European Journal of Clinical Nutrition (2007) 61, 1167–1173 & 2007 Nature Publishing Group All rights reserved 0954-3007/07 $30.00 www.nature.com/ejcn

ORIGINAL ARTICLE Effect of the main dietary (, c-tocopherol, , and C) on a-tocopherol absorption

E Reboul1,2,3, S Thap1,2,3, E Perrot1,2,3, M-J Amiot1,2,3, D Lairon1,2,3 and P Borel1,2,3

1INSERM, U476 ‘Nutrition Humaine et Lipides’, Marseille, France; 2INRA, UMR1260, Marseille, France and 3Univ Me´diterrane´e Aix- Marseille 2, Faculte´ de Me´decine, IPHM-IFR 125, Marseille, France

Objective:(R,R,R)-a-tocopherol is a -soluble vitamin generally ingested with other dietary antioxidants. The objective of this study was to assess whether the main dietary antioxidant classes, that is carotenoids, polyphenols, and g-tocopherol, affect the intestinal absorption of a-tocopherol. Methods, design and subjects: We evaluated first the effect of different combinations of antioxidants on (R,R,R)-a-tocopherol absorption by a human intestinal cell line (Caco-2 clone TC7). Then we compared the effect of two doses of a dietary antioxidant (lutein) on the postprandial chylomicron a-tocopherol responses to an a-tocopherol-rich meal. Eight healthy men ate two similar meals in a random order at a 1 month interval. The meals contained 24 mg a-tocopherol in sunflower oil plus either 18 or 36 mg lutein. Blood samples were collected during the postprandial periods to compare chylomicron a-tocopherol responses. Results: A mixture of polyphenols (gallic acid, caffeic acid, ( þ )- and naringenin) and a mixture of carotenoids (, b-carotene and lutein) significantly impaired a-tocopherol absorption in Caco-2 cells (Po0.001 and Po0.0001, respectively). The inhibitory effect of g-tocopherol was close to significance (P ¼ 0.055). In contrast, vitamin C had no significant effect (P ¼ 0.158). Naringenin was the only that significantly impaired a-tocopherol absorption. Postprandial a- tocopherol response was weakest at the highest dose of lutein (6167280 nmol/l h vs 10017287 nmol/l h). The observed extent of reduction (À38%, P ¼ 0.069) supported the inhibitory effect of carotenoids observed in the Caco-2 experiments. Conclusion: Naringenin, carotenoids and probably g-tocopherol can impair a-tocopherol absorption whereas vitamin C and phenolic acids have no effect. European Journal of Clinical Nutrition (2007) 61, 1167–1173; doi:10.1038/sj.ejcn.1602635; published online 31 January 2007

