Article

pubs.acs.org/JAFC

Absorption, Metabolism, and Effects at Transcriptome Level of a Standardized French Oak Wood Extract, Robuvit, in Healthy Volunteers: Pilot Study † § † § † † † Fausta Natella, , Guido Leoni, , Mariateresa Maldini, Lucia Natarelli, Raffaella Comitato, ‡ † † Frank Schonlau, Fabio Virgili, and Raffaella Canali*, † Consiglio per la Ricerca e Sperimentazione in Agricoltura, Food and Nutrition Research Centre, via Ardeatina 546, 00178 Roma, Italy ‡ Horphag Research (U.K.) Ltd. 28 Old Brompton Road, Suite 393, South Kensington, London SW7 3SS, United Kingdom

ABSTRACT: The consumption of wine and spirits, traditionally aged in oak barrels, exposes humans to roburin ingestion. These molecules belong to a class of (ETs), and their only known source is oak wood. Very little is currently known about roburin bioavailability and biological activity. We reported for the first time human absorption of roburins from a French oak wood (Quercus robur) water extract (Robuvit) by measuring the increase of total phenols (from 0.63 ± 0.06 to 1.26 ± 0.18 μg GAE equiv/mL plasma) and the appearance of roburin metabolites (three different glucoronidate and ), in plasma, after 5 days of supplementation. Robuvit supplementation induced also the increase of plasma antioxidant capacity from 1.8 ± 0.05 to 1.9 ± 0.01 nmol Trolox equiv/mL plasma. Moreover, utilizing a combined ex vivo cell culture approach, we assessed the effect of Q. robur metabolites (present in human serum after supplementation) on gene expression modulation, utilizing an Affymetrix array matrix, in endothelial, neuronal, and keratinocyte cell lines. The functional analysis reveals that Robuvit metabolites affect ribosome, cell cycle, and spliceosome pathways KEYWORDS: ellagitannins, roburins, Quercus robur, human study, gene array

■ INTRODUCTION The French oak water extract Robuvit is a registered In the past decades, epidemiological surveys pointed out the proprietary water extract obtained from fresh Quercus robur fi wood, standardized to provide a specific profile of and association between a signi cantly lower risk for cardiovascular, ≥ − metabolic, and cancer diseases, cognitive decline, and neuro- in particular a mixture of 20% roburins (A E) including degeneration and the consumption of diets rich in plant grandinin. The main ETs in Q. robur are two diastereoisomers, 1,2 fi vescalagin and , which were originally isolated and . The bene cial role of polyphenols was initially 6 attributed to a “nonspecific” antioxidant capacity. After a first described by Mayer et al. Sugar derivatives of vescalagin with aldopentoses xylose and lyxose are defined as roburin E and enthusiasm, it became apparent that health contributions of 7 − polyphenols were not (only) associated with cellular protection grandinin, respectively (Figure 1). Roburins A D are dimers, from oxidation but are due to a more complex mechanism with roburin D consisting of vescalagin and castalagin and involving the induction of cellular response to potentially of two vescalagin subunits. Roburins B and C harmful stimuli finally resulting in a number of specific adaptive resemble the structure of roburin A with additional xylose and defensive responses. Despite a consensus about this (roburin C) or lyxose (roburin B) bound to one vescalagin unit “hormetic” (para-hormetic) activity,3 comprehensive knowl- (Figure 2). Further to the roburins, Robuvit contains edge about the absorption, metabolism, and specific biological monomeric vescalagin and castalagin, as well as ellagic acid activities of the different classes of polyphenols is still scarce. (EA) and gallic acid (GA). Oak wood is currently the only fi Tannins are a specific family of polyphenols having a known source of roburins, and according to this speci city, the relatively large molecular structure. Tannins are considered to major source of roburins in human diets results from the be “secondary” compounds of plant physiology, not functioning consumption of wine and spirits (cognac and whiskey) in “primary” metabolism such as biosynthesis or energy traditionally matured, aged, and stored in oak barrels. In fact, conversions, although they serve biological purposes such as oak wood constituents contribute to the improvement of the protecting plants from herbivores and microbial infections. aroma, taste, and color and to the building up of the sensation “ ” Tannins are divided into two groups on the basis of their of mouth-fullness . As a result of this technological practice, structural characteristics: hydrolyzable and condensed tannins. humans have been exposed to roburins for centuries. In wines In the hydrolyzable tannins, a carbohydrate is partially or totally and spirits, roburins (including grandinin and the monomers esterified with phenolic groups such as gallic acid (gallotannins, GTs) or ellagic acid (ellagitannins, ETs). These compounds are Received: August 7, 2013 present in significant amounts in many berries, nuts, Revised: December 16, 2013 pomegranates, muscadine grapes, and edible dark-colored Accepted: December 19, 2013 fruits.4,5 Published: December 19, 2013

© 2013 American Chemical Society 443 dx.doi.org/10.1021/jf403493a | J. Agric. Food Chem. 2014, 62, 443−453 Journal of Agricultural and Food Chemistry Article

Figure 1. Chemical structures of the monomers, castalgin, vescalgin, and their sugar derivatives roburin E and grandinin.

Figure 2. Chemical structure of roburins A−D. castalagin and vescalagin) have been found to range between on gene expression in cultured endothelial, neuronal, and 1.2 mg/L (bourbon whiskey) and 9.4 mg/L (red wines).8 A keratinocyte cells to obtain a solid background for further number of indications suggest that, similarly to other studies addressing the understanding of possible beneficial polyphenols, ETs and their metabolites may exert beneficial effects of this family of molecules on human health. effects on human health.9 Despite the traditional consumption of roburins with wine and spirits, very little is currently known ■ MATERIALS AND METHODS about their bioavailability in humans. In general, ET from Chemicals. MeOH, HOCOOH, gallic acid, ellagic acid, Trolox, berries, pomegranate, walnuts, pecans, and pistachios are fluorescein, perchloric acid, and 2,2-azobis(2-amidinopropane) dihy- fl drochloride (AAPH) were purchased from Sigma-Aldrich (Milano, known to require the involvement of the colonic micro ora β 10 Italy); randomly methylated -cyclodextrin (RMCD) was purchased to be metabolized into dibenzopyranone species (urolithins). from Cyclolab (Budapest, Hungary). The consumption of oak-aged red wine was found to result in In Vivo Study. Experimental Design. Three healthy, nonsmoker, only trace amounts of -B aglycone and glucuronide in volunteers participated in this study, which was approved by the CRA- urine.11 Furthermore, several kinds of grapes have been found NUT ethical committee. The investigation conforms to the principles to contain EA, and therefore excreted urolithins may not outlined in the Declaration of Helsinki and its amendments. None of necessarily be attributed to oak constituents extracted following subjects was taking any medication or vitamin supplementation for the 2 weeks before the beginning of the experiment. The day before the wine aging in barrels. The aim of our study is to investigate the experiment subjects abstained from drinking coffee, tea, wine, and absorption and metabolism of Robuvit extract in healthy beer, and they ate no chocolate, fruits, and vegetables. The first day of volunteers and to evaluate the effect of Robuvit consumption the in vivo experimental protocol, blood was withdrawn in fasting

