Article pubs.acs.org/JAFC
Pilot Walnut Intervention Study of Urolithin Bioavailability in Human Volunteers Beate Pfundstein, Roswitha Haubner, Gerd Würtele, Nicole Gehres, Cornelia M. Ulrich, and Robert W. Owen* Division of Preventive Oncology, National Center for Tumor Diseases, Im Neuenheimer Feld 460, 69120 Heidelberg, Germany German Cancer Research Center (DKFZ), Im Neuenheimer Feld 581, 69120 Heidelberg, Germany ABSTRACT: A pilot intervention study was conducted in human volunteers (n = 4) to establish the bioavailability of urolithins, which are the terminal end-products of ellagitannin metabolism by the gastrointestinal microflora. Biospecimens (blood, feces, and urine) along with urolithins purified therefrom were analyzed for their antioxidant capacity in a range of in vitro assays. Urolithin metabolites were identified and quantitated in the biospecimens by negative ion mode HPLC-ESI-MS analysis. The data in this pilot study show that the metabolism of ellagitannins in the four volunteers gave rise to a diverse profile and a highly variable concentration of urolithins in urine. The concentration of glucuronidated urolithins in blood and urine did not correlate with antioxidant capacity. However, the antioxidant capacity of urine, but not plasma biospecimens, was highly correlated with uric acid concentration. The antioxidant capacity of fecal extracts correlated positively with the concentration of urolithin D in both the DPPH and FRAP assays, but not in the ORAC assay, which was entirely consistent with the in vitro assays for pure urolithin D. KEYWORDS: ellagitannins, DPPH assay, FRAP assay, ORAC assay, uric acid, urolithins, walnuts
■
INTRODUCTION Previous walnut intervention trials in other laboratories, using both animal1,2 and human models,3,4 have established that the intestinal bacterial metabolites of the major phenolic components (ellagitannins) present in walnuts are urolithins. Ellagitannins and urolithins (Figure 1) have been shown to display potent anti-inflammatory,5 antioxidant,6 antiproliferative, and antiaromatase7 properties and to also inhibit the growth of human breast8 and prostate cancer cells.9 Therefore, it is likely that ingestion of foodstuffs with high concentrations of ellagitannins may be beneficial to health, as reviewed recently.10 The logic is that antioxidants11 provided by the diet decrease oxidative stress by lipid peroxidation, thereby reducing DNA damage, which ultimately leads to carcinogenesis.12 A comprehensive review on cancer chemoprevention through dietary antioxidants has been published.13 To confirm the ellagitannin data, it is necessary to conduct human intervention studies. However, before embarking on large-scale screening and clinical intervention trials in our laboratories, it was essential to establish and justify the criteria necessary to conduct such studies. To this end, four individual volunteers agreed to enter the study, and the following parameters were measured: compliance, biospecimen collection and storage, biospecimen workup, analyses of urolithins in blood, urine, and feces, and identification of glucuronidated urolithins in blood and urine. Likewise, free urolithins in feces, purification of both glucuronidated and free urolithins from the biospecimens for definitive identification by HPLC-ESI-MS (high performanceelectrospray ionization-mass spectrometry), and evaluation of antioxidant capacity in a range of in vitro assays [2,2-diphenyl1-picrylhydrazyl (DPPH), ferric reducing ability of plasma © 2014 American Chemical Society
(FRAP), and oxygen radical absorbance capacity (ORAC)] were investigated. The variety of ellagitannin metabolites was studied in the urine of the four volunteers, and an individual displaying the most dynamic profile of urolithin metabolites was selected for an in-depth study in blood, feces, and urine. Extracts of biospecimens, along with purified urolithins therefrom, were evaluated in a range of in vitro antioxidant assays to elucidate the potential health beneficial effects of walnut consumption.