Keywords: ; enterocyte; naringenin; ( þ )-catechin; gallic acid; caffeic acid

Introduction

Vitamin E is the main fat-soluble dietary antioxidant. Of the eight dietary forms of vitamin E (a-, b-, g-, d-tocopherols and Correspondence: Dr P Borel, UMR 476 INSERM/1260 INRA, Faculte´ de a-, b-, g-, d-), two represent the vast majority of Me´decine, 27, Boulevard Jean-Moulin, 13385 Marseille Cedex 5, France. vitamin E intake: (R,R,R)-a and (R,R,R)-g-tocopherol. Vitamin E-mail: [email protected] Contributors: PB and ER contributed to the conception of the cell experiments. E absorption is assumed to be affected by several factors and PB and DL contributed to the design of the human study. M-JA provided the effect of on vitamin E absorption remains significant advice about the choice and the use of the different microcon- poorly understood (Borel, 2003). stituents. ER carried out the practical aspects of the study, with help from ST The metabolism of a-tocopherol, which is naturally non- for the cellular aspect and from EP for the postprandial study. ER and PB carried out the statistical analyses, and initially interpreted the data before esterified, may start in the stomach where foods are writing the manuscript. PB wrote the first draft of the manuscript with help processed by acidic conditions and gastric enzymes. In fact, from ER. All authors participated in the writing of the final draft of the no significant degradation of a-tocopherol was observed in manuscript and in the final interpretation of the data. None of the authors had the human stomach (Borel et al., 2001). In the duodenum, any conflicts of interest. Received 26 October 2006; revised 28 November 2006; accepted 29 it is assumed that a-tocopherol is transferred from vegetable November 2006; published online 31 January 2007 oil (in which it is generally embedded) into mixed micelles Dietary antioxidants and a-tocopherol absorption E Reboul et al 1168 during triacylglycerol hydrolysis by pancreatic lipase. Mixed Biomedia (Issy-les-Moulineaux, France) and non-essential micelles carry a-tocopherol to the enterocyte where it is, at amino acids and penicillin/streptomycin were purchased least partially, absorbed via scavenger-receptor class B type I from Gibco BRL (Cergy-Pontoise, France). (Reboul et al., 2005c). After cellular uptake, a significant proportion of a-tocopherol is incorporated into chylo- microns and transported to the bloodstream, via the Preparation of microconstituent-rich media for cell experiments lymphatics. For delivery of fat-soluble microconstituents (i.e. (R,R,R)-a- In a normal meal-containing plant-derived foods, a- and (R,R,R)-g-tocopherol, lutein, b-carotene and lycopene) to tocopherol is necessarily ingested with other antioxidants: cells, mixed micelles rich in each of the microconstituents carotenoids, g-tocopherol, polyphenols and vitamin C. It has were separately prepared as described previously (Reboul been suggested that these antioxidants may either (i) protect et al., 2005a) to obtain the following final concentrations: a-tocopherol against oxidative degradation in the upper 0.04 mM phosphatidylcholine, 0.16 mM lysophosphatidyl- gastrointestinal tract (Beyer, 1994; Halliwell et al., 2000; Sen choline, 0.3 mM monoolein, 0.1 mM free cholesterol, et al., 2000; Mortensen et al., 2001), thus enhancing its 0.5 mM oleic acid, 5 mM taurocholate and 0.1–5 mM of the absorption, or (ii) compete with a-tocopherol for incorpora- relevant microconstituent. Microconstituent concentrations tion into micelles or uptake by putative transporters, thus in the micellar solutions were checked before each experi- impairing its absorption. Recent data support both these ment. The different stock solutions of micellar micro- hypotheses. Loest et al. (2002) have shown that a poly- constituents were mixed to obtain the final concentrations -rich green tea extract inhibits the lymphatic absorp- required. Vitamin C was directly dissolved in the culture tion of a-tocopherol in rats. Another team has observed a medium. Polyphenols were added to the culture medium in decrease in plasma a-tocopherol concentrations following dimethylsulfoxide (DMSO) (a same volume of DMSO alone g-tocopherol supplementation (Yoshikawa et al., 2005). was added when there was no polyphenols). The micro- Conversely, data from Kama-Eldin et al. (1995, 2000), suggest constituent concentrations used are summarized in Table 1. that the polyphenol sesamin increases a-tocopherol absorption . This study aimed to specify which classes of dietary Tocopherol absorption by intestinal cells antioxidants significantly affect a-tocopherol absorption. Cell culture. Caco-2 clone TC7 cells (Salvini et al., 2002) The purified antioxidants and their relative concentrations were a gift from Dr M Rousset (U178 INSERM, Villejuif, were designed to mimic as closely as possible those found in France). Cells were cultured in the presence of DMEM a standard Western diet. A clinical study was performed with supplemented with 20% heat-inactivated FBS, 1% non- an antioxidant microconstituent (lutein) to validate the essential amino acid and 1% antibiotics (complete medium), results obtained with the enterocyte cell model. as described previously (Reboul et al., 2005a). For each experiment, cells were seeded and grown on semi-permeable filters as described previously (Reboul et al., 2005a) to obtain Materials and methods differentiated confluent cell monolayers. Twelve hours before each experiment, the medium of apical and baso- Chemicals lateral chambers was changed to a serum-free complete (R,R,R)-a-T (X99% pure) and (R,R,R)-g-T (X97% pure) were medium. During preliminary tests, the integrity of the cell purchased from Flucka (Vaulx-en-Velin, France). Lutein (96% monolayers was checked by measuring transepithelial elec- pure), lycopene (95.5% pure) and b-carotene (95.6% pure) trical resistance using a voltohmmeter fitted with a chopstick were kindly provided by DSM LTD (formerly F Hoffmann-La Roche, Basel, Switzerland). Naringenin, catechin, gallic acid, caffeic acid and eriodictyol were purchased from Extra- Table 1 Concentration of antioxidant microconstituents in the Caco-2 synthe`se (Genay, France). Tocol, which was used as internal cell experiments of the factorial design

standard for high-performance liquid chromatography a Microconstituent Concentration (mM) (HPLC) analysis, was purchased from Lara Spiral (Couternon, France). 2-Oleoyl-1-palmitoyl-sn-glycero-3-phosphocholine Lutein 0.63 (phosphatidylcholine), 1-palmitoyl-sn-glycero-3-phospho- b-carotene 0.19 choline (lysophosphatidylcholine), monoolein, free choles- Lycopene 0.33 (R,R,R)-a-tocopherol 4.4 terol, oleic acid, sodium taurocholate and pyrogallol were (R,R,R)-g-tocopherol 3.0 purchased from Sigma-Aldrich (Saint-Quentin-Fallavier, Naringenin 25 France). ( þ )-Catechin 25 Dulbecco’s modified Eagle medium (DMEM) containing Gallic acid 50 Caffeic acid 50 4.5 g/l glucose and trypsin-EDTA (500 and 200 mg/l, respec- Vitamin C 75 tively) was purchased from Bio Whittaker (Fontenay-sous- Bois, France). Fetal bovine serum (FBS) was purchased from aMeans of concentrations measured before each experiment.

