extract, dissolved in 250 mL of hot water. The isotopic phenolic acids were dosed for detailed pharmacokinetic assessment, and results are reported elsewhere.31 Identical capsule formulations were tested to comply with the criteria for content uniformity and immediate release characteristics. BBRL extract was a brown, dry, fine powder, which was nearly fully soluble in hot water. The composition of this extract was measured before and after the study for total catechins, gallic acid, and caffeine according to ISO 14502 to assess product stability over the course of the study. The tea extract contained 26% total polyphenols, 22% total flavonoids, 1.2% theaflavins, 3% catechins, 0.3% theogallin, 2.7% gallic acid, 8.7% caffeine, and 0.5% theobromine. Besides gallic acid, no other phenolic acids were detected.
MATERIALS AND METHODS
Participants
Fourteen males, aged 18 to 65 years, were recruited from the general public and invited to participate. They were screened for being generally healthy by means of a screening questionnaire, being nonsmoking, body mass index assessment (18.5 to 25 kg/m2), standard hematology, and urine testing. Excluded were volunteers that had donated blood at least 4 weeks prior to the start of the study or were using NSAIDS and/or antibiotics. Two subjects did not meet criteria during a selection visit: one did not pass the general medical examination, and the other had a BMI that was too high. The remaining 12 subjects were randomly allocated to the two treatment orders. The age of the volunteers ranged from 19 to 65 and their body mass index from 18.6 to 24.2 kg/m2 (at the time of inclusion). Body weight varied from 51.3 to 79.9 kg. All subjects received oral and written information about the experimental procedures and provided written informed consent. All subjects completed the intervention without adverse effects, and compliance was 100%. Upon a blind review performed immediately after the intervention phase, all 12 subjects were included in the analysis. The study was performed in accordance with the Declaration of Helsinki, the recommendations on Good Clinical Practice (ICH E6), and the applicable French regulatory requirements. The study received ethical approval from the Commité de Protection des Personnes SUD-EST IV (Sud-Est IV Ethics Committee,
Plasma
In the morning, after overnight fasting, a baseline sample was taken at 15 min before consuming the test products, and further samples were collected 0.5, 1, 2, 3, 4, 5, 6, 7, 8, 10, 14, 24, and 30 h afterwards. For pharmacokinetic modeling, the baseline sample was taken as t = 0 h. At each time point, a 6 mL blood sample was withdrawn into EDTA-2K Vacutainer tubes previously stored at 4 °C. The blood samples were gently inverted a few times for complete mixing with the coagulant. Within 30 min following blood collection, each sample was 2669
dx.doi.org/10.1021/pr5001253 | J. Proteome Res. 2014, 13, 2668−2678
Journal of Proteome Research
Article
centrifuged at 1500g for 10 min at 4 °C. Immediately after centrifugation, the top layer of plasma was aliquoted into four polypropylene tubes (at least 500 uL per tube) and stored at −80 °C.
min. An additional 15 min was used to wash and equilibrate the column before next injection. The MS analysis was carried out in negative electrospray ionization mode at a source voltage of 4.5 kV. The spectra were collected at the mass range m/z 90−1200 at a resolution of 60 000 (at m/z 400) in full-scan centroid mode. The capillary temperature was 300 °C, sheath gas flow 50 arbitrary units (au), auxiliary gas flow 10 au, and the sweep gas flow 5 au. The Orbitrap was externally calibrated in negative mode using sodium formate clusters in the range m/z 150−1200 and automatically tuned on m/z 554.26. The AGC target value of the ion trap was set at 30 000 charges. The ion tube was cleaned after each sample batch to obtain constant ionization and sensitivity.