■
MATERIALS AND METHODS
Chemicals. Ascorbic acid, L-cysteine HCl, 2,2-diphenyl-1-picrylhydrazyl, fluorescein, perchloric acid, sodium carbonate, sodium formaldehyde sulfoxylate, sodium hydroxide, sodium metabisulfate, trifluoroacetic acid (TFA), HCl, FeCl3·6H2O, and Trolox were obtained from Sigma-Aldrich (Deisenhofen Germany); ellagic acid and gallic acid from Extrasynthese (Lyon Nord, Genay, France); ethanol, n-hexane, FeSO4··7H2O, methanol, 2,4,6,-tripyridyl-s-triazine complex (TPTZ), acetic acid, and acetonitrile from Fluka/Riedel de Haen (Seelze, Germany); and acetone, DMSO, sodium hydrogen phosphate, potassium dihydrogen phosphate, sodium acetate, brain− heart infusion (BHI), and uric acid from Merck (Darmstadt, Germany). Sephadex LH-20 was obtained from Amersham Biosciences (Uppsala, Sweden). Sep-Pak C18 cartridges for solid-phase extraction (500 and 5000 mg) were obtained from Supelco (Bellefonte, PA, USA); 2,2′-azobis(2-amidinopropane) dihydrochloride (AAPH) was obtained from Wako Chemicals (Neuss, Germany); anaerobic jars were obtained from neoLAB (Heidelberg, Germany); Received: Revised: Accepted: Published: 10264
April 29, 2014 October 1, 2014 October 2, 2014 October 2, 2014 dx.doi.org/10.1021/jf5040652 | J. Agric. Food Chem. 2014, 62, 10264−10273
Journal of Agricultural and Food Chemistry
Article
Figure 1. Structures of the urolithin glucuronides and free urolithins: urolithin B glucuronide, 1; urolithin A glucuronide, 2; urolithin C glucuronide, 3; urolithin D glucuronide, 4; urolithin B, 5; urolithin A, 6; urolithin C, 7; urolithin D, 8; urolithin E, 9. and AnaeroGen sachets for generating CO2 and BR55 anaerobic indicators were obtained from Oxoid Ltd. (Basingstoke, UK). Source of Walnuts. California walnuts (Sweet Valley brand) were purchased from a local supermarket. Study Design. Four healthy volunteers were entered into the study. No restriction with regard to general diet was enforced, only ingestion of supplements or medication and walnuts was forbidden for 3 weeks prior to the study. The volunteers consumed peeled walnuts (200 g) starting at 9:00 a.m. on day 1 and completed ingestion within 4 h. Morning urine samples were collected 1 day later. On the basis of urine urolithin excretion data, one volunteer (V-1) displaying the most dynamic profile of ellagitannin metabolites was selected for a further in-depth study of blood, feces, and urine. V-1 was female, aged 43, a nonsmoker and nonvegetarian, without a history of gastrointestinal or any other chronic diseases, and was not involved in a weight-reducing dietary regimen. On day 1, 24 h urine and fecal samples were collected. On day 2 of the study, 200 g of commercially available peeled walnuts was consumed over the course of the morning between 9:00 a.m. and 12:00 p.m. Successive urine and fecal samples were collected separately over the following 7 days. Additionally, plasma samples were collected on days 2, 3, and 4 of the
study in 2−4 h periods. On day 8, the ingestion of 200 g of peeled walnuts was repeated, and successive urine and fecal samples were again collected up to day 13 of the study. Collection and Storage of Urine, Blood, and Fecal Samples for V-1. Urine. Following ingestion of a further 200 g of walnuts by V1, all subsequent urine samples were collected in measuring cylinders, noting the time and volume. Aliquots of 2 and 10 mL were pipetted into plastic test tubes (15 mL), and the rest was discarded. During the day, samples were placed immediately at −80 °C, whereas outside working hours, samples were placed at −20 °C overnight before transport to the laboratory and subsequent storage at −80 °C. Feces. For V-1, all fecal samples were collected throughout the two intervention phases, in preweighed round plastic containers of 10 cm diameter. The plastic containers plus feces were reweighed for wet weight calculations. During the day, samples were placed immediately at −80 °C, whereas after working hours samples were placed at −20 °C overnight before transport to the laboratory and subsequent storage at −80 °C. The samples were freeze-dried within 8 days of collection to obtain dry weights and were pulverized to a fine homogeneous powder prior to Soxhlet extraction. 10265
dx.doi.org/10.1021/jf5040652 | J. Agric. Food Chem. 2014, 62, 10264−10273