European Journal of Clinical Nutrition Dietary antioxidants and a-tocopherol absorption E Reboul et al 1169 electrode (Millicell ERS; Millipore, Saint-Quentin-en-Yvelines, Table 2 Characteristics and intake of the volunteers 5 days France). before the beginning of the study Item Mean7s.e.m. Experiment design. At the beginning of each experiment, cell monolayers were washed with PBS with1 ml at the apical side Age (years) 30.373.2 7 and 2 ml at basolateral side. Then, the apical side of the cell Height (m) 1.80 0.03 Weight (kg) 72.173.4 monolayers received either 1 ml micellar a-tocopherol or Glucose (mmol/l) 4.6470.14 1 ml micellar a-tocopherol plus other micellar microconsti- Triacylglycerols (mmol/l) 0.7670.04 tuents, whereas the basolateral side received 2 ml FBS-free Cholesterol (mmol/l) 4.6370.27 7 medium. Cell monolayers were incubated at 371C for Energy (kJ/day) 9345 652 Carbohydrate, % of energy 53.072.8 30 min. After the incubation period, media from each side Protein, % of energy 15.970.9 of the monolayer were harvested. Cell monolayers were Fat, % of energy 31.172.3 washed twice with 0.5 ml PBS containing 10 mmol/l tauro- Vitamin E (mg/day) 8.672.1 cholate to eliminate potentially adsorbed fat-soluble micro- Nutrient intake was assessed by a 5-day food diary analyzed using Score-AN constituents, then scraped and collected in 500 ml PBS. software (Avantage Nutrition, Marseille, France). Absorbed a-tocopherol was estimated as a-tocopherol in the scraped cells plus a-tocopherol in the basolateral chambers, if any. perform a power analysis to calculate the number of subjects required to observe a significant effect. We therefore used an usual number of eight subjects with whom significant Identification of the antioxidant microconstituents that differences in absorption had been observed significantly affected a-tocopherol absorption previously (Cardinault et al., 2003; Riso et al., 2003; Tesoriere A full factorial design was performed to specify the et al., 2004; Blum et al., 2005; Reboul et al., 2005b). Eight antioxidant classes of microconstituents that significantly non-smoking, healthy male subjects, 20–40 years of age and affected a-tocopherol absorption, and to identify potential with a body mass index under 24 kg/m2 were selected by the interactions between antioxidant classes. The design was Clinical Investigation Center (Sainte Marguerite, Marseille). constructed using Trial Run 1.0 (SPSS Inc., Chicago, IL, USA) Serum glucose, triacylglycerols and cholesterol were deter- software. It comprised two levels for each factor, that is, with mined by enzymatic procedure (Trinder, 1969; Fossati and or without antioxidant microconstituents. The first factor Prencipe, 1982; Siedel et al., 1983). The subjects were not was carotenoids (we used a mixture of b-carotene, lycopene taking medication and had no history of gastrointestinal and lutein). The second factor was vitamin C. The third disease or disorders of metabolism. Subject character- factor was vitamin E (we used (R,R,R)-g-tocopherol, the istics and their nutrient daily intakes are reported in Table 2. other main vitamin E species found in the western diet). The The daily intake was assessed by a 5-day food diary and was fourth factor was polyphenols (we used a mixture of gallic analyzed using Score-AN software (set by Avantage Nutri- acid, caffeic acid, ( þ )-catechin and naringenin). The 16 tion, Marseille, France). The software nutrient database was experiments generated by the design were performed in obtained from Souci’s nutrition table (Souci et al., 2000). The triplicate, and the full factorial design was performed twice. study was approved by the regional ethics committee in Data analysis was based on the general linear model using Marseille. The objectives and requirements were fully analysis of variance (ANOVA). explained to the participants and informed written consent After having identified which antioxidant classes signifi- was obtained for each subject. Each subject received each cantly affected a-tocopherol absorption as well as any meal in a random order at a 1-month interval (cross-over potential interactions (see Results section), a second series design with a 1-month washout period). On the evening of experiments was performed to identify which antioxidant before the experiment, the subjects were asked to consume microconstituents among the polyphenol and carotenoid a light meal and to fast overnight. In the morning, an mixtures were responsible for these effects. intravenous catheter was inserted into one forearm of each All the samples obtained in these experiments were stored subject. A first blood sample was obtained at fasting (baseline at À801C under nitrogen with 0.5% pyrogallol as a sample), then the volunteers consumed either meal #1 (reducing agent) before extraction and HPLC (containing 24 mg a-tocopherol plus 18 mg lutein) or meal analysis (see below). #2 (containing 24 mg a-tocopherol plus 36 mg lutein). The amount of a-tocopherol in the test meal was chosen to accurately detect a-tocopherol in the chylomicrons. This Postprandial experiment to assess the effect of a microconstituent amount remains nutritional, as it is only twice the adult antioxidant (lutein) on a-tocopherol bioavailability in healthy US-RDA for vitamin E. The lower dose of lutein (18 mg) was subjects selected to be close to the daily total carotenoid intake, As there were no available literature data on the effect of which is estimated at 14 mg/day in Europe (ONeill et al., lutein on a-tocopherol absorption, we were unable to 2001). The higher dose was set at 36 mg to match intake

