mercredi 25 septembre 2013

Flavonoids and cardiovascular disease

The flavanonols, also called dihydroflavonols, are found only in trace amounts in plants, and are therefore not important constituents in the human diet (Pierpoint, 1986). The contribution offlavones to the human intake of flavonoids is generally limited; however, some spices and herbs contain high amounts of flavones. Parsley, for instance, contains large amounts of apigenin (Justesen et al., 1998), and the highly methylated flavone tangeretin is found in the peel of citrus fruits, and can thus occur in juices made from whole fruits (Pierpoint, 1986).

The flavonols are one of the major groups of flavonoids present in the human diet. Quercetin is the most abundant flavonol, found ubiquitously in fruits and vegetables and is especially present in high amounts in onions, cruciferous, apples, wine and tea. Kaempferol, found in broccoli, kale and tea, and myricetin, found in tea and wine, are also major flavonols found in the human diet (Hertog et al., 1992, 1993b).

The isoflavones are present only in legumes, especially in soybean. The European intake of isoflavones is therefore limited, but in Asia and especially in Japan, where soy are consumed in large amounts, the intake is considerable.

Catechins are quantitatively a quite large group within the human diet. They either occur as free catechins or are derivatised with gallic acid, and are found mainly in green and black tea, chocolate and wine (Forsyth, 1955; Rimm et al., 1996a). In countries with a high intake of tea, such as Japan (mainly green tea) and the United Kingdom (black tea), the daily intake of catechins is substantial. The average intake of the strongly coloured anthocyanins may also be extensive, in particular for regular consumers of red wine, blackcurrant juice, berries, and red grapes.

The estimated average daily intake of flavonoids including catechins and anthocyanins is thus well above 50 mg for all countries and the real intake is probably higher than 100 mg/day if data on all flavonoid subgroups were available. The daily intake of other dietary antioxidants such as vitamin C (80 mg/day), vitamin E (8.5 mg/day) and 3-carotene (1.9 mg/day) (Nielsen, 1999a) is comparable to or considerably lower than the intake of the flavonoids, so these compounds certainly constitute an important part of the daily intake of dietary antioxidants.

Until recently, it was generally accepted that the flavonoid aglycons had to be liberated from the glycosides in the large intestine, prior to absorption (KuÈhnau, 1976). It was thought that the hydrophilic nature of the glycosides precludes absorption in the small intestine, and that the flavonoid O-glycosides resists intestinal hydrolysis. Consequently, flavonoid glycosides would pass unaltered into the large intestine, being hydrolysed by micro-organisms to the free aglycon

and sugar moiety (Hertog et al., 1997). This was supported by the findings that incubations of flavonoid glycosides with intestinal micro-flora resulted in the release of the free aglycons, and in germ-free rats the flavonoid glycosides were excreted unchanged with faeces (Griffiths, 1982). It was later demonstrated that human intestinal bacteria are in fact capable of hydrolysing flavonoid glycosides to the free aglycons and sugar moieties (Bokkenheuser and Winter, 1988). The intact flavonoid aglycons are then absorbed from the large intestine and enter the systemic circulation. The liver is the main organ for metabolism, and here also the flavonoid aglycons are hydrolysed or demethylated by the cytochrome P450 enzyme system (CYP450) (Griffiths, 1982; Nielsen et al., 1998, 2000a), methylated by catechol-O-methyl transferase (COMT) (Zhu et al., 1994) and conjugated to glucuronic acid and sulphate esters by the Phase II enzymes (Gee et al., 2000, Boersma et al., 2002).

The micro-organisms in the large intestine are also capable of degrading the flavonoid aglycons by ring fission of the C-ring, resulting in a variety of phenolic acids (e.g. phenylpropionic and phenylacetic acid derivatives) and phloroglucinol (from the A-ring) (Griffiths, 1982). These phenolic acids can thus also be absorbed from the gut, and further metabolised in the liver by the CYP450 enzyme system, resulting in either hydroxylation or methylation and further conjugated with glucuronic acid or sulphate.

During the past decade the absorption of flavonoid glycosides has been the subject of several controversies. Some research groups reported that intact flavonoid glycosides were being absorbed in humans after intake of a flavonoid­rich meal (Hollman et al., 1995). However, the majority of the following studies on this matter were not able to verify any absorption of intact flavonoid glycosides in humans. In a study on naringin from grapefruit juice by Fuhr and Kummert (1995), naringin was not detectable in urine; only the aglycon naringenin was determined. Also in a study on diosmin, the 7-rutinoside of diosmetin (4'-OMe-Luteolin), failed to reveal any diosmin in urine or plasma after oral administration of diosmin to healthy volunteers (Cova et al., 1992). Only the aglycone diosmetin was detected as glucuronic and sulphate conjugates. Furthermore, no apiin (apigenin-7-apiosylglycoside), the major apigenin glycoside in parsley, was detected in the urine after parsley consumption (Nielsen et al., 1999b) and in a study with a fruit and vegetable mixture, only the flavonoid aglycons could be identified in urine by a liquid chromatography mass spectrometry (LC-MS) method able to reveal both flavonoid aglycons and glycosides if present (Nielsen et al., 2000b, 2002). In a human study using onions as the quercetin source, Moon and colleagues (2000) were unable to determine quercetin glycosides in plasma by a very sensitive method using electrochemical detection, whereas quercetin was detectable, supporting the fact that no intact glycosides were absorbed.

Recently, several researchers have demonstrated the difficulties in distin­guishing between the flavonoid glycosides and glucuronides when performing chemical analyses of urine and plasma samples. Both types of conjugates have UV-absorption spectra that are almost identical with quercetin itself (Day and Williamson 2001), and very similar retention times even with the most specific gradient systems. Using a highly sensitive and specific high-performance liquid chromatography (HPLC) method with coulometric detection and confirmation by LC-MS, it was demonstrated, that only the glucuronides and not the glucosides are present in human plasma after consumption of pure quercetin-3­ glucoside or quercetin-4'-glucoside (Sesink et al., 2001).

In 1995±1997 Hollman and co-workers performed several studies on the bioavailability of quercetin in humans (Hollman et al., 1995, 1996b, 1997), and it was observed that the amount of quercetin absorbed from quercetin glucosides was higher than from the free aglycon or other glycosides with different sugars than glucose as the terminal sugar, such as, for example, rutin. Furthermore, quercetin was also more rapidly absorbed when given as quercetin glucosides with peak plasma levels already within the first 30±60 min, whereas other quercetin glycosides from apple had peak plasma levels several hours later (9.3 h for pure rutin) (Hollman et al., 1997). This was not in accordance with the general concept that the glycosides have to pass unaltered all the way to the large intestine prior to hydrolysis and absorption, since this would have required a much longer transit time than the 30±60 min observed for the quercetin mono­glucosides. Thus, the mono-glucosides must have already been absorbed in the small intestine. This was supported by the findings by Day et al. in 1998, demonstrating that both human and rat small intestine have 3-glucosidase activities capable of hydrolysing flavonoid glucosides. Furthermore, studies by Gee et al. (1998, 2001) using isolated preparations of rat small intestine suggested a possible involvement of the sodium-dependent glucose transporter SGLT1 in the absorption of quercetin monoglucosides. These studies confirmed a more rapid absorption of the 3-or 4'-monoglucosides of quercetin compared to the 3,4'-diglucoside or quercetin itself. The authors concluded that there probably are two mechanisms for the transport of the quercetin monoglucosides

the sugar transporter SGLT1 transport the intact quercetin glucoside into the epithelial cells, where the glucose moiety are released by ,i-glucosidases, and the free quercetin is then absorbed, orthe extracellular enzyme LPH (lactase phlorizin hydrolase), which was also demonstrated to be able to hydrolyse quercetin glucosides (Day et al., 2000), deglycosylate the flavonoid glucoside, prior to passive diffusion of the aglycon into the epithelial cells

Furthermore, these studies verified that no intact quercetin glycosides were able to cross the intestinal epithelium, and only quercetin or quercetin glucuronides were determined after transfer. Recent studies on quercetin glucosides by the same research group has further showed that the major products produced in the small intestine are the 3- and 7-glucuronides of quercetin (O’Leary et al., 2003). Based on their results, they suggest, that from a normal dietary intake of quercetin only glucuronic conjugates reach the liver via the hepatic portal vein (O’Leary et al., 2003). They have furthermore demonstrated that in the liver, the flavonoid conjugates are further metabolised either by methylation by COMT or by intracellular deglucuronidation by 0-glucuronidase followed by sulphation to the mono-sulphate conjugate. In this figure, all possible routes are illustrated for deglycosylation, absorption and metabolism of a flavonoid glucoside as it is presently thought to occur in the body. Flavonoids with other sugars attached than glucose, pass unaltered down to the large intestine, where the micro-flora can cleave the glycosidic bond and degrade the resulting flavonoid aglycone  as reviewed by KuÈhnau (1976).

Although an extensive number of studies have reported effects of flavonoids on enzymatic, biological and physiological processes, only very few researchers have attempted to determine the actual compound/metabolite responsible for the observed effects. It has generally been assumed that the biological activities originated from the flavonoids investigated, although they may be bio­transformed into one or more structurally quite different compound in vivo. Investigations on in vitro metabolism of flavonoids have so far been limited. The synthetic flavonoids, a- and 0-naphthoflavone, well-known inducers and inhibitors of monooxygenase activities, have been shown to be extensively hydroxylated by cytochrome P450 (Vyas et al., 1983). The polymethoxylated flavone tangeretin was shown to be demethylated in vitro by the cytochrome P450 system, although the structures of the metabolites were not elucidated in this study (Canivenc-Lavier et al., 1993).

The first systematic investigation of the structural requirements for metabolism of flavonoid aglycons by the cytochrome P450 enzyme system was provided from our laboratory only a few years ago in a study, where 16 different flavonoids were incubated with rat liver microsomes (Nielsen et al., 1998). It was shown that the flavonoids naringenin, hesperetin, chrysin, apigenin, tangeretin, kaempferol, galangin and tamarixetin all were extensively metabolised by Aroclor-induced rat liver microsomes and to a minor extent by uninduced microsomes. All metabolites were isolated and their structures elucidated by LC-MS and 1H NMR (nuclear magnetic resonance). The identity of the metabolites was consistent with a general metabolic pathway leading to the corresponding 3',4'-dihydroxylated flavonoid either by hydroxylation or demethylation. No metabolites were, however, detected from eriodictyol, taxifolin, luteolin, quercetin, myricetin, fisetin, morin or isorhamnetin. Structural requirements for microsomal hydroxylation by the cytochrome P450 enzyme system thus appeared to be only a single or no hydroxyl group on the B-ring of the flavan nucleus. The presence of two or more hydroxyl groups on the B-ring seemed to abolish further hydroxylation. Furthermore, the results indicated that demethylation only occurs in the B-ring, when the methoxyl group is positioned at the 4'-position as in tamarixetin, and not in the 3'-position as in isorhamnetin. The CYP1A isozymes were found to be the main enzymes involved in flavonoid hydroxylation, whereas other cytochrome P450 isozymes seemed to be involved in the flavonoid demethylation (Nielsen et al., 1998). These findings with rat liver microsomes were later verified in mouse and human liver microsomal preparations, where identical metabolic patterns were observed for the flavonoids (Breinholt et al., 2002).

Most of the research on in vivo metabolism and disposition of flavonoids in experimental animals was performed in the 1960s and 1970s. These investigations have been thoroughly reviewed by Griffiths (1982), Hackett (1986) and by Hollman and Katan (1997). Except for the ring cleavage products, the only metabolites identified from flavonoid aglycons in rodents have been: isorhamnetin and tamarixetin (3'- or 4'-methoxyquercetin) from administration of quercetin (Ueno et al., 1983), 4?-hydroxy- and 3?,4?-dihydroxy-flavone from flavone, apigenin (3?-hydroxychrysin) from chrysin, and eriodictyol (3?- hydroxynaringenin) from naringenin (Hackett, 1986).

We recently investigated the in vivo metabolism of the polymethoxylated flavonoid tangeretin in order to evaluate the relevance of the identified in vitro metabolic pathways (Nielsen et al., 2000a; Rasmussen and Breinholt 2003). Urine collected consecutively during 24 hours was enzymatically hydrolysed to release the flavonoid aglycons from glucuronic acid or sulphate ester conjugates, and, by means of LC-MS and proton NMR, the presence of ten metabolites of tangeretin were identified. The metabolites were either demethylated or hydroxylated derivatives of the parent compound. The changes were again found primarily to occur in the B-ring of the compound as also observed in the in vitro metabolic studies (Nielsen et al., 1998, 2000a). Thus although the metabolism of flavonoids by CYP450 and COMT is limited for the majority of the dietary flavonoids, where only minor amounts of metabolites are produced, it is important that this endogenous metabolism is taken into account for some flavonoids, e.g. quercetin and tangeretin, where the metabolism is extensive. The metabolites produced are chemically very different from the parent compound and thus have the potential to exert biological effects other than those produced by the parent compound. The methylated derivatives of quercetin, isorhamnetin and tamarixetin, are both less polar and have a lower antioxidative potential than quercetin itself. On the contrary, the demethylated metabolites of tangeretin presumably have a higher antioxidative potential than the fully methoxylated compound, tangeretin.

It is furthermore important to bear in mind that the majority of the flavonoids that reach the systemic circulation are conjugated by the Phase II enzyme system to glucuronic and sulphate conjugates, and that it thus are these compounds that have the potential to exert biological effects in the body.

The red, violet or blue anthocyanins, found in most berries and fruits, belong to the group of flavonoids. The anthocyanins consist of an aglycon, the antho­cyanidin, linked to a sugar moiety. The six most frequently found aglycons in fruits and berries. These aglycons may be glycosylated or acylated by different sugars and acids in different positions. The most common glycoside moieties found in anthocyanins are the 3-monosides, 3-biosides, 3­ triosides and 3,5-diglycosides (Strack and Wray, 1986).

Anthocyanins are as the other classes of flavonoids present in most higher plants. They are biosynthesised in the vacuola with naringenin as the flavanone precursor. Naringenin is converted by a monooxygenase to the flavononol, dihydroxykaempherol, which can be hydroxylated further in the B­ring, forming the precursors of the 3?,4? or 3?,4?,5? hydroxylated or methoxylated anthocyanins. The resulting dihydroxyflavonols are reduced in the 4-position, forming leucoanthocyanins. The leucoanthocyanins are transformed by antho­cyanidin synthase to the corresponding anthocyanidins. The anthocyanidins is then glucosylated by 3-O-glucosyltransferase. This anthocyanidin-3-O­glucosides can further be either glycosylated by glycosyltransferases or methylated in the B-ring by methyltransferases (Delgado-Vargas et al., 2000).

The content of anthocyanins in coloured fruits and berries varies a lot, from around 3.4 mg in reddish apples, 232 mg in strawberries, 1064 mg in blackcurrants up to 3090 mg in blueberries per 100g dried fruit or berry (KaÈhkoÈnen et al., 2001). The anthocyanins are responsible for the colour difference between red and white wine (Mazza and Miniati, 1993). One glass of red wine can contain up to 80 mg anthocyanin, depending on the grape variety and processing of the wine (Waterhouse, 2002). The average Danish dietary intake of anthocyanins has been estimated to around 6± 60 mg per day (Dragsted et al., 1997) and the Finnish average intake to be 82.5 mg per day (Heinonen, 2001). Other rich dietary sources, apart from red grapes and red wine, are cherries, elderberries, blueberries and blackcurrants (Macheix et al., 1990).

In contrast to the other classes of flavonoids, the anthocyanins are absorbed as intact glycosides in both animals and humans, although in very low amounts (Cao et al., 2001; Nielsen et al., 2001, 2003; Wu et al., 2002). We recently investigated the absorption and excretion of blackcurrant anthocyanins in both humans and rabbits (Nielsen et al., 2003). Here we found as others also have reported, that the absorption of anthocyanins in humans is fast and proportional with dose, with peak plasma concentrations after about 1 hour (Bub et al., 2001; Matsumoto et al., 2001; Cao et al., 2001; Nielsen et al., 2003). A recent study by Passamonti et al. (2003) suggests that the anthocyanins are absorbed already in the stomach, and they demonstrate that malvidin 3-glucoside appeared in both portal and systemic plasma only 6 min after dosage.

In our study on blackcurrant anthocyanins, we found no aglycone-dependent differences in the absorption or excretion of the cyanindin and delphinidin anthocyanins in either rabbits or humans (Nielsen et al., 2003). However, in both species a significantly larger absorption was observed for the anthocyanin rutinosides than for the anthocyanin glucosides. This was also reflected in the urinary data, where the rutinosides were excreted to a significantly higher extent than the glucosides. These observations were supported in the literature, although the investigators did not address this matter specifically (Matsumoto et al., 2001; Netzel et al., 2001). An explanation for the lower bioavailability of the two anthocyanin glucosides than of the rutinosides might be that part of the anthocyanin glucosides are cleaved by the 3-glucosidases in the small intestine, resulting in the formation of their corresponding aglycone, as observed for other flavonoids (Day et al., 1998). The aglycons of anthocyanins are quickly degraded at elevated pH and are absorbed only in small quantities, if at all (Cao et al., 2001; Wu et al., 2002). Cleavage of part of the anthocyanin glucosides thus results in a larger proportion of the rutinosides being intact and accessible for absorption and distribution in plasma and urine.

Recently, Wu et al. (2002) observed trace amounts of anthocyanin meta­bolites in human urine after treatment of humans with large doses of an elder­berry extract containing 720 mg anthocyanins. The detected metabolites were the methylated and glucuronidated derivatives of the ingested anthocyanins, which is similar to the metabolic products of other flavonoids (Sesink et al., 2001, Zhu et al., 1994). In the same study by Wu et al., high doses of blueberries containing 690 mg anthocyanin did not generate detectable amounts of anthocyanin metabolites. It was speculated by the authors that this was due to the large number of different anthocyanins (at least 25) in blueberries, resulting in lower doses of each anthocyanin and consequently a metabolite concentration below the limit of detection. It is thus likely that too low doses and the low absorption of anthocyanins are responsible for the lack of positive identification of anthocyanin metabolites in other studies. This is supported by a recent study on strawberry anthocyanins containing mainly pelargonidin-3-glucoside (Felgines et al., 2003). In addition to pelargonidin-3-glucoside, HPLC-ESI­MS-MS (high-performance liquid chromatography — electron spray ionisation ­ mass spectroscopy — mass spectroscopy) studies revealed five anthocyanin metabolites in urine: three monoglucuronides of pelargonidin, one sulpho­conjugate of pelargonidin and pelargonidin itself. Total urinary excretion of strawberry anthocyanin metabolites corresponded to 1.80 Ô 0.29 per cent (mean Ô SEM, n = 6) of pelargonidin-3-glucoside ingested. More than 80 per cent of this excretion was present as a monoglucuronide. As soon as 4 hours after the meal, more than two-thirds of the anthocyanin metabolites had been excreted in urine, although the excretion of the metabolites continued until the end of the 24-h experiment.

The urinary excretion of anthocyanins in humans has only been followed up to 24h at the most (Cao et al., 2001), showing a total excretion of intact unmetabolised anthocyanins between 0.033 and 0.28 per cent (Netzel et al., 2001; Matsumoto et al., 2001; Frank et al., 2003). In animals, a total excretion of around 0.36±1 per cent has been observed (Morazzoni et al., 1991; Nielsen et al., 2003). We recently demonstrated that the biokinetics in both rabbits and humans were comparable, and that the urinary excretion correlated well with the amount determined in plasma (R2 = 0.773) (Nielsen et al., 2003). These data confirmed that very limited amounts of the anthocyanins are bioavailable in humans, presumably well below 0.5 per cent of a given anthocyanin dosage. If the metabolites, like the glucuronides observed by Felgines et al. (2003) are taken into account, the total absorption of anthocyanins is still around only a few per cent of the ingested dose. Based on this low bioavailability and rapid systemic elimination, a significant contribution to health protection of dietary anthocyanins thus seems questionable.

The flavonoids have been shown to exert a number of health beneficial properties in in vitro studies and in animal studies (for review see Harborne and Williams 2000; Nijveldt et al., 2001; Kris-Etherton et al., 2002). Epidemio­logical studies have, however, not been able to give a clear picture of the protective impact of dietary flavonoids against such diseases as cancer or coronary heart disease. In most cases, the limitation in the outcome of epidemiological studies on the health protective effects of flavonoids is the lack of precise flavonoid intake data.

The first and also one of the best attempts to generate precise flavonoid intake data was the Zutphen Elderly Study investigating the association between flavonoid intakes and risk of coronary heart diseases (Hertog et al., 1993a). This study was based on a thorough chemical analysis of the content of quercetin, kaempferol, myricetin, apigenin and luteolin in 28 vegetables and 9 fruits, tea infusions, wines, and fruit juices commonly consumed in the Netherlands (Hertog et al., 1992, 1993a). However, since the study included only these five dietary flavonoids, it covered only a fraction of the total amount of flavonoids found in our diet. The authors found that the total average intake of the flavonoids determined was 23 mg/day, with the flavonol quercetin as the most important dietary flavonoid (mean intake 16 mg/day) and with tea as the major flavonoid source (48 per cent of total intake), followed by onions (29 per cent), and apples (7 per cent). That the inclusion of especially the citrus flavonoids, but also of the catechins and anthocyanins in this study would have revealed other major flavonoid sources and altered the estimated flavonoid intake dramatically. Thus a stronger conclusion on the health effects of dietary flavonoid intake might have been drawn.

The majority of the epidemiological studies that followed this first attempt to investigate the health protective effects of dietary flavonoids have all been based on the same set-up or with even poorer tools to estimate the habitual flavonoid intake. For example, in the Seven Country Study, the average flavonoid intake was estimated by analyses of a few food samples that represented the average daily intake of flavonoid-containing foods in each country.

Additional flavonoid compounds, such as the citrus flavonoids and the catechins, were included in later epidemiological studies by Knekt et al. (2002) and Arts et al. (2001a,b), which investigated the association of these flavonoids with the risk of cardiovascular disease. However, in the study by Arts et al. on a cohort of postmenopausal women from Iowa, the estimated catechin intake was merely based on older analyses of catechin content in Dutch food (Arts et al., 2001b). Furthermore, the semi-quantitative food frequency questionnaire (FFQ) used, with 127 items, was not designed to evaluate catechin intake. Some items in the FFQ referred to more than one food, e.g. `fresh apples or pears’, and evaluation of the FFQ’s ability to asses catechin intake showed correlation coefficients between 0.45 and 0.83 for the main sources of catechins.

The drawback in prospective studies is thus often that the FFQ used at baseline is insufficient and too unspecific at the time of follow up, and this may cause misclassifications of the dietary exposure. Also the variation in the content of, for example, flavonoids in different types of foods due to different cultivars, seasonal variation, cooking and food production methods, is largely missed by the FFQs and this may also lead to misclassifications.

An alternative to calculated estimates of the intake of flavonoids in epidemiological studies would be the use of a biomarker that could assess the flavonoid intake in each individual by a simple measure in a blood or urine sample.

We recently developed a very selective and sensitive LC-MS methodology that is able to quantify 12 dietary flavonoids simultaneously in urine from unsupplemented subjects (Nielsen et al., 2000b). The methodology was applied on urine samples collected from 94 subjects on their habitual diet or eating a controlled diet either high or low in fruits, berries and vegetables for 6 weeks (Nielsen et al., 2002). In the intervention period with controlled dietary intakes, we found highly significant differences between the urinary excretion of all measured flavonoid aglycons on the high fruit and vegetable diet compared with the low. The correlation between the habitual intake of fruits and vegetable, determined by 3 days of food registration, with the total excretion of flavonoids was 0.35, p < 0.001.

In the same study, the traditional biomarker for fruit and vegetable intakes, the plasma carotenoids, showed a correlation of only 0.213 with intake of fruits and vegetables (Nielsen et al., 2002). Thus urinary flavonoids may not only be a valid marker for flavonoid intake, but also a useful biomarker for fruit and vegetable intake in general.

We are presently trying to further validate the use of the flavonoid biomarker, both as a marker for flavonoid intake and as a biomarker for fruit and vegetable intake in both dietary intervention studies and in cohort studies (Brevik et al., 2004; Krogholm et al., 2004). A study investigating the concentration of flavonoids in fasting plasma samples found that this parameter correlated significantly with the intake of flavonoids estimated by a 7-day dietary record (Radtke et al., 2002). The authors in this study concluded that the combination of dietary estimates and biomarker determinations may be the best approach for epidemiological research on the health effects of flavonoids. Two other dietary intervention studies using high doses of onions and tea also investigated the use of flavonoid concentrations in plasma and urine as a biomarker of intake and both concluded that the dietary intake of flavonoids can be estimated by these parameters (de Vries et al., 1998; Noroozi et al., 2000).

The epidemiological studies mainly investigated the protective effect of flavonols predominated by the intake of quercetin from tea. However, our studies and the study by Radtke et al. demonstrate the importance of including other flavonoids than the flavonols, especially the citrus flavonoids when investigating the intake of flavonoids, since they account for a major part of the daily flavonoid intake (Nielsen et al., 2002; Radtke et al., 2002).

Atherosclerosis is the primary cause of many cardiovascular diseases. Mortality from cardiovascular disease is the leading cause of death in the U.S., numbering 41.2 per cent of all deaths in 1997 (Reed, 2002). Atherosclerosis is an inflammatory disease that is characterised by the accumulation of lipids in the innermost layer, the intima, of the walls of large and medium-sized arteries (for reviews see Thompson, 1994; Steinberg, 1997). The process initiates by accumulation of modified low-density lipoproteins (LDL) in the intima and uptake of the modified LDL by macrophages. The LDL may be modified by oxidation by free radicals of enzyme-mediated oxidation resulting in aggregation within the intima (Brown and Goldstein, 1983). The trapped oxidised LDL initiates an inflammatory response of the endothelial cells that attracts monocytes to the area. The monocytes adhere to the endothelium, cross into the intima and differentiate into macrophages. Other macrophages engulf the oxidised LDL, leading to unregulated accumulation of LDL. These lipid-loaded macrophages are called foam cells and are the characteristic component of early atherosclerotic lesions called fatty streaks (Fuster, 1994). The foam cells are incapable of escaping the intima and then they undergo cell death (apoptosis). This may result in an oxidative burst that further contributes to the oxidation of LDL, thus aggravating the inflammatory process. The progression of the lesion involves smooth muscle cells, collagen and platelets in addition to further lipid accumulation (Steinberg, 1997). The advanced lesion may also develop a fibrous cap. The atherosclerotic lesions or plaques reduce the endothelial function, limit the effective diameter of the vessels, and thereby restrict the blood flow and the supply of oxygen to tissues. When the plaques grow larger, they have an increased tendency to rupture and cause thrombosis and sudden death by myocardial infarction (Fuster, 1994).

The uptake of oxidised LDL by macrophages and the subsequent formation of foam cells is an early event in the development of atheroscleroses. Antioxidants that can inhibit the oxidation of LDL may therefore potentially protect against atherosclerosis.

Flavonoids are potent antioxidants and have been shown to inhibit the oxidation of lipids both in vitro and ex vivo (De Whalley et al., 1990; Miura et al., 2000). Since the first study in 1990, where flavonoids were shown to be effective in protecting LDL from oxidation in vitro (De Whalley et al., 1990), at least 130 studies of the effects of flavonoids on lipoprotein oxidation have been published (Leake, 2001). Some flavonoids have been shown to be active in inhibiting LDL oxidation even at low concentrations (< 1 µM) that are physiologically achievable in plasma (De Whalley et al., 1990; Miura et al., 1994). Flavonoids and other antioxidants in black or green tea (Kasaoka et al., 2002), fruit juices (Aviram et al., 2002; Stein et al., 1999) and cocoa (Osakabe et al., 2000, 2001; Kondo et al., 1996) have been shown to inhibit LDL oxidation. Anthocyanins and anthocyanin rich red wine polyphenol fractions have also been shown to protect LDL against oxidation in vitro, and red wine has been found to limit the uptake of oxidised LDL by macrophages (SatueÂ-Gracia et al., 1997; Kerry and Abbey, 1997).

The flavonoids probably act in part as chain-breaking antioxidants, thereby converting lipid peroxyl or alkoxyl radicals to lipid hydroperoxides or hydroxides, respectively (Leake, 2001). In addition to this effect, the flavonoids are thought to protect a-tocopherol, the main endogenous antioxidant in LDL, from being consumed during LDL oxidation. The flavonoids are also able to convert the a-tocopheroxyl radical back into a-tocopherol (Leake, 2001).

In the past decade there has been a major shift in the paradigm of our understanding of the pathogenesis of atherosclerosis. It is now generally accepted that inflammatory mechanisms play a central role in mediating all phases of the development of atherosclerosis (Blake and Ridker, 2002). Since several of the potentially anti-atherogenic compounds in our diet have anti-inflammatory properties, their mechanisms of action may be by inhibiting or blocking the inflammatory processes of atherosclerosis.

Aggregation of platelets is known to contribute to the development of atherosclerosis by several mechanisms (Fuster et al., 1992) and the inhibition of platelet aggregation is thus regarded as beneficial. Platelets produce the pro-inflammatory mediators such as thromboxane A2, PAF and serotonin, and are thus key participants in the atherogenesis (Ross, 1993).

Several studies have investigated the effect of flavonoids on platelet activation and aggregation. These studies have been reviewed by Middleton and Kandascami (1994) and later by Harborne and Williams (2000).

Recent studies on the flavonoids in cocoa have shown that epicatechin and its related oligomers, the procyanidins, also have potent anti-inflammatory properties (Steinberg et al., 2003). The low-molecular weight procyanidins and epicatechin itself were shown to be a potent inhibitor of human 5­ lipoxygenase (Schewe et al., 2002) and procyanidins from cocoa were demonstrated to decrease platelet function significantly in vivo in humans (Murphy et al., 2003). Furthermore, another study showed that the combination of quercetin and catechin synergistically inhibited platelet function in collagen-induced platelet aggregation by antagonising the intracellular production of hydrogen peroxide (Pignatelli et al., 2000). A recent study on flavonoids and the platelet-activating factor (PAF) and related phospholipids in endothelial cells during oxidative stress showed that the flavonoids hesperedin, naringenin and quercetin were able to mediate these enzymes, and thereby limit the inflammatory response (Balestrieri et al., 2003).

Studies on anthocyanins have also demonstrated, that they are able to inhibit platelet aggregation. Treatment of humans with blueberry anthocyanins for 60 days was found to reduce the ex vivo platelet aggregation (Pulliero et al., 1989). This observation was supported in a study by Keevil et al. (2000) who after one week of treatment found a reduced platelet aggregation by red grape juice, but not by orange or grapefruit juice. Further indications of a beneficial effect were observed by Demrow et al. (1995) and Folts (1998) who found inhibitory effects on platelet aggregation in dogs and humans by red wine and red grape juice, but not by white wine, which could point to an anti-artherosclerotic effect of the anthocyanins.

The flavonoids may furthermore mediate other anti-inflammatory mechanisms involved in the development of cardiovascular disease. Studies indicate that they are implicated in the modulation of the monocyte adhesion in the inflammatory process of atherosclerosis. The expression of intercellular adhesion molecule-1 (ICAM-1), playing a pivotal role in the inflammatory response, was, for example, shown to be mediated by quercetin in human endothelial cells (Kobuchi et al., 1999). Koga and Meydani (2001) investigated the effects of plasma metabolites of (+)-catechin and quercetin on the modulation of monocyte adhesion to human aortic endothelial cells, and found that the plasma metabolites of catechin, but not of quercetin were potent inhibitors. This underlines the importance of investigating the biological effects of the metabolites present in vivo, e.g. the flavonoid glucuronic and sulphate conjugates instead of the parent compounds.

The endothelial function plays an important role in regulating the vascular function, and endothelial dysfunction is associated with increased cardiovascular disease risk. Several animal and human studies have shown that flavonoids also may have favourable effects on the vascular endothelial function as recently reviewed by Duffy and Vita (2003). Other biological effects of flavonoids in relation to cardiovascular disease have recently been reviewed by Nijveldt et al. (2001) and Kris-Etherton et al. (2002).

Since the Zutphen Study by Hertog et al. (1993a), a number of epidemiological studies have been undertaken on the association between dietary flavonoid intake and the risk of cardiovascular disease (CVD). The epidemiological studies show that the majority of the studies revealed an inverse association with the risk of CVD, although the outcomes of some of the studies are conflicting. Overall, the protective effect of flavonoids was strongest against the mortality of coronary heart disease (CHD), whereas the effect on risk of nonfatal incidences of CVD was weaker or non-existing.

The average daily flavonoid intake in the studies ranged from 2.6 to 28.6 mg/day, with quercetin as the dominating flavonoid in most of the studies. However, as discussed in section 9.5.1, the flavonoid intake in these epidemiological studies was based mainly on the food composition tables generated by Hertog et al. (1992, 1993b), covering only the content of selected flavonols and flavones in the food. If intake data on additional flavonoids had been included in these studies, e.g. the citrus flavonoids, the catechins, the anthocyanins and the isoflavonoids, the flavonol quercetin would probably not have been the major dietary flavonoid in the cohorts , tea would perhaps be a less important flavonoid source, and the outcome of these studies would then possibly have been different.

The quercetin intake originated mainly from tea intake, but apples and onions were also important sources of quercetin in some studies (Hertog et al., 1993a; Rimm et al., 1996a). The early studies by Hertog et al. showed a highly protective effect of both quercetin and tea against CVD (Hertog et al., 1993a, 1995; Keli et al., 1996). However, some of the later and larger cohort studies, trying to confirm these early studies, found no association or even aggravating effects of flavonoids and especially of tea consumption (Rimm et al., 1996a; Hertog et al., 1997; Hirvonen et al., 2000; Sesso et al., 2003).

The association of tea and incidences of CVD was later further investigated in several cohort studies. These studies have all been reviewed in a recent meta-analysis on the relationship between tea consumption and stroke, myocardial infarction and all coronary heart disease in 10 cohort studies and seven case-control studies (Peters et al., 2001). The incidence rate of myocardial infarction was concluded to be weakly inversely associated (11 per cent) with an increase in tea consumption of three cups per day. However, the authors stress that the heterogeneity of the studies and the risk of bias due to the larger number of smaller studies showing a protective effect, urge caution in interpreting this result. The mechanism of the protective effect of tea has recently been investigated and does, however, support a beneficial effects of tea intake. For example, consumption of 900 ml black tea for 4 weeks reversed the endothelial vasomotor dysfunction in patients with proven coronary artery disease (Duffy et al., 2001).

It has been suggested that the catechin content in tea could be the protective factor, and Arts et al. (2001a) thus estimated the catechin intake to 72 Ô 47.8 mg/ day in the Zutphen Elderly Study and found a significant negative association between ischaemic heart disease and intake of tea catechins. However, in another study on catechin intake by the same authors, in postmenopausal women from Iowa, a protective effect of catechins was seen from only dietary sources other than tea (Arts et al., 2001b).

The intake of red wine has been postulated to explain the French paradox, i.e. the low incidence of coronary heart disease in France despite the main risk factors for this disease being similar to those in northern European countries (Renaud and de Lorgeril, 1992). Anthocyanins are present in red wine, and several cohort studies have in fact suggested that wine drinkers have a lower mortality from CVD than others (Nanji, 1985; Renaud and de Lorgeril 1992; Gronbaek et al., 1995; Theobald et al., 2000). Other cohort studies have, however, found equally beneficial effects of all alcoholic beverages, and there is no general agreement on this matter as stated in the review by Rimm et al. (1996b). It has been proposed that the possible lower mortality by CVD in wine drinkers could be due in part to differences in lifestyle, e.g. in dietary habits and exercise, since factors such as dietary fat composition, little exercise and hypertension are major risk factors on the development of atherosclerosis (Tjonneland et al., 1999).

Flavonoid and anthocyanin intake may thus be strongly influenced by socio­economic factors, and these factors are themselves strongly associated with coronary heart disease. Residual confounding may therefore be a major problem in epidemiological studies on flavonoids (Leake, 2001). Overall, the picture of the health effects of flavonoids in relation to cardiovascular disease is somewhat inconsistent, although there seems to be an emerging body of evidence for a protective effect of dietary flavonoids. However, attempts to reveal the specific food item or flavonoid compound that may exert the protective effect have yet failed to give conclusive results.

The overall picture of the flavonoids as a protective agent against cardiovascular disease has been consolidated during the past decade. The mechanism of action of the flavonoids is, however, still unknown, but recent studies have moved the focus away from the antioxidant properties of the compounds towards a broader view on the potential mechanisms of action including especially the anti-inflammatory effects of flavonoids.

Furthermore, several studies have shown the importance of investigating the metabolism of the flavonoids and of elucidating the biological significance of these metabolites rather than of the parent compounds, the flavonoid aglycons that are of minor importance in vivo.

The early research on the dietary protective action of flavonoids has mainly focused on the flavonols, especially on quercetin, in part because of limitations in the available analytical methods at that time, which merely restricted the investigations to this class of compounds. However, within the past few years, flavonoid research has produced evidence for the importance of other dietary flavonoid classes and subgroups with potential health protective properties and with a similar or even greater impact on our total daily flavonoid intake. Examples are the citrus flavonoids, the red-coloured anthocyanins, the tea catechins and the procyanidins present in cocoa and wine. Furthermore, there are indications of the importance of a diet rich in a range of different flavonoids, rather than containing a high concentration of an individual compound, since some studies have shown additive or even synergistically effects of flavonoids (Pignatelli et al., 2000).

The inclusion of a broader range of the dietary flavonoids in future epidemiological studies, either by use of intake biomarkers and/or advanced food composition tables, will further elaborate on the importance of all these dietary compounds in relation to the risk of development of cardiovascular disease.

The research on flavonoids in relation to cardiovascular disease merits more and improved epidemiological studies, including additional flavonoid compounds, than the few flavonols and flavones that have been extensively investigated during the past decade. The majority of these studies found that the major dietary flavonoid was quercetin from tea, and these studies thus totally overlooked all the other flavonoids originating from fruits and vegetables. An increasing number of food composition tables on flavonoid content in foods and the development of new biomarkers would thus strengthen the epidemiological research on flavonoids and disease prevention. Biomarkers are very precise and often neglected tools for investigating biological effects of dietary compounds. A flavonoid biomarker would have the potential to reveal both the effect of the total intake of flavonoids, the effect of individual flavonoids, perhaps as markers of specific food items, and be useful to investigate the importance of flavonoids in combination with other dietary components.

Furthermore, disease-related biomarkers e.g. in relation to inflammatory mechanisms, should be included in future dietary intervention, case-control or cohort studies on flavonoids to further elaborate on the potential disease preventive mechanisms of these compounds.

An important issue for future research is to gain more information on what developmental stages of the cardiovascular disease the flavonoids are able to prevent or delay. This is crucial information for the planning of future epidemiological studies on the disease preventive effects of flavonoids, and may explain part of the inconsistence in the outcome of previous studies.

Aucun commentaire:

Enregistrer un commentaire