Involvement of thromboxane A2 in the endothelium-dependent contractions induced by myricetin in rat isolated aorta


Department of Pharmacology, School of Pharmacy, University of Granada, 18071 Granada, Spain; E-mail:


  • The present study was undertaken to analyse the mechanism of the contractile response induced by the bioflavonoid myricetin in isolated rat aortic rings.

  • Myricetin induced endothelium-dependent contractile responses (maximal value=21±2% of the response induced by 80 mM KCl and pD2=5.12±0.03). This effect developed slowly, reached a peak within 6 min and then declined progressively.

  • Myricetin-induced contractions were almost abolished by the phospholipase A2 (PLA2) inhibitor, quinacrine (10 μM), the cyclo-oxygenase inhibitor, indomethacin (10 μM), the thromboxane synthase inhibitor, dazoxiben (100 μM), the putative thromboxane A2 (TXA2)/prostaglandin endoperoxide receptor antagonist, ifetroban (3 μM). These contractions were abolished in Ca2+-free medium but were not affected by the Ca2+ channel blocker verapamil (10 μM).

  • In cultured bovine endothelial cells (BAEC), myricetin (50 μM) produced an increase in cytosolic free calcium ([Ca2+]i) which peaked within 1 min and remained sustained for 6 min, as determined by the fluorescent probe fura 2. This rise in [Ca2+]i was abolished after removal of extracellular Ca2+ in the medium.

  • Myricetin (50 μM) significantly increased TXB2 production both in aortic rings with and without endothelium and in BAEC. These increases were abolished both by Ca2+-free media and by indomethacin.

  • Taken together, these results suggests that myricetin stimulates Ca2+ influx and subsequently triggers the activation of the PLA2 and cyclo-oxygenase pathways releasing TXA2 from the endothelium to contract rat aortic rings. The latter response occurs via the activation of Tp receptors on vascular smooth muscle cells.

British Journal of Pharmacology (1999) 127, 1539–1544; doi:10.1038/sj.bjp.0702694


bovine aortic endothelial cells




HEPES-buffered saline solution




nitric oxide


prostaglandin H2


phospholipase A2


thromboxane A2


thromboxane B2


Flavonoids are a large group of polyphenolic compounds, diverse in structure and characteristics (Cook & Samman, 1996; Rice-Evans & Miller, 1997). They occur naturally in a variety of foods from vegetable origin, mainly apples, onions, tea and red wine and are an integral part of the human diet (Hertog et al., 1993). The estimated diet intake of major flavonoids ranges between 23 and 170 mg day−1 (expressed as aglycones) (Hertog et al., 1993; Cook & Samman, 1996). In the cardiovascular system, flavonoids have been reported to exhibit antiarrhythmic and antiischaemic effects in the heart (Occhiuto et al., 1991). They also produce endothelium-dependent (Andriambeloson et al., 1997) and -independent vasorelaxant effects in different blood vessels including the rat thoracic aorta (Duarte et al., 1993a,1993b; Herrera et al., 1996) and inhibit platelet aggregation and lipid peroxidation (Tzeng et al., 1991; Cook & Samman, 1996). These effects may explain why flavonoid intake appears inversely related with mortality from coronary heart disease in epidemiological studies (Hertog et al., 1993).

Myricetin (3,5,7,3′,4′,5′-hexahydroxyflavone) is a major antioxidant food flavonoid which has been shown to exhibit a biphasic contractile response in pre-contracted rat thoracic aorta (Berger et al., 1992; Herrera et al., 1996). At low concentrations (<50 μM), it potentiates the responses to different contractile agents such as noradrenaline, high KCl and phorbol 12-myristate 13-acetate in rat aortic rings, whereas at higher concentrations, it exerts a vasorelaxant effect on vessels precontracted with these agents (Herrera et al., 1996). The potentiating effect of myricetin has been reported in a number of arteries from different species such as the rat tail and femoral arteries (Berger et al., 1992) or the rabbit pulmonary artery (Russell & Rohrbach, 1989). However, the mechanisms involved in the vasoconstrictor response remains unclear.