European Journal of Clinical Nutrition Dietary antioxidants and a-tocopherol absorption E Reboul et al 1170 Table 3 Test-meal composition at 325 nm after light emission at 292 nm, and identified by retention time compared with pure (495%) standards. Meal #1 Meal #2 Quantification was performed using Chromeleon software K 35 g egg white (version 6.50 SP4 Build 1000, Dionex, Aix-en-Provence, K 60 g semolina/180 ml hot water France) comparing peak area with standard reference curves. K 40 g sunflower oila All solvents used were HPLC grade from SDS (Peypin, K 50 g white bread France). K 125 g whole-milk sugared yoghurt (Yoplait)

Macronutrient content: K Energy: 3710.4 kJ Calculations and statistical analysis K Proteins: 21.6 g Results are expressed as means7s.e.m. Calculation of the K Carbohydrates: 98.5 g K : 45.1 g postprandial chylomicron concentrations was performed by subtracting the baseline concentration value measured at the Added antioxidant model: Added antioxidant model: zero time. If this value was negative, it was considered as zero. Elaboration and data analysis of the full factorial design K Luteinb:18mg K Luteinb:36mg designed to identify which antioxidant classes affect aThis amount of sunflower oil provided around 24 mg a-tocopherol and 1 mg a-tocopherol absorption were performed using Trial Run g-tocopherol (Souci et al., 2000). 1.0. Results obtained in the experiment designed to assess b Free lutein from marigold was purchased from Holland & Barrett (Nuneaton, the effect of different polyphenols on a-tocopherol absorp- Warwickshire, UK). tion were analyzed by the Kruskall–Wallis test followed by the Mann–Whitney U-test, used as a post hoc test when the when a standard diet is supplemented with lutein supple- first test showed significant differences between groups. For ments (three pills of 6 mg). Foods were purchased from a the clinical study, differences between the two groups of local supermarket. Lutein pills were purchased from Holland paired data (human subject data, n ¼ 8) were tested using the and Barrett (Nuneaton, Warwickshire, UK). Meal composi- non-parametric Wilcoxon test. Statistical analyses were tions are given in Table 3. Additional blood samples were performed using Statview software, version 5.0 (SAS Insti- drawn 1, 2, 3, 4, 6 and 8 h after the beginning of the meal. tute, Cary, NC, USA). Differences with Po0.05 were Blood samples were stored at 41C and rapidly centrifuged considered significant. (at 610 g for 10 min at 41C) to isolate plasma fractions. Chylomicrons were isolated according to Luchoomun and Hussain (Luchoomun and Hussain, 1999). Chylomicron Results triacylglycerols were measured using the PAP 150 Biome´rieux kit (Charbonnie`re-les-Bains, France). Chylomicron a-toco- Effect of antioxidant microconstituents on a-tocopherol absorption pherol was measured as described above. by Caco-2 cells Identification of the antioxidant classes that affect a-tocopherol absorption. Data analysis of the results obtained in the Tocopherol analysis factorial design showed that the mixture of polyphenols Tocopherol was extracted from 500 ml aqueous samples using (gallic acid, caffeic acid, ( þ )-catechin and naringenin) the following method. Distilled water was added to sample (Po0.001) and the mixture of carotenoids (lutein, lycopene volumes below 500 ml to reach a final volume of 500 ml. Tocol and b-carotene) (Po0.0001), significantly impaired a-toco- used as internal standard was added to the samples in pherol absorption (À10 to À30%). The inhibitory effect of 500 ml ethanol. The mixture was extracted once with 1 vol of g-tocopherol was close to significance (P ¼ 0.055), In con- hexane. The hexane phase obtained after centrifugation trast, vitamin C had no significant effect (Table 4). No (500 g for 10 min at 41C) was evaporated to dryness under interaction was observed between the different classes of nitrogen, and the dried extract was dissolved in 100 ml antioxidants in terms of a-tocopherol absorption, although methanol. A volume of 5–20 ml of the methanol mixture was the effect of the combination of all antioxidants was close to used for HPLC analysis. significance (P ¼ 0.052). a-Tocopherol, g-tocopherol and tocol were separated using

a 250 Â 4.6 nm RP C18,5mm Zorbax column (Interchim, Effects of individual carotenoids. Each carotenoid at dietary Montluc¸on, France) and a guard column. The mobile phase concentrations (0.14 mM for b-carotene, 0.46 mM for lycopene was 100% methanol. Flow rate was 1.5 ml/min, and the and 0.55 mM for lutein) decreased the absorption of a- column was kept at a constant temperature (301C). The tocopherol when co-incubated for 30 min (data not shown). HPLC system consisted in a Dionex separation module (P680 Lycopene exhibited the most pronounced effect (À29.8%, HPLC Pump and ASI-100 Automated Sample Injector, Po0.05), lutein showed an intermediate effect (À13.8%, Dionex, Aix-en-Provence, France) and a Jasco fluorimetric P40.05), and b-carotene had the lowest effect (À3.7%, detector (Jasco, Nantes, France). Compounds were detected P40.05).