Reagents and Chemicals
Methanol and acetonitrile (HPLC grade) were purchased from BioSolve BV (Valkenswaard, The Netherlands), and formic acid (FA, p.a. grade) was from Merck (Darmstadt, Germany). Baicalin was purchased from Extrasynthesis (Lyon, France) and Genistein 7-ß-D-O-glucuronide was from Toronto Research Chemicals (Ontario, Canada). Commercial human plasma (batch X1707; K2-EDTA stabilized) was purchased from Innovative Research (Novi, MI). Standards Identified and Quantified by NMR
Experimental Design for Sample Injection into LC−MS
Standards of conjugated (poly)phenolics were purified from concentrated urine samples collected after black and green tea consumption using LC−MS-SPE and identified by MSn and NMR.29 Seventeen compounds were obtained in sufficient quantities, that is, more than 60 ng in a NMR tube, to be used in dilution series for constructing calibration curves (Table S2 in the Supporting Information), with number corresponding to the Supporting Information supplied with a previous study:29 (Epi)catechin-O-methyl-O-sulfate IV (UT0001), (epi)catechinO-sulfate-O-methyl II (UT0004), 5-(3′,4′,5′-trihydroxyphenyl)γ-valerolactone-4′-O-glucuronide (UT0047), 5-(3′,4′-dihydroxyphenyl)-γ-valerolactone-3′-O-glucuronide (UT0055), 5-(3′,4′dihydroxyphenyl)-γ-valerolactone-4′-O-glucuronide (UT0056), 5-(3′,4′-dihydroxyphenyl)-γ-valerolactone-O-sulfate (UT0058), 5-(3′,5′-dihydroxyphenyl)-γ-valerolactone-3′-O-glucuronide (UT0059), 5-(3′,5′-dihydroxyphenyl)-γ-valerolactone-3′-O-sulfate (UT0060), 5-(3′-hydroxyphenyl)-γ-valerolactone-3′-O-glucuronide (UT0065), hippuric acid (UT0097), m-hydroxyhippuric acid (UT0112), pyrogallol-2-O-glucuronide (UT0125), pyrogallol-2-O-sulfate (UT0126), urolithin A-3-O-glucuronide (UT0131), urolithin A-8-O-glucuronide (UT0132), urolithin B-O-glucuronide (UT0134), and vanillic acid-COOH-glucuronide (UT0138).
Plasma extracts of each of the 12 volunteers were prepared and injected batch-wise in which the injection order was randomized. One batch consisted of 33 samples including all 28 study samples per volunteer, corresponding to the 14 time points of both the placebo and the BTP treatments. In addition, three commercial plasma samples, simultaneously prepared with each plasma volunteer batch and injected at the beginning, middle, and end of each batch, were used as quality controls (QCs). Finally, each batch series ended with one repeated injection for both a treated and a placebo sample from a randomly selected time point. Recovery Experiments of NMR Standards in Plasma
To determine the recovery of (poly)phenolic compounds from the plasma of each individual volunteer, 50 μL from each placebo plasma sample per volunteer was pooled, and 400 μL was combined with 25 ng of each of the nine NMR standards UT0047, UT0055, UT0056, UT0097, UT0112, UT0125, UT0138, UT0134, and UT0132 (Table S2 in the Supporting Information). Pooled plasma from volunteers consuming the placebo without the addition of NMR standards served as background level in each volunteer. Samples were prepared as described for the study samples and injected in a separate batch.
Sample Preparation
Calibration Curves of NMR Standards
The plasma samples were thawed on ice, and extracts were prepared as follows: 0.4 mL of plasma was mixed with 1.3 mL of acetonitrile containing 2% (v/v) of FA, 4 mmol/L of vitamin C (Vit C), and two polyphenol internal standards (ISs, Baicalin at 0.3 μg/mL and Genistein 7-ß-D-O-glucuronide at 0.1 μg/ mL). The samples were allowed to precipitate on ice for 10 min and then sonicated for 5 min and centrifuged for 10 min at 1405g. Exactly 1.4 mL of supernatant was taken and dried in a N2 gas flow for 4 h and subsequently redissolved in 100 μL of 50% methanol with 1% FA and 2 mmol/L Vit C, briefly vortexed, rested on ice for 10 min, sonicated for 5 min, and centrifuged again for 15 min. The supernatant was filtered through a 0.45 μm inorganic membrane filter and transferred to an HPLC vial with glass insert.