444 dx.doi.org/10.1021/jf403493a | J. Agric. Food Chem. 2014, 62, 443−453 Journal of Agricultural and Food Chemistry Article

Figure 3. LC-MS total ion current (TIC) chromatograms of standards (top) and a representative plasma sample (bottom). Peaks: 1, gallic acid; 2, methyl gallate; 3, ellagic acid; 4, urolithin C diglucuronide; 5, urolithin A glucuronide; 6, glucuronide. condition and serum isolated and stored at −80 °C until its utilization. with 1 mL of MeOH acidified with 0.1% HCOOH, conditioning with fi Volunteers were then supplemented with one capsule of Robuvit 1mLofH2O acidi ed with 0.1% HCOOH, loading of the sample, fi (containing 100 mg of Q. robur extract) three times a day for 5 days. washing with H2O acidi ed with 0.1% HCOOH (2 mL) and with During the 5 days of the supplementation, the subjects were water/methanol (95:5) (2 mL), and finally elution of metabolites with encouraged to reduce as much as possible the consumption of MeOH acidified with 0.1% HCOOH (2 mL). The eluted fraction was -rich foods. The last day of the supplementation, subjects evaporated under nitrogen, and the residue was reconstituted with were administered 300 mg of Robuvit in fasting condition, about 1 h mobile phase to 0.3 mL and filtered through a 0.2 mm PVDF filter before blood withdrawal, and serum was collected and isolated again as (Whatman) into a vial insert for LC-MS/MS analyses. We determined described above. only the principal compounds that were extracted in MeOH/H2O. Q. robur Capsule Analyses. Robuvit capsule was extracted with 50 However, several ETs are insoluble and should be measured after mL of MeOH/H2O 80:20 (v/v) in a shaking bath overnight in hydrolysis (by detection of the produced EA). Moreover, our MS has a reduced light and temperature conditions. Subsequently, the extract maximum limit of m/z 1500; therefore, ETs that have higher was centrifuged for 10 min at 2000g and 4 °C. The supernatant was molecular weights (such as roburin dimers) cannot be measured. filtered through a 0.20 mm PVDF filter (Whatman, USA) and diluted ESI-MS, ESI-MS/MS, LC-ESI-MS, and LC-ESI- MS/MS Analyses. 1:1 with methanol before the analyses. Quantitative online HPLC-ESI-MS/MS analysis of the capsule extract Plasma Sample Preparation. Plasma samples were acidified and plasma sample were performed using a Perkin-Elmer series 200 immediately to pH 3 with 1 N HCl immediately after blood HPLC system interfaced to an Applied Biosystems (Foster City, CA, withdrawal and stored at −80 °C. Once thawed, samples were mixed USA) API 3200 instrument equipped with a Turbo ionspray source. A with acetonitrile 1:4 (v/v) to precipitate plasma proteins and then gradient elution was performed by using a mobile phase A represented centrifuged for 10 min at 25000 rpm and 4 °C. The supernatant was by water acidified with formic acid (0.1%) and a mobile phase B fi recovered, dried under nitrogen (N2), and reconstituted with 0.5 mL represented by acetonitrile acidi ed with formic acid (0.1%). The fi of H2O acidi ed with 0.1% HCOOH. Extraction of metabolites was gradient started from 0% of eluent B, remained at 0% of B for 10 min, performed using solid-phase extraction (SPE) OASIS HLB 1 cm3 (30 then rose from 0 to 60% B in 40 min and subsequently to 100% B in mg) extraction cartridges (Waters, Milford, MA, USA). The SPE 25 min. The flow (0.300 mL min−1) generated by chromatographic procedure consisted of the following steps: activation of SPE cartridges separation was directly injected into the electrospray ion source. The