Therefore, the present study was undertaken to better characterize the contractile effect induced by myricetin in rat thoracic rings. The role of endothelium was investigated. As the endothelium can release different factors including those from the cyclo-oxygenase products, the involvement of this pathway and the underlying mechanism(s) were also studied.


Tissue preparation

Wistar rats (either sex, 250–300 g) were killed by a blow on the head. The descending thoracic aorta was quickly dissected and placed in a Krebs' solution of the following composition (mM): NaCl 118, KCl 4.75, NaHCO3 25, MgSO4 1.2, CaCl2 2.0, KH2PO4 1.2, and glucose 11. After excess fat and connective tissue were removed, the aortae were cut into rings (2 mm length) and mounted under a basal tension of 2 g in 20 ml organ baths containing Krebs' solution and attached to a force-displacement transducer (Letigraph 2000, Letica) to measure isometric contraction as previously described (Herrera et al., 1996). The tissue bath was maintained at 37°C and bubbled with 95% O2-5% CO2. For experiments in which Ca2+ free Krebs' solution was used, Ca2+ was omitted and 0.5 mM EGTA was added. Each preparation was allowed to equilibrate for at least 90 min prior to initiation of experimental procedures and during this period the incubation medium was changed every 20 min. In some experiments, the endothelium was mechanically removed by gently rubbing the ring intimal surface. The presence of functional endothelium was assessed in all preparations by determining the ability of acetylcholine (1 μM) to induce more than 50% relaxation of rings precontracted with phenylephrine (1 μM). Vessels were considered to be denuded of functional endothelium when there was no relaxant response to acetylcholine.

Characterization of the contractile effect of myricetin

After equilibration, aortic rings were challenged with 80 mM KCl Krebs' solution (in which the concentration of NaCl was replaced with an equimolar amount of KCl) in order to test the maximal contractile response of the vessels. Then, rings were washed several times until the contraction had returned to its original baseline. After a 30 min washout period, myricetin was added to the bath in different concentrations (1–100 μM). Preliminary experiments showed that the contractile response to myricetin was not reproducible when it was added in either a cumulative or a non cumulative manner to the bath. Therefore, each vessel was challenged with a single concentration of myricetin.

In another series of experiments, the possible signalling pathways involved in the contractile responses induced by myricetin were analysed in intact aorta. In order to characterize the involvement of the nitric oxide (NO) pathway, some arteries were exposed to the NO synthase inhibitor, NG - nitro - L - arginine - methyl-ester (L-NAME, 100 μM), added to the bath 30 min prior to myricetin. The sources of Ca2+ involved in the contraction induced by 50 μM myricetin, were analysed by testing its effect in Ca2+-free Krebs' solution or in the presence of the calcium channel blocker verapamil (10 μM) added 30 min prior to myricetin. The involvement of arachidonic acid metabolism in the contractile response to myricetin was also analysed. The role of phospholipase A2 (PLA2) pathway was investigated using its inhibitor quinacrine (10 μM). The involvement of cyclo-oxygenase pathway was studied using the cyclo-oxygenase inhibitor, indomethacin (10 μM), the thromboxane synthase inhibitor, dazoxiben (100 μM) or the thromboxane A2 (TXA2)/prostaglandin endoperoxide receptor blocker, ifetroban (0.3 or 3 μM). All the inhibitors were added to the bath 20 min prior to myricetin (50 μM).

Cell cultures

Bovine aortic endothelial cells (BAEC) were isolated from bovine aortae as described previously by Kessler & Lugnier (1995). Cells were cultured in plastic flasks using as culture medium a mixture of DMEM and HAM F12 mediums (50/ 50) supplemented with 10% foetal calf serum, 2 mM glutamine, 100 mg ml−1 heparin, 10,000 U ml−1 penicillin, 10,000 U ml−1 streptomycin and 10 μM vitamin C. The cultures were maintained at 37°C in a humidified atmosphere with 5% CO2. Cells were used for measurements of the cytosolic free Ca2+ concentration [Ca2+]i and the release of thromboxane B2 (TXB2) after the first or second passages when confluent.

