Supplementation with Wine Phenolic Compounds Increases The

Home , Caftaric acid

Supplementation with wine phenolic compounds increases the antioxidant capacity of plasma and vitamin E of low-density lipoprotein without changing the lipoprotein Cu2-oxidizability: Possible explanation by phenolic location

M-A Carbonneau1,CLLeÂger1, L Monnier2, C Bonnet1, F Michel1, G Fouret1, F Dedieu3 and B Descomps1

1Laboratoire de Biologie et Biochimie des Lipides, Institut de Biologie, UFR de MeÂdecine, Boulevard Henri IV, 34060 Montpellier, France; 2Laboratoire des Maladies MeÂtaboliques, HoÃpital Lapeyronie, 371 Avenue Doyen Gaston Giraud, 34295 Montpellier Cedex 9, France; and 3Centre d'ênologie, UFR de Pharmacie, 35 Chemin des Maraichers, 31000 Toulouse, France

Objectives: To evaluate the effect of the red wine phenolic compound (RWPC) dietary supplementation without alcohol interference on: (1) some of the biochemical characteristics of LDL, (2) the oxidative susceptibility of LDL and (3) the antioxidant capacity of total plasma (Pl-AOC). In order to account for discrepancies between the three series of data, the in vitro stability of the association of phenolic compounds and LDL was tested. Design: An intervention study with 20 volunteers. Each served as his own control. Cu2-oxidizability of LDL and Pl-AOC were tested on blood samples before and after dietary supplementation. Cu2-oxidizability of LDL was also tested by co-incubation in the presence of RWPC or phenolic acids with or without extensive dialysis. Setting: The Laboratory of Lipid Biochemistry and Biology, School of Medicine, and the Laboratory of Metabolic Diseases, Lapeyronie Hospital, University of Montpellier, France. Subjects: Healthy males, nonsmokers and moderate drinkers, submitted to a dietary regimen deprived of vitamin E and C for a period of 10 d before supplementation. They also abstained from alcohol, wine, fruit juices, coffee, tea and cola beverages during this period. Intervention: Six 0.33 g capsules=d (namely two capsules at each meal) of a preparation of red wine phenolic compounds in a dry powder form were given to the volunteers over a period of two weeks. Blood samples were drawn in fasting conditions at day 0 and day 14 of the supplementation period. Results: Supplementation led to: (1) in LDL, a signi®cant increase in vitamin E content (n 20, P 0.01) or vitamin E=total fatty acid bis-allylic carbon number ratio (n 20, P 0.006) without modi®cation in the other biochemical characteristics or Cu2-oxidizability; (2) in plasma, a signi®cant increase in the antioxidant capacity (n 11, P 0.01). In vitro studies showed that RWPC or sinapic, caffeic or ferulic acids incubated in the presence of LDL increased the protection of the lipoparticle against oxidation (caffeic > sinapic > ferulic). This effect, however, was totally lost after extensive dialysis. Conclusions: The enhancing effect of the RWPC supplementation on Pl-AOC may be due to a phenolic- compound action both in the aqueous phase of plasma and at the surface of lipoprotein particles. Surface location possibly explains the enhancing-sparing effect of supplementation on LDL vitamin E and the absence of effect on dialysed-LDL oxidizability. Sponsorship: Supported by a grant from the Scienti®c Committee `Vin et SanteÂ: Pathologies Vasculaires'. Descriptors: phenolic compounds; red wine; low-density lipoprotein; oxidizability; plasma antioxidant capacity; vitamin E; dietary supplementation

Introduction cells resulting in the release of several factors and proteins which promote both monocyte adhesion to endothelium It is commonly acknowledged that oxidized low-density (Quinn et al, 1987; Rajavashisth et al, 1990) and monocyte- lipoprotein (ox-LDL) plays an important role in atherogen- derived macrophage transformation into foam cells (Agel et esis (Steinbrecher et al, 1990). This is supported particu- al, 1984) and (3) the stimulation of the superoxide anion larly by the presence of high quantities of ox-LDL at the production by cells present at the lesion site, this produc- site of the atherosclerotic lesion (YlaÈ-Herttuala et al, 1989), tion being probably implicated in the LDL oxidative by the role of ox-LDL in the formation of foam cells and modi®cation (Li & Cathcart, 1994). Taken together, these the so-called fatty streak (Ross, 1993). An array of addi- events could be able to initiate, maintain or amplify the tional actions of ox-LDL should also be mentioned, espe- atherogenic process. cially: (1) the increase in LDL endothelium permeability Dry extracts of red wine contain phenolic compounds (Steinberg et al, 1989); (2) the activation of endothelium that have been shown in vitro to be strong reactive oxygen species scavengers (Bauman et al, 1980; Torel et al, 1986; Roback & Griglewski, 1988). Red wine-extracted ¯avo- Correspondence: Dr CL LeÂger. noids protect LDL against oxidation in vitro (De Walley et Received 16 August 1996; revised 9 June 1997; accepted 13 June 1997 al, 1990; NeÁgre-Salvayre & Salvayre, 1992; Frankel et al, Supplementation with wine phenolic compounds M-A Carbonneau et al 683 1993; Frankel et al, 1995). Recent studies suggest that Subjects people who are fed a rather well de®ned diet usually called Twenty men (20±45 y old) took part in this study. They the `Mediterranean diet' (Willett et al, 1995) statistically were normo-cholesterolemic (3.04±6.64 mmol=L), moder- show low incidence of coronary heart disease (CHD) ate drinkers and nonsmokers, and none was receiving any (Renaud & De Lorgeril, 1992), with an epidemiological pharmacological treatment. For each subject the sequential link between ¯avonoid intake and decreasing risk of CHD experiment was as follows: (1) a deprivation period of ten (Hertog et al, 1993). It is postulated that French people are days (day 7 10 to day 7 1) during which vitamin E and protected against CHD risk by the larger amounts of wine C, alcohol, wine, fruit juices, coffee, tea and cola were not (particularly red wine) they drink as compared to North allowed to be consumed; (2) a supplementation period of European or American people. Moreover, decreasing car- fourteen days (day 0 to day 13) during which the dietary diovascular risks occur from the North to the South of restrictions were maintained and the subjects received six France (Ruidavets et al, 1993). Altogether, these results capsules daily (two at each meal) corresponding to 1.98 g seem to indicate the anti-atherogenic role of the wine of RWPC dry powder. It can be assumed from the above phenolic compounds by preventing ox-LDL formation. chemical data that 1.98 g of the dry powder contained Whether the putative cardiovascular protective effect of 222 mg of catechins proanthocyanidins which is usually red wine is brought about by an in vivo action of the supplied by 1.17 L of the red wine chosen for this experi- RWPC, by the alcoholic component, or by a synergistic ment. Thus, the supplementation was equivalent to the action of both is currently open to question. consumption of approximately 1 L of wine daily on the A bene®cial effect of a moderate red wine consumption basis of the catechin proanthocyanidin fraction. The has been attributed to ethanol (Rimm et al, 1996) regarding equivalence of the weight concentration of RWPC with CHD prevention in humans, possibly due to an enhancing the concentration of a standard gallic acid solution was also effect on the HDL-cholesterol fraction (Criqui & Ringel, determined according to Singleton & Rossi (1965). 1994). The aim of the present work was to study the effect A dietary diary in a questionnaire form was to be ®lled of ingestion of a RWPC dry extract, in order to eliminate out by each volunteer; it was checked by dieticians of the possible interferences with alcohol, both on the biochem- Laboratory of Metabolic Diseases in order to examine ical characteristics and the oxidative susceptibility of LDL, whether the regimen was in compliance with the required and on the antioxidant capacity of total plasma, in healthy restrictions. human volunteers. In order to specify why the plasma Blood samples were drawn by venepuncture on 1 g=L antioxidant capacity was found to be increased while EDTA after overnight fasting at day 0 and day 14. As we LDL oxidative susceptibility was not affected, we tested decided to assess the Pl-AOC when this work was in the hypothesis that phenolic compounds can be released progress, we were only able to include 11 of 20 volunteers from LDL by the extensive dialysis usually included in in this part of the study. LDL preparation. The study was approved by the University of Mont- pellier Ethics Committee and subjects gave informed, written consent. Methods LDL oxidation Chemicals Plasma lipoproteins (triglyceride rich lipoprotein, TRL, The RWPC dry powder was prepared from a red wine d < 1.019 g=mL; low density lipoprotein, LDL, Cabernet-Sauvignon grape variety and analysed (Institut d 1.019±1.063 g=mL; high density lipoprotein, HDL, National de la Recherche Agronomique, Narbonne, d 1.063±1.21 g=mL) were isolated by sequential ultra- France). Brie¯y, this preparation involved three steps: (1) centrifugation (Schumaker & Puppione, 1986) in a TFT phenolic adsorption on a preparative column; (2) alcoholic 65±13 rotor (Kontron Instruments SA, Montigny Le Bre- desorption and (3) alcoholic eluent concentration by a tonneux, France) for 4 h per run at 300 000 6 g gentle evaporation. The concentrated residue was then (65 000 rpm) with density adjustments using potassium sprayed to obtain a RWPC dry powder which was given bromide (0.71 mol=L KBr for LDL isolation). LDL was to volunteers in 0.33 g capsule form. One liter of red wine then extensively dialyzed (with at least a 100=1 volume contained 190 mg=L (expressed as catechin equivalent or ratio, repeated three times, at 4C in the dark, for 24 h) CATeq) of monomeric, dimeric and oligomeric ¯avonols against deoxigenated phosphate-buffered saline (PBS: (the `catechin proanthocyanidin' fraction) and produced 10 mmol phosphate buffer=L, 150 mmol NaCl=L, pH 7.4) 1.3 g of dry powder or 1.2 g (as CATeq) of total poly- containing 10 mmol=L diethylenetriamine pentacetic acid phenols. The dry powder contained 112 mg=g of cate- (DTPA) and sterilized through a 0.22 mm ®lter (Millex-GS, chins proanthocyanidins with only 1.0% of each Millipore, Saint Qentin-Yvelines, France). Apoproteins B monomeric form catechin and epicatechin; 64 mg=g of and A1, lipid fractions and electrochemical detection of total anthocyanins with a major part (36%) of malvidin-3- vitamin E were measured as previously described (Cachia glucoside; 5 mg=g of total ¯avonols with 85% of quercetin et al, 1995). After lipid extraction of LDL, fatty acids were and quercetin-3-glucoside; and 8.7 mg=g of total phenolic transesteri®ed and analyzed according to the usual labora- acids with 19.5% of caftaric acid. tory procedure (Sadou et al, 1995). The total fatty acids Sinapic, caffeic and ferulic acids were synthetized by the were also expressed as the sum of bis-allylic carbon group of S. Labidalle (Centre d'ênologie, UFR de Phar- number (Cachia et al, 1995) per LDL particle (as the bis- macie, Toulouse, France). Horseradish peroxidase (HRP allylic carbon number=apoB molar ratio). LDL oxidative EC 1.11.1.7, type VI A) was from Sigma (Saint Quentin susceptibility was investigated separately at day 0 and day Fallavier, France). Luminol (3-amino phthalhydrazide) and 14 just after ultracentrifugation and dialysis of the lipopro- Trolox (6-hydroxy 2,5,7,8-tetramethyl chroman 2-car- tein fraction. To do so, we measured the conjugated dienes boxylic acid) were from Aldrich Chemicals (Strasbourg, formed (234 nm absorbance) at 15±20 min intervals during France). copper chloride (5 mmol Cu2=L)-mediated oxidation of a Supplementation with wine phenolic compounds M-A Carbonneau et al 684 LDL solution (0.1 mmol apoB=L) prepared by dilution in 150 6 4.6 mm (5 mm diameter particles) column (a C18 air-saturated PBS without DTPA, at 37C and under inverse phase, Interchim, MontlucËon, France). A 50-min magnetic stirring. This allowed us to assess the lag time, linear elution gradient was used with the following mixture: the maximal rate and the maximal production of the H2O=acetonitrile=1%-H3PO4, from 20 : 0 : 80 (v=v=v) at the conjugated dienes (Esterbauer et al, 1989), which are beginning to 0 : 30 : 70 (v=v=v) at the end of the elution, at indicative of the earliest step in