European Journal of Clinical Nutrition Dietary antioxidants and a-tocopherol absorption E Reboul et al 1171 0.5 AUC Table 4 Results of statistical analysis of data obtained from the factorial 2 design on the effect of antioxidants on a-tocopherol absorption by Caco-2 cells 0.4 1

Microconstituent Probability value 0.3 mmol.h/L 0 Meal #1 Meal #2

CAR o0.0001 0.2

POLYP 0.001 chylomicron ∆ GAMMA 0.055 0.1

VITC 0.158 triacylglycerols (mmol/L) CAR*POLYP 0.287 0 0 2468 CAR*GAMMA 0.121 CAR*VITC 0.787 Time (h) POLYP*GAMMA 0.455 POLYP*VITC 0.830 Figure 2 Chylomicron triacylglycerol concentrations in healthy & a GAMMA*VITC 0.187 males after meal no. 1 ( ,24mg -tocopherol þ 18 mg lutein) and ’ a D CAR*POLYP*GAMMA 0.994 meal no. 2 ( ,24mg -tocopherol þ 36 mg lutein). Data ( from fasting values) are expressed as means7s.e.m., n ¼ 8. Insets: CAR*POLYP*VITC 0.977 7 CAR*GAMMA*VITC 0.744 mean s.e.m. of the chylomicron triacylglycerol responses (AUC) POLYP*GAMMA*VITC 0.905 after meal no. 1 and meal no. 2. CAR*POLYP*GAMMA*VITC 0.052

Abbreviations: CAR: mixture of carotenoids (b-carotene, lycopene and lutein), POLYP: mixture of polyphenols (gallic acid, caffeic acid, ( þ )-catechin and a AUC 400 1500 g naringenin), GAMMA: -tocopherol, VITC: vitamin C. *Indicates studied 1000 interactions. A probability value o0.05 states that the antioxidant(s) or the 500

300 nmol.h/L combination of antioxidants significantly affected a-tocopherol absorption. 0 Meal #1 Meal #2 200 140

120 chylomicron 100 ∆ -tocopherol nmol/L

100 α 9 0 80 * 7 -tocopherol 1 0 2468 α 8 60 * 4 5 6 10 Time (h) 40 2 (% of control) AUC b 20 80 20 3 Absorbed 40 0

nmol.h/L 0 Figure 1 Effect of polyphenols on a-tocopherol absorption by Meal #1 Meal #2 differentiated Caco-2 TC7 monolayers. The apical sides of the cell 10 monolayers were rinsed with PBS and then received FBS-free medium containing either a-tocopherol-rich mixed micelles at 5.5 mM alone (1) chylomicron ∆

or 5.5 mM a-tocopherol-rich micelles plus (2) 25 mM naringenin, (3) -tocopherol nmol/L 150 mM naringenin, (4) 25 mM ( þ )-catechin, (5) 150 mM ( þ )-catechin,  0 (6) 50 mM gallic acid, (7) 150 mM gallic acid, (8) 50 mM caffeic acid, (9) 246 mM mM 150 caffeic acid and (10) a mixture of polyphenols (25 Time (h) naringenin plus 25 mM ( þ )-catechin plus 50 mM gallic acid plus 50 mM caffeic acid). The basolateral sides received complete medium. 7 Figure 3 Chylomicron vitamin E concentrations in healthy males Incubation time was 30 min. Data are means s.e.m. of three assays. after meal no. 1 (J,24mga-tocopherol þ 1mg g-tocopherol þ An asterisk indicates a significant difference (Po0.05)withthecontrol 18 mg lutein) and meal no. 2 (K,24mga-tocopherol þ 1mg g- (assay performed without polyphenol). tocopherol þ 36 mg lutein). (a) Chylomicron a-tocopherol concen- trations, (b) chylomicron g-tocopherol concentrations. Data 7 Effects of individual polyphenols. This experiment showed (D from fasting values) are expressed as means s.e.m., n ¼ 8. Insets: mean7s.e.m. of the chylomicron vitamin E responses (AUC) that naringenin was the only polyphenol able to signifi- after meal no. 1 and meal no. 2. cantly impair a-tocopherol absorption (about 25% at 25 mM and 50% at 150 mM naringenin, Po0.05, Figure 1). The mixture of polyphenols, which contained 25 mM naringenin, had a similar effect to 25 mM naringenin alone. prandial 0 to 8 h curves) following the ingestion of meals containing 18 or 36 mg lutein (Figure 2). In contrast, both a- and g-tocopherol responses were lower after the meal #2 Effect of lutein on postprandial chylomicron triacylglycerol and (containing 36 mg lutein) than after the meal #1 (containing vitamin E responses 18 mg lutein): 6167280 vs 10017287 nmol h/l, P ¼ 0.0687 There was no significant difference between chylomicron and 48714 vs 54712 nmol h/l P ¼ 0.4008 for a- and triacylglycerol responses (expressed as area under the post- g-tocopherol, respectively (Figure 3a and b).