[Ca2+]i measurements

[Ca2+]i measurements were performed with the fluorescent Ca2+-sensitive probe fura-2. Cells were washed and incubated with 5 μM fura-2/AM (the membrane permeant acetoxymethylester derivative) for 1 h at room temperature in HEPES-buffered saline solution (HBSS) of the following composition in mM: NaCl, 119; KCl, 4.75; CaCl2, 1.25; MgSO4, 1.2; KH2PO4, 1.2; NaHCO3, 25; glucose, 5 and HEPES 20; pH 7.4. BAEC were then washed twice with phosphate-buffered saline Ca2+/Mg2+-free solution and dispersed using 1% trypsin. After 5 min centrifugation at 60×g, cells were again washed and suspended at a density of 2×106 cells ml−1 in HBSS. Cells were then transferred to a quartz cuvette in final volume of 2.5 ml, which was constantly stirred and maintained at 37°C. Fluorimetric readings were performed with a F-2000 Hitachi spectrofluorimeter system, with excitation alternating between 340 and 380 nm (10 Hz) and emission at 510 nm. In each preparation, the maximum and minimum fluorescences were sequentially determined by the addition of 10 μM ionomycin in the presence of 2 mM Ca2+, followed by the addition of 10 mM EGTA at pH 8. The [Ca2+]i was calculated according to the equation described by Grynkiewitz et al. (1985).

Thromboxane B2 production in aortic rings and endothelial cells

To determine the production of TXB2, rat aortic rings (5 mm length) with and without endothelium were placed in 1 ml of Krebs' solution plus vehicle (DMSO with a final concentration of 0.05% v v−1) or myricetin (50 μM) for 15 min. The tissues were bubbled with a 95% O2-5% CO2 gas mixture and kept at 37°C. At the end of this period, 500 μl of the medium was collected and TXB2 was measured by an enzymeimmunoassay (EIA) (Amersham Life Science, Buckinghamshire, U.K.).

TXB2 production was assessed in BAEC (106 cells ml−1) resuspended in HBSS with or without CaCl2 plus 0.5 mM EGTA. After 15 min incubation with myricetin (50 μM) or myricetin plus indomethacin (10 μM), cells were centrifuged (10 min at 900 r.p.m.) and the medium was collected for TXB2 measurement by EIA.


The following drugs were used: myricetin (Extrasynthese, Genay, France), indomethacin, L-NAME, prazosin, quinacrine and verapamil (Sigma, Madrid, Spain), dazoxiben and ifetroban (Pfizer, New York, U.S.A.). Myricetin was initially dissolved in dimethylsulphoxide (DMSO) to prepare a 10 mM stock solution. Indomethacin was prepared in 2 mM Na2CO3 immediately before use and ifetroban was dissolved in absolute ethanol to prepare a 1 mM stock solution. Other drugs were dissolved in distilled water such that volumes of <0.2 ml were added to the organ chambers.

Statistical analysis

For the contractile experiments, the results were expressed as a percentage of the contractile response induced by 80 mM KCl. Results are expressed as means±s.e.mean, n representing the number of aortic rings from different animals. Statistical analysis was performed by means of a two way analysis of variance (ANOVA) followed by a Newman Keuls' test. The differences between control and experimental values were considered significant when P<0.05.