445 dx.doi.org/10.1021/jf403493a | J. Agric. Food Chem. 2014, 62, 443−453 Journal of Agricultural and Food Chemistry Article chromatographic separation was carried out using a Restek Ultra C18 Ex Vivo Study. Serum isolated before and after the supplementa- column (100 × 2.1 mm, 3 μm). tion of Robuvit from the three healthy subjects was used to enrich the The API 3200 ES source was tuned by infusing standard solutions medium of different cell cultures, that is, a permanent human of EA, GA, and methylgallate (1 mg/mL in methanol) into the source endothelial cell, obtained by cell fusion of HUVEC, primary human at a flow rate of 0.01 mL/min. The mass spectrometer was operated in endothelial cells, and immortalized human alveolar epithelial cells the negative ion mode under the following conditions: declustering (EA.hy 926), human keratinocytes (HaCaT), and human neuro- potential, −75 eV; entrance potential, −6.2 eV; collision energy, −45 blastoma cells (SH-SY5Y). Cells were cultured in the absence of eV; and collision cell exit potential, −23.3 eV for EA; declustering bovine serum in culture media supplemented with 10% (v/v) of the potential, −40 eV; entrance potential, −4.2 eV; collision energy, −20 isolated human serum for 3 h. eV; and collision cell exit potential, −18.4 eV for GA; declustering RNA Extraction and Gene Array. At the end of the incubation, potential, −40 eV; entrance potential, −4.2 eV; collision energy, −35 RNA was isolated using an RNAeasy Plus Mini kit (Qiagen). The eV; and collision cell exit potential, −19 eV for methyl gallate. MS and concentration, purity, and integrity were analyzed using the Agilent MS-MS daughter spectra were measured from m/z 150 to 500 for 2100 bioanalyzer, before gene expression analysis was performed. Two plasma sample and from m/z 150 to 1500 for Q. robur capsule. The micrograms of each isolated RNA samples was sent to the FIRC spectrometer was used in the MS/MS mode with multiple reaction Institute of Molecular Oncology Foundation (IFOM), Milano, Italy, to monitoring (MRM) of fragmentation reactions selected for each be processed and hybridized on Gene 1.0 ST array chips according to ff compound. Volume injection for capsule extract was 0.01 mL and that the A ymetrix protocol. These arrays allow the assessment of the for plasma samples, 0.03 mL. Because of the absence of commercially expression level of 28869 genes represented by approximately 26 available standards for metabolites, the same MS/MS conditions of EA probes spread across the full length of the gene, providing a more complete and accurate picture of gene expression than the classical 3′ were used for other compounds. For calibration curves, stock solutions fi of EA (1 mg 5 mL−1) and GA standards were prepared by dissolving -based expression array designs. Cell intensity les for each Gene 1.0 each compound in MeOH. Eight different solutions, containing, ST chip processed were generated utilizing Command Console −1 software. Microarray statistical analysis was performed utilizing the respectively, 0.1, 0.5, 1, 5, 10, 25, 50, and 100 mg mL of each 16 compound (external standards), were prepared in MeOH and used for OneChannelGUI R package. Raw signal intensities were normalized utilizing the gene-chip robust multiarray average (GCRMA) method method development. A volume of 0.01 mL was injected in the employing the empirical Bayes approach for background correction chromatographic system. The calibration curves for each standard followed by quantile normalization. Differentially expressed genes were were made by linear regression by a graph reporting the area ratio of identified using the “limma package”, applying linear models and external standard against the known concentration of external moderated t statistics that implement empirical Bayes regularization of standard. The result represents the average of curves performed by standard errors. Differences in expression values were expressed by three injection of each concentration. All quantitative data were ff means of log2 fold change (FC). A minimum di erence threshold of elaborated with the aid of Analyst software (Applied Biosystems). ETs FC ± 0.5 with a p value ≤0.05 was selected as gene differentially and ET metabolites were identified according to their ion mass spectra, 12 fi expressed in cells incubated with serum isolated after the retention times, and MS-MS fragments. For quanti cation purposes supplementation with respect to cells incubated with serum isolated data were collected in MRM tracking the transition of parent and before the supplementation. To highlight small variations on a product ions specific for each compound. Because of the absence of fi quantitative scale, we decided to decrease the threshold of FC. The commercially available standards for metabolites, quanti cation was lists of significantly differentially expressed genes in association with calculated on the basis of EA calibration curve. In Figure 3 the HPLC- the administration of Robuvit-enriched serum were analyzed by using TIC-MS chromatograms of standards and a representative post- DAVID17 and FIDEA18 Web servers to identify statistically over- supplementation plasma sample are shown. represented biological processes annotated in gene ontology (GO). − Plasmatic Total Phenols. Folin Ciocalteu colorimetry was used to Enriched GO biological processes (level 5) and Kegg pathways were measure the total phenolic amount in plasma isolated from the three identified according to p value threshold <0.05. subjects before and after Q. robur supplementation. To remove Both tools produced overlapping lists of enriched terms, and over- plasmatic protein interference, 0.5 M perchloric acid (1:1) was used, represented biological processes highlighted by DAVID were further whereas the hydrolysis of the conjugated forms of polyphenols was clustered utilizing the functional annotation clustering utility included 13 performed according to Serafini. Results are expressed as micrograms in the Web server. Most representative terms were selected and are of GA equivalent (GAE) per milliliter of plasma. This assay was also listed under Results. carried out on Q. robur capsule after extraction with 80% MeOH. Statistical Analysis. Data obtained by total phenol analysis and Antioxidant Capacity Analysis. The total antioxidant capacity has antioxidant capacity analysis were analyzed using a paired, two-tailed, been evaluated using the oxygen radical absorbance capacity (ORAC) Student’s t test. The threshold for significance was set at p < 0.05. assay. Plasma was treated according to the method of Prior14 to obtain the hydrophilic and lipophilic components of the plasma and to ■ RESULTS fl μ remove the interference of protein fraction. Brie y, 100 L of plasma Q. robur was mixed with 200 μL of ethanol and 100 μL of water. Then 400 μL Phenolic Composition of Capsules. Each of hexane was added twice. The hexane layers were removed, Robuvit capsule contains a total phenol amount equivalent to combined, and dried under nitrogen. For the lipophilic antioxidant 59.91 ± 14.67 mg of GA equivalents (GAE). On the basis of assay, the dried hexane extract was dissolved in 250 μL of acetone and these data, each volunteer ingested approximately 179 mg of then diluted with a 7% randomly methylated β-cyclodextrin (RMCD) total phenols per day, expressed as GAE, for 5 days. The solution as a solubility enhancer. For the hydrophilic antioxidant assay, content of the principal soluble ETs in Robuvit capsule is the residual aqueous plasma sample was treated with 400 μL of 0.5 M presented in Table 1. The capsule contains EA and GA and perchloric acid to precipitate proteins and then dissolved in phosphate peculiar ETs, such as vescalagin, castalagin, or vescalin, buffer. ORAC assay was performed in a microplate fluorescence reader 15 fl and grandinin or roburin E. The soluble ETs were about 1.8 (TECAN) according to the method of Huang et al. Brie y, ORAC is mg/capsule. a method that utilizes the peroxyl radical generator 2,2-azobis(2- Plasmatic Total Phenols and ET Metabolites. amidinopropane) dihydrochloride (AAPH) and fluorescein as an Five days of Robuvit supplementation are associated with doubling of oxidizable substrate. AAPH induces oxidative damage to the substrate, μ which is inhibited by the presence of reductants. The inhibition total plasmatic phenolic concentration up to 1.26 g GAE/mL measured is related to the antioxidant capacity of the reductants under (Table 2). Table 3 shows the identification and relative analysis. Results are expressed as nanomoles of Trolox equivalent per quantification of six ET metabolites obtained by HPLC-MS2 milliliter of plasma. analysis of plasma isolated after supplementation. EA, GA, and

446 dx.doi.org/10.1021/jf403493a | J. Agric. Food Chem. 2014, 62, 443−453 Journal of Agricultural and Food Chemistry Article

Table 1. Analysis of Soluble ETs in Robuvit and Milligrams Table 4. Plasmatic Antioxidant Capacity Analysis (ORAC) a of Polyphenols Consumed Each Day by the Volunteers Partitioned in Hydrophilic and Lipophilic Components

μg/capsule mg/day (three capsules per day) nmol Trolox equiv/mL plasma gallic acid 643 ± 152 1.9 ± 0.4 hydrophilic component lipophilic component ± ± ellagic acid 491 52 1.5 0.1 before supplementation 1.8 ± 0.05 1.2 ± 0.07 a ± ± castalin/vescalin 306 69 0.9 0.2 after supplementation 1.9 ± 0.01b 1.2 ± 0.06 a grandinin/roburin E 120 ± 20 0.4 ± 0.1 a a ± ± Data are expressed as nmol Trolox equiv/mL plasma and measured HHDP-galloyl-glucose 87 5 0.3 0.1 before and after supplementation. bSignificantly different p < 0.05. castalagina 84 ± 1 0.2 ± 0.1 vescalagina 76 ± 4 0.2 ± 0.1 a the spliceosome that removes introns from a transcribed pre- HHDP-glucose traces RNA) was modulated with a FC ≥ 1, and 13 genes were a Data are expressed as μg ellagic acid equivalent (EAE)/tablet. modulated with a FC between ±1 and ±0.5 (Table 5). In EA.hy.926 (permanent human endothelial cells), the Table 2. Increase of Plasmatic Total Phenols Measured after genes RPL30 and RPL24 (60S ribosomal proteins) have been Supplementation found to be up-regulated with a FC > 1 and 12 genes up- μg GAE equiv/mL plasma regulated with a FC between 0.5 and 1. Eight genes were down- regulated with a FC between −0.5 and −1 (Table 6). before supplementation 0.63 ± 0.06 In SHSY5Y cells (human bone marrow neuroblastoma), two after supplementation 1.26 ± 0.18a genes (SMNDC1, a component of the spliceosome that a fi ff Signi cantly di erent, p < 0.005. removes introns from a transcribed pre-RNA, and FAU, involved in translation) were down-regulated with a FC ≤ 1 four EA-derived metabolites, conjugated with glucuronic acid, and 12 genes were up-regulated with a FC number between 0.5 namely, urolithin B-glucuronide, urolithin A-glucuronide, and 1. Conversely, 23 genes have been found to be down- urolithin C-diglucuronide, and EA methyl ether glucuronide, regulated with a FC between −0.5 and −1 (Table 7). Despite were detected. The total concentration of metabolites in plasma these cell-specificdifferences, the genes modulated by Robuvit was in the order of 0.2 μM. With the exception of EA, none of supplementation are mainly involved in translation, cell cycle, these metabolites was present in plasma collected before RNA splicing, and ribosome biogenesis in all cell types, Robuvit supplementation. No sulfate derivatives were detected according to the statistical pathway enrichment analysis. in our analysis. The interindividual variability observed among On the basis of the evidence of a non-negligible the volunteers was quite high. interindividual difference of Q. robur metabolite profile we Antioxidant Capacity Analysis. To assess if the considered that the effect of Robivit supplementation on gene administration of Robuvit resulted in an increase of plasmatic expression modulation could be flattened out. We therefore antioxidant capacity, we performed a classical “antioxidant assessed the effect of sera isolated from each subject (1,2 or 3) capacity test”, according to the ORAC assay. Five days of separately on the different cell lines. Table 8 shows the total Robuvit supplementation was associated with a significant number of genes modulated within a FC≤≥1. However, increase of the ORAC value related to the hydrophilic plasmatic because most of the genes regulated are pseudogenes, not component, but not to the lipohilic fraction (Table 4). When mapped in any GO term, the pathway analysis is only able to the assay was directly performed on a Robuvit capsule, an capture a percentage ranging from 30 to 50% of the total ORAC value equal to 648 nmol Trolox eq/mg Q. robur was number of the modulated genes measured (shown in the obtained. second column). Moreover, the table shows the list of the Microarray Results. As mentioned under Materials and genes, up- and down-regulated by each serum, that were used Methods, human serum isolated before and after 5 days of by the software to enrich the GO pathways and the resulting Robuvit supplementation was used to enrich the culture over-represented pathway in the three cell lines: ribosome, medium of cells at a concentration of 10% v/v. aminoacyl tRNA biosynthesis, and spliceosome. These results Microarray results show that, in this experimental condition, show that, independently of the individual profile metabolites the presence of Robuvit metabolites induces a differential resulting by Robuvit supplementation, the over-represented expression of a limited number of genes. In HaCaT (stabilized pathways associated with the three sera are similar in the three human keratinocites), only one gene (PRPF8, a component of cell lines considered in the study.