European Journal of Clinical Nutrition Dietary antioxidants and a-tocopherol absorption E Reboul et al 1172 Discussion second series of experiments. Results clearly showed that lycopene inhibited a-tocopherol absorption. The effect of The present study was carried out to evaluate the effects lutein was lower and the few effect of b-carotene can be of the main dietary classes of antioxidant molecules explained by its low micellar concentration. This inhibitory (carotenoids, polyphenols, vitamin C and vitamin E) on a- effect of carotenoids on a-tocopherol absorption, which has tocopherol absorption. As a clinical study to assess the effect never been described before can be explained by competition of all combinations of antioxidants on a-tocopherol absorp- between a-tocopherol and other carotenoids for uptake tion was unfeasable, we opted for an approach using the through the SR-BI receptor, which facilitates apical uptake human intestinal Caco-2 cell model. This model was selected of both classes of antioxidants (Reboul et al., 2005a, c). because it has been successfully used to study the intestinal Concerning polyphenols, the second set of experiments absorption of vitamin E (Traber et al., 1990; Kiyose et al., showed that only naringenin impaired a-tocopherol absorp- 1995) and it expresses Scavenger receptor class B type I tion. The specific effect exhibited by naringenin needs to be (SR-BI), which is involved in a-tocopherol uptake (Reboul elucidated by additional experiments, but it is striking that et al., 2005c). The TC7 clone was used because it is much naringenin is the most lipophilic of all the polyphenols more homogenous than the parent Caco-2 cell line (Gres tested (log P ¼ 2.52 vs 0.86, 0.82 and 0.38 for gallic acid, et al., 1998). As it was not possible to assess the effect of every caffeic acid and ( þ )-catechin, respectively (Cooper et al., individual antioxidant species (thousands of carotenoid 1997)). Thus, a first hypothesis is that naringenin affects and polyphenol species are recovered in human diets), we a-tocopherol absorption through interaction with SR-BI, assessed the effect on a-tocopherol absorption of the main which is known to transport lipophilic molecules with low dietary carotenoids (b-carotene, lycopene and lutein), the substrate specificity. A second hypothesis might be related to main dietary polyphenols (flavonoids and non-flavonoids, an interaction between naringenin and membrane lipids mainly phenolic acids), vitamin C and g-tocopherol (which (Tachibana et al., 2004), which may affect the invagination of is the other main vitamin E species found in Western diets). lipid raft domains containing putative vitamin E receptors. We chose the relative proportions of these antioxidants that Once we had determined which classes of antioxidants best mimicked those observed in a standard diet (Table 5). affect a-tocopherol absorption, we aimed to validate the Analysis of the data from the factorial design performed on results obtained in the Caco-2 model by performing a Caco-2 showed that carotenoids, polyphenols and probably clinical study. Lutein was chosen as the antioxidant to (P ¼ 0.055) g-tocopherol significantly impaired a-tocopherol compete with a-tocopherol for absorption because both absorption. Vitamin C had no significant effect. The fact that vitamin E and lutein supplements are becoming increasingly g-tocopherol seemed to impair a-tocopherol absorption is in popular owing to their potential ability to prevent age- good agreement with a recent study (Reboul et al., 2005c) related (Belda et al., 1999; Delcourt and could be explained by competition between the two et al., 1999; Mares-Perlman et al., 2001; Olmedilla et al., 2003; vitamin E species for SR-BI or other putative vitamin E Richer et al., 2004). Data from the clinical study suggest transporters. This phenomenon could explain the competi- that increasing dietary lutein from 18 to 36 mg reduces tion between these species with regard to their bioavail- a-tocopherol absorption (area under the curve (AUC) down ability (Huang and Appel, 2003; Yoshikawa et al., 2005). by 38%). This effect was not, however, statistically signifi- Concerning carotenoids and polyphenols, the mixtures used cant, likely because of the high interindividual variability in in the factorial design were unable to identify which species chylomicron a-tocopherol responses. In fact, on the basis of were responsible for this effect. We therefore performed a the data obtained, that is, standard deviations and between- group differences, we calculated that 25 subjects would be required to establish a significant decrease in chylomicron Table 5 Relative proportions of antioxidant microconstituents in the a-tocopherol response with a power efficiency of 80%. western diet and in the Caco-2 cell experiments performed in this study In conclusion, the results obtained suggest that dietary RDA or daily RR Concentrations in RR antioxidants behave differently with regard to a-tocopherol intake (mg) the Caco-2 cells absorption. Vitamin C and phenolic acids apparently have no experiments (mM) significant effect on a-tocopherol absorption, at least in the a-Tocopherol 12a 1 4.4 1 relative proportions found in standard diets, whereas carote- g-Tocopherol 14b 1.2 3.0 0.7 noids and certain flavonoids (naringenin and possibly other c Carotenoids 14 1.2 1.1 0.3 flavanons) can impair it. These results should be taken into a Vitamin C 90 7.5 75 17.0 account when antioxidant supplements are regularly eaten. Polyphenols 1000d 83.3 150 34.1

