Flavonoids are polyphenolic compounds ubiquitously distributed in plants and are consumed regularly in the diet in considerable amounts in the form of fruits, vegetables, nuts and derived products such as red wine, tea and chocolate. Based on their chemical structure, flavonoids can be classified into several subclasses such as flavonols, flavones, flavanones, flavan-3-ols (also referred to as catechins), anthocyanidins, isoflavones, dihydroflavonols and chalcones. Flavan-3-ols and flavonols are the most abundant and are widely distributed in food. The average daily intake in the occidental diet of flavonols is estimated to be approximately 20–25 mg, with quercetin contributing 60–75 % of the total(Reference Hertog, Feskens and Hollman1, Reference Sampson, Rimm and Hollman2). The average intake of flavan-3-ols is within a similar range (approximately 18–31 mg), with (+)-catechin (henceforth referred to as catechin) and epicatechin being the most abundant(Reference Manach, Scalbert and Morand3).
Prospective epidemiological studies have shown an inverse correlation between dietary flavonoid intake and mortality from CHD(Reference Hertog, Feskens and Hollman1, Reference Sampson, Rimm and Hollman2). Several studies using various animal models have provided support for the observed protective effects of dietary flavonoids with respect to CVD(Reference Perez-Vizcaino, Bishop-Bailley and Lodi4–Reference Perez-Vizcaino, Duarte and Jimenez6). However, the profiles of biological effects show marked differences among the different classes of flavonoids(Reference Loke, Proudfoot and Hodgson7). For example, flavonols exert systemic and coronary vasodilatation(Reference Duarte, Perez Vizcaino and Utrilla8, Reference Ibarra, Perez-Vizcaino and Cogolludo9), while catechins exhibit a weak vasodilator effect(Reference Duarte, Perez Vizcaino and Utrilla8). Both flavonoid groups show antioxidant effects in vitro and can prevent endothelial dysfunction in vivo (Reference Duarte, Perez-Palencia and Vargas10, Reference Ihm, Lee and Kim11).
Due to their ubiquitous distribution, flavonoids from different classes are commonly present together in foods and/or are consumed in the same meal. In fact, prototypical flavonoid-rich foods contain a large number of different flavonoids in variable amounts. However, little is known about the interactions between them. Quercetin and catechin have been reported to act synergistically in reducing platelet recruitment via the inhibition of protein kinase C-dependent NADPH oxidase activation(Reference Pignatelli, Pulcinelli and Celestini12, Reference Pignatelli, Di Santo and Buchetti13). Pharmacokinetic interactions have also been described(Reference Silberberg, Morand and Manach14).
In the present study, we have analysed the interactions between quercetin and catechin, the representative compounds of the two major classes of flavonoids, on vascular function in vitro.
Experimental methods
The present study was conducted according to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication no. 85-23, revised 1996) and approved by our Institutional Committee (Comité de de Experimentación animal, Universidad Complutense) for the ethical care of animals. Male Wistar rats were obtained from Harlan Laboratories.
Materials and solutions
All drugs and chemicals were purchased from Sigma (St Louis, MO, USA). Flavonoids were dissolved in dimethyl sulphoxide, and all the other compounds were dissolved in distilled, deionised water. Quercetin–catechin mixtures were prepared by mixing the stock solutions, and the concentrations of mixtures are expressed as the sum of both, e.g. a 10 μm-quercetin–catechin (1:10) mixture contains 0·91 μm-quercetin and 9·09 μm-catechin. To prepare the NO solutions used for biological activity, a vial containing 20 ml of Krebs solution at 37°C was initially bubbled with N2 for 15 min and then continuously bubbled with NO (450 parts per million from Air Liquide España, Paris, France), resulting in a concentration of 0·9–1 μm (as measured by an ISO-NO meter electrochemical electrode; WPI, Stevenage, Herts, UK). The Krebs solution contained 118 mm-NaCl, 4·75 mm-KCl, 25 mm-NaHCO3, 1·2 mm-MgSO4, 2 mm-CaCl2, 1·2 mm-KH2PO4 and 11 mm-glucose. The HEPES-buffered solution contained 130 mm-NaCl, 5 mm-KCl, 1·2 mm-MgCl2, 10 mm-glucose and 10 mm-HEPES (pH adjusted to 7·4 with NaOH).