polyunsaturated fatty acid the constant ¯ow rate of 1 mL=min. A 1040 A UV-visible peroxidation. We always respected the same time period diode array detector (Hewlett-Packard, Evry, France) was from the plasma sample collection to the ®nal LDL- employed, equipped with a DOS Chemstation. The quanti- oxidizability determination. Under these conditions, for tative measurement was performed by integrating the peak ten consecutive oxidation experiments, the variation coef®- area at 320 nm. cients of lag time, maximal rate and maximal production of the conjugated dienes were 5.8, 13.7 and 8.4%, respec- Chemiluminescence assay for plasma antioxidant capacity tively. LDL peroxides were analyzed as previously Pl-AOC was determined by means of a chemiluminescence described (El-Saadani et al, 1989) and thiobarbituric acid reaction (Misra & Squatrito, 1982; Henley & Worwood, reactive substances (TBARS) (Lynch & Frei, 1993) were 1994) using a mixture of luminol, hydrogen peroxide and measured as indicative of malondialdehyde and peroxida- horseradish peroxidase (HRP). The light emission is sup- tion break-down product formation from numerous sub- pressed in the presence of antioxidative substances acting strates including polyunsaturated fatty acids with more than as radical scavengers and restored when antioxidants are two double bounds. consumed (Whitehead et al, 1992). The duration of the Omitting dialysis in the LDL preparation cannot be light suppression Ð indicative of the antioxidant content envisaged because of the presence of Cu2-oxidation- of the medium Ð was measured with a 1251 Luminometer impairing EDTA in the blood collection tubes and the (LKB, Paris, France). Brie¯y, a 1 g=L stock solution of high ionic strength provided by potassium bromide. We luminol in 1 mol=L NaOH was preparated and diluted daily have checked that this led to non measurable lag times in a 0.1 mol=L Tris-HCl (pH 7.4) buffer for obtaining a ( > 10 h). Conversely, the particularly long-lasting lag time 10 mg=L luminol working solution. A 0.1% hydrogen with non-dialyzed LDL cannot be attributed to higher peroxide solution in 10 mmol=L NaOH was daily prepared. vitamin E content of LDL as already reported on the A 0.32 g=L stock solution of HRP was prepared and diluted basis of vitamin E=total protein ratio of LDL (Scheek et daily in 0.9% NaCl for obtaining a 32 mg=L working al, 1995). In our conditions and on the basis of vitamin solution. A 4 mmol=L Trolox and a 0.1 mGAE=L RWPC E=apoB ratio we only found a decrease of 11.3% stock solutions in ethanol were freshly prepared and diluted (11.51 Æ 1.09 vs 10.29 Æ 1.06, n 10; P < 0.05) which is in 0.1 mol Tris±HCl=L (pH 7.4) buffer for obtaining work- of particular interest because such a variation in vitamin E ing solutions. Last, 1=10-diluted plasma solutions in content of LDL particule is known to have a negligible 0.1 mol=L Tris-HCl (pH 7.4) buffer were prepared. Che- incidence on the LDL resistance to oxidation (Esterbauer et miluminescence reaction was initiated in the reaction cuv- al, 1992). ette by addition of the following working solutions: luminol To investigate the participation of LDL-associated phe- (200 mL), HRP (200 mL), to which 0.1 mol Tris-HCl=L nolic compounds in lipoprotein antioxidant protection in (pH 7.4) buffer (180 mL) and either diluted plasma, Trolox vitro, LDL (®nal concentration: 1 mmol apoB=L) was or RWPC working solution (20 mL) were ®nally added. preincubated in 1 mL of PBS at 37 C under N2 with The cuvette was placed in the luminometer and 500 mL of RWPC (®nal concentrations: 0±20 mg=L) or sinapic, caf- the solution of hydrogen peroxide was then added. The Pl- feic or ferulic acid (®nal concentrations: 0±100 mmol=L) AOC was calculated by determining the time needed to previously dissolved in ethanol=deionized water (10 : 90, reach 50% of the maximal light emission in the presence of v=v) for 1 h. LDL solution was then adjusted to 0.1 mmol plasma (tpl) and this time was compared to the Trolox apoB=L with PBS and submitted to copper-mediated oxi- concentration or the RWPC concentration (expressed as m dation without or after usual extensive dialysis. Residual gallic acid equivalent per L or mGAE=L) needed to obtain concentration of DTPA (0.5 mmol=L) was veri®ed to be the same time. This was easily performed by plotting time without effect on LDL oxidation process. ttr or tGAE obtained with increasing concentrations of Trolox or RWPC, respectively. The curve obtained was Sinapic acid distribution into lipoprotein then used to determine the Trolox or RWPC concentration and non-lipoprotein fractions corresponding to tpl. It had previously been veri®ed that Plasma was incubated under N2 in the presence of increas- ethanol amounts used in the medium had no in¯uence on ing concentrations (0±1 mg=L) of sinapic acid at 37C for the light suppression time. 1 h. TRL, LDL, HDL and `very high density lipoprotein' Several light suppression time determinations were (VHDL) non lipoprotein (nonLP) fraction (d > 1.21 performed using the same plasma sample provided by g=mL) were then isolated by sequential ultracentrifugation one single volunteer in order to check the reproducability and the sinapic acid concentration was determined in each of the luminol±HRP±H2O2 luminescence method. The lipoprotein fraction by HPLC procedure, before and after variation coef®cient was found to be 10.1% for n 20. usual extensive dialysis. This means that the probability to ®nd a 15% difference to be signi®cant for a risk of ®rst species a 0.05 was 80% HPLC assay for sinapic acid for n 10. Lipoprotein fraction (1 mL) was diluted with PBS contain- Plasma samples submitted to chemiluminescence assay ing 10 mmol DTPA=L (1 : 1, v=v) and extracted with had been stored at 7 20C for 22±40 weeks before use, 4.2 mL of a mixture containing diisopropylether=3N HCl because sample collection was not performed at the same (20 : 1, v=v). After evaporation, the dry residue was dis- time. It has been veri®ed that there was no correlation solved in 2 mL of an ethanol=deionized water mixture between the decrease in Pl-AOC and the time of conserva- (10 : 90, v=v) and 20 mL was submitted to HPLC using an tion (r20.04, regression line with a regression