Table 3. Analysis of the Metabolites Contained in Plasma Isolated before and after Supplementation from Each Subject

before supplementation after supplementation subject subject 1 2 3 mean 1 2 3 mean gallic acid 0 0 0 0 ± 0 0 2.9 6.5 3 ± 3 ellagic acid 13.5 4.5 0 6 ± 7 14.5 14.5 9.5 13 ± 11 urolithin A glucuronidea 0000± 0 8.1 11.8 12.2 11 ± 2 urolithin B glucuronidea 0000± 0 96.1 7.1 7.9 37 ± 51 urolithin C diglucuronidea 0000± 0 0 24.1 46.9 24 ± 23 EA methyl ether glucuronide 0 0 0 0 ± 0 5.6 0 0 2 ± 3 aData are expressed as ng of ellagic acid equivalent (EAE)/mL plasma.

447 dx.doi.org/10.1021/jf403493a | J. Agric. Food Chem. 2014, 62, 443−453 Journal of Agricultural and Food Chemistry Article

a Table 5. Gene Expression Regulation by the Metabolites of Q. robur in HaCaT

official gene symbol gene name FC p value representative pathways USP22 ubiquitin specific peptidase 22 0.51 0.007 cell cycle EIF4G2 eukaryotic translation initiation factor 4 γ 2 −0.55 0.003 cell cycle NPM1 nucleophosmin −0.77 0.004 ribosome biogenesis; cell cycle RPS7 40S ribosomal protein S7 0.71 0.002 ribosome biogenesis; RNA processing C1D C1D nuclear receptor corepressor −0.60 0.034 ribosome biogenesis; RNA processing PRPF8 pre-mRNA processing factor 8 homologue 1.00 0.018 RNA processing TARDBP TAR DNA-binding protein 43 −0.77 0.027 RNA processing; cell cycle TP53TG3 TP53TG3 0.58 0.002 LYZL2 lysozyme 2 0.53 0.007 CD69 CD69 0.51 0.032 EIF3D eukaryotic translation initiation factor 3 subunit D −0.51 0.012 CAMTA2 calmodulin binding transcription activator 2 −0.56 0.022 SNX3 sorting nexin-3 −0.76 0.041 FMN1 formin 1 −0.83 0.023 aFC threshold is ≤≥ 0.5; p value ≤0.05.

a Table 6. Gene Expression Regulation by the Metabolites of Q. robur in EA.hy.926

official gene symbol gene name FC p value representative pathways NPM1 nucleophosmin 0.82 0.042 cell cycle RPL24 60S ribosomal protein L24 1.08 0.011 cell cycle; M phase mitotic cycle SEPT2 septin 2 0.52 0.028 cell cycle; M phase mitotic cycle SMC4 structural maintenance of chromosomes protein 4 0.50 0.007 cell cycle; M phase mitotic cycle RPL30 60S ribosomal protein L30 1.28 0.001 ELMO1 engulfment and cell motility protein 1 0.95 0.047 MT-ND2 NADH-ubiquinone oxidoreductase chain 2 0.75 0.037 HNRNPC heterogeneous nuclear ribonucleoproteins C1/C2 0.59 0.045 NONO non-POU domain-containing octamer-binding protein 0.58 0.045 EIF4E3 eukaryotic translation initiation factor 4E member 3 0.57 0.005 ADAM21 ADAM metallopeptidase domain 21 0.54 0.034 B3GNT5 βGal β-1,3-N-acetylglucosaminyltransferase 5 0.53 0.004 EPCAM epithelial cell adhesion molecule −0.83 0.013 IK IK cytokine −0.51 0.008 CYP1A1 cytochrome P450, family 1 −0.53 0.006 MYST2 histone acetyltransferase −0.72 0.011 CD69 CD69 −0.73 0.004 C8orf73 −0.73 0.002 C22orf28 0.73 0.021 AC007688.1 −0.61 0.002 LOC642947 −0.51 0.002 HOMER2 0.55 0.007 aFC threshold is ≤≥ 0.5; p value ≤0.05. ■ DISCUSSION produce the derivative urolithins. Similarly to isoflavone fi metabolism, where there is a quite clear-cut separation between This study represents the rst investigation on standardized 22 French oak wood extract, Robuvit, absorption/metabolism and equol and nonequol producers, it has been proposed that on the effect of Robuvit metabolites on gene expression humans can be divided into high and low urolithin producers due to their microbiota, supporting the large variability in the modulation, by utilizing a combined ex vivo cell culture 23,24 approach.19 To the best of our knowledge we have reported for metabolism of ETs. The appearance of three glucuronated the first time absorption of ETs from Q. robur in humans; in urolithin (urolithin B-glucuronide, urolithin A-glucuronide, fact, to date, metabolites resulting from human consumption of urolithin C-diglucuronide) moieties in plasma isolated after oak wood constituents remained unknown. the supplementation, absent at baseline, suggests that oak wood The HPLC-MS2 analysis of Robuvit capsule shows that it is ETs are bioavailable. Although glucuronidated urolithin A was mainly composed of GA and EA and of several ETs such as present in all three subjects, only one subject showed high castalagin, vescalagin, and roburin E. Our data show that the urolithin B (about 0.1 μg/mL) but not detectable urolithin C consumption of about 180 mg of total phenols per day (three levels. ET bioavailability has been already studied addressing capsules a day of Robuvit) resulted in a significant increase of the consumption of different foods and beverages.11 EA and GA and EA in plasma. GA is generally rapidly absorbed and has urolithin glucoronides were detected in human plasma at 6 h a good bioavailability,20,21 whereas ETs are hydrolyzed in vivo from pomegranate juice (PJ) administration (providing 318 mg to EA, which is further metabolized by gut microbiota mostly to of and 12 mg of free EA)9,25 or pomegranate