Calibration curves were constructed for each of the 19 purified compounds that were quantified by NMR spectroscopy (NMR). To do so, the content of each NMR tube (varying between 60 and 1500 ng, depending on the compound) (Table S2 in the Supporting Information) was completely transferred into Eppendorf tubes with two extra washing steps, dried under N2 gas, and redissolved in 75% methanol to obtain stock solutions of 1000 ng/mL in 75% methanol. Then, an 11-point dilution series (0, 0.78, 1.56, 3.13, 6.25, 12.5, 25, 50, 100, 200, and 500 ng/mL) was prepared for each compound and injected after the last study sample. Identification and Annotation of Masses and Retention Times
To annotate in the plasma LC−MS chromatograms those metabolites previously identified in urine, the same urine sample was run under the UPLC conditions used for the plasma data acquisition.29 On the basis of their calculated accurate [M−H]− m/z values, the retention times of compounds were determined for the specific UPLC conditions used. All raw LC−MS data files were preprocessed (peak picking by baseline correction and noise correction) using MetAlign software,32 and the resulting .redms files were browsed for the specific masses and retention time information
LC−MS Analysis
An Accela U-HPLC connected to an Accela photodiode array (PDA) detector and subsequently to an Ion trap−Orbitrap FTMS hybrid mass spectrometer (Thermo Fisher Scientific) was used for the LC−MS profiling. Five μL of extract was injected, and compounds were separated using a 1.7 μm AQUITY UPLC BEH C18 column (2.1*150 mm; Waters) held at 40 °C and a linear 20 min gradient of 5−35% acetonitrile (acidified with 0.1% FA) at a flow rate of 400 μL/ 2670
dx.doi.org/10.1021/pr5001253 | J. Proteome Res. 2014, 13, 2668−2678
Journal of Proteome Research
Article
Table 1. Pharmokinetic Parameters of Conjugated Metabolites of Directly Absorbed BTP’s and Microbial Catabolites Thereofa precursor Directly Absorbed Polyphenols and Phenolics Catechins (epi)catechin
(epi)gallocatechin (epi)catechingallate
(epi)gallocatechingallate
Other Polyphenols kaempferol Phenolic Acids gallic acid Gut Microbial Catabolites Valerolactones (δ-(3,4-dihydroxyphenyl)-γ-valerolactone)
5-(3′-hydroxyphenyl)-γ-valerolactone 5-(4′-hydroxyphenyl)-γ-valerolactone 5-(3′,4′,5′-trihydroxyphenyl)-γ-valerolactone
5-(3′,5′-dihydroxyphenyl)-γ-valerolactone
Valeric Acids 4-hydroxy-5-(hydroxyphenyl)-valeric acid 4-hydroxy-5-(dihydroxyphenyl)-valeric acid
4-hydroxy-5-(3′,4′-dihydroxyphenyl)-valeric acid 4-hydroxy-5-(3′,5′-dihydroxyphenyl)-valeric acid 4-hydroxy-5-(3′,4′,5′-trihydroxyphenyl)-valeric acid Phenols and Phenolic Acids pyrogallol (1,2,3-triOH-benzene)
code
O-sulfate O-sulfate-O-methyl II O-sulfate-O-methyl III O-sulfate-O-methyl V O-sulfate-O-methyl VI O-sulfate-di-O-methyl II O-methyl-O-sulfate IV O-methyl-O-sulfate II O-methyl-O-sulfate IV unconjugatedp O-methyl-O-sulfatep O-methyl-O-sulfate-O-glucuronidep O-sulfatep O-sulfate-O-glucuronide unconjugatedp O-sulfatep O-methyl-O-sulfatep
UT0002 UT0004 UT0005 UT0006 UT0085 UT0012 UT0001 UT0008 UT0010 UT0142 UT0144 UT0145 UT0143 UT0146 UT0139 UT0140 UT0141
0 0 0 0 0 0 0 1 0 0 1 0 0 0 0 1 1
0 0 0 0 0 0 0 1 0 0 1 0 0 0 0 1 1
2 2 2 2 2 2 2 2 2 1 2 4 3 3 1 2 2
O-glucunoride
UT0106
1±2
0,5
O-methyl-O-sulfate I O-methyl-O-sulfate II
UT0109 UT0110
1±2 0, 5 ± 2