Auto-oxidation
Absorption spectra of quercetin and catechin in the Krebs solution, bubbled with 95 % O2 and 5 % CO2 at 37°C, pH 7·4, were analysed in a spectrophotometer (6405; Jenway, BioWorld, Bibby Scientific Ltd, Staffs, UK)(Reference Lopez-Lopez, Moreno and Cogolludo15). Spectra were obtained just after dilution and at 10, 20, 60 and 120 min.
Nitric oxide scavenging
NO was monitored with the ISO-NO meter in a 20 ml chamber at 37°C filled with the HEPES-buffered solution as described previously(Reference Lopez-Lopez, Moreno and Cogolludo15). At the beginning of the experiment, the system was closed with no headspace; under constant rapid stirring, NO concentration was approximately 100 nm and O2 concentration was 20 %, and the pH was 7·4.
Aortic contractile tension recording
Thoracic aortic rings (3 mm in length) from Wistar rats were mounted in organ chambers for isometric force recording(Reference Jimenez, Andriambeloson and Duarte16). In some experiments, aortae were mechanically denuded of endothelium by rubbing the luminal surface with a metal rod. Aortic rings were stretched to 2 g of tension and equilibrated for 90–120 min. During this period, Krebs solution was changed every 30 min, and the rings were re-stretched as needed to maintain a final tension of 2 g. Endothelium denudation was confirmed by the absence of a relaxant response to acetylcholine ( < 10 %). To analyse the effects on endothelial function and its inhibition by the superoxide dismutase inhibitor diethyldithiocarbamate (DETCA), endothelium-intact rings were pre-incubated in the absence or presence of DETCA (1 mm) and with vehicle (dimethyl sulphoxide 0·1 %), quercetin, catechin or the quercetin–catechin mixture (1:1; 10 μm) for 10 min. After stimulation with phenylephrine (100 nm), a concentration–response curve analysis to acetylcholine was performed.
Contraction in mesenteric and pulmonary resistance arteries
Mesenteric and pulmonary arterial rings (internal diameter 300–500 μm) were mounted in a wire myograph in Krebs solution and stretched to give an equivalent transmural pressure of 100 mmHg(Reference Cogolludo, Moreno and Frazziano17). After equilibration, arteries were exposed to U46619 (0·1 μm), and after a steady-state constrictor response was obtained, the flavonoids quercetin, catechin or their mixtures were added cumulatively. The NO synthase inhibitor G-nitro-l-arginine-methyl ester (0·1 mm) was added at the end of the experiment to test the reversal of flavonoid-induced relaxation.
Pro-oxidant effect in rat aortic smooth muscle cells in culture
Primary cultures of smooth muscle cells were prepared from rat aortic explants as described previously(Reference Lodi, Jimenez and Moreno18). Dihydroethidium (DHE) enters the cells and is oxidised by superoxide to yield ethidium, which binds to DNA to produce bright red fluorescence. Cells cultured in ninety-six-well plates (passages 2–3) were incubated in HEPES-buffered solution containing DHE (10 μm) at 37°C for 30 min, and then exposed for another 30 min to various concentrations of quercetin, catechin or their mixtures. DHE fluorescence was read in a fluoroskan fluorometer (Thermoscientific, Madison, WI, USA). Excitation and emission were filtered at 530 and 620 nm, respectively).
Statistical analysis
Results are expressed as means with their standard errors, and n describes the number of measurements made from different animals. Differences between experimental groups were analysed using the unpaired Student's t test or, for multiple comparisons, using one-way ANOVA followed by Dunnett's post hoc test. P values < 0·05 were considered statistically significant. Individual cumulative concentration–response curves were analysed, and IC30, i.e. the concentration producing a relaxation of 30 % of the previous tone, was calculated. For the isobolographic analysis, the calculated IC30 value of quercetin or catechin reflected the estimated concentration present in the mixture when a relaxation of 30 % was achieved.