Journal of Agricultural and Food Chemistry
Article
Plasma. During the first intervention period of the two-stage intervention study with V-1, blood samples (4−5 mL) were taken from the basilic vein into heparin prepared tubes at different time points; five samples on day 2 (ingestion of walnuts), five samples on days 3 and, and four samples on day 4 of the study. The blood samples were centrifuged (13000 rpm) at 5 °C for 10 min, and plasma was frozen at −80 °C. Column Chromatography on Sephadex LH-20. Dried urine extraction residues, after suspension in absolute ethanol, were adsorbed on Sephadex LH-20 by freeze-drying, prior to addition to the free volume at the head of glass columns (38 cm × 4.5 cm i.d.) filled to ca. 5 cm from the top with Sephadex LH-20 in ethanol. After bedding down of the gel material, fractionation was conducted by successive elution with ethanol followed by increasing concentrations of methanol (5−50%) in ethanol followed by 100% methanol (250 mL of each solvent). Fractions (250 mL) were collected, and the solvent was removed by rotary evaporation in vacuo at 40 °C. Column Chromatography on C18. The dried fractions obtained by Sephadex LH-20 chromatography were dissolved in 2% acetic acid (2.0 mL) and applied to C18 columns for subfractionation. The columns were preconditioned with methanol (4.0 mL) and 2% acetic acid (4.0 mL) and were not allowed to dry. Elution was performed with solvent mixtures containing increasing concentrations of methanol (5−50%) in 2% acetic acid followed by 100% methanol. The solvent was removed on a SpeedVac (Bachhofer, Reutlingen, Germany). Individual phenolic compounds, in relevant fractions, were purified by semipreparative HPLC for structure elucidation by spectroscopic analysis. Isolation of Urolithin Glucuronides from Urine. A week later, three of the healthy volunteers again consumed 200 g of walnuts, acccording to the same protocol as before. The following morning, first-pass urine was collected, and the volume of urine was reduced in vacuo on a rotary evaporator. The samples were all worked up separately. Sephadex LH-20 was added, and the remaining urine was adsorbed on this substrate by freeze-drying prior to addition to the free volume at the head of individual glass columns (38 cm × 4.5 cm i.d.), filled to ca. 5 cm from the top, with Sephadex LH-20 in ethanol and fractionated as described above. The dried fractions were dissolved in methanol (5.0 mL) prior to analyses by analytical reverse-phase HPLC and HPLC-ESI-MS. The fractions obtained from the Sephadex column were subfractionated on C18 columns as described above. Dried subfractions were suspended in methanol (2 mL) and diluted when necessary prior to reverse-phase analytical HPLC and HPLCESI-MS. Phenolic compounds were purified by semipreparative HPLC for spectroscopic analysis. Generation of Free Urolithins by Fermentation of Ellagic Acid in Vitro. Reference-free urolithins were unavailable commercially and therefore were produced by bacterial fermentation according to the following protocol. Urolithins were produced by adding ellagic acid (50.0 mg) in ethanol (5.0 mL) to BHI broth (100 mL) in Duran bottles (100 mL) fortified with the reducing agents, sodium formaldehyde sulfoxylate (300 mg/L) and L-cysteine HCl (500 mg/ L), to a final concentration of 1 mg/mL. After previous sterilization at 120 °C for 20 min, the medium was steamed for 30 min, immediately prior to use, to remove residual oxygen, and allowed to cool before inoculation with freshly voided fecal matrix (100 mg) from three of the volunteers. The bottles with open caps were placed in an anaerobic churn (neoLAB, Heidelberg, Germany) containing an indicator of anaerobiosis (Anaerobic indicator, BR55), which turned from pink to white on generation of an anaerobic atmostphere, by two freshly opened sachets of ascorbic acid (AnaeroGen). The fermentation broths were centrifuged and metabolites extracted on two C18 (5 g) (octadecyl-modified silica, Chromabond) columns. The columns were preconditioned with methanol (50 mL) and 2% acetic acid (50 mL). The fermentation solutions were eluted, and the columns were washed twice with distilled water (50 mL). The columns were allowed to dry and eluted with methanol (50 mL). Solvent from the combined eluants was removed by rotary evaporation, and the dried extracts were dissolved in methanol (2.0 mL). The metabolites