coef®- Supplementation with wine phenolic compounds M-A Carbonneau et al 685 cient1.2 mmol=L=week for both day 0 and day 14 plasma; supplementation, P 0.02). It is of interest to notice that this value is largely lower than the variation obtained there was no signi®cant modi®cation in the bis-allylic between day 0 and day 14 ascribed to the volunteers' carbon number per LDL particle, which accounts for the consumption of phenolic compounds). carbon atoms being highly susceptible to the proton abstraction initiating the polyunsaturated fatty acid perox- Statistics idation, whereas the vitamin E=bis-allylic carbon number Results (means Æ s.e.m.) were statistically analysed by ratio was signi®cantly higher after supplementation using a Wilcoxon non parametric test (StatViewtm, Alsyd, (3.89 Æ 0.38 vs 3.07 Æ 0.25 mmol=mol, P 0.006). Meylan, France). Changes in LDL oxidative susceptibility due to supplementation Results The time-course of conjugated diene formation was similar Changes in biochemical characteristics of LDL due to to that already observed (Esterbauer et al, 1989). The lag supplementation time, maximal rate and maximal amount of conjugated Biochemical characteristics of plasma at day 14 compared diene formation were drawn from the curve and reported in to day 0 (total cholesterol: 4.23 Æ 0.28 vs 4.04 Æ Table 2 together with peroxide and TBARS productions. It 0.18 mmol=L; triglycerides: 0.66 Æ 0.06 vs 0.67 Æ 0.07 appears that treatment changed only the conjugated diene mmol=L; apoB: 0.74 Æ 0.03 vs 0.78 Æ 0.04 g=L; apoA1: production (from 444.6 Æ 17.1 at day 0 to 489.5 Æ 16.7 at 1.04 Æ 0.04 vs 1.06 Æ 0.04 g=L) and those of HDL (total day 14, P 0.01). However, when corrected for bis-allylic cholesterol: 94.5 Æ 9.2 vs 99.2 Æ 9.8; triglycerides: 10.6 Æ carbon number, the conjugated diene production was no 2.2 vs 10.6 Æ 2.6; phospholipids: 60.6 Æ 4.7 vs 71.4 Æ 7.6; longer signi®cantly higher after supplementation vitamin E: 0.42 Æ 0.03 vs 0.42 Æ 0.03, all results expressed (0.19 Æ 0.02 vs 0.16 Æ 0.01 mol=mol, P 0.1). There was as mol=mol of apoA1) were not affected by the RWPC no signi®cant change in the lag time (65.4 Æ 4.1 and supplementation. 61.4 Æ 2.9, P 0.08). No correlation was observed between As shown in Table 1, the increase in vitamin E content lag time and LDL vitamin E content (r 0.25 and 0.27 was the unique statistical modi®cation (P 0.01) in LDL before and after supplementation, respectively). brought about by the RWPC supplementation. The vitamin E per LDL particle (the vitamin E=apoB molar ratio) was Consequence of phenolic-compound incubation on 9.98 Æ 0.49 after and 8.63 Æ 0.67 before supplementation. oxidative susceptibility of LDL In the treated group, LDL particles showed no modi®cation RWPC-incubated LDL without subsequent dialysis was in total lipids (3969.5 Æ 194.1 and 3813.5 Æ 226.4 mol=mol used ®rst in order to assess oxidative susceptibility. As of apoB at day 14 and day 0, respectively, namely 4.1%, shown by curve A in Figure 1, lag time and RWPC P 0.2). There was no statistical change in the vitamin concentration used for incubation were highly correlated E=total lipids molar ratio (2.69 Æ 0.22 after against (r 0.98; P < 0.001) with a regression coef®cient of 2.62 Æ 0.43 before treatment, expressed as mmol=mol) 12.2%=mg=L. Likewise, sinapic, caffeic or ferulic acid- together with no signi®cant correlation between vitamin incubated LDL was tested (Figure 2a and Table 3). Phe- E and total lipids (r 0.03 and 0.36). nolic acids also appeared to protect LDL against Cu2- Table 1 also shows the total fatty acid composition of mediated oxidation. In addition, a growing protection LDL. No signi®cant change was observed after supplemen- capacity was found: ferulic acid < sinapic acid < caffeic tation, except an increase in stearic acid (the C18 : 0=apoB acid. Curve B in Figure 1 and Figure 2b clearly established molar ratio was of 441.5 Æ 25.3 vs 389.1 Æ 21.4 before that dialysis after LDL incubation resulted in suppression of the relationship between the increased lag time and Table 1 Effect of a two weeks RWPC supplementation on LDL incubated RWPC or phenolic acids, showing that LDL biological parameters and fatty acid composition (mol=mol of apoB) in phenolic compounds were lost by dialysis. Addition of healthy subjects sinapic acid to the total plasma and subsequent isolation day 0 day 14 P of lipoprotein fractions were also performed in order to study a model of phenolic-compound distribution into the T Chol 2497.7 Æ 153.5 2569.0 Æ 134.8 0.2 different lipoprotein and the non-lipoprotein fractions. TG 534.7 Æ 37.5 544.2 Æ 60.9 0.4 Figure 3 shows that the total recovered amount of sinapic PL 786.1 Æ 59.7 856.6 Æ 47.3 0.1 T Lipids 3813.5 Æ 226.4 3969.5 Æ 194.1 0.2 acid and the sinapic acid content of each lipoprotein Vitamin E 8.63 Æ 0.67 9.98 Æ 0.49 0.01* 2 C14 : 0 94.2 Æ 9.3 111.9 Æ 24.8 0.3 Table 2 Cu -Oxidation parameters of LDL from healthy subjects C16 : 0 1251.9 Æ 74.1 1337.2 Æ 64.2 0.2 before (day 0) and after (day 14) RWPC supplementation C18 : 0 389.1 Æ 21.4 441.5 Æ 25.3 0.02* C18 : 1 (n±9) 793.2 Æ 45.8 794.7 Æ 53.1 0.4 day 0 day 14 P C18 : 2 (n±6) 1775.7 Æ 120.0 1664.1 Æ 107.7 0.3 T lag (min) 65.4 Æ 4.1 61.4 Æ 2.9 0.08 C20 : 4 (n±6) 305.9 Æ 19.7 276.3 Æ 22.0 0.09 Oxidation ratea 10.5 Æ 0.6 10.4 Æ 0.4 0.4 C20 : 5 (n±3) 32.3 Æ 4.3 26.9 Æ 3.6 0.2 Dienesb 444.6 Æ 17.1 489.5 Æ 16.7 0.01* C22 : 6 (n±3) 64.7 Æ 5.3 60.0 Æ 6.2 0.3 P P Dienes= of C alc 0.16 Æ 0.01 0.19 Æ 0.02 0.1 of C al 3136.6 Æ 181.5 2842.7 Æ 208.8 0.1 Peroxidesb 562.3 Æ 44.1 598.8 Æ 67.4 0.2 P Vit E= of C ala 3.07 Æ 0.25 3.89 Æ 0.38 0.006* TBARSb 55.1 Æ 6.4 56.1 Æ 4.4 0.3

Results are X Æ s.e.m. (n 20), Results are X Æ s.e.m. (n 20), aOxidation rate is expressed as mol of dienes=mol of apoB=min; T Chol total cholesterol;P TG triglycerides; PL phospholipids; bProducts of Cu2-mediated oxidation are expressed as mol=mol of apoB; T LipidsP total lipids; of C al sum of bis allylic carbon number; P aVit E= of C al expressed as mmol=mol. cDienes= of C al are expressed as mol=mol. *Values at day 0 and day 14 are signi®cantly different. *Values at day 0 and day 14 are signi®cant. Supplementation with wine phenolic compounds M-A Carbonneau et al 686 Table 3 In vitro inhibitory effect of phenolic acids on LDL Cu2- mediated oxidation susceptibility without subsequent dialysis