Results
Vasodilator effects in the rat aorta
As reported previously, quercetin-induced a concentration-dependent relaxant effect (Fig. 1(a)). Catechin was less potent, and the 50 % relaxation was not achieved at the maximal concentration (100 μm) tested. The 1:0·1 quercetin–catechin mixture (90·9 % quercetin and 9·1 % catechin) produced an effect similar to quercetin, and the 1:10 mixture (9·1 % quercetin and 90·9 % catechin) produced a similar effect to catechin. To analyse the interactions between the two compounds, we performed an isobolographic analysis(Reference Perez-Vizcaino, Cogolludo and Zaragoza-Arnaez19, Reference Tallarida20) from the data shown in Fig. 1(a) by plotting the IC30 values of the pure compounds and their mixtures (Fig. 1(b)). The dotted line in this figure represents the line of additivity, so that deviations above and below this line indicate negative and positive interactions, respectively. It can be observed that the points of the quercetin–catechin mixtures were not significantly different from the line of additivity, indicating that there were no interactions between the two flavonoids, i.e. their effects were additive.
Fig. 1 Quercetin (Quer) and catechin (Cat) induce additive vasodilatory effects. Vasorelaxant responses induced by quercetin, catechin and the 1:0·1, 1:1 and 1:10 quercetin–catechin mixtures in the endothelium-denuded rat aorta stimulated with noradrenaline (1 μm). (a) Cumulative concentration–response curves. The concentrations of the mixtures are expressed as the sum of both compounds, e.g. in the 1:0·1 quercetin–catechin mixture at 10 μm (log[quercetin 1–0·1] = − 5), the concentrations of each compound are 9·09 μm for quercetin and 0·91 μm for catechin. ○, Catechin; □, quercetin; ▲, Quer–Cat 1:0·1; ■, Quer–Cat 1:1; •, Quer–Cat 1:10; ◆, DMSO. (b) Isobologram for the interaction between quercetin and catechin in causing relaxation in endothelium-denuded aortic rings. Each point represents the estimated concentration of the drug present in the mixture when a 30 % relaxation is achieved, calculated from the data in (a). Results are means from six to eight experiments (except dimethyl sulphoxide (DMSO), where n 3) with standard errors represented by vertical bars. The dotted line represents the line of additivity; deviations above and below the line would show negative and positive interactions, respectively.
Vasodilator effects in mesenteric and pulmonary resistance arteries
In endothelium-intact small arteries mounted in a myograph, the contractile responses were not as sustained as in the aortae, i.e. after 70–90 min, the tone usually declined by more than 30 %. Therefore, only three concentrations of the flavonoids (1, 10 and 100 μm) could be tested. As expected(Reference Perez-Vizcaino, Ibarra and Cogolludo21), quercetin produced a stronger relaxant response in the small arteries than in the aorta that was also significantly more potent in systemic (mesenteric) than in pulmonary arteries, with 10 μm producing approximately 90 and 50 % relaxation, respectively (Fig. 2). Addition of G-nitro-l-arginine-methyl ester at the end of the experiment could not reverse the vasodilator effects of quercetin. In contrast, 100 μm-catechin produced only approximately 20 % relaxation in both arteries. Despite the use of only three concentrations and the weak effect of catechin precluded the isobolographic analysis of the interactions, the results obtained suggest the lack of the interaction between the two flavonoids, e.g. the 1:0·1 quercetin–catechin mixture produced effects similar to those of quercetin, and the effects of the 1:1 mixture were weakly less pronounced than those of quercetin in pulmonary arteries.