Abbreviations: RR: relative ratio to a-tocopherol. aUS-RDA: US recommended dietary allowances for male adults. Acknowledgements bEstimated intake from a study on 482 subjects (El-Sohemy et al., 2001). cMedian total carotenoid intake in five European countries (ONeill et al., 2001). This work was supported by the INSERM and the INRA (ATC dEstimated intake (Scalbert and Williamson, 2000). Nutrition 2002, project #A02256AS).

European Journal of Clinical Nutrition Dietary antioxidants and a-tocopherol absorption E Reboul et al 1173 References Mortensen A, Skibsted LH, Truscott TG (2001). The interaction of dietary carotenoids with species. Arch Biochem Biophys 385, Belda JI, Roma J, Vilela C, Puertas FJ, Diaz-Llopis M, Bosch-Morell F 13–19. et al. (1999). Serum vitamin E levels negatively correlate with Olmedilla B, Granado F, Blanco I, Vaquero M (2003). Lutein, but not severity of age- related macular degeneration. Mech Ageing Dev alpha-tocopherol, supplementation improves visual function in 107, 159–164. patients with age-related : a 2-y double-blind, placebo- Beyer RE (1994). The role of ascorbate in antioxidant protection of controlled pilot study. Nutrition 19, 21–24. biomembranes: interaction with vitamin E and coenzyme Q. ONeill ME, Carroll Y, Corridan B, Olmedilla B, Granado F, Blanco I J Bioenerg Biomembr 26, 349–358. et al. (2001). A European carotenoid database to assess carotenoid Blum S, Aviram M, Ben-Amotz A, Levy Y (2005). Effect of a intakes and its use in a five-country comparative study. Br J Nutr mediterranean meal on postprandial carotenoids, paraoxonase 85, 499–507. activity and C-Reactive protein levels. Ann Nutr Metab 50, 20–24. Reboul E, Abou L, Mikail C, Ghiringhelli O, Andre M, Portugal H Borel P (2003). Factors affecting intestinal absorption of highly et al. (2005a). Lutein transport by Caco-2 TC-7 cells occurs partly lipophilic food microconstituents (fat-soluble , carote- by a facilitated process involving the scavenger receptor class B noids and ). Clin Chem Lab Med 41, 979–994. type I (SR-BI). Biochem J 387, 455–461. Borel P, Pasquier B, Armand M, Tyssandier V, Grolier P, Alexandre- Reboul E, Borel P, Mikail C, Abou L, Charbonnier M, Caris-Veyrat C Gouabau MC et al. (2001). Processing of and E in the et al. (2005b). Enrichment of tomato paste with 6% tomato peel human gastrointestinal tract. Am J Physiol 280, G95–G103. increases lycopene and beta-carotene bioavailability in men. J Nutr Cardinault N, Tyssandier V, Grolier P, Winklhofer-Roob BM, Ribalta J, 135, 790–794. Bouteloup-Demange C et al. (2003). Comparison of the post- Reboul E, Klein A, Bietrix F, Gleize B, Malezet-Desmoulins C, prandial chylomicron carotenoid responses in young and older Schneider M et al. (2005c). Scavenger receptor class B type I (SR- subjects. Eur J Nutr 42, 315–323. BI) is involved in vitamin E transport across the enterocyte. J Biol Cooper DA, Webb DR, Peters JC (1997). Evaluation of the potential Chem 281, 4739–4745. for olestra to affect the availability of dietary . Richer S, Stiles W, Statkute L, Pulido J, Frankowski J, Rudy D et al. J Nutr 127, S1699–S1709. (2004). Double-masked, placebo-controlled, randomized trial of Delcourt C, Cristol JP, Tessier F, Leger CL, Descomps B, Papoz L lutein and antioxidant supplementation in the intervention of (1999). Age-related macular degeneration and antioxidant status atrophic age-related macular degeneration: the Veterans LAST in the POLA study. POLA Study Group. Pathologies Oculaires Liees study (Lutein Antioxidant Supplementation Trial). Optometry 75, a l’Age. Arch Ophthalmol 117, 1384–1390. 