Phenolic acids T Lag (% of initial value)

0 100 Sinapic acid 1 100 10 120.9 100 469.9 Caffeic acid 10 181.4 100 527.4 Ferulic acid 10 100 100 200

LDL (1 mmol apoB=L) was incubated at 37 C for 1 h in PBS under N2 without or with phenolic acids (1±100 mmol=L for sinapic acid, 10± 100 mmol=L for caffeic and ferulic acids). It was then diluted to 1=15 with air-saturated PBS and oxidized by Cu2=L. T Lag lag time. Figure 1 Participation of RWPC in LDL protection against the Cu2- mediated oxidation. LDL (1 mmol apoB=L) was preincubated in PBS in the presence of increasing concentrations of RWPC (0±20 mg=L) at 37C for 1 h. LDL was then adjusted to 0.1 mmol=L with PBS and submitted to decreased with increasing incubated amount (77.2% and Cu2-mediated oxidation without (A) or with (B) extensive dialysis prior 60.6% for 300 mg=L and 1 mg=L, respectively). to the oxidative procedure. The vertical axis is the percentage lag time (T Lag), with 100% in the absence of RWPC. Changes in the plasma antioxidant capacity due to supplementation fraction were related to the added amount. In the presence Addition of diluted plasma to the luminescence producing of 100 mg=L of sinapic acid, the total amount recovered system triggered a rapid inhibition of light emission which after incubation was of 97.5% with 20.8% in VLDL, 10.3% was recovered after a time period dependent on the plasma in LDL, 4.7% in HDL and 64.1% in the VHDL nonLP added (see the typical curves shown in Figure 4A). Figure fraction. Only 35% of sinapic acid was associated to 4B shows the linear relationship between the light suppres- lipoproteins with d1.21 g=mL. As mentioned above for sion time and the amount of either Trolox (a water soluble RWPC, the sinapic acid part associated to LDL was tocopherol analogue) or RWPC added to the reaction released during dialysis (not shown). It is to notice that medium as external standards. Figure 5 illustrates the percentage of total recovered amount of sinapic acid signi®cant increase in Pl-AOC after RWPC ingestion (170.0 Æ 10.4 vs 156.4 Æ 10.4 mmol=L and 76.3 Æ 7.7 vs 66.0 Æ 4.5 mGAE=L for non diluted plasma, when expressed as Trolox concentration or RWPC concentration; P 0.01). The absolute increase appeared to be of 13.6 mmol=L and 10.3 mGAE=L, respectively.

Discussion It is widely acknowledged that alcohol, phenolic compounds or both may be responsible for a cardiovascular protective effect of wine. Without being exhaustive, it is worth recalling that as early as 1981, Klurfeld & Kritch- evski strongly suspected the more bene®cial role of red and white wine consumption as compared with that of all other alcoholic beverages in terms of protection against atherosclerosis in atherogenic rabbits. More recently, in rat the protection of wine against the platelet rebound effect speci®cally due to cessation of alcohol drinking has been attributed to its polyphenolic content, presumably by increasing the plasma antioxidant capacity in vivo (Ruf et al, 1995). Research into the effect of wine phenolic compounds in humans is therefore a timely avenue of investigation. The protection against oxidative modi®cation of LDL can be provided either by LDL itself, by the plasma or by both. The approach by in vitro co-incubation of LDL and phenolic compounds (De Whalley et al, 1990; NeÁgre- Salvayre & Salvayre, 1992; Frankel et al, 1993; Frankel et al, 1995) has been crucial for establishing the protective effect of these substances towards LDL oxidation but is of Figure 2 Participation of phenolic acids in LDL protection against the limited physiological interest because it does not take into Cu2-mediated oxidation. The protocol was the same as in Figure 1. LDL account the early stages, namely bioconversion in the was submitted to Cu2-mediated oxidation in the presence of intestinal lumen, selective intestinal absorption, which 100 mmol=L sinapic (SA), ferulic acid (FA) or caffeic acid (CA) or in the absence of phenolic acid (±PA) without (A) or with (B) extensive determine to a great part the phenolic forms actually dialysis. present in the plasma. The purpose of the present paper Supplementation with wine phenolic compounds M-A Carbonneau et al 687

Figure 3 Sinapic acid partition into lipoprotein and non-lipoprotein fractions after incubation with plasma. Plasma was incubated with increasing concentrations of sinapic acid at 37C for 1 h and submitted to sequential ultracentrifugation for lipoprotein isolation. Recovered sinapic acid was then determined in each lipoprotein fraction by HPLC procedure on a column (150 6 4.6 mm) of C18 inverse phase after extraction with diisopropylether=3N HCL (20 : 1, v=v), evaporation and dilution in a 10% hydroethanolic mixture. For the gradient of elution, see text. was consequently to examine the in vivo `functional' no precise information about proanthocyanidins and antho- effects of an alcohol-free phenolic supplementation in cyanins of red wine used by these authors, it seems that the human volunteers. The question was whether RWPC inges- daily ingested amounts of `catechins' and ¯avonols were tion over a long term period is ef®cient enough to prevent similar to those of the present paper and, interestingly, that LDL and=or plasma from oxidation in vivo. Indeed, the the amounts present in white wine were extremely low. In absence of suf®ciently reliable and sensitive methods for these conditions, it is interesting that De Rijke et al. assessing wine phenolic or wine phenolic-deriving products showed no change in LDL vitamin E in relation with in the plasma at the beginning of the present study led to LDL cholesterol. It is worth mentioning that our study the decision to evaluate the functional effect brought about also supports no signi®cant change in vitamin E as com- by phenolic-compound ingestion. In order to assess more pared to LDL cholesterol (to LDL total lipids as well). particularly the remanent antioxidant action of RWPC, A meta-analysis clearly indicates that a 40% minimal blood sampling was made in fasting conditions after two increase in the particle charge of vitamin E is needed to