Fig. 2 Vasorelaxant responses induced by quercetin (Quer), catechin (Cat) and the 1:0·1 and 1:1 quercetin–catechin mixtures in (a) mesenteric and (b) pulmonary resistance arteries stimulated with U46619 (100 nm). After the last concentration of the flavonoids, the nitric oxide synthase inhibitor G-nitro-l-arginine-methyl ester (l-NAME; 0·1 mm) was added to test the possible reversion of flavonoid-induced vasodilation. Results are means from five to eight experiments, with standard errors represented by vertical bars.
Effects on endothelial function and dysfunction
In endothelium-intact aortic rings pre-contracted with phenylephrine, acetylcholine induced a relaxant response due to the endothelial release of NO. In the presence of quercetin, catechin or the 1:1 quercetin–catechin mixture at 10 μm, this relaxant response was unaffected (Fig. 3(a)). The Cu2+ chelator DETCA (1 mm) increases superoxide by inhibiting endogenous Cu/Zn superoxide dismutase activity and hence induces NO inactivation. As expected, DETCA induced a marked impairment of endothelial-derived NO response as indicated by the reduction in the relaxant response to acetylcholine (Fig. 3(b)). Under these conditions, 10 μm-quercetin increased acetylcholine-induced relaxation, i.e. it prevented oxidative stress-induced endothelial dysfunction. Catechin produced a weak, borderline significant (P = 0·06) increase, and the 1:1 quercetin–catechin mixture produced an intermediate effect between the pure compounds.
Fig. 3 (a) Effects of quercetin, catechin and the 1:1 quercetin–catechin (Quer–Cat) mixture on the endothelium-dependent vasorelaxation of acetylcholine (ACh) under control conditions and (b) after inducing endothelial dysfunction with the superoxide dismutase inhibitor diethyldithiocarbamate. Endothelium-intact aortic rings were treated for 10 min with the flavonoids (10 μm), then stimulated with phenylephrine (100 nm), and a concentration–response curve to Ach (0·01–100 μm) was carried out in a cumulative fashion. Results are means of five to eight experiments, with standard errors represented by vertical bars. ●, Dimethyl sulphoxide (DMSO); □, quercetin; △, catechin; 
, Quer–Cat 1:1.
Pro-oxidant and nitric oxide-scavenging effects in cell-free systems
Quercetin is oxidised in aqueous buffers, generates superoxide and inactivates NO(Reference Lopez-Lopez, Moreno and Cogolludo15). Its oxidation can be monitored by the changes in UV spectra, i.e. an increase in absorbance at 320 nm and a decrease at 385 nm. Catechin (10 μm), which has negligible absorbance in this wavelength range, was not able to prevent the oxidation of quercetin in the media (data not shown). Fig. 4(a) shows the inactivation of NO induced by quercetin. In contrast, the rate of decay of NO in the presence of catechin was similar to that observed in the presence of the vehicle dimethyl sulphoxide. Moreover, catechin did not alter the NO-scavenging effect induced by quercetin (Fig. 4(b)).
Fig. 4 Catechin does not scavenge nitric oxide and does not prevent quercetin-induced nitric oxide scavenging. (a) Time-course of nitric oxide decay in the presence of quercetin (□), catechin (○) and 0·1 % dimethyl sulphoxide (DMSO (■); vehicle). (b) Nitric oxide decay induced by quercetin in the presence of DMSO (vehicle) (●) and catechin (○). Results are means of three experiments, with standard errors represented by vertical bars.
Pro-oxidant effect in cultured aortic smooth muscle cells
In the absence of DHE, neither 30 μm-quercetin nor 100 μm-catechin induced any change in red fluorescence. Addition of quercetin to aortic smooth muscle cells produced a concentration-dependent increase in DHE red fluorescence, indicating an increase in intracellular superoxide (Fig. 5). In contrast, catechin produced a negligible effect (not significant compared with dimethyl sulphoxide) at concentrations up to 100 μm. The 1:10 quercetin–catechin mixture produced a concentration-dependent increase in DHE fluorescence. This latter effect was equivalent to that produced by one-tenth of pure quercetin, e.g. 100 μm of the mixture containing 9·1 μm of quercetin produced an effect similar to 10 μm of pure quercetin, indicating that catechin does not interfere with the pro-oxidant effects of quercetin in smooth muscle cells.