were analyzed by analytical HPLC and HPLC-ESI-MS and purified by semipreparative HPLC prior to spectroscopic analyses. Semipreparative HPLC. Semipreparative HPLC was conducted on a HP 1100 liquid chromatograph (Agilent Technologies, Waldbronn, Germany) fitted with a C18 column (10 mm i.d.) similar to that used for analytical HPLC. For the separation of individual compounds in the extracts, the mobile phase (3 mL/min) consisted of 0.2% acetic acid in distilled water (solvent A) and acetonitrile (solvent B), utilizing the following solvent gradient profile over a total run time of 50 min: initially 95% A for 1 min; reduced to 90% A over 9 min; to 85% A over 10 min; to 80% A over 10 min; to 0% A over 5 min; and continuing at 0% A until completion of the run. Peaks eluting from the column were collected on a HP 220 microplate sampler and subsequently lyophilized. Electrospray Ionization Mass Spectrometry (ESI-MS). HPLCESI-MS was conducted on an Agilent 1100 HPLC, coupled to a HP 1101 single-quadrupole, mass-selective detector (Agilent Technologies). The column used was a 250 mm × 4.5 mm i.d., 5 μm, RP-18 with a 4 mm × 4 mm i.d. guard column of the same material (Latek, Eppelheim, Germany). The mobile phase consisted of 2% acetic acid in water (solvent A) and acetonitrile (solvent B) with the following gradient profile: initially 95% A for 10 min; reduced to 90% A over 1 min; to 60% A over 9 min; to 80% A over 10 min; to 60% A over 10 min; to 0% A over 5 min; and continuing at 0% A until completion of the run. Volumes (10 μL) were injected into the HPLC, and phenolic compounds in the eluant were detected at 278 and 340 nm with a diode array UV detector (HP 1040M). Mass spectra in negative-ion selected ion monitoring (SIM) mode were generated under the following conditions: fragmentor voltage, 100 V; capillary voltage, 2500 V; nebulizer pressure, 30 psi; drying gas temperature, 350 °C; mass range, 100−1500 Da. The ions used for the detection of glucuronidated and free urolithins are listed in Table 1. Instrument control and data handling were performed with the same software as for analytical HPLC.
Table 1. Details of the Parameters Used for Detection of the Urolithins by HPLC-ESI-MS in Negative Ion Selected Ion Monitoring Mode phenolic compound 1 2 3 4 5 6 7 8 9
urolithin B glucuronide urolithin A glucuronide urolithin C glucuronide urolithin D glucuronide urolithin B urolithin A urolithin C urolithin D urolithin E
formula
exact mass M (calcd)
selected [M − H]− ion
C19H16O9
388.079
387
33.74
C19H16O10
404.074
403
C19H16O11
420.069
419
25.82, 26.18, 27.11, 27.42 27.77
C19H16O12
436.064
435
C13H8O3 C13H8O4 C13H8O5 C13H8O6 C13H8O7
212.047 228.042 244.037 260.032 276.027
211 227 243 259 275
HPLC-MS ret time (min)
23.37, 25.57, 29.34, 33.64 42.68 41.94 37.84, 39.67 36.24, 37.74 27.98
Total Polyphenol Content of Walnuts Evaluated by the Folin−Ciocalteu Assay. Preparation of a cold walnut extract was conducted according to the method of Anderson et al.,14 whereby 50 g of powdered walnuts (frozen for 24 h) was incubated in a solution of 75% aqueous acetone containing sodium metabisulfite (10 mg/L) at 4 °C for 4 days. The acetone was decanted and evaporated under reduced pressure, and methanol (50% aqueous, v/v) was added. The solution was extracted three times with n-hexane. The defatted solution was freeze-dried and dissolved in 100 mL of 10% aqueous dimethyl sulfoxide and used for the analysis of total polyphenols. To evaluate the total polyphenol content of walnuts, 20 μL of gallic acid standard solution or walnut extract was mixed with 1.58 mL of double-distilled water and 100 μL of Folin−Ciocalteu phenol reagent 10266
dx.doi.org/10.1021/jf5040652 | J. Agric. Food Chem. 2014, 62, 10264−10273
Journal of Agricultural and Food Chemistry
Article
were diluted 1:10 and 1:100 with methanol, respectively. Plasma samples (100 μL) were added to 100 μL of perchloric acid. The samples were centrifuged at 13000 rpm for 5 min, and the supernatants (150 μL) were neutralized with 10 M sodium hydroxide (5 μL), and 20 μL of this solution was used in the assay. Each batch of samples was measured simultaneously with Trolox at concentrations of 50−1000 μM. Using the percentage degradation of DPPH, at the different Trolox concentrations, a standard curve was constructed, which was always exponential. To obtain linear regression curves, the data points were subdivided into two parts. With Trolox concentrations of 250−1000 μM, a linear regression curve was calculated and used for the calulation of all Trolox equivalents for all data points ≥10% DPPH value. With Trolox concentrations of 50− 250 μM a linear regression curve was calculated and used for the calulation of all Trolox equivalents for all data points ≤10% DDPH. The Trolox equivalents calculated were corrected for dilution factors and, in the case of urine and feces, calculated in relation to the amount of urine or fecal extracts. FRAP Assay. The FRAP assay was conducted as previously described.16 Raw urine and fecal extracts were diluted 1:10 and 1:100 with acetate buffer (300 mM sodium acetate, pH, 3.6). Plasma samples were prepared as described for the DPPH assay. Each batch of samples was measured simultaneously with Trolox at concentrations of 50−1000 μM to obtain a linear regression curve. Using this linear regression curve, the Trolox equivalents for each sample of fecal extracts, plasma, and urine were calculated and corrected for dilution factors. ORAC Assay. The ORAC assay was conducted as described previously.16 To precipitate proteins, plasma samples (100 μL) were mixed with 0.5 M perchloric acid (100 μL) and centrifuged. Of the upper layer, 100 μL was mixed with 10 M sodium hydroxide solution (2 μL). The plasma extracts, diluted 1:100 or 1:200 with ORAC buffer, and aliquots (10 μL) were placed in quadruplicate in a 96-well plate to measure peroxy radical scavenging capacity. Raw urine and fecal extracts were diluted 1:50 and 1:200, respectively, with ORAC buffer prior to analysis. Statistics. The Trolox equivalent values of pure compounds and biospecimens in the DPPH assay were correlated with the Tablecurve program (Jandel Scientific, Chicago, IL, USA). Correlations between urolithin and uric acid concentrations and antioxidant capacity in the biospecimens were evaluated by linear regression using the Origin program (version 7.5).
N2 (Sigma-Aldrich, Deisenhof, Germany) in a 2 mL cuvette. After 8.5 min, 300 μL of sodium carbonate solution (200 g/L) was added, and the solution was stirred and incubated for 30 min at 40 °C, following the procedure of Waterhouse.15 The absorbance was measured at 765 nm on a spectrophotometer (Pharmacia Biotech, Ultraspec 3000) against a blank. Different concentrations of gallic acid (0.5−10 mM) in 10% aqueous dimethyl sulfoxide were used to generate a standard curve with a slope of 0.1534 (SD = 0.00329, P < 0.0001). The total polyphenol content was calculated as gallic acid equivalents (GAE). Total and Individual Content of Phenolic Compounds in Walnuts As Evaluated by Soxhlet Extraction and HPLC-ESI-MS. Freeze-dried walnuts (5 g) were delipidated by extraction with hexane (3 × 150 mL) for 3 h. The walnut cake was dried under a stream of nitrogen and further extracted with methanol (3 × 150 mL) for 3 h. The methanol extracts were combined, and solvent was removed by rotary evaporation. The dried residue was immobilized on Sephadex LH-20 and fractionated by successive elution with increasing concentrations of methanol in ethanol as described. Phenolic compounds in the fractions were identified by HPLC-ESI-MS and quantitated against standard curves of the purified metabolites. Sample Analyses. Urine. Urine samples (20 μL) were analyzed for urolithin metabolites of ellagitannins by HPLC-ESI-MS operating in negative ion SIM mode, using the ions listed in Table 1. Calculation of the amounts of individual metabolites were conducted with the relevant standard curves listed in Table 2.
Table 2. Standard Curves of Purified Glucuronidated Urolithins, Free Urolithins, and Uric Acid Generated by HPLC-ESI-MS in Selected Ion Monitoring Negative Ion Mode substance with mass urolithin urolithin urolithin urolithin urolithin urolithin urolithin urolithin uric acid
B glucuronide A glucuronide C glucuronide D glucuronide B A C D
slope
R
P
5791 9101 738 7418 4318 12035 657 9618 7.710
0.999 0.999 0.999 0.998 0.999 0.996 0.997 0.999 0.999