weeks of supplementation. signi®cantly improve the LDL resistance to oxidation It can be clearly drawn from the present results that (Dieber±Rotheneder et al, 1991; Reaven et al, 1993; LDL, in its own right, does not gain a higher protection Belcher et al, 1993; Suzukawa et al, 1995; Jialal et al, from oxidation (at least, in the assessment conditions used) 1995; Princen et al, 1993). Therefore, the lack of LDL- after a 14 d period of RWPC supplementation. This is true oxidizability effect of the 15.7% increase in vitamin E of in spite of a 15.7% vitamin E signi®cant enrichment and a LDL in the present study is not surprising. On the other more highly signi®cant increase in the vitamin E=bis-allylic hand, the ®nding that the supplementation led to an carbon molar ratio of LDL, which is expected to be the enhanced diene production by LDL-oxidation procedure, mark of an improved protection against oxidation (since an which was noted to be the only change in measurement of antioxidant parameter at the numerator is associated to a LDL oxidative susceptibility, is likely to be related partly to pro-oxidant parameter at the denominator). This lack of the slight concomitant change in bis-allylic carbon number. improved protection is not in agreement with other results Regarding the effect of RWPC on the total plasma (Kondo et al, 1994; Fuhrman et al, 1995) where the action (namely including lipoproteins in their natural plasma of red or white wine consumption was analyzed using environment) of treated volunteers, it was found a statisti- similar LDL preparative procedure. The difference, how- cally signi®cant increase in Pl-AOC. This is in accordance ever, may be explained by the supply of an integrate wine. with the changes reported by others using a slightly It is indeed questionable whether wine phenolic compounds different system of chemiluminescence production (White- are readily absorbed in the absence of alcohol, which is a head et al, 1992; Lissi et al, 1995; Maxwell et al, 1994; natural stabilizing agent for polyphenols. Nevertheless, our Whitehead et al, 1995). The present results, however, add results agree with those of Sharpe et al (1995). It is also new information because the enhanced antioxidant capacity worth noting that De Rijke et al (1996) reported no was achieved by ingesting a phenolic extract and not wine difference in LDL protecting effects of red wine as com- itself, and by using fasting rather than postprandial plasma pared with white wine in the particular case of lower (see Maxwell et al, 1994; Whitehead et al, 1995). Thus, alcohol content (3.5%). In this respect, although we have they support a real remanence of the antioxidative potency Supplementation with wine phenolic compounds M-A Carbonneau et al 688

Figure 5 Plasma antioxidant capacity (Pl-AOC) determined by the chemiluminescence assay on volunteers before (day 0) and after (day 14) RWPC supplementation. Chemiluminescence assay was as in Figure 4. Pl-AOC was expressed as mmol Trolox=L.

This is in accordance with other studies showing that intestinal absorption of quercetin, quercetin glucosides and phenolic acids (these last ones being present in wine and possibly deriving from ¯avonoids by the action of the intestinal micro¯ora) occurs in humans (Hollman et al, 1995) and dogs (S. Labidalle, personal communication). In order to explain the discrepancy between the results obtained with isolated LDL and total plasma, we decided to proceed in vitro LDL oxidation after incubation in the presence of RWPC or phenolic acids present in red wine. One dialysis was performed after incubation and just before the oxidation procedure, or this dialysis was omitted. The results clearly showed that these phenolic compounds, without dialysis, readily increased the defense of the lipoparticle against the oxidative stress. The higher protection provided by caffeic acid (3,4-dihydroxy-cinnamic acid) as compared with sinapic and ferulic acids (4- hydroxy-3,5-dimethoxy-cinnamic and 4-hydroxy-3-meth- Figure 4 Typical time course of light emission with 1=10 diluted plasma from one subject before (day 0) and after (day 14) RWPC supplementation oxy-cinnamic acids, respectively) underlines the negative (A) and linear relationship between the time of light suppression (t) and the effect of the presence of methoxyl group(s) on the aromatic trolox or RWPC concentration (B). 20 mL of plasma once diluted at 1=10 cycle near the phenyl hydroxyl in preventing oxidation in PBS or 20 mL of standard solutions (trolox or RWPC) was added to the (Rafat et al, 1987; Nardini et al, 1995) as well as the luminescence producing system (luminol=HRP=H2O2). The time of positive effect of the orthodiphenol con®guration. How- plasma (tpl) or standards (t) light suppression was assessed by the time needed to reach 50% of the maximum value. Trolox or RWPC concentra- ever, the antioxidative protection of phenolic compounds tions were expressed as mmol=L or mGAE=L, respectively. was totally lost after extensive LDL dialysis. This strongly suggests that RWPC or phenolic acids are loosely associated to LDL, possibly remaining at least partially, by of RWPC. The increase in the antioxidant capacity was virtue of their mainly hydrophilic properties, at the surface equivalent to one third of the vitamin E content of plasma of the particle where they could achieve their protective in people who live in the Montpellier, France, area, action. This is at variance with another work (Fuhrman et (namely 42 mmol=L, Cachia et al, 1995), suggesting that al, 1995) which found that dialysis before the oxidation it is not a negligible increase. It is worth noting that this procedure did not affect the phenolic content of LDL as increase was also equivalent to 10.3 mGAE=L which might assessed by a procedure which unfortunately interferes with represent a rather realistic plasma content in polyphenols in proteins. We propose that the effect of dialysis presently cases of red wine consumption. However, we have to keep shown explains why LDL (once isolated) and original in mind that only some of the RWPC components (the low plasma have different antioxidative capacities. Some ®nd- molecular weight polyphenols) might be able to pass