2 0,5
3′-O-sulfate 3′-O-glucuronide 4′-O-glucuronide O-glucuronide-O-sulfate 3′-O-sulfate 3′-O-glucuronide 4′-O-glucuronide 3′-O-glucuronide 4′-O-glucuronide O-glucuronide-O-sulfate 3′-O-sulfate O-methyl-3′/4″-O-glucuronide O-methyl-3′/5′-O-’glucuronide O-methyl-O-sulfate I O-methyl-O-sulfate II 3′-O-glucuronide 3′-O-glucuronide-O-methyl I 3′-O-sulfate 3′-O-sulfate-O-methyl I 3′-O-sulfate-O-methyl II
UT0058 UT0055 UT0056 UT0057 UT0067 UT0065 UT0066 UT0045 UT0047 UT0049 UT0046 UT0050 UT0052 UT0053 UT0054 UT0059 UT0061 UT0060 UT0063 UT0064
2 2 2 3 3 3 3 4 3 4 4 4 5 3 4 3 2 3 3 4
± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
1 1 1 2 1 2 1 2 1 2 2 2 2 1 2 1 1 2 2 3
5 5 5 6 6 8 5 5 5 8 6 6 7 6 6 8 5 8 3 6
O-sulfate, -O-methyl I O-methyl-O-sulfate II O-sulfate I O-sulfate II O-sulfate II O-glucuronide I O-glucuronide II O-glucuronide I O-methyl-O-glucuronide I O-methyl-O-glucuronide III
UT0036 UT0032 UT0033 UT0034 UT0035 UT0022 UT0023 UT0026 UT0027 UT0021
4 4 2 2 3 3 5 3 -
±2
6 5 8 6 6 6 6 8 5 8
2-O-sulfate 2-O-glucuronide
UT0125 UT0126
3±1 3±1
2671
delay [h]
Tmax [h]
conjugated metabolite
± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
± ± ± ± ± ± ±
2 2 2 2 2 1 2
8 5
Cmax [μmol/L]
DIFF AUC [nmol h−1 L−1]
0.008
30 ± 15
0.008
20 ± 9
0.034 0.034 0.055
198 ± 138 164 ± 94 23 ± 15
0.001
4±5
0.005
19 ± 22
0.004
34 ± 25
0.015
150 ± 124
2.6 0.019
24528 ± 11047 90 ± 54
dx.doi.org/10.1021/pr5001253 | J. Proteome Res. 2014, 13, 2668−2678
Journal of Proteome Research
Article
Table 1. continued precursor Phenols and Phenolic Acids catechol (1,2-diOH benzene) resorcinol (1,3-diOH benzene) benzoic acid 3-OH-benzoic acid homovanillic acid
conjugated metabolite
code
O-sulfate-O-methyl I O-sulfate-O-methyl II O-glucuronide IIb O-sulfateb hippuric acid m-hydroxyhippuric acid O-sulfate
UT0115 UT0116 UT0094 UT0095 UT0097 UT0112 UT0100
delay [h] 3 4 3 3 2 3 3
± ± ± ± ± ± ±
1 1 2 1 1 2 2
Tmax [h] 6 10 6 6 6 6 7
Cmax [μmol/L]
DIFF AUC [nmol h−1 L−1]
2.0 0.046
9754 ± 4427 308 ± 287
a
Metabolites were previously identified in urine (confidence MSI = 1 or 229), except for those indicated with (P) that were identified directly in plasma (MSI = 2). bPreviously assigned as diphenol sulfate29 with uncertain substitution pattern of hydroxyl groups, here associated with 1,3 diOH benzene.
diet (DIFF AUC > 0 and RSD (DIFF AUC) = 0−230) for more than two subjects.
on the annotated metabolites using the dedicated SearchLCMS tool of Metalign.32 Because of the faster gradient, some previously resolved isomeric compounds coeluted in the LC− MS data. In addition to the aforementioned list,29 all Metalignderived .redms files were screened for the presence of epi(gallo)catechingallates and their methylated, sulfated, or glucuronidated conjugates on the basis of their theoretical masses [M−H]− masses. Using the Search-LCMS database, the generated output file contained relative signal intensities (peak height) for 131 target compounds and the 2 IS in all samples including the compound calibration samples.
■
RESULTS
Analytical Performance of U(H)PLC-Orbitrap-FTMS
Qualitative and quantitative analysis using U(H)PLC-LTQOrbitrap proved to be very efficient in obtaining comprehensive profiles of even low abundant (low μmol/L to nmol/L range) metabolites present in a complex matrix-like human plasma. Comprehensive accurate mass (