Contractile responses

KCl (80 mM) induced a sustained contraction in aortic rings which averaged 831.1±25.9 mg (n=29) and 833.3±68.6 mg (n=9) in arteries with and without functional endothelium, respectively. In intact rings, myricetin (1–100 μM) induced a concentration-dependent contractile response (pD2 value=5.12±0.03), the maximum response was reached at 50 μM (21.0±2.3% of the response induced by 80 mM KCl, Figure 1A). As shown in Figure 1B, the increase in tension produced by 50 μM myricetin developed slowly, reached a peak within 6 min and then declined progressively within 20 min. In aortic rings without functional endothelium, myricetin (50 or 100 μM) failed to produce any contraction suggesting that its effect depends exclusively on the presence of functional endothelium (Figure 2). Inhibition of NO-synthase with L-NAME (100 μM) increased by approximately 50% the contractile response to myricetin (Figure 2).

Figure 1.

Contractile responses induced by myricetin in endothelium-intact rat aortic rings. (A) The concentration-response curve was constructed in a non cumulative manner by incubation of each ring with a single concentration of myricetin. Each symbol represents the mean of 6–29 arteries; vertical lines show s.e.mean. (B) Time-course of the contractile response induced by 50 μM myricetin. The results in both pannels are expressed as a percentage of the 80 mM KCl-induced contractile response. Each symbol represents the mean of 29 arteries; vertical lines show s.e.mean.

Figure 2.

Role of endothelium and nitric oxide in the contractile response induced by 50 μM myricetin. The experiments were performed in endothelium-intact (+E) or endothelium denuded (−E) aortic rings incubated in the absence (CTRL) or in the presence of L-NAME (100 μM) for 20 min before the addition of 50 μM myricetin. The results are expressed as a percentage of the 80 mM KCl-induced contractile response. Each symbol represents the mean of 6–29 arteries; vertical lines show s.e.mean. **P<0.01 vs CTRL.

Figure 3 shows that the PLA2 inhibitor, quinacrine (10 μM), the cyclo-oxygenase inhibitor, indomethacin (10 μM), the thromboxane synthase inhibitor, dazoxiben (100 μM) and the putative TXA2/prostaglandin endoperoxide receptor antagonist, ifetroban (0.3 and 3 μM) significantly reduced the contractile response induced by 50 μM myricetin.

Figure 3.

Effects of arachidonic acid pathway inhibitors on the contractile response induced by 50 μM myricetin. Aortic rings with functional endothelium were incubated in the absence (CTRL) or in the presence of quinacrine (QUIN, 10 μM), indomethacin (INDO, 10 μM), dazoxiben (DZB, 100 μM) or ifetroban (IFET, 0.3 or 3 μM) for 20 min before the addition of myricetin. The results are expressed as a percentage of the 80 mM KCl-induced contractile response. Each symbol represents the mean of 6–29 arteries; vertical lines show s.e.mean. *P<0.05, **P<0.01 vs CTRL.

To study the source of Ca2+ involved in the endothelium-dependent contractile response to myricetin, experiments were performed in Ca2+-free medium or in normal medium in the presence of the L-type Ca2+ channel blocker verapamil (10 μM). After incubation of aortic rings in Ca2+-free medium the contractile response induced by 50 μM myricetin was abolished. In contrast, verapamil did not significantly affect the contraction induced by myricetin (21.8±3.7%, n=10; P>0.05). Moreover, it should be noted that the response to myricetin was not altered in the presence of alpha1 antagonist, prazosin (1 μM) (18.0±4.2%, n=6, P>0.05).

[Ca2+]i measurements in cultured endothelial cells

In fura2-loaded BAEC resuspended in HBSS the basal [Ca2+]i was 226±37 nM (n=3). Myricetin, at concentration at which it produced maximal endothelium-dependent contractile response (50 μM), produced a transient increase in [Ca2+]i which peaked within 40 s (Figure 4). In Ca2+-free HBSS, the basal [Ca2+]i was significantly reduced (101±2 nM, n=4, P<0.01) as compared to the basal level obtained in normal HBSS. Under these conditions, myricetin failed to induce an increase in [Ca2+]i (Figure 4).

Figure 4.