ings of our group (unpublished data) have recently been through the intestinal wall and to be present in the circulat- obtained using gel chromatography which entirely con®rm ing blood. This topic is of crucial importance for the that phenolic compounds are easily released from LDL understanding of wine polyphenol action and remains in a particles. In addition, the data on the sinapic acid partition great part to be explored. into plasma fractions also show that this compound is The present ®nding, however, provides indirect but mainly located in the protein, protein-rich lipoprotein rather convincing evidence that at least one part of the (d > 1.21 g=mL) and aqueous part of plasma. antioxidant materials from red wine phenolic extracts can It could be assumed that ¯avonoids play a role similar to be absorbed and enter the plasma in the absence of alcohol. that of vitamin C at the surface of LDL particles (Niki, Supplementation with wine phenolic compounds M-A Carbonneau et al 689 1986; Frei, 1991) by regenerating the reduced form of Frei B (1991): Ascorbic acid protects lipids in human plasma and low- vitamin E from its oxidized form chromanoxyl and by density lipoprotein against oxidative damage. Am. J. Clin. Nutr. 54, 1113S±1118S. generating phenoxyl radicals Ph-O (Belyakov et al, 1995), Fuhrman B, Lavy A & Aviram M (1995): Consumption of red wine with benzofuranic substances and other new radical-scavenging meals reduces the susceptibility of human plasma and low-density phenolic compounds (Tsuda et al, 1996). A sparing effect lipoprotein to lipid peroxidation. Am J. Clin. Nutr. 61, 549±554. of tannins toward the plasma vitamin E of moderate Henley R & Worwood M (1994): The enhancement of iron-dependent luminol peroxidation by 2,20-dipyridyl and nitrilotriacetate. J. Biolum. drinkers has already been suggested from examination of Chemilumin. 9, 254±250. epidemiological data (Leconte et al, 1994). Based on the Hertog MGL, Feskens EJM, Hollman PCH, Katan MB & Kromhout D present increase in LDL vitamin E, it appears that the (1993): Dietary antioxidant ¯avonoids and risk of coronary heart sparing effect may essentially involve the LDL particle as disease: the Zutphen elderly study. Lancet 342, 1007±1011. HDL was not modi®ed. Moreover, it also suggests that the Hollman PCH, De Vries JHM, Van Leeuwen SD, Mengelers MJB & Katan MB (1995): Absorption of dietary quercetin glycosides and quercetin in increase in Pl-AOC is of physiological relevance in terms healthy ileostomy volunteers. Am. J. Clin. Nutr. 62, 1276±1282. of LDL antioxidative protection. Jialal I, Fuller C & Huet BA (1995): The effect of a-tocopherol supple- In spite of a better understanding of the ¯avonoid mentation on LDL oxidation. A dose-response study. Arterioscler. reactions leading to an antioxidative protection in physio- Thromb. Vasc. Biol. 15, 325±333. logical conditions, further investigations are still needed to Kondo K, Matsumoto A, Kurata H, Tanahashi H, Koda H, Amachi T & Ikura H (1994): Inhibition of oxidation of LDL with red wine. Lancet determine which forms of ¯avonoid-derived products in 344, 1152. vivo would be effective in humans in preventing oxidative Klurfeld DM & Kritchevsky D (1981): Differential effects of alcoholic deleterious alterations, which are assumed to result in an beverages on experimental atherosclerosis in rabbits. Exp. Mol. Pathol. anti-atherogenic effect (work in progress in a collaborative 34, 62±71. Leconte E, Herbeth B, Pirollet P, Chancerelle Y, Arnaud J, Musse N, project). Paille F, Siest G, Artur Y (1994): Effect of alcohol consumption on blood antioxidant nutrients and oxidative stress indicators. Am. J. Clin. Nutr. 60, 255±261. Li Q & Cathcart MK (1994): Protein kinase C activity is required for lipid References oxidation of low density lipoprotein by activated human monocytes. J. Biol. Chem. 269,17508±17515. Agel NM, Ball RY, Walman H & Mitchinson MJ (1984): Monocytic Lissi E, Salim-Hanna M, Pascual C & Del Castillo MD (1995): Evaluation origin of foam cells in human atherosclerotic plaques. Atherosclerosis of total antioxidant potential (TRAP) and total antioxidant reactivity 53, 265±271. from luminol-enhanced chemiluminescence measurements. Free Rad. Bauman J, Bruchhausen F & Wurm G. (1980): Favonoids and related Biol. Med. 18, 153±158. compounds as inhibitors of arachidonic acid peroxidation. Prostaglan- Lynch SM & Frei B (1993): Mechanisms of copper- and iron-dependent dins 20, 627±639. oxidative modi®cation of human low density lipoprotein. J. Lipid Res. Belcher JD, Balla J, Balla G, Jacobs DR Jr, Gross M, Jacob HS, Vercelotti 34, 1745±1753. GM (1993): Vitamin E, LDL and endothelium. Brief oral vitamin Maxwell S, Cruickshank A & Thorpe G (1994): Red wine and antioxidant supplementation prevents oxidized LDL-mediated vascular injury in activity in serum. Lancet 344, 193±194. vitro. Arteriosc. Thromb. 13, 1779±1789. Misra PH & Squatrito PM (1982): The role of superoxide anion in Belyakov VA, Roginsky VA & Bors W (1995): Rate constants for the peroxidase-catalysed chemiluminescence of luminol. Arch. Biochem. reaction of peroxyl free radical with ¯avonoids and related compounds Biophys. 215, 59±65. as determined by kinetic chemiluminescence method. J. Chem. Soc. Nardini M, D'Aquino M, Tomassi G, Gentili V, Di Felice M & Scaccini C Perkin Trans. 2, 2319±2326. (1995): Inhibition of human low-density lipoprotein oxidation by Cachia O, Leger CL, Boulot P, Vernet MH, Michel F, Drastes de Paulet A caffeic acid and other hydroxycinnamic acid derivatives. Free Rad. & Descomps B (1995): Red blood cell vitamin E concentrations in Biol. Med. 19, 541±552. fetuses are related to but lower than those in mothers during gestation: NeÁgre-Salvayre A & Salvayre R (1992): Quercetin prevents the cytotoxi- A possible association with maternal lipoprotein (a) plasma levels. Am. city of oxidized LDL on lymphoid cell lines. Free Rad. Biol. Med. 12, J. Obstet. Gynecol. 173, 42±51. 