Fig. 5 Intracellular superoxide generation in aortic smooth muscle cells in primary culture. Cells incubated with the red fluorescent dye dihydroethidium were exposed to different concentrations of quercetin, catechin and the 1:10 quercetin–catechin (Quer–Cat) mixture. Fluorescence was measured at 30 min. Results are means of two to three experiments performed in triplicate, with standard errors represented by vertical bars. AU, arbitrary units; DMSO, dimethyl sulphoxide.
Discussion
Endothelial, smooth muscle cells and platelets are three main targets for the actions of flavonoids regarding CVD. A synergistic interaction between catechin and quercetin to inhibit platelet aggregation has been described(Reference Pignatelli, Pulcinelli and Celestini12, Reference Pignatelli, Di Santo and Buchetti13). In the present study, we show that compared with quercetin, catechin produced weak vasodilator effects in conductance and resistance arteries, a weak ability to modulate oxidative stress-induced endothelial dysfunction and did not show pro-oxidant effects. Importantly, catechin did not influence the effects of quercetin on these parameters.
A direct endothelium-independent vasodilator effect of flavonols is one of the potential mechanisms involved in their anti-hypertensive effects, which might also contribute to prevent other forms of CVD such as IHD and stroke. Interestingly, quercetin is a more potent vasodilator in smaller arteries, which is involved in the control of vascular resistance and therefore regulates tissue flow and blood pressure(Reference Perez-Vizcaino, Ibarra and Cogolludo21). This difference may reflect the diverse signalling involved in the contraction in large and small arteries. Recently, pulmonary arterial pressure has been implicated as a cardiovascular risk factor(Reference Lam, Borlaug and Kane22). Herein, we report for the first time that quercetin also relaxes pulmonary resistance arteries. However, the vasodilator potency of quercetin was higher in pulmonary than in the aorta but yet lower than in systemic (mesenteric) resistance arteries, indicating a low selectivity for the pulmonary circulation. In contrast, catechin had a weak vasodilator effect. Despite some structural similarities between catechin and quercetin, the latter is planar but the former is not, which may explain some of the differences found in the biological activity. Moreover, catechin did not modify the vasodilator effects of quercetin in the three vessels studied. In the aorta, the isobolographic analysis revealed that the effects of both flavonoids were additive.
The mechanisms of drug interactions (synergism, additivity or antagonism) are most often analysed intuitively, without a proper mathematical method. The isobolographic analysis is an easy and appropriate method for the study of drug–drug interactions when both drugs show activity(Reference Tallarida20). In the aorta, the isobologram revealed that the vasodilator effects of both flavonoids were additive because the points for their combinations were near the line of additivity (within the standard errors of the points). In mesenteric and pulmonary arteries, the isobologram could not be plotted because of the weak effect of catechin. However, the data also suggest a lack of interaction.
Endothelial dysfunction is an early and independent predictor of poor prognosis in most forms of CVD. One of the key mechanisms of endothelial dysfunction involves the vascular production of reactive oxygen species, particularly superoxide derived from NADPH oxidase, which reacts rapidly with NO and induces NO inactivation. Quercetin restores the impaired endothelial function in vitro (Reference Sanchez, Lodi and Vera23) and in vivo in several animal models of hypertension(Reference Sanchez, Galisteo and Vera24) by at least two mechanisms. Acutely, quercetin directly scavenges superoxide and inhibits the activity of superoxide-generating enzymes. In the chronic phase, quercetin can prevent the expression of genes involved in superoxide generation such as the NADPH oxidase subunits induced by pathological stimuli such as angiotensin II or endothelin-1. Herein, we used a protocol to induce acute endothelial dysfunction based purely on oxidative stress by means of inhibiting the degradation of endogenous superoxide. Compounds scavenging intracellular superoxide or inhibiting the synthesis are expected to protect this (and possibly other) form of endothelial dysfunction. Catechin at physiologically relevant concentrations of 10 μm had a weak protective effect on the impairment of acetylcholine-induced relaxation by DETCA. The combination of catechin and quercetin both at 5 μm produced less effective protection than 10 μm-quercetin, ruling out a synergistic interaction between the two compounds.