Time course of the effects of 50 μM myricetin on intracellular calcium measured in fura2-loaded BAEC resuspended in HBSS or in Ca2+-free HBSS. Results are expressed as increases in [Ca2+]i above baseline. Each symbol represents the mean of 3–4 experiments; vertical lines show s.e.mean. Myricetin was added at time 0.

Production of TXB2

The effects of myricetin on the production of TXB2 was studied in aortic rings and in BAEC (Figure 5). The basal TXB2 production was significantly higher (P<0.05) in aortic rings with endothelium as compared to that in arteries without endothelium. Myricetin (50 μM) significantly increased the TXB2 production both in vessels with and without endothelium. However, TXB2 production was significantly higher (P<0.01) in intact vessels treated with myricetin as compared to denuded vessels.

Figure 5.

TXB2 production stimulated by myricetin in (A) aortic rings and (B) BAEC. (A) shows the TXB2 production in resting (CTRL) or myricetin-stimulated (50 μM for 15 min, MYR) aortic rings with (+E) or without (−E) endothelium. Each column represents the mean of six arteries; vertical lines show s.e.mean. **P<0.01 vs CTRL. (B) shows the TXB2 production in BAEC resuspended in HBSS or a Ca2+-free HBSS. BAEC were incubated for 15 min in the absence (CTRL) or in the presence of myricetin (50 μM, MYR) or myricetin plus indomethacin (10 μM, INDO). Each column represents the mean of five experiments; vertical lines show s.e.mean. **P<0.01 vs CTRL.

The TXB2 production in non stimulated BAEC resuspended in HBSS was 4239±453 pg ml−1 (n=5). Myricetin produced a 5 fold increase in TXB2 production and this effect was abolished in the presence of indomethacin (10 μM). In Ca2+-free HBSS, the basal TXB2 level did not significantly differ to that obtained in normal HBSS. However, under these conditions myricetin did not increase significantly TXB2 production (Figure 5B).


The present study indicates that the contraction of isolated rat aorta induced by myricetin requires the presence of endothelium. The results provide evidence that myricetin, through the activation of phospholipase A2 pathway, induces the release of TXA2 following the activity of cyclo-oxygenase eliciting vascular smooth muscle contraction subsequent to Tp receptor stimulation. Likewise, myricetin is able to release TXA2 through a cyclo-oxygenase inhibitor sensitive pathway in cultured endothelial cells. All of the above mechanisms implicate an increase of [Ca2+]i in the endothelial cells through an extracellular Ca2+-dependent pathway.

The endothelium contributes to the local regulation of vascular smooth muscle tone by releasing endothelium-derived relaxing factors (NO, prostacyclin and endothelium-derived hyperpolarizing factor) and endothelium-derived contracting factors such as endothelins, vasoconstrictor prostanoids and superoxide anions (Moncada et al., 1991; Luscher & Barton, 1997). The contractions induced by myricetin were abolished in endothelium denuded rings indicating that myricetin either removed an endothelium derived vasodilator or released an endothelium-dependent vasoconstrictor. The augmented contractile responses to myricetin in the presence of the NO synthase inhibitor L-NAME indicates that these responses are not mediated by inhibition of NO synthesis and supports the idea that basal or myricetin-induced NO release partly inhibits the increase of vascular tone induced by myricetin. Cleavage of membrane lipids by PLA2 causes the release of arachidonic acid that can be metabolized via the cyclo-oxygenase and TXA2 synthase enzymes into TXA2 (Mentha & Roberts, 1983). TXA2 is produced mainly by the media layer of intact blood vessels (Brunkwall et al., 1987), although some is synthesized by vascular endothelial cells (Ingermam-Wojenski et al., 1981). The contractions induced by myricetin were inhibited either in the presence of the phospholipase A2 inhibitor, quinacrine, or the cyclo-oxygenase inhibitor, indomethacin. These findings suggest that the metabolism of arachidonic acid through the cyclo-oxygenase pathway plays a key role in the endothelium-dependent responses to myricetin. Prostaglandin H2 (PGH2), the unstable precursor of prostaglandin F, prostaglandin E2 and TXA2 induces a contractile response in rat aorta (Förstermann et al., 1984). In the present study, we found that the activation of TXA2-PGH2 receptors on vascular smooth muscle is implicated in the endothelium-dependent contractions induced by myricetin. This conclusion is based on the observation that the TXA2-PGH2 receptor antagonist ifetroban (Ogletree et al., 1992) abolished this contractile response. Moreover, we found that myricetin, at a concentration at which it induces its maximal contractile response, stimulated TXB2 production in intact aortic rings and cultured BAEC. However, we found that TXB2 accumulation was only decreased by 50% by endothelial denudation while the contractile responses were abolished. The thromboxane synthase inhibitor dazoxiben partly inhibited myricetin-induced contractions suggesting that TXA2 is involved in these effects. However, PG endoperoxides (PGG2 and PGH2) released from the endothelium may also participate, accounting for the endothelial dependence of the contractile response. A number of reports have shown that flavonoids can modulate arachidonic acid metabolism (for a review see Alcaraz & Ferrándiz, 1987). While most flavonoids were able to inhibit both the platelet cyclo-oxygenase and lipoxygenase pathways at relatively high concentrations (50 μM), only myricetin (10 μM) increased the conversion of arachidonic acid into TXB2 (Landolfi et al., 1984). Furthermore, we cannot exclude that an additional source of TXA2 in our study is from platelets adhered to the intact aorta and that physical endothelium-denudation would also remove the platelets.

The contractile response induced by myricetin was abolished in aortic rings incubated in a Ca2+-free medium. Extracellular Ca2+ may be required either for smooth muscle contraction or for endothelial production of TXA2. Activation of TXA2-PGH2 receptors by TXA2 mimetics, such as U46619, produces vascular smooth muscle contractions by increasing the calcium sensitivity of the contractile apparatus through an activation of protein kinase C with no significant change or relatively small increases in [Ca2+]i (Himpens et al. 1990; Jiang et al., 1994). Thus, an extracellular source of Ca2+ does not appear to be a requisite for TXA2-induced contraction. Moreover, PLA2 isoforms involved in signal transduction are regulated by [Ca2+]i (Kramer & Sharp, 1997) and, therefore, an increase in [Ca2+]i in the endothelial cells might be the triggering signal for myricetin-induced TXA2 production. In fact, myricetin increased [Ca2+]i in BAEC. Furthermore, in Ca2+-free solution myricetin failed to increase both [Ca2+]i and TXB2 levels, indicating that myricetin stimulated extracellular Ca2+ influx into the endothelial cells rather than its release from intracellular stores. The inability of verapamil to inhibit myricetin evoqued contractions suggests that myricetin stimulates Ca2+ entry to endothelial cells through a pathway insentitive to this L-type Ca2+ channel blocker.

To the best of our knowledge, there is no information concerning the bioavalability and plasma levels of myricetin in humans. However, myricetin is present in the diet in smaller amounts than the more common flavonoid quercetin which is detected as such or as conjugated active metabolites in plasma from non-supplemented humans at concentrations up to 1.6 μM (Paganga & Rice-Evans, 1997). Thus, it is unlikely that subjects on a normal diet reach plasma concentrations of myricetin as high as those used in the present study. However, it cannot be excluded that these levels might be reached after selected meals (e.g. broad beans and red wine) with a high content of myricetin.

In conclusion, our findings suggest that an activation of PGH2-TXA2 receptors on vascular smooth muscle by the TXA2 released from endothelium by a Ca2+-sensitive activation of the arachidonic acid metabolism is the main mechanism involved on the contractile response induced by myricetin in rat aortic rings.


This work was supported by a DGICYES (PR95-334) and CYCIT Grants (SAF 94–0528, SAF98-0160 and SAF 96–0042).