101±106. Criqui MH & Ringel BL (1994): Does diet or alcohol explain the French Niki E (1986): Antioxidants in relation to lipid peroxidation. Chem. Phys. paradox? Lancet 344, 1719±1723. Lipids 44, 227±253. De Rijke YB, Demacker PNM, Assen NA, Sloots LM, Katan MB & Princen HMG, van Duyvenwoorde W, Buytenhek R, van der Laarse A, Stalenhoef AFH (1996): Red wine consumption does not affect oxidiz- van Poppel G, Geversleuven JA & van Hinsbergh VWM (1995): ability of low-density lipoproteins in volunteers. Am. J. Clin. Nutr. 63, Supplementation with low doses of vitamin E protects LDL from 329±334. lipid peroxidation in men and women. Arterioscler. Thromb. Vasc. De Whalley CV, Rankin SM, Hoult JRS & Jessup W (1990): Flavonoids Biol. 15, 325±333. inhibit the oxidative modi®cation of low-density lipoproteins by macro- Quinn MF, Parthasarathy S, Fong LG & Steinberg D (1987): Oxidatively phages. Biochem. Pharmacol. 11, 1743±1750. modi®ed low density lipoproteins: a potential role in recruitment and Dieber-Rotheneder M, Puhl H, Waeg G, Striegl G & Esterbauer H (1991): retention of monocyte=macrophage during atherogenesis. Proc. Natl. Effect of oral supplementation with d-alpha tocopherol on the vitamin E Acad. Sci. 84, 2995±2998. content of human low-density lipoproteins and resistance to oxidation. Rafat Husain S, Cillard J & Cillard P (1987): Hydroxyl radical scavenging J. Lipid Res. 32, 1325±1332. activity of ¯avonoids. Phytochemistry 26, 2489±2491. El-Saadani M, Esterbauer H, El-Sayed M, Goher M, Nassar AY & Jurgens Rajavashisth TB, Andalibi A, Territo ML, Berliner JA, Navab M, Fogel- G (1989): A spectrophotometric assay for lipid peroxides in serum man AM & Lusis AJ (1990): Induction of endothelial cell expression of lipoproteins using a commercially available reagent. J. Lipid Res 30, granulocyte and macrophage colony-stimulating factors by modi®ed 627±630. low-density lipoproteins. Nature 344, 244±257. Esterbauer H, Gebicki J, Puhl H & JuÈrgens G (1992): The role of lipid Renaud S & De Lorgeril M (1992): Wine alcohol platelets and the French peroxidation and antioxidants in oxidative modi®cation of LDL. Free Paradox for coronary heart disease. Lancet 339, 1523±1526. Rad. Biol. Med. 13, 341±390. Reaven PD, Khouw A, Beltz WF, Parthasarathy S & Witztum JL (1993): Esterbauer H, Striegl G, Puhl H & Rotheneder M (1989): Continuous Effect of dietary antioxidant combinations in humans. Protection of monitoring of in vitro oxidation of human low density lipoprotein. Free LDL by vitamin E but not by b-caroten. Arteriosc. Thromb. 13, 590± Rad. Res. Comms. 6, 67±75. 600. Frankel EN, Kanner J, German JB, Parks E & Kinsella JE (1993): Rimm EB, Klatsky A, Grobbee D & Stampfer MJ (1996): Review of Inhibition of oxidation of human low-density lipoprotein by phenolic moderate alcohol consumption and reduced risk of coronary heart substances in red wine. Lancet 341, 454±457. disease; is the effect due to beer, wine or spirits? BMJ 312, 731±736. Frankel EN, Waterhouse AL & Teissedre PL (1995): Principal phenolic Roback J & Griglewski RJ (1988): Flavonoids are scavenger of superoxide phytochemicals in selected California wines and their antioxidant anion. Biochem. Pharmacol. 37, 837±841. activity in inhibiting oxidation of human low-density lipoproteins. J. Ross R (1993): The pathogenesis of atherosclerosis: a perspective for the Agric. Food Chem. 43, 890±894. 1990s. Nature 362, 801±809. Supplementation with wine phenolic compounds M-A Carbonneau et al 690 Ruidavets JB, Cambou JP, Arveiler D, Nuttens MC, Bingham A, Richard Steinbrecher UP, Zhang H & Lougheed M (1990): Role of oxidatively JL, Schaffer P, Salomez JL & Ducimetiere P (1993): Differences inter- modi®ed LDL in atherosclerosis. Free Rad. Biol. Med. 9, 155±168. reÂgionales de consommation alimentaire: le Projet MONICA. Cah. Suzukawa M, Ishikawa T, Yoshida H & Nakamura H (1995): Effect of in Nutr. Diet. 28, 105±109. vivo supplementation with low dose of vitamin E on susceptibility of Ruf JC, Berger JL & Renaud S (1995): Platelet rebound effect of alcohol low-density lipoprotein and high-density lipoprotein to oxidation mod- withdrawal and wine drinking in rats. Arterioscler. Thromb. Vasc. Biol. i®cation. J. Am. Col. Nutr. 14, 46±52. 15, 140±144. Torel J, Cillard J & Cillard P (1986): Antioxidant activity of ¯avo- Sadou H, LeÂger CL, Descomps B, Barjon JN, Monnier L & Crastes de noids and reactivity with peroxyl radical. Phytochemistry 25, Paulet A (1995): Differential incorporation of ®sh oil eicosapentaenoate 383±387. and docosahexaenoate into lipids of lipoprotein fractions as related to Tsuda T, Oshima K, Kawakishi S & Osawa T (1996): Oxidation products their glyceryl esteri®cation: a short-term (postprandial) and long-term of cyanidin 3-O-b-glucoside with a free radical initiator. Lipids 31, study in healthy humans. Am. J. Clin. Nutr. 62, 1193±1200. 1259±1263. Schumaker VN & Puppione DL (1986): Sequential ¯otation ultracentri- Whitehead TP, Thorpe GHG & Maxwell SRJ (1992): Enhanced chemilu- fugation. Methods in Enzymol. 128, 155±170. minescence assay for antioxidant capacity in biological ¯uids. Anal. Scheek LM, Wiseman SA, Tijburg LBM & von Tol A (1995): Dialysis of Chim. Acta. 266, 265±277. isolated low density lipoprotein induces a loss of lipophilic antioxidants Whitehead TP, Robinson D, Allaway S, Syms J & Hale A (1995): Effect of and decreases the susceptibility to oxidation in vitro. Atherosclerosis red wine ingestion on the antioxidant capacity of serum. Clin. Chem. 41, 117, 139±144. 32±35. Sharpe PC, McGrath LT, McClean E, Young IS & Archbold GPR (1995): Willett WC, Sacks F, Trichopoulou A, Drescher G, Ferro-Luzzi A, Helsing Effect of red wine consumption on lipoprotein (a) and other risk factors E & Trichopoulos D (1995): Mediterranean diet pyramid: a cultural for atherosclerosis. Q. J. Med. 88, 101±108. model for healthy eating. Am. J. Clin. Nutr. 61, 1402S±1406S. Singleton VL & Rossi JA (1965): Colorimetry of total phenolics with YlaÈ-Herttuala S, Palinski W, Rosenfeld ME, Parthasarathy S, Carew TE, phosphomolybdic-phosphotungstic acid reagents. Am. J. Enol. Vitic. 16, Butler S, Witztum JL & Steinberg D (1989): Evidence for the presence 144±158. of oxidatively modi®ed low density lipoprotein in atherosclerotic Steinberg D, Parthasarathy S, Carew TE, Khoo JC & Witztum JL (1989): lesions of rabbit and man. J. Clin. Invest. 84, 1086±1095. Beyond cholesterol. Modi®cation of low-density lipoprotein that increase its atherogenicity. N. Engl. J. Med. 320, 915±924.