216–230. El-Sohemy A, Baylin A, Ascherio A, Kabagambe E, Spiegelman D, Riso P, Brusamolino A, Ciappellano S, Porrini M (2003). Comparison Campos H (2001). Population-based study of alpha- and gamma- of lutein bioavailability from vegetables and supplement. Int J tocopherol in plasma and adipose tissue as biomarkers of intake in Vitam Nutr Res 73, 201–205. Costa Rican adults. Am J Clin Nutr 74, 356–363. Salvini S, Charbonnier M, Defoort C, Alquier C, Lairon D (2002). Fossati P, Prencipe L (1982). Serum triglycerides determined color- Functional characterization of three clones of the human imetrically with an enzyme that produces peroxide. Clin intestinal Caco-2 cell line for dietary lipid processing. Br J Nutr Chem 28, 2077–2080. 87, 211–217. Gres MC, Julian B, Bourrie M, Meunier V, Roques C, Berger M et al. Scalbert A, Williamson G (2000). Dietary intake and bioavailability of (1998). Correlation between oral drug absorption in humans, and polyphenols. J Nutr 130, 2073S–2085S. apparent drug permeability in TC-7 cells, a human epithelial Sen G, Mukhopadhaya R, Ghosal J, Biswas T (2000). Interaction of intestinal cell line: comparison with the parental Caco-2 cell line. ascorbate and alpha-tocopherol enhances antioxidant reserve of Pharm Res 15, 726–733. erythrocytes during in visceral leishmaniasis. Life Sci 67, Halliwell B, Zhao K, Whiteman M (2000). The gastrointestinal tract: a 3181–3190. major site of antioxidant action? Free Radic Res 33, 819–830. Siedel J, Hagele EO, Ziegenhorn J, Wahlefeld AW (1983). Reagent for Huang HY, Appel LJ (2003). Supplementation of diets with alpha- the enzymatic determination of serum total cholesterol with tocopherol reduces serum concentrations of gamma- and delta- improved lipolytic efficiency. Clin Chem 29, 1075–1080. tocopherol in humans. J Nutr 133, 3137–3140. Souci SW, Fachmann W, Kraut H (2000). Food Composition and Kamal-Eldin A, Frank J, Razdan A, Tengblad S, Basu S, Vessby B Nutrition Tables. 6th revised and completed edn. Medpharm (2000). Effects of dietary phenolic compounds on tocopherol, GmbH Scientific publishers. CRC Press: Boca Raton, London, cholesterol, and fatty acids in rats. Lipids 35, 427–435. New York, Washington, DC. Kamal-Eldin A, Pettersson D, Appelqvist LA (1995). Sesamin (a Tachibana H, Fujimura Y, Yamada K (2004). Tea polyphenol compound from sesame oil) increases tocopherol levels in rats fed epigallocatechin-3-gallate associates with plasma membrane lipid ad libitum. Lipids 30, 499–505. rafts: Lipid rafts mediate anti-allergic action of the catechin. Kiyose C, Muramatsu R, Fujiyama-Fujiwara Y, Ueda T, Igarashi O Biofactors 21, 383–385. (1995). Biodiscrimination of alpha-tocopherol stereoisomers dur- Tesoriere L, Allegra M, Butera D, Livrea MA (2004). Absorption, ing intestinal absorption. Lipids 30, 1015–1018. excretion, and distribution of dietary antioxidant in Loest HB, Noh SK, Koo SI (2002). Green tea extract inhibits the LDLs: potential health effects of betalains in humans. Am J Clin lymphatic absorption of cholesterol and alpha-tocopherol in Nutr 80, 941–945. ovariectomized rats. J Nutr 132, 1282–1288. Traber MG, Goldberg I, Davidson E, Lagmay N, Kayden HJ (1990). Luchoomun J, Hussain MM (1999). Assembly and secretion of Vitamin E uptake by human intestinal cells during lipolysis in chylomicrons by differentiated Caco-2 cells. Nascent triglycerides vitro. Gastroenterology 98, 96–103. and preformed phospholipids are preferentially used for lipopro- Trinder P (1969). Determination of blood glucose using 4-amino tein assembly. J Biol Chem 274, 19565–19572. phenazone as oxygen acceptor. J Clin Pathol 22, 246. Mares-Perlman JA, Fisher AI, Klein R, Block G, Millen AE, Wright JD Yoshikawa S, Morinobu T, Hamamura K, Hirahara F, Iwamoto T, (2001). Lutein and zeaxanthin in the diet and serum and their Tamai H (2005). The effect of gamma-tocopherol administration relation to age-related maculopathy in the Third National Health on alpha-tocopherol levels and metabolism in humans. Eur J Clin and Nutrition Examination Survey. Am J Epidemiol 153, 424–432. Nutr 59, 900–905.

European Journal of Clinical Nutrition