448 dx.doi.org/10.1021/jf403493a | J. Agric. Food Chem. 2014, 62, 443−453 Journal of Agricultural and Food Chemistry Article

a Table 7. Gene Expression Regulation by the Metabolites of Q. robur in SHSY-5Y

official gene symbol gene name FC p value representative pathways EIF4G2 eukaryotic translation initiation factor 4 γ 2 0.91 0.000 translation RBMY1D RNA binding motif protein 0.56 0.000 RNA splicing HNRNPK heterogeneous nuclear ribonucleoprotein K −0.52 0.005 RNA splicing SNRNP200 small nuclear ribonucleoprotein 200 kDa −0.69 0.029 RNA splicing HNRNPM heterogeneous nuclear ribonucleoprotein M −0.91 0.001 RNA splicing SMNDC1 survival motor neuron domain containing 1 −1.10 0.003 RNA splicing EIF3D eukaryotic translation initiation factor 3 subunit D 0.69 0.014 translation RPL17 ribosomal protein L17 −0.58 0.001 translation; ribosome RPL4 60S ribosomal protein L4 0.58 0.029 translation; ribosome RPS13 40S ribosomal protein S13 0.56 0.013 translation; ribosome FAU Finkel−Biskis−Reilly murine sarcoma virus −1.00 0.002 translation; ribosome PSMB2 proteasome subunit, β type, 2 0.85 0.002 EFCAB4B calcium release-activated calcium channel regulator 2A 0.75 0.002 ACAA1 acetyl-eoenzyme A acyltransferase 1 0.66 0.003 OR5L1 olfactory receptor 5L1 0.62 0.031 STARD7 StAR-related lipid transfer (START) domain containing 7 0.55 0.005 TSPAN33 tetraspanin 33 0.53 0.017 FLJ25758 0.50 0.025 BBS9 Bardet−Biedl syndrome 9 −0.53 0.039 FOLH1 folate hydrolase −0.54 0.036 FOXK2 forkhead box −0.56 0.034 CR769776.1 ankyrin repeat domain 20 family −0.59 0.008 LIAS lipoic acid synthetase −0.60 0.010 AC068775.1 −0.61 0.025 ZNF846 zinc finger protein 846 −0.62 0.015 ANKRD20A4 ankyrin repeat domain 20 family −0.63 0.003 ZNF117 zinc finger protein 117 −0.65 0.013 ARF1 ADP-ribosylation factor 1 −0.66 0.047 ARF3 ADP-ribosylation factor 3 −0.67 0.023 CITED1 Cbp/p300-interacting transactivator −0.69 0.008 XCL1 chemokine (C motif) −0.69 0.013 SLC2A14 solute carrier family 2 (facilitated glucose transporter) −0.76 0.004 CBX3 chromobox homologue 3 −0.80 0.027 HSD3B1 hydroxy-δ-5-steroid dehydrogenase, 3 β- and steroid δ-isomerase 1 −0.85 0.039 C19orf50 −0.87 0.042 USP22 ubiquitin specific peptidase 22 −0.92 0.000 SART3 squamous cell carcinoma antigen recognized by T cells 3 −1.01 0.027 aFC threshold is ≤≥ 0.5; p value ≤0.05. extract (providing 330 mg of and 21 mg of EA).26 urine9 and plasma28 of subjects supplemented with PJ and in Glucuronides were the only urolithin metabolites found in the plasma of rats supplemented with 6% punicalagin.30 In the plasma of subjects supplemented for 5 weeks with PJ present study, we detected a low amount of methylellagic consumption.27 Urolithin glucuronides and EA derivatives glucuronide in plasma isolated after the supplementation in were detected after 3 days of walnut consumption.28 Stoner only one of three subjects. Also in agreement with our study, a and co-workers found EA in human plasma at 2 h after the high interindividual variability of metabolic profile has been administration of 45 g of freeze-dried black rasberries already frequently reported.9,24,27,28 (containing 13.5 mg of EA).29 However, neither EA nor any Moreover, as expected, the total plasma polyphenol amount of ET metabolites have been found in detectable quantities in increased significantly by on average 100% in all volunteers. human plasma over a 24 h period after 300 g of rasberry The consumption of the aqueous oak wood extract Robuvit led, consumption.12 According to the authors, this negative result is also, to a statistically significant increase of hydrophilic blood not a real proof of absence as it could be due to the high oxygen radical absorbance capacity (ORAC) in healthy detection limit of the method used in the study. In agreement volunteers, even if the small change observed may not have a with our data, no sulfate metabolites have been found in biological relevance. humans after ET-rich food supplementation.10 The methodologies utilized to assess plasma polyphenol EA structure presents two ο-dihydroxy groups, and it may concentration and plasma antioxidant capacity are known to be therefore be transformed by the activity of catechol-O- non specific. In fact, it has been reported that the Folin assay methyltransferase (COMT) (in the liver) in mono- and/or provides polyphenol estimation consistently 2−4 times greater dimethyl ether of EA. These metabolites can be further than that obtained by HPLC methods,31 because it measures metabolized by enzymatic glucuronic acid derivatization. In the reducing capacity of phenolic components by detecting fact, dimethylellagic acid glucuronide has been found in the nonspecific hydroxyl groups. GA and EA contain three and four

449 dx.doi.org/10.1021/jf403493a | J. Agric. Food Chem. 2014, 62, 443−453 a Chemistry Food and Agricultural of Journal Table 8. Gene Expression Regulation Associated with Each Serum Isolated from the Healthy Subjects (Sera 1, 2, and 3) in HaCaT, EA.hy.926, and SHSY-5Y Cell Lines

total genes serum modulated up-regulated down-regulated over-represented pathway HaCaT 1 36 APOBE3CA, USP22, ZNF259 CANX, EIF3F, FNTA, HNRNPM, ILF2, NPM1, RPL11, SART3, SEPT2, TCEB2, TARDBP, ZNF737, 2 29 CANX, NARS, PRPF8, RPS7, SART3 ANAPC5, EIF3D, FNTA, HRNPC, IK, ZNF259, ZNF737 spliceosome 3 52 IL10RA, LYZL2, NPM1, PRPF8, RPL11, RPL17, SEPT2, ZNF259 ARF3, C22orf28,CBX3, EIF3D, HNRNPC, IL18R1, IK, KARS, NONO, spliceosome, ribosome SMNCD1, SNX3, SNX32 EA.hy.926 1 47 ANAPC5, ARF3, CASZ1, C1D, EIF3F, FNTA, MT-ND2, NPM1, DEPDC5, EPCAM, HNRNPC, RPL34, TCEB2 ribosome RHOA,RPL11, RPL30, SMNDC1, SART3, SEPT2, SNX32, TARDBP 2 48 CANX, ELMO1, IL10RA, KARS, NARS, RPL11, RPL17, RPL24, RPL30, CD69, CYP1A1,EPCAM, PDE5A, HSD3B1, IK, ILF2, ILF2, RASA4 ribosome, aminoacyl-tRNA RPL34, SART1, SART3 biosynthesis, steroid hormone biosynthesis 3 50 AREG, ANAPC5, CANX, CBX3, EIF4E3, IK, MT-CYB, NARS, NONO, RPL34, ARF3, EIF3F, HNRNPK, SART3, TCEB2 ribosome RPL24, RPL30, SEPT2, SERPIND2,SMNDC1, TAF1D, ZNF28 SHSY-5Y 1 38 CDC26, EPCAM, JTB ACTR3C, C1D, EIF3D, FAU, GUK1, HNRNPK, EFCAB4B, IK, IL10RA, KARS, aminoacyl-tRNA biosynthesis NARS, NPM1, RPL24, SART1, TARDBP, TCEB2 2 101 ANAPC5, AREG, CANX, CBX3, EIF3F, FNTA, NPM1, RBMYD1D, ARF3, ATP6 V0B, C1D, c22orf28, CCL4, EIF3D, HNRNPK, HNRNPM, ILF2, SLC25A43, SMNDC1, TMED2, ZNF259, ZNF487P, ZNF492 NARS, NONO, RHOA, RPL4, RPS27A, SEPT2, SNX3, SRP14, USP22, XCL1 3 95 ARF3, C1D, CBX3, EIF3F, GPC2, HSD3B1, KARS, MT-ND3, PTCH2, C22orf28, CANX, HNRNPC, HNRNPM, IK, INF2, NPM1, PSMB2,PTH2R, spliceosoma

450 RABGGTB, RPNS1, SART1 RPL11, SART3, SLCO4C1 SMNDC1, TARDBP, TMED2 a FC threshold is ≤≥ 1. dx.doi.org/10.1021/jf403493a | .Arc odChem. Food Agric. J. 04 2 443 62, 2014, Article − 453 Journal of Agricultural and Food Chemistry Article hydroxyl groups, respectively, and they strongly contribute to putative effect of these botanical extracts in human health and Folin and possess high antioxidant properties. However, in our disease. results the extent of the increase of plasmatic GA and EA, after Specific interindividual differences in the metabolization of the supplementation, is not sufficient to explain the observed Q. robur extract components may have differently influenced increase of ORAC and Folin. It is known that urolithins the final outcome in terms of the single gene expression. exhibited a significant antioxidant action correlated with the However, surprisingly, the functional analysis reveals that the number of hydroxyl groups,32 whereas no antioxidant activity three sera, independently of the individual profile metabolites, has been reported for the glucuronide derivative forms.24,33 modulate similar pathways, ribosome, aminoacyl tRNA biosyn- Therefore, we have to hypothesize that a synergic activity of all thesis, and spliceosome, suggesting that the ET metabolites Q. robur metabolites occurs. display a uniform behavior, in a cell-type independent way. Dietary EA and its metabolites, either independently or However, the small differences in FC of most of the genes do synergistically, are good candidates to be among the most active not let us assign to the modulated pathways either a positive or molecules responsible for a wide spectrum of molecular effects. a negative health concern. Moreover, it is important to note In vivo and in vitro studies have shown that EA and urolithins that the rate of cell proliferation was not affected by serum have antiproliferative effects in human bladder cancer cells and containing Q. robur metabolites compared to control sera (data antitumoral properties in several lines of cultured human cancer not shown). cells,34,35 in hamster buccal carcinoma36 and anti-inflammatory In conclusion, we demonstrate that Robuvit, a proprietary properties in colon fibroblasts,4 and in human primary patented extract from Q. robur, is (at least in part) bioavailable endothelial cells.37 EA and urolithin A have been demonstrated to humans and that despite the very low concentration of the to down-regulate the expression of COX-2 and iNOS and other metabolite in serum, its consumption is associated with a inflammatory markers and to modulate the expression of genes statistically significant increase of the antioxidant capacity and involved in cell cycle in colonic mucosa in a colitis rat model.4 to specific changes of gene expression profile. Moreover, CaCo2 cells exposed to concentrations of EA and urolithins achievable in the colon from the diet arrested cell ■ AUTHOR INFORMATION fi growth at the S and G2/M phases. Transcriptional pro ling Corresponding Author obtained by microarray and functional analysis revealed * fi (R. Canali) Phone: +39 06 51494519. Fax: +39 06 51494550. signi cant changes in the expression levels genes involved in E-mail: raff[email protected]. MAPK signaling such as growth factor receptors (FGFR2, Author Contributions EGFR), oncogenes (K-Ras, c-Myc), and tumor suppressors § (DUSP6, Fos) as well as the modulation of genes involved in F.N. and G.L. equally contributed. cell cycle (CCNB1,CCNB1IP1).38 Similarly, GA has been Funding shown to inhibit carcinogenesis in animal models and in vitro This study was supported by Horphag Research, Geneva, cancerous cell lines. The inhibitory effect of GA on cancer cell Switzerland, and by Italian Ministry of Agriculture, Food and growth has been reported to be mediated by the modulation of Forestry (MiPAAF) Grants “NUME” (DM 3688/7303/08). the expression of genes encoding for proteins involved in cell Notes 39,40 cycle, metastasis, angiogenesis, and apoptosis. Upstream to The authors declare the following competing financial gene expression, GA has been reported to inhibit activation of interest(s): F.S. worked at Horphag Research Ltd. The other κ NF- B and Akt signaling pathways along with the activity of authors declare no competing financial interest. COX and inflammatory cytokines.41,42 The present study provides an original background to the ACKNOWLEDGMENTS ff ■ understanding of possible e ects of Q. robur extract metabolites ́ on the modulation of gene expression in endothelial, neuronal, We are grateful to Prof. Michael Jourdes (Universite Bordeaux Segalen, Facultéd’Oenologie, Unitéde recherche Oenologie, and keratinocyte cell lines. Our methodological approach allows ’ the assessment of the effects of Q. robur metabolites at Villenave d Ornon, France) for his help with the chemical concentrations physiologically achievable in vivo and in a structure drawing. profile prevailing in humans. The transcriptome of each cell line was identified utilizing an Affymetrix customized array that ■ ABBREVIATIONS USED recognizes a total number of 28869 genes. ETs, ellagitannins; EA, ellagic acid; GA, gallic acid Subsequent, functional analysis revealed that culturing cells in serum obtained from subjects administered Q. robur extract ■ REFERENCES fi results in a moderate although signi cant modulation of the (1) Arts, I. C.; Hollman, P. C. Polyphenols and disease risk in expression of genes related to specific GO categories, namely, epidemiologic studies. Am. J. Clin. Nutr. 2005, 81, 317S−325S. cell cycle, ribosome biogenesis, and RNA processing pathways (2) Queen, B. L.; Tollefsbol, T. O. Polyphenols and aging. Curr. in HaCaT (stabilized human keratinocytes) cells; cell cycle, M Aging Sci. 2010, 3,34−42. phase mitotic cycle in Ea.Hy.926 cells (a permanent human (3) Forman, H. J.; Davies, K. J.; Ursini, F. How do nutritional endothelial cells); translation, RNA splicing, and ribosome antioxidants really work: nucleophilic tone and para-hormesis versus pathways in SHSY5Y cells (human bone marrow neuro- free radical scavenging in vivo. Free Radical Biol. Med. 2013, blastoma). Obviously, changes of gene expression at the level of DOI: 10.1016/j.freeradbiomed.2013.05.045. (4) Larrosa, M.; Gonzalez-Sarrias, A.; Yanez-Gascon, M. J.; Selma, M. mRNA do not necessarily imply that corresponding changes V.; Azorin-Ortuno, M.; Toti, S.; Tomas-Barberan, F.; Dolara, P.; Espin, occur at the level of protein. Nonetheless, an understanding of J. C. Anti-inflammatory properties of a pomegranate extract and its ff the e ect of dietary bioactive molecules at the transcriptional metabolite urolithin-A in a colitis rat model and the effect of colon level is important to the building up of a conceptual inflammation on phenolic metabolism. J. Nutr. Biochem. 2010, 21, background to be exploited in future studies addressing the 717−725.

