Novel Insights into the Mode of Action of Vasorelaxant Synthetic Polyoxygenated Chalcones

Polyphenols consumption has been associated with a lower risk of cardiovascular diseases (CVDs) notably through nitric oxide (NO)- and estrogen receptor α (ERα)-dependent pathways. Among polyphenolic compounds, chalcones have been suggested to prevent endothelial dysfunction and hypertension. However, the involvement of both the NO and the ERα pathways for the beneficial vascular effects of chalcones has never been demonstrated. In this study, we aimed to identify chalcones with high vasorelaxation potential and to characterize the signaling pathways in relation to ERα signaling and NO involvement. The evaluation of vasorelaxation potential was performed by myography on wild-type (WT) and ERα knock-out (ERα-KO) mice aorta in the presence or in absence of the eNOS inhibitor Nω-nitro-L-arginine methyl ester (L-NAME). Among the set of chalcones that were synthesized, four (3, 8, 13 and 15) exhibited a strong vasorelaxant effect (more than 80% vasorelaxation) while five compounds (6, 10, 11, 16, 17) have shown a 60% relief of the pre-contraction and four compounds (12, 14, 18, 20) led to a lower vasorelaxation. We were able to demonstrate that the vasorelaxant effect of two highly active chalcones was either ERα-dependent and NO-independent or ERα-independent and NO-dependent. Thus some structure-activity relationships (SAR) were discussed for an optimized vasorelaxant effect.


Introduction
According to the World Health Organization (WHO), cardiovascular diseases are the first leading cause of death worldwide [1]. Most cardiovascular diseases (CVDs) are initially due to hypertension and/or atherosclerosis that could be initiated by endothelial dysfunction [2,3]. These pathological conditions lead to alterations of the endothelium resulting in imbalanced endothelial factors, including a decrease in nitric oxide (NO) bioavailability, vascular constriction, and platelet activation.
Interestingly, epidemiological studies indicate that the prevalence of CVDs is lower in premenopausal women compared to age-matched men and significantly increases in women after menopause, suggesting a protective role of estrogens in premenopausal women [4,5]. Both estrogen receptor α (ERα) and estrogen receptor β (ERβ) are involved in cardiovascular protection with a predominant role of ERα at the vascular level [6,7]. ERα, expressed by endothelial cells and co-localized with endothelial NO synthase (eNOS), has been identified as an eNOS activator through an Src/ERK1/2-dependent pathway both in human primary endothelial cells and mice [8,9]. These

Chemistry and cLogP
Synthetic polyhydroxylated chalcones 3, 6, 8 and 10 were synthesized through a two-step procedure starting from the corresponding methoxymethyl (MOM)-protected acetophenones and benzaldehydes. Although this strategy required an extra deprotection step compared to Claisen-Schmidt condensation of phenolic starting materials, it has led to a better overall yield as observed for the synthesis of chalcone 8 compared to literature data [20,21]. Due to reactivity and efficiency issues in basic medium, chalcones 11, 13 and 17 bearing a hydroxyl (OH) group in the C-4 position have been prepared through an acid-catalyzed coupling mechanism using thionyl chloride (SOCl 2 ) in ethanol (EtOH). Thus it avoided the need for a synthesis strategy with a MOM protection-deprotection approach [22]. Nevertheless, in situ generated hydrochloric acid (HCl) catalysis was unsuccessful when applied to the synthesis of diphenolic chalcone 18. Thus, it has been prepared under microwave (MW) irradiations in the presence of piperidine [23]. In this work two other synthetic chalcones 12 and 14, bearing a phenol function on the C-2 ' position, have been prepared through selective demethylation of the corresponding 2'-methoxychalcones 13 and 15 in the presence of aluminum chloride (AlCl 3 ) [24]. It is worth mentioning that the demethylation of hexamethoxychalcone 16 led only to 17 (52% yield) without any trace of 18 (Scheme 1). Finally, the calculