Epidemiologic (1–4) and clinical (5,6) evidence suggests that a plant-centered Mediterranean-type diet is associated with a lower incidence of cardiovascular diseases compared with an animal-centered Western-style diet. Because the reduced intake of saturated fats and cholesterol and the lower glycemic load (7) associated with a Mediterranean diet had no major impact on cardiovascular mortality (3–6,8), research focused on specific ingredients of plant-bearing foods that might explain the beneficial effects. One of these ingredients is the polyphenols, which are widespread among all higher plants but completely lacking in animals. They consist of flavonoids, phenolic acid derivatives, and their polymers, the tannins (9–11). Cardioprotective effects of flavonoids have been demonstrated already with relatively low dietary intakes (10–40 mg/d) in several epidemiologic studies (12–17), though not in all (18,19).
Usually, the anti-oxidant properties of the plant phenols (20) are thought to be responsible for the reported pharmacologic effects. As lipoprotein oxidation is believed to play a key role in the development of cardiovascular events (21,22) and in vitro findings unequivocally confirmed that polyphenols prolong the lag-time of low-density lipoprotein (LDL) oxidation (23–25), it has been suggested that the inhibition of the LDL oxidation would be the relevant mechanism underlying the in vivo activity of the polyphenols, as well (26). Although some clinical trials revealed a delayed ex vivo LDL oxidation after ingestion of foods rich in plant phenols (27–35), these results could not be reproduced by similarly conducted studies (25,36–48). Moreover, recently the principal relevance of the lipoprotein oxidation in the pathogenesis of atherosclerosis has been questioned (49–51). However, because the epidemiologic evidence strongly supports cardioprotective effects of the polyphenols, alternative modes of action have been considered.
Reports of several research groups show that extracts of plant foods as well as defined plant phenols are capable of relaxing isolated arterial vessels (52–66). In most cases vasorelaxations were dependent on the integrity of the vascular endothelium or diminished in the presence of inhibitors of the endothelial nitric oxide synthase, indicating that (certain) phenolic compounds may enhance the nitric oxide release of the vascular endothelium. Providing a sufficient bioavailability of the dietary phenols (67), this mechanism might contribute to the protective activity of plant-dominated diets, as on the one hand the constitutive endothelial nitric oxide production is known to be pivotal for keeping up the vascular homeostasis and on the other hand an impaired endothelial nitric oxide formation is one of the first symptoms appearing in the pathogenesis of vascular diseases (68–70). This is supported by clinical studies showing an improved endothelium-dependent relaxation in patients with endothelial dysfunction after ingestion of polyphenol-rich nutrients (71,72).
However, evidence from exploratory work is limited, because phenol-induced nitric oxide release has been measured only indirectly so far through its functional effects (e.g., vasorelaxation (59)) or biochemical effects (e.g., l-arginin-l-citrullin conversion (53), increase of the cGMP content in the vasculature (63)). Likewise, no consistent relationship between phenolic structures and the nitric oxide–dependent vasorelaxation has been established.
We present a method allowing the direct real-time registration of the phenol-stimulated nitric oxide release from isolated vessels that could be reliably correlated with the extent of vasorelaxation. Screening common plant phenols of different classes in this way, we were able to identify structural features that are prerequisites for a high nitric oxide activity.
Measurement of nitric oxide
Nitric oxide release was monitored amperometrically in real time by a nitric oxide–selective microsensor (ISO-NO-Meter, World Precision Instruments, Sarasota, FL, U.S.A.). Segments of porcine coronary arteries (PCAs) of approximately 2 cm in length were longitudinally cut and mounted in an organ bath (10 ml, 25°C) containing 2-[4-(2-hydroxyethyl)-1-piperazinyl]-ethenesulfonic acid (HEPES)/Krebs solution (in m M: NaCl 140.0, KCl 5.0, CaCl2 2.0, MgCl2 1.0, HEPES 10.0, d-glucose 10.0, adjusted to pH 7.4 with NaOH). The sensor was placed in a distance of 0.2–0.4 mm over the endothelial surface (half of the thickness of the vessel wall). After the amperometric signal had reached a baseline (15–30 min), individual reactivity of each vessel preparation was assessed by application of 15 n M of substance P, yielding an average peak nitric oxide concentration of 8.5 ± 0.2 n M (n = 658) (this value includes only PCAs in which the substance P–induced nitric oxide release exceeded the detection limit of 2.2 n M nitric oxide that we determined for this method); higher single doses of substance P caused no further increases in the nitric oxide release. Vessels were then washed by changing buffer solution and were left to be equilibrated for another 15–30 min before a test compound was added. Experiments were performed by cumulatively adding increasing concentrations of the test compound to the organ bath. Each compound was tested at a separate vessel. The concentration-response curves reflect the induced nitric oxide releases relative to the individual substance P reference of 15 n M (= 100%). After preincubation with 100 μM of the nitric oxide synthase inhibitor NG-monomethyl-l-arginine (l-NMMA), the maximal substance P–elicited response to the addition of cumulative increasing concentrations was reduced from Emax = 156.7 ± 29.4% (n = 15) to Emax = 57.1 ± 9.9% (n = 7).
Measurement of vascular tone
Changes in vascular tone were monitored isometrically on coronary ring segments of approximately 3–4 mm in length using a force transducer (TF6V5, W. Fleck, Mainz, Germany) and were recorded after amplification on a MacLab/8 system with the Chart V 3.2 software. Rings were mounted between two stainless steel triangles in constant-temperature (37°C) organ chambers filled with 10 ml of Tyrode solution (composed of in m M): Na+ 161.11, K+ 5.36, Ca2+ 1.80, Mg2+ 1.05, Cl− 147.95, HCO3− 23.80, H2PO4− 0.42, and d-glucose 10.0) and continuously aerated with carbogen. Resting tension was adjusted to 20 mN by mechanical stretching. After tension remained constant (≈60 min), the PCAs were precontracted with 25 μM prostaglandin F2α (PGF2α) and the individual reactivity of each vessel was tested by endothelium-dependent relaxation with 90 n M of substance P, yielding an average vasorelaxation (expressed as percent reversal of the PGF2α-induced precontraction) of 67.3 ± 2.7% (n = 191) PCAs, in which substance P–induced relaxations of < 30% were excluded. Higher concentrations of substance P elicited no further increases in relaxation, whereas 15 n M of substance P caused an average vasorelaxation of just ≈50%. After a washout period during which the incubation buffer was replaced two or three times, vessels equilibrated for an additional 30–60 min; then, following precontraction with 25 μM PGF2α, increasing concentrations of the test compound were added in a cumulative manner. The relaxation of each coronary ring in response to a concentration of a test compound is expressed in percent of the individual relaxation induced by the reference concentration of 90 n M substance P and finally standardized on the average substance P–elicited relaxation of 67.3% (by multiplication of the individual values with 0.673). The rationale for this standardization is to avoid maximum average values for relaxation of > 100% by more potent relaxing agents than substance P. After preincubation with 100 μM l-NMMA, the cumulative response to substance P was reduced from Emax = 71.6 ± 8.3% (n = 12) to Emax = 32.5 ± 8.0% (n = 8). Further contractions induced by a test compound are expressed as percentage of the precontraction with 25 μM PGF2α. Some rings were denuded of the endothelium by rubbing the luminal surface with a circular file; the absence of functional endothelium was tested by the lack of the substance P–elicited relaxation (< 10%).
The experimental parameters were set in such way as to mimic the physiologic pressure of 100 ± 20 mm Hg by loading the PCAs with a total force of just 25–35 mN (20 mN mechanical preload plus 5–15 mN isometric precontraction with PGF2α), corresponding to 30–40% of the isometric maximum (73), whereas in most prior experiments vessels were adjusted to higher preloads (≈80% of the maximum) (52,59,61).
All substances were of the highest analytic grade commercially available. The oligomeric proanthocyanidins are a standardized extract of Vitis vinifera seeds (Leucoselect, Indena SpA, Milan, Italy), mainly (80%) composed of dimers, trimers, and tetramers, partly esterfied with gallic acid (mean molecular weight 1,100 ± 80 D (CAS 85594–37–2). Leucocyanidol was provided by the Institut fuer Pharmazeutische Biologie (Universitaet Duesseldorf, Germany). The other plant phenols were purchased from Carl Roth GmbH KG (Karlsruhe, Germany). PGF2α (Dinoprost F2α) was from Pharmacia and Upjohn GmbH (Erlangen, Germany). All the other chemicals were from Sigma-Aldrich Chemie GmbH (Deisenhofen, Germany), and the solvents from E. Merck KGaA (Darmstadt, Germany). The tested substances were generally dissolved in HEPES/Krebs- or Tyrode-buffer containing maximal 30% (vol/vol) ethanol or 10% (vol/vol) dimethylsulfoxide to increase solubility; the final dilution contained not more than 0.15% (vol/vol) ethanol or 0.05% (vol/vol) dimethylsulfoxide, which did not affect nitric oxide formation or vascular tension. To avoid precipitation all solutions of plant phenols were prepared directly before usage. The anthocyan(idin)s were dissolved in HCl-acidified water (pH 3) to prevent carbinol or chalcon formation.
PCAs (Arteria coronaria dextra) were freshly dissected from the hearts of (castrated) male and female pigs (7–9 months old) at the local slaughterhouse.
Cumulative log dose-response data for nitric oxide and tonus measurements were fitted by sigmoidal regression using SigmaPlot for Windows version 5.00 software (SPSS, Inc., Richmond, CA, U.S.A.). Emax and pD2 values are derived from the regression curves and expressed as 95% CIs. Differences were considered statistically significant when CIs were not overlapping corresponding to a p value in a two-sided t test of p ≤ 0.01. Additionally, identity of Emax values under control conditions and after incubation with l-NMMA was determined by a two-sided t test. Average responses are principally given as arithmetic means ± 95% CIs. Usage of different statistical measures is specially indicated.
Effects of plant phenols on vascular nitric oxide formation
As summarized in Table 1 and Fig. 1, the screened phenols differently affected the nitric oxide concentration at the endothelial surface of PCAs: According to their maximal extent increasing the nitric oxide concentration, they could be separated into three categories. Category 1 (highly active) comprised those phenols inducing a maximal nitric oxide release of Emax > 100%, i.e., exceeding the nitric oxide formation caused by the substance P reference; this included the flavonols quercetin and myricetin, the flavan-3,4-diol leucocyanidol, and the oligomeric proanthocyanidins. As category 2 (moderately active) such phenols were classified as causing maximal nitric oxide levels lower than the substance P reference value but higher than the substance P–induced nitric oxide release in the presence of l-NMMA (i.e., 57.1% < Emax < 100%); these were caffeic acid, fisetin, and the quercetin-glycosids hyperosid and isoquercitrin. All the other tested phenols elicited less or nonsignificant increases of the endothelial nitric oxide release (i.e., Emax < 57.1%); they were assigned to category 3 (weak or nonactive). Corresponding Emax and pD2 values (= negative logarithm of the concentration eliciting half of the substance's own maximum) of all phenolic compounds with respect to their nitric oxide–releasing activity are presented in Table 1. As also demonstrated in Table 1, the nitric oxide formation induced by all category 1 and 2 phenols was significantly attenuated after pretreatment with the nonselective nitric oxide synthase inhibitor l-NMMA (100 μM). Concentration-response curves for the nitric oxide formation induced by representatives of each group of phenols are presented in Fig. 2A.
Effects of plant phenols on vascular tone
Those phenols, which had been identified as highly active concerning nitric oxide release (category 1 phenols), were also shown to be potent vascular relaxing agents (Table 2). In PGF2α-precontracted endothelium intact porcine coronary rings quercetin, myricetin, leucocyanidol, and oligomeric proanthocyanidins caused maximal relaxations, exceeding the relaxation induced by the substance P reference (i.e., Emax > 67.3%). The phenols moderately increasing the nitric oxide release (caffeic acid, fisetin, hyperosid, and isoquercitrin) caused maximal relaxations smaller than the substance P reference value but more than the substance P–induced relaxation in the presence of l-NMMA (i.e., 67.3% < Emax < 32.5%). The majority of the category 3 phenols induced either no changes in vascular tone (transresveratrol, morin, apigenin, naringenin, eriodictyol, epicatechin) or only weak, partly nonsignificant relaxations of maximal 10% (cinnamic acid, coumaric acid, ferulic acid, kaempferol, isorhamnetin, delphinidin) or minimal further increases in vascular tone up to 13.5% (chlorogenic acid, rutin, luteolin, epicatechingallate). Concentration-response curves for relaxation induced by representatives of each group of phenols are shown in Fig. 2B.
Previous exposure of the vessel rings to l-NMMA (100 μM) significantly attenuated the relaxing effects of the category 1 and 2 phenols (Table 2), whereas subsequent addition of l-NMMA (100 μM) reversed the phenol-induced relaxations, and a further dose of 100 μM l-NMMA caused a further increase of the vascular tone (data not shown). To the contrary, subsequent addition of indomethacin (10 μM) or miconazole (10 μM) was without effect on the maximal (cumulative) relaxation of the phenols (n = 3–4). In endothelium-denuded rings the phenol-induced relaxations were almost completely abolished (e.g., the residual relaxation to quercetin, then, was Emax = 8.2 ± 4.7%, n = 6) and to the proanthocyanidins Emax = 3.3 ± 1.2%, n = 6).
Some of the nitric oxide–inactive polyphenols, however, caused vasorelaxations (catechin, gallic acid) or further vasocontractions (malvidin, oenin), independently of the existence of a functional endothelium (data not shown).
The results of the nitric oxide registrations indicate that some of the phenols commonly occurring in plant-derived foods are capable of enhancing the nitric oxide release of endothelium-preserved PCAs. Correlating the nitric oxide with the tonus measurements reveals that all phenols inducing a nitric oxide release above a certain threshold (category 1 and 2 phenols) may cause vasorelaxations. The increase in bioavailabilty of nitric oxide from the vascular endothelium seems to be responsible for the relaxations by these phenols, because the relaxations were significantly attenuated in the presence of l-NMMA (i.e., Emax values were reduced by at least half); l-NMMA diminished relaxations to a similar extent as the peak nitric oxide formation (by 50–75%); and relaxations were completely abolished on endothelium-denuded rings. The residual phenol-induced relaxations evolving after preincubation with 100 μM l-NMMA are likely due to an incomplete inhibition of the nitric oxide synthase, as the nitric oxide stimulation, likewise, could not be entirely prevented by 100 μM l-NMMA, and postincubation with an additional 100 μM l-NMMA (after the phenol-induced relaxation and a first dose of l-NMMA) exhibited further elevations of vascular tone. However, these findings do not exclude the possibility that other endothelium-derived autacoids, in particular prostacyclin (PGI2) or endothelium-derived hyperpolarizing factors (EDHFs), may contribute to the vasorelaxing activity of plant phenols. The stimulation of the synthesis of relaxing prostanoids (PGI2) appears not to be involved, as the cyclooxygenase inhibitor indomethacin did not reverse the phenol-induced relaxations. In porcine and some other (including bovine and human) coronary arteries cytochrome P-450 2C–derived epoxyeicosatrienoic acids (EETs) are supposed to function as the EDHF opening large-conductance, Ca2+-dependent K+ channels in vascular smooth muscle cells, thereby mediating nitric oxide– and PGI2-independent vasorelaxations (74,75). This pathway, too, seems not to be induced by plant phenols, because miconazole, which has previously been shown to inhibit EET-dependent vasorelaxations (76), did not reduce the relaxations to the phenols. Moreover, because nitric oxide was found to attenuate the formation of EET, it was supposed that EDHF-mediated relaxations may only occur under conditions of impaired nitric oxide production (such as inhibition with l-NMMA) (77,78). Thus, the rapid plant phenol–induced elevation of nitric oxide concentrations should inhibit, rather than stimulate, EDHF release.
Previous work suggested that activation of the cGMP-dependent vasorelaxation under physiologic conditions requires a minimal nitric oxide concentration in the media of 5 n M (= Km value of the soluble guanylate cyclase) (79), which corresponds in our in vitro measuring system to a nitric oxide formation of 60% of the substance P control. As nitric oxide underlies an unhindered isotropic diffusion (80) and the nitric oxide scavenger hemoglobin (81) was absent, the luminal nitric oxide concentrations we registered in a distance of half of the vessel wall thickness to the endothelial surface should be in the same order of magnitude as the abluminal nitric oxide concentration in the media. Because in the myographic experiments the coronary rings were set to a physiologic tension, the height of a plant phenol–induced nitric oxide release should be a reliable predictor for its relaxing properties (possibly also under in vivo conditions). This theoretical deduction is in good agreement with our experimental data. Phenols causing maximal nitric oxide stimulations in the range of 40–50% (like cinnamic acid, kaempferol, or epicatechingallate) showed just marginal (endothelium-dependent) relaxations < 10%, whereas phenols leading to Emax (nitric oxide) values in the range of 60% (e.g., isoquercitrin as well as substance P or quercetin after inhibition of the nitric oxide synthase with l-NMMA) already showed relaxations of > 30%. Accordingly, for the active phenols (categories 1 and 2) a close linear correlation between Emax values of nitric oxide release and relaxation (r = 0.93, p < 0.01) was observed and also (to a lesser extent) between corresponding pD2 values (r = 0.77, p < 0.05). However, the half-maximal concentrations exerting nitric oxide elevations were three to five orders of magnitude lower than those causing relaxations; a minor difference was observed for substance P (pD2 NO = 9.12 ± 0.17 versus pD2 relaxation = 7.98 ± 0.07, −log[M]). The major reason may be that low doses of a phenol (although stimulating a significant endothelial nitric oxide release) are not sufficient to elevate nitric oxide levels in the vasculature above the relaxing threshold of 5 n M. Furthermore, the lower incremental increase of the nitric oxide release induced by increasing phenol concentrations compared with substance P (curve not shown) requires higher doses to generate a diffusion gradient yielding a steady state of ≥ 5 n M nitric oxide (80). Additionally, the different conditions of nitric oxide and tonus measurements may cause differences in sensitivity. One point is the lower temperature of 25°C used in the nitric oxide assay. Although peak nitric oxide levels were similar at 37°C, the nitric oxide signal tended to relax to baseline more rapidly, presumably due to the lower nitric oxide gas solubility at higher temperatures. Hence, the reduced duration of the nitric oxide flux might be too short for a significant accumulation in the media when low phenol concentrations were used (however, routine nitric oxide measurements could not be performed at 37°C because of the high noise and low baseline stability of the amperometric signal; tonus measurements, in turn, could not be performed at temperatures < 37°C because relaxations were considerably attenuated). Another point is the higher oxygen concentration in the tonus assay due to aeration with carbogen, which might accelerate the oxidation of nitric oxide and thereby diminish its bioavailability (82).
Our results are in the line with findings of Fitzpatrick et al. (56–59) and Andriambeloson et al. (60,61), who also identified quercetin, leucocyanidol, and oligomeric proanthocyanidins as endothelium-dependent relaxing ingredients of plant nutritions. Additionally, we demonstrated nitric oxide–stimulating and -relaxing activities for the quercetin glycosides hyperosid and isoquercitrin, the flavonol myricetin, and caffeic acid.
It is evident from the examination of different polyphenols that the nitric oxide–releasing and therefore the relaxing activity are confined to certain structural features. The highly active phenols (quercetin, myricetin, leucocyanidol, and oligomeric proanthocyanidins) share a basic structure: a flavan moiety with free hydroxyl residues at C3, C3′ C4′, C5, and C7 and a hydroxyl-, oxo-, or phenolic substituent at C4 (Fig. 3). The o-dihydroxy-group in ring B (C3′-, C4′-OH) emerged as indispensable for the flavonoid-induced nitric oxide production, because removal (kaempferol) or methylation (isorhamnetin) of one OH group or meta-positioning of the OH groups as, in morin, resulted in an almost complete loss of activity, when compared with quercetin. In contrast, introduction of an additional OH group at C5′ of ring B (o-trihydroxy structure) enhanced the activity (myricetin). The OH substituent at C3 of ring C was also shown to be necessary for the nitric oxide–stimulating effects of the flavonoids; its replacement by hydrogen (luteolin) was associated with a lack in nitric oxide response. This finding corresponds with results of Chan et al. (65), who reported that only the flavonols (i.e., flavonoids with hydroxyl group at C3) but not related flavones (without hydroxyl group at C3) exhibited endothelium-dependent vasorelaxations at the rat aorta. However, unlike Chan et al. (65), we and others (55,56,61,64,66) could not confirm a major endothelium-independent relaxing potency for flavonols or flavones, which may be due to different experimental protocols or different species. The glycosylation of the C3-OH function of the quercetin with a monosaccharide (hyperosid, isoquercitrin) resulted in a reduction of the nitric oxide activity by 30–50%; when a longer sugar chain was introduced (rutin), the nitric oxide response declined under the threshold where a relaxation could be induced (Emax < 5 n M nitric oxide). An attenuated relaxation due to glycosylation of flavonoids has been also reported by Hammad and Abdalla (83) using isolated rat ileum as functional target. This may be a result of the decrease in hydrophobicity (membrane permeability) due to the introduction of hydrophilic sugar residues in the hydrophobic molecules (84). Contribution of the hydroxyl-group at C5 of ring A to the nitric oxide activity of the flavonols seems to be marginal, as their replacement by hydrogen (fisetin), in comparison with quercetin, was associated with a decrease in the maximal effect of only 20%. To the contrary, replacement of the oxygen function or the phenolic moiety (proanthocyanidins) at C4 of ring C by hydrogen resulted in the loss of nitric oxide activity (flavonols, anthocyanidins). The latter is in principal agreement with the observations of Andriambeloson et al. (61). However, they registered for delphinidin a strong endothelium-dependent relaxation, which was not seen in our experiments. The divergent finding may be explained by the instability of nonacidified anthocyanidin solutions. At physiologic pH values those are rapidly transformed in uncharged carbenol-2- and 4-adducts (85). The carbenol-4-adduct of delphinidin exerted a nitric oxide–dependent vasorelaxant activity (data not shown), as was to be expected due to the structural analogy to myricetin. Flavonols (quercetin) and flavan-3,4-diols (leucocyanidol) induced similar nitric oxide releases, indicating that the enone structure with the α-positioned OH group of the flavonols is equieffective to the vicinal hydroxyl structure of the flavandiols. The reason may be that flavonols are in a tautomeric equilibrium to the vicinal endiol (9). Comparison between quercetin and leucocyanidol further suggests that a double bond in ring C may not be relevant for the nitric oxide activity of flavonoids, which is also supported by the finding that no differences in nitric oxide activity were observed between anthocyanidins (1→2- and 3→4-double bond), flavones (2→3-double bond), and flavanones (saturated C-ring). In contrast to the monomeric flavan-3-ols (catechin, epicatechin, epicatechingallate), their oligomeric condensates, the proanthocyanidins, revealed themselves as strong nitric oxide–stimulating agents. A cause may be that in the proanthocyanidins the flavonol monomers are connected through C4-C8- or C4-C6-bonds, resulting in the formation of structural unities with phenylogic or vinylogic OH groups at C4 position, which correspond to the leucocyanidol structure. In relaxation experiments with proanthocyanidins from grape seeds performed on rat aorta, Fitzpatrick et al. (59) also measured strong endothelium-dependent relaxations only with the oligomeric compounds; trimers, tetramers, and pentamers showed the highest activities with medium effective concentrations (EC50s) between 1–10 μM, which are in the range of the value (EC50 = 1.3 μM) we determined for a total extract of dimeric, trimeric, and tetrameric grape seed proanthocyanidins.
Interestingly, caffeic acid still showed a moderate nitric oxide activity, indicating that the flavan moiety is not a necessary structural prerequisite for nitric oxide–stimulating plant phenols. Accordingly, neither ring A with the OH-substituents at C5 and C7 nor the oxa bridge in ring C are essential, whereas the catechol-ring of caffeic acid may mimic ring B and the α,β-unsaturated carbonic acid structure the α-hydroxylated enone structure of the active flavonols, as can be inferred from superimposing caffeic acid with quercetin (Fig. 4). Removal or derivatization of one catechol-OH group (coumaric acid, ferulic acid) or of the carbonyl function (chlorogenic acid) of caffeic acid (or of both as in the case of resveratrol) was associated with a complete loss in vasoactivity, the same as we observed after related alterations of the quercetin molecule. As the removal of the vinyl bridge (gallic acid) resulted in a similar decrease in the nitric oxide activity, it is tempting to propose that the caffeic acid molecule represents minimally structural prerequisites for a native nitric oxide–stimulating plant phenol.
Several mechanisms have been proposed to explain the plant phenol–induced nitric oxide release of the vascular endothelium. A prominent one is that phenols exert an anti-oxidative potential to prevent the degradation of nitric oxide by superoxide (O2•−) leading to elevations of the basal nitric oxide levels (55,86,87). Although the category 1 and 2 phenols were previously shown to exhibit strong O2•−-scavenging activities in in vitro experiments (88–90), in the same studies most of the weak nitric oxide–stimulating phenols were revealed to be similarly effective or even stronger O2•− scavengers, indicating that oxygen radical scavenging alone does not substantially contribute to the nitric oxide–increasing activity of polyphenols. One reason may be that the rate constants of the reaction between the phenols and O2•− are much lower (data not shown) than for the rapid, nearly diffusion-controlled degradation of O2•− with superoxide dismutase (91), which is normally abundant at high concentrations around 10 μM inside the endothelium (92) as well as in the extracellular matrix of the vessel wall (93), thereby limiting the steady-state concentrations of O2•− to values < 0.1 n M (94). Hence we suppose that the phenols are increasing the nitric oxide bioavailability by stimulating the nitric oxide production rather than by inhibiting its degradation (at least under conditions of saturating superoxide dismutase concentrations). Our findings allow some further considerations: Because the nitric oxide elevation was confined to such phenols with a common structure and almost abolished after minimal alterations, the mode of the phenolic action seems to be highly specific. As picomolar concentrations of the phenols were sufficient to stimulate a nanomolar nitric oxide release and the nitric oxide signal increased within a few seconds after a phenol was added, they might stimulate a pathway of the amplifying transduction cascade, leading to the activation of endothelial nitric oxide synthase. The acute onset of the nitric oxide release and its rapid attenuation (t1/2 ≤ 5 min) may also exclude the possibility that an enhanced expression of the nitric oxide synthase protein is involved in the stimulating effect of the phenols. Nevertheless, the exact mechanism of the phenol-induced nitric oxide stimulation is still elusive.
The clinical relevance of the nitric oxide–releasing and -relaxing activity of plant phenols clearly depends on their systemic availability. Although polyphenols show a limited intestinal absorption and are intensively metabolized, evidence from recent intervention studies indicates that after consumption of 10–100 mg of an individual phenol (a usual dose in a plant-rich meal), maximum concentrations of 0.1–1 μM are recovered unchanged in plasma (67). These concentrations exceed the EC50 values we determined for the nitric oxide activity and are in the range in which the PCAs began to relax. Moreover, it was demonstrated that repeated application of the same phenol dose yields (cumulative) plasma levels which are four- to fivefold higher compared with the single application (44). Conversely, it must be considered that natural polyphenols predominantly occur as glycosides, which were shown to be also more efficiently absorbed (by active transport from the small intestine) than the aglycones and therefore reach higher plasma concentrations (95,96), whereas the aglycones were the more vasoactive compounds. Additionally, it is questionable whether oligomeric phenols, like the proanthocyanidins, which exhibited the highest vasoactivity and are abundant in the highest concentrations in most plant foods (97), may enter the systemic circulation intact: There are reports suggesting complete decomposition before absorption (98,99) but also a report recovering high plasma levels of proanthocyanidins (≈10 μM) in rats fed a proanthocyanidin-rich extract (100). Consequently, a transmission of our in vitro findings to putative in vivo effects of plant phenols should be undertaken with extreme care and requires an evaluation in clinical trials.
This study demonstrates a strong correlation between the plant phenol–induced nitric oxide release in the vascular endothelium and the extent of vasorelaxation. As endothelium-independent relaxations (after pharmacologic inhibition of the nitric oxide synthase or mechanical removal of the endothelium) were considerably smaller, most phenols are supposed to dilate blood vessels predominantly through activation of the nitric oxide/cGMP-pathway. However, a significant nitric oxide release alone is not sufficient to trigger vasorelaxation; only phenols increasing the nitric oxide level in the vasculature above a threshold of about 5 n M exhibited a relaxing activity. Such a supraliminal nitric oxide induction is confined to phenols with common structural elements very sensitive to derivatizations and therefore probably independent of the anti-oxidant properties of the phenols. Despite their very low bioavailability (≤1%), the recovered plasma concentrations after ingestion of nutritive doses of phenols are reported to be in the nanomolar to low micromolar range, which may be sufficient to cause significant nitric oxide elevations and even vasorelaxations. Considering the limitations of in vitro data for the prediction of eventual clinical outcomes, the structure-activity relations presented here may serve as a basis for recommendations of plant nutrients rich in nitric oxide–active phenols (e.g., grape juice, black tea, apples, or chocolate), direct supplementations with those phenols, or for the development of structurally related drugs to prevent conditions involving endothelial dysfunction.
1. Knekt P, Reunanen A, Jarvinen R, et al. Antioxidant vitamin intake and coronary mortality in a longitudinal population study. Am J Epidemiol 1994; 139:1180–9.
2. Gaziano JM, Manson JE, Branch LG, et al. A prospective study of consumption of carotenoids in fruits and vegetables and decreased cardiovascular mortality in the elderly. Ann Epidemiol 1995; 5:255–60.
3. Liu S, Manson JE, Lee IM, et al. Fruit and vegetable intake and risk of cardiovascular disease
: the Women's Health Study. Am J Clin Nutr 2000; 72:922–8.
4. Liu S, Lee IM, Ajani U, et al. Intake of vegetables rich in carotenoids and risk of coronary heart disease in men: the Physicians' Health Study. Int J Epidemiol 2001; 30:130–5.
5. Singh RB, Rastogi SS, Niaz MA, et al. Effect of fat-modified and fruit- and vegetable-enriched diets on blood lipids in the Indian Diet Heart Study. Am J Cardiol 1992; 70:869–74.
6. de Lorgeril M, Salen P, Martin JL, et al. Mediterranean diet
, traditional risk factors, and the rate of cardiovascular complications after myocardial infarction: final report of the Lyon Diet Heart Study. Circulation 1999; 99:779–85.
7. Jenkins DJ, Kendall CW, Axelsen M, et al. Viscous and nonviscous fibers, nonabsorbable and low glycaemic index carbohydrates, blood lipids and coronary heart disease. Curr Opin Lipidol 2000; 11:49–56.
8. Hooper L, Summerbell CD, Higgins JP, et al. Dietary fat intake and prevention of cardiovascular disease
: systematic review. BMJ 2001; 322:757–63.
9. Harborne JB. Plant phenolics. Methods in plant biochemistry
, vol. 1. London: Academic Press, 1989.
10. Harborne JB. The flavonoids: advances in research since 1986.
London: Chapman and Hall, 1994.
11. Bravo L. Polyphenols: chemistry, dietary sources, metabolism, and nutritional significance. Nutr Rev 1998; 56:317–33.
12. Hertog MG, Feskens EJ, Hollman PC, et al. Dietary antioxidant flavonoids
and risk of coronary heart disease: the Zutphen Elderly Study. Lancet 1993; 342:1007–11.
13. Knekt P, Jarvinen R, Reunanen A, et al. Flavonoid intake and coronary mortality in Finland: a cohort study [see comments]. BMJ 1996; 312:478–81.
14. Yochum L, Kushi LH, Meyer K, et al. Dietary flavonoid intake and risk of cardiovascular disease
in postmenopausal women. Am J Epidemiol 1999; 149:943–9.
15. Sesso HD, Gaziano JM, Buring JE, et al. Coffee and tea intake and the risk of myocardial infarction. Am J Epidemiol 1999; 149:162–7.
16. Geleijnse JM, Launer LJ, Hofman A, et al. Tea flavonoids
may protect against atherosclerosis: the Rotterdam Study. Arch Intern Med 1999; 159:2170–4.
17. Hirvonen T, Pietinen P, Virtanen M, et al. Intake of flavonols and flavones and risk of coronary heart disease in male smokers. Epidemiology 2001; 12:62–7.
18. Rimm EB, Ascherio A, Giovannucci E, et al. Vegetable, fruit, and cereal fiber intake and risk of coronary heart disease among men. JAMA 1996; 275:447–51.
19. Hertog MG, Sweetnam PM, Fehily AM, et al. Antioxidant flavonols and ischemic heart disease in a Welsh population of men: the Caerphilly Study. Am J Clin Nutr 1997; 65:1489–94.
20. Rice-Evans CA, Miller NJ, Bolwell PG, et al. The relative antioxidant activities of plant-derived polyphenolic flavonoids
. Free Radic Res 1995; 22:375–83.
21. Witztum JL, Steinberg D. Role of oxidized low density lipoprotein in atherogenesis. J Clin Invest 1991; 88:1785–92.
22. Diaz MN, Frei B, Vita JA, et al. Antioxidants and atherosclerotic heart disease. N Engl J Med 1997; 337:408–16.
23. Abu-Amsha R, Croft KD, Puddey IB, et al. Phenolic content of various beverages determines the extent of inhibition of human serum and low-density lipoprotein oxidation in vitro: identification and mechanism of action of some cinnamic acid derivatives from red wine. Clin Sci (Colch) 1996; 91:449–58.
24. Teissedre PL, Frankel EN, Waterhouse AL, et al. Inhibition of in vitro human LDL oxidation by phenolic antioxidants from grapes and wines. J Sci Food Agriculture 1996; 70:55–61.
25. O'Reilly JD, Sanders TA, Wiseman H. Flavonoids
protect against oxidative damage to LDL in vitro: use in selection of a flavonoid rich diet and relevance to LDL oxidation resistance ex vivo? Free Radic Res 2000; 33:419–26.
26. Fuhrman B, Aviram M. Flavonoids
protect LDL from oxidation and attenuate atherosclerosis. Curr Opin Lipidol 2001; 12:41–8.
27. Kondo K, Matsumoto A, Kurata H, et al. Inhibition of oxidation of low-density lipoprotein with red wine. Lancet 1994; 344:1152.
28. Fuhrman B, Lavy A, Aviram M. Consumption of red wine with meals reduces the susceptibility of human plasma and low-density lipoprotein to lipid peroxidation [see comments]. Am J Clin Nutr 1995; 61:549–54.
29. Miyagi Y, Miwa K, Inoue H. Inhibition of human low-density lipoprotein oxidation by flavonoids
in red wine and grape juice. Am J Cardiol 1997; 80:1627–31.
30. Nigdikar SV, Williams NR, Griffin BA, et al. Consumption of red wine polyphenols reduces the susceptibility of low-density lipoproteins to oxidation in vivo [see comments]. Am J Clin Nutr 1998; 68:258–65.
31. Natella F, Ghiselli A, Guidi A, et al. Red wine mitigates the postprandial increase of LDL susceptibility to oxidation. Free Radic Biol Med 2001; 30:1036–44.
32. Day AP, Kemp HJ, Bolton C, et al. Effect of concentrated red grape juice consumption on serum antioxidant capacity and low-density lipoprotein oxidation. Ann Nutr Metab 1997; 41:353–7.
33. Aviram M, Dornfeld L, Rosenblat M, et al. Pomegranate juice consumption reduces oxidative stress, atherogenic modifications to LDL, and platelet aggregation: studies in humans and in atherosclerotic apolipoprotein E-deficient mice. Am J Clin Nutr 2000; 71:1062–76.
34. Ishikawa T, Suzukawa M, Ito T, et al. Effect of tea flavonoid supplementation on the susceptibility of low-density lipoprotein to oxidative modification. Am J Clin Nutr 1997; 66:261–6.
35. Miura Y, Chiba T, Miura S, et al. Green tea polyphenols (flavan 3-ols) prevent oxidative modification of low density lipoproteins: an ex vivo study in humans. Nutr Biochem 2000; 11:216–22.
36. Sharpe PC, McGrath LT, McClean E, et al. Effect of red wine consumption on lipoprotein (a) and other risk factors for atherosclerosis. Q J Med 1995; 88:101–8.
37. de Rijke YB, Demacker PN, Assen NA, et al. Red wine consumption does not affect oxidizability of low-density lipoproteins in volunteers. Am J Clin Nutr 1996; 63:329–34.
38. Carbonneau MA, Leger CL, Monnier L, et al. Supplementation with wine phenolic compounds increases the antioxidant capacity of plasma and vitamin E of low-density lipoprotein without changing the lipoprotein Cu(2+)-oxidizability: possible explanation by phenolic location. Eur J Clin Nutr 1997; 51:682–90.
39. van Golde PH, Sloots LM, Vermeulen WP, et al. The role of alcohol in the anti low density lipoprotein oxidation activity of red wine. Atherosclerosis 1999; 147:365–70.
40. Caccetta RA, Croft KD, Beilin LJ, et al. Ingestion of red wine significantly increases plasma phenolic acid concentrations but does not acutely affect ex vivo lipoprotein oxidizability. Am J Clin Nutr 2000; 71:67–74.
41. Abu-Amsha Caccetta R, Burke V, Mori TA, et al. Red wine polyphenols, in the absence of alcohol, reduce lipid peroxidative stress in smoking subjects. Free Radic Biol Med 2001; 30:636–42.
42. McAnlis GT, McEneny J, Pearce J, et al. Black tea consumption does not protect low density lipoprotein from oxidative modification. Eur J Clin Nutr 1998; 52:202–6.
43. van het Hof KH, de Boer HS, Wiseman SA, et al. Consumption of green or black tea does not increase resistance of low-density lipoprotein to oxidation in humans. Am J Clin Nutr 1997; 66:1125–32.
44. van het Hof KH, Wiseman SA, Yang CS, et al. Plasma and lipoprotein levels of tea catechins following repeated tea consumption. Proc Soc Exp Biol Med 1999; 220:203–9.
45. Princen HM, van Duyvenvoorde W, Buytenhek R, et al. No effect of consumption of green and black tea on plasma lipid and antioxidant levels and on LDL oxidation in smokers. Arterioscler Thromb Vasc Biol 1998; 18:833–41.
46. Hodgson JM, Puddey IB, Croft KD, et al. Acute effects of ingestion of black and green tea on lipoprotein oxidation. Am J Clin Nutr 2000; 71:1103–7.
47. O'Reilly JD, Mallet AI, McAnlis GT, et al. Consumption of flavonoids
in onions and black tea: lack of effect on F2-isoprostanes and autoantibodies to oxidized LDL in healthy humans. Am J Clin Nutr 2001; 73:1040–4.
48. Young JF, Dragsted LO, Daneshvar B, et al. The effect of grape-skin extract on oxidative status. Br J Nutr 2000; 84:505–13.
49. Witting P, Pettersson K, Ostlund-Lindqvist AM, et al. Dissociation of atherogenesis from aortic accumulation of lipid hydro(pero)xides in Watanabe heritable hyperlipidemic rabbits. J Clin Invest 1999; 104:213–20.
50. Heinecke JW. Is the emperor wearing clothes? Clinical trials of vitamin E and the LDL oxidation hypothesis. Arterioscler Thromb Vasc Biol 2001; 21:1261–4.
51. Halliwell B. Lipid peroxidation, antioxidants and cardiovascular disease
: how should we move forward? Cardiovasc Res 2000; 47:410–8.
52. Cishek MB, Galloway MT, Karim M, et al. Effect of red wine on endothelium-dependent relaxation in rabbits. Clin Sci (Colch) 1997; 93:507–11.
53. Karim M, McCormick K, Kappagoda CT. Effects of cocoa extracts on endothelium-dependent relaxation. J Nutr 2000; 130:2105S–8S.
54. Duarte J, Perez Vizcaino F, Utrilla P, et al. Vasodilatory effects of flavonoids
in rat aortic smooth muscle: structure-activity relationships. Gen Pharmacol 1993; 24:857–62.
55. Duarte J, Jimenez R, Villar IC, et al. Vasorelaxant effects of the bioflavonoid chrysin in isolated rat aorta. Planta Med 2001; 67:567–9.
56. Fitzpatrick DF, Hirschfield SL, Coffey RG. Endothelium-dependent vasorelaxing activity of wine and other grape products. Am J Physiol 1993; 265:H774–8.
57. Fitzpatrick DF, Hirschfield SL, Ricci T, et al. Endothelium-dependent vasorelaxation
caused by various plant extracts. J Cardiovasc Pharmacol 1995; 26:90–5.
58. Fitzpatrick DF, Bing B, Rohdewald P. Endothelium-dependent vascular effects of Pycnogenol. J Cardiovasc Pharmacol 1998; 32:509–15.
59. Fitzpatrick DF, Fleming RC, Bing B, et al. Isolation and characterization of endothelium-dependent vasorelaxing compounds from grape seeds. J Agric Food Chem 2000; 48:6384–90.
60. Andriambeloson E, Kleschyov AL, Muller B, et al. Nitric oxide
production and endothelium-dependent vasorelaxation
induced by wine polyphenols in rat aorta. Br J Pharmacol 1997; 120:1053–8.
61. Andriambeloson E, Magnier C, Haan-Archipoff G, et al. Natural dietary polyphenolic compounds cause endothelium-dependent vasorelaxation
in rat thoracic aorta. J Nutr 1998; 128:2324–33.
62. Andriambeloson E, Stoclet JC, Andriantsitohaina R. Mechanism of endothelial nitric oxide
induced by wine polyphenols in rat thoracic aorta. J Cardiovasc Pharmacol 1999; 33:248–54.
63. Flesch M, Schwarz A, Bohm M. Effects of red and white wine on endothelium-dependent vasorelaxation
of rat aorta and human coronary arteries. Am J Physiol 1998; 275:H1183–90.
64. Chen ZY, Zhang ZS, Kwan KY, et al. Endothelium-dependent relaxation induced by hawthorn extract in rat mesenteric artery. Life Sci 1998; 63:1983–91.
65. Chan EC, Pannangpetch P, Woodman OL. Relaxation to flavones and flavonols in rat isolated thoracic aorta: mechanism of action and structure-activity relationships. J Cardiovasc Pharmacol 2000; 35:326–33.
66. Kim SH, Kang KW, Kim KW, et al. Procyanidins in crataegus extract evoke endothelium-dependent vasorelaxation
in rat aorta. Life Sci 2000; 67:121–31.
67. Scalbert A, Williamson G. Dietary intake and bioavailability of polyphenols. J Nutr 2000; 130:2073S–85S.
68. Vanhoutte PM. Endothelial dysfunction and atherosclerosis. Eur Heart J 1997; 18:E19–29.
69. Shimokawa H. Primary endothelial dysfunction: atherosclerosis. J Mol Cell Cardiol 1999; 31:23–37.
70. Kelm M, Rath J. Endothelial dysfunction in human coronary circulation: relevance of the l -arginine-NO pathway. Basic Res Cardiol 2001; 96:107–27.
71. Stein JH, Keevil JG, Wiebe DA, et al. Purple grape juice improves endothelial function and reduces the susceptibility of LDL cholesterol to oxidation in patients with coronary artery disease. Circulation 1999; 100:1050–5.
72. Duffy SJ, Keaney Jr, JF Holbrook M, et al. Short- and long-term black tea consumption reverses endothelial dysfunction in patients with coronary artery disease. Circulation 2001; 104:151–6.
73. Lalli J, Harrer JM, Luo W, et al. Targeted ablation of the phospholamban gene is associated with a marked decrease in sensitivity in aortic smooth muscle. Circ Res 1997; 80:506–13.
74. Fisslthaler B, Popp R, Kiss L, et al. Cytochrome P450 2C is an EDHF synthase in coronary arteries. Nature 1999; 401:493–7.
75. Fisslthaler B, Popp R, Michaelis UR, et al. Cyclic stretch enhances the expression and activity of coronary endothelium-derived hyperpolarizing factor synthase. Hypertension 2001; 38:1427–32.
76. Campbell WB, Gebremedhin D, Pratt PF, et al. Identification of epoxyeicosatrienoic acids as endothelium-derived hyperpolarizing factors. Circ Res 1996; 78:415–23.
77. Bauersachs J, Popp R, Hecker M, et al. Nitric oxide
attenuates the release of endothelium-derived hyperpolarizing factor. Circulation 1996; 94:3341–7.
78. Campbell WB, Falck JR, Gauthier K. Role of epoxyeicosatrienoic acids as endothelium-derived hyperpolarizing factor in bovine coronary arteries. Med Sci Monit 2001; 7:578–84.
79. Ignarro LJ. Haeme-dependent activation of guanylate cyclase and cyclic GMP formation by endogenous nitric oxide
: a unique transduction mechanism for transcellular signaling. Pharmacol Toxicol 1990; 67:1–7.
80. Lancaster Jr. JR Diffusion of free nitric oxide
. Methods Enzymol 1996; 268:31–50.
81. Kelm M, Schrader J. Control of coronary vascular tone by nitric oxide
. Circ Res 1990; 66:1561–75.
82. Ford PC, Wink DA, Stanbury DM. Autoxidation kinetics of aqueous nitric oxide
. FEBS Lett 1993; 326:1–3.
83. Hammad HM, Abdalla SS. Pharmacological effects of selected flavonoids
on rat isolated ileum: structure-activity relationship. Gen Pharmacol 1997; 28:767–71.
84. Brown JE, Khodr H, Hider RC, et al. Structural dependence of flavonoid interactions with Cu2+
ions: implications for their antioxidant properties. Biochem J 1998; 330:1173–8.
85. Brouillard R. The in vivo epression of anthocyanin color in plants. Phytochemistry 1983; 22:1311–23.
86. Girard P, Sercombe R, Sercombe C, et al. A new synthetic flavonoid protects endothelium-derived relaxing factor-induced relaxation in rabbit arteries in vitro: evidence for superoxide scavenging. Biochem Pharmacol 1995; 49:1533–9.
87. Huk I, Brovkovych V, Nanobash Vili J, et al. Bioflavonoid quercetin scavenges superoxide and increases nitric oxide
concentration in ischemia-reperfusion injury: an experimental study. Br J Surg 1998; 85:1080–5.
88. Yuting C, Rongliang Z, Zhongjian J, et al. Flavonoids
as superoxide scavengers and antioxidants. Free Radic Biol Med 1990; 9:19–21.
89. Hu JP, Calomme M, Lasure A, et al. Structure-activity relationship of flavonoids
with superoxide scavenging activity. Biol Trace Elem Res 1995; 47:327–31.
90. Bors W, Michel C. Antioxidant capacity of flavanols and gallate esters: pulse radiolysis studies. Free Radic Biol Med 1999; 27:1413–26.
91. Fridovich I. Superoxide radical: an endogenous toxicant. Annu Rev Pharmacol Toxicol 1983; 23:239–57.
92. Keaney Jr, JF Frei B. Antioxidant protection of low-density lipoprotein and its role in prevention of atherosclerotic vascular disease In: Frei B, ed. Natural antioxidants in human health and disease. San Diego: Academic Press, 1994:303–52.
93. Stralin P, Karlsson K, Johansson BO, et al. The interstitium of the human arterial wall contains very large amounts of extracellular superoxide dismutase. Arterioscler Thromb Vasc Biol 1995; 15:2032–6.
94. Brawn K, Fridovich I. Superoxide radical and superoxide dismutases: threat and defense. Acta Physiol Scand Suppl 1980; 492:9–18.
95. Hollman PC, van Trijp JM, Buysman MN, et al. Relative bioavailability of the antioxidant flavonoid quercetin from various foods in man. FEBS Lett 1997; 418:152–6.
96. Ader P, Block M, Pietzsch S, et al. Interaction of quercetin glucosides with the intestinal sodium/glucose co-transporter (SGLT-1). Cancer Lett 2001; 162:175–80.
97. Hammerstone JF, Lazarus SA, Schmitz HH. Procyanidin content and variation in some commonly consumed foods. J Nutr 2000; 130:2086S–92S.
98. Spencer JP, Chaudry F, Pannala AS, et al. Decomposition of cocoa procyanidins in the gastric milieu. Biochem Biophys Res Commun 2000; 272:236–41.
99. Deprez S, Brezillon C, Rabot S, et al. Polymeric proanthocyanidins are catabolized by human colonic microflora into low-molecular-weight phenolic acids. J Nutr 2000; 130:2733–8.
100. Yamakoshi J, Kataoka S, Koga T, et al. Proanthocyanidin-rich extract from grape seeds attenuates the development of aortic atherosclerosis in cholesterol-fed rabbits [see comments]. Atherosclerosis 1999; 142:139–49.
101. Burkert U, Allinger NL. Molecular mechanics.
Washington, DC: American Chemical Society, 1982.