451 dx.doi.org/10.1021/jf403493a | J. Agric. Food Chem. 2014, 62, 443−453 Journal of Agricultural and Food Chemistry Article

(5) Wu, S. B.; Dastmalchi, K.; Long, C.; Kennelly, E. J. Metabolite (22) Shor, D.; Sathyapalan, T.; Atkin, S. L.; Thatcher, N. J. Does profiling of jaboticaba (Myrciaria cauliflora) and other dark-colored equol production determine soy endocrine effects? Eur. J. Nutr. 2012, fruit juices. J. Agric. Food Chem. 2012, 60, 7513−7525. 51, 389−398. (6) Mayer, W.; Gabler, W.; Riester, A.; Korger, H. On tannins (23) Cerda, B.; Periago, P.; Espin, J. C.; Tomas-Barberan, F. A. compounds from the wood of chestnut and oak wood−IIs. Isolation of Identification of urolithin a as a metabolite produced by human colon castalagin, vescalagin, castalin and vescaline. Liebigs Ann. Chem. 1967, microflora from ellagic acid and related compounds. J. Agric. Food 707, 177−181. Chem. 2005, 53, 5571−5576. (7) García-Estevez,́ I.; Escribano-Bailon,́ M. T.; Rivas-Gonzalo, J. C.; (24) Cerda, B.; Espin, J. C.; Parra, S.; Martinez, P.; Tomas-Barberan, Alcalde-Eon, C. Development of a fractionation method for the F. A. The potent in vitro antioxidant ellagitannins from pomegranate detection and identification of oak ellagitannins in red wines. Anal. juice are metabolised into bioavailable but poor antioxidant hydroxy- Chim. Acta 2010, 660, 171−176. 6H-dibenzopyran-6-one derivatives by the colonic microflora of (8) Glabasnia, A.; Hofmann, T. Sensory-directed identification of healthy humans. Eur. J. Nutr. 2004, 43, 205−220. taste-active ellagitannins in American (Quercus alba L.) and European (25) Seeram, N. P.; Lee, R.; Heber, D. Bioavailability of ellagic acid in oak wood (Quercus robur L.) and quantitative analysis in bourbon human plasma after consumption of ellagitannins from pomegranate whiskey and oak-matured red wines. J. Agric. Food Chem. 2006, 54, (Punica granatum L.) juice. Clin. Chim. Acta 2004, 348,63−68. 3380−3390. (26) Mertens-Talcott, S. U.; Jilma-Stohlawetz, P.; Rios, J.; Hingorani, (9) Seeram, N. P.; Henning, S. M.; Zhang, Y.; Suchard, M.; Li, Z.; L.; Derendorf, H. Absorption, metabolism, and antioxidant effects of Heber, D. Pomegranate juice metabolites are present in pomegranate (Punica granatum L.) polyphenols after ingestion of a human plasma and some persist in urine for up to 48 h. J. Nutr. 2006, standardized extract in healthy human volunteers. J. Agric. Food Chem. 136, 2481−2485. 2006, 54, 8956−8961. (10) Espın,́ J. C.; Gonzalez-Barrio,́ R.; Cerda,́ B.; Lopez-Bote,́ C.; (27) Cerda, B.; Soto, C.; Albaladejo, M. D.; Martinez, P.; Sanchez- Rey, A. I.; Tomas-Barberá n,́ F. A. Iberian pig as a model to clarify Gascon, F.; Tomas-Barberan, F.; Espin, J. C. Pomegranate juice obscure points in the bioavailability and metabolism of ellagitannins in supplementation in chronic obstructive pulmonary disease: a 5-week humans. J. Agric. Food Chem. 2007, 55, 10476−10485. randomized, double-blind, placebo-controlled trial. Eur. J. Clin. Nutr. (11) Cerda, B.; Tomas-Barberan, F. A.; Espin, J. C. Metabolism of 2006, 60, 245−253. antioxidant and chemopreventive ellagitannins from strawberries, (28) Gonzalez-Sarrias, A.; Gimenez-Bastida, J. A.; Garcia-Conesa, M. raspberries, walnuts, and oak-aged wine in humans: identification of T.; Gomez-Sanchez, M. B.; Garcia-Talavera, N. V.; Gil-Izquierdo, A.; biomarkers and individual variability. J. Agric. Food Chem. 2005, 53, Sanchez-Alvarez, C.; Fontana-Compiano, L. O.; Morga-Egea, J. P.; 227−235. Pastor-Quirante, F. A.; Martinez-Diaz, F.; Tomas-Barberan, F. A.; (12) Gonzalez-Barrio, R.; Borges, G.; Mullen, W.; Crozier, A. Espin, J. C. Occurrence of urolithins, gut microbiota ellagic acid Bioavailability of anthocyanins and ellagitannins following consump- metabolites and proliferation markers expression response in the tion of raspberries by healthy humans and subjects with an ileostomy. human prostate gland upon consumption of walnuts and pomegranate J. Agric. Food Chem. 2010, 58, 3933−3939. juice. Mol. Nutr. Food Res. 2010, 54, 311−322. (13) Serafini, M.; Maiani, G.; Ferro-Luzzi, A. Alcohol-free red wine (29) Stoner, G. D.; Sardo, C.; Apseloff, G.; Mullet, D.; Wargo, W.; enhances plasma antioxidant capacity in humans. J. Nutr. 1998, 128, Pound, V.; Singh, A.; Sanders, J.; Aziz, R.; Casto, B.; Sun, X. 1003−1007. Pharmacokinetics of anthocyanins and ellagic acid in healthy (14) Prior, R. L.; Hoang, H.; Gu, L.; Wu, X.; Bacchiocca, M.; volunteers fed freeze-dried black raspberries daily for 7 days. J. Clin. Howard, L.; Hampsch-Woodill, M.; Huang, D.; Ou, B.; Jacob, R. Pharmacol. 2005, 45, 1153−1164. Assays for hydrophilic and lipophilic antioxidant capacity (oxygen (30) Cerda, B.; Llorach, R.; Ceron, J. J.; Espin, J. C.; Tomas- radical absorbance capacity (ORAC(FL))) of plasma and other Barberan, F. A. Evaluation of the bioavailability and metabolism in the biological and food samples. J. Agric. Food Chem. 2003, 51, 3273− rat of punicalagin, an antioxidant polyphenol from pomegranate juice. 3279. Eur. J. Nutr. 2003, 42,18−28. (15) Huang, D.; Ou, B.; Hampsch-Woodill, M.; Flanagan, J. A.; Prior, (31) Robbins, R. J.; Kwik-Uribe, C.; Hammerstone, J. F.; Schmitz, H. R. L. High-throughput assay of oxygen radical absorbance capacity H. Analysis of flavanols in foods: what methods are required to enable (ORAC) using a multichannel liquid handling system coupled with a meaningful health recommendations? J. Cardiovasc. Pharmacol. 2006, microplate fluorescence reader in 96-well format. J. Agric. Food Chem. 47, S110−S121. 2002, 50, 4437−4444. (32) Bialonska, D.; Kasimsetty, S. G.; Khan, S. I.; Ferreira, D. (16) Sanges, R.; Cordero, F.; Calogero, R. A. oneChannelGUI: a Urolithins, intestinal microbial metabolites of pomegranate ellagitan- graphical interface to Bioconductor tools, designed for life scientists nins, exhibit potent antioxidant activity in a cell-based assay. J. Agric. who are not familiar with R language. Bioinformatics 2007, 23, 3406− Food Chem. 2009, 57, 10181−10186. 3408. (33) Ito, H.; Iguchi, A.; Hatano, T. Identification of urinary and (17) Huang da, W.; Sherman, B. T.; Tan, Q.; Kir, J.; Liu, D.; Bryant, intestinal bacterial metabolites of ellagitannin in rats. J. Agric. D.; Guo, Y.; Stephens, R.; Baseler, M. W.; Lane, H. C.; Lempicki, R. A. Food Chem. 2008, 56, 393−400. DAVID Bioinformatics Resources: expanded annotation database and (34) Qiu, Z.; Zhou, B.; Jin, L.; Yu, H.; Liu, L.; Liu, Y.; Qin, C.; Xie, S.; novel algorithms to better extract biology from large gene lists. Nucleic Zhu, F. In vitro antioxidant and antiproliferative effects of ellagic acid Acids Res. 2007, 35, W169−175. and its colonic metabolite, urolithins, on human bladder cancer T24 (18) D’Andrea, D.; Grassi, L.; Mazzapioda, M.; Tramontano, A. cells. Food Chem. Toxicol. 2013, 59, 428−437. FIDEA: a server for the functional interpretation of differential (35) Losso, J. N.; Bansode, R. R.; Trappey, A., 2nd; Bawadi, H. A.; expression analysis. Nucleic Acids Res. 2013, 41, W84−88. Truax, R. In vitro anti-proliferative activities of ellagic acid. J. Nutr. (19) Canali, R.; Ambra, R.; Stelitano, C.; Mattivi, F.; Scaccini, C.; Biochem. 2004, 15, 672−678. Virgili, F. A novel model to study the biological effects of red wine at (36) Vidya Priyadarsini, R.; Kumar, N.; Khan, I.; Thiyagarajan, P.; the molecular level. Br. J. Nutr. 2007, 97, 1053−1058. Kondaiah, P.; Nagini, S. Gene expression signature of DMBA-induced (20) Shahrzad, S.; Aoyagi, K.; Winter, A.; Koyama, A.; Bitsch, I. hamster buccal pouch carcinomas: modulation by chlorophyllin and Pharmacokinetics of gallic acid and its relative bioavailability from tea ellagic acid. PLoS One 2012, 7, e34628. in healthy humans. J. Nutr. 2001, 131, 1207−1210. (37) Gimenez-Bastida, J. A.; Gonzalez-Sarrias, A.; Larrosa, M.; (21) Manach, C.; Williamson, G.; Morand, C.; Scalbert, A.; Remesy, Tomas-Barberan, F.; Espin, J. C.; Garcia-Conesa, M. T. Ellagitannin C. Bioavailability and bioefficacy of polyphenols in humans. I. Review metabolites, urolithin A glucuronide and its aglycone urolithin A, of 97 bioavailability studies. Am. J. Clin. Nutr. 2005, 81, 230S−242S. ameliorate TNF-alpha-induced inflammation and associated molecular