Antioxidants often also show pro-oxidant effects when analysed under specific experimental conditions. In fact, besides the antioxidant effect, flavonoids also show pro-oxidant effects(Reference Lopez-Lopez, Moreno and Cogolludo15, Reference Halliwell25), which may be involved in some of their biological properties. Under certain experimental conditions, the generation of superoxide by quercetin can scavenge NO and hence reduce NO-induced vasodilation(Reference Lopez-Lopez, Moreno and Cogolludo15). In addition, the activation of large conductance potassium channels by quercetin involves the generation of H2O2(Reference Cogolludo, Frazziano and Briones26), and this effect contributes at least partially to the coronary vasodilator effect. Quercetin is auto-oxidised and generates superoxide that rapidly reacts with NO(Reference Lopez-Lopez, Moreno and Cogolludo15). In contrast, catechin at 10 μm lacked NO-scavenging effects. The effects of quercetin on NO can be prevented by superoxide scavengers such as superoxide dismutase. It could be expected that catechin could prevent the effects of quercetin via an antioxidant effect. However, at least at this concentration of 10 μm, catechin did not prevent quercetin auto-oxidation nor the NO-scavenging effect. Intracellular reactive oxygen species can be detected by using fluorescent dyes such as DHE, which detects superoxide, or dichlorofluorescein, which detects mainly H2O2. Confirming our previous report using dichlorofluorescein in coronary arteries(Reference Cogolludo, Frazziano and Briones26), in the present study, we also found that quercetin increases intracellular reactive oxygen species in aortic smooth muscle. Consistent with a weak NO-scavenging effect in cell-free systems, catechin did not significantly increase intracellular superoxide. Synergistic effects of polyphenolic mixtures (including catechin, quercetin, hesperidin and ferulic acid) in preventing human LDL oxidation have been found(Reference Cirico and Omaye27). However, catechin alone did not prevent the pro-oxidant effect of quercetin either in cell-free systems or in cultured smooth muscle cells.
A limitation of the present study is that the concentration range of plasma quercetin levels in human subjects consuming flavonoid-rich foods is 0·3–7·6 μm(Reference Manach, Williamson and Morand28), which is below the range of effective concentrations used herein (1 μm had almost no effect and 10 μm had almost full inhibitory effect in the mesentery). Moreover, quercetin is mostly present in the plasma in the form of glucuronidated and sulphated metabolites(Reference Manach, Williamson and Morand28), which lack a direct acute vasorelaxant effect(Reference Lodi, Jimenez and Moreno29). Recent evidence has suggested that glucuronides deconjugate and the free quercetin (aglycone) is accumulated in cells at variable concentrations(Reference Bieger, Cermak and Blank30), but its biological relevance is yet unknown.
In conclusion, the present in vitro study shows that catechin has weak effects on vascular function acutely and, when present, its effects are additive with those of quercetin, excluding synergistic interactions of these two flavonoids at this level.
Acknowledgements
The present study was supported by grants from the Ministerio de Ciencia e Innovación (AGL2007-06685, SAF2008-03948, Red HERACLES RD06/0009 and C. Menendez FPI studentship), Junta de Andalucía (P06-CTS-01555 and R. Jimenez ‘Retorno de Doctores’ contract) and Mutua Madrileña. The authors are grateful to Enrique Moreno and Bianca Barreira for their excellent technical assistance. The authors declare no conflicts of interest. F. P.-V. and J. D. designed the experiments, F. P.-V. wrote the manuscript, and all others, especially C. M., performed and analysed the experiments and contributed to the discussion.