452 dx.doi.org/10.1021/jf403493a | J. Agric. Food Chem. 2014, 62, 443−453 Journal of Agricultural and Food Chemistry Article markers in human aortic endothelial cells. Mol. Nutr. Food Res. 2012, 56, 784−796. (38) Gonzalez-Sarrias, A.; Azorin-Ortuno, M.; Yanez-Gascon, M. J.; Tomas-Barberan, F. A.; Garcia-Conesa, M. T.; Espin, J. C. Dissimilar in vitro and in vivo effects of ellagic acid and its microbiota-derived metabolites, urolithins, on the cytochrome P450 1A1. J. Agric. Food Chem. 2009, 57, 5623−5632. (39) Ho, H. H.; Chang, C. S.; Ho, W. C.; Liao, S. Y.; Lin, W. L.; Wang, C. J. Gallic acid inhibits gastric cancer cells metastasis and invasive growth via increased expression of RhoB, downregulation of AKT/small GTPase signals and inhibition of NF-kappaB activity. Toxicol. Appl. Pharmacol. 2013, 266,76−85. (40) Yeh, R. D.; Chen, J. C.; Lai, T. Y.; Yang, J. S.; Yu, C. S.; Chiang, J. H.; Lu, C. C.; Yang, S. T.; Yu, C. C.; Chang, S. J.; Lin, H. Y.; Chung, J. G. Gallic acid induces G(0)/G(1) phase arrest and apoptosis in human leukemia HL-60 cells through inhibiting cyclin D and E, and activating mitochondria-dependent pathway. Anticancer Res. 2011, 31, 2821−2832. (41) Verma, S.; Singh, A.; Mishra, A. Gallic acid: molecular rival of cancer. Environ. Toxicol. Pharmacol. 2013, 35, 473−485. (42) Al-Halabi, R.; Bou Chedid, M.; Abou Merhi, R.; El-Hajj, H.; Zahr, H.; Schneider-Stock, R.; Bazarbachi, A.; Gali-Muhtasib, H. Gallotannin inhibits NFkB signaling and growth of human colon cancer xenografts. Cancer Biol. Ther. 2011, 12,59−68.

453 dx.doi.org/10.1021/jf403493a | J. Agric. Food Chem. 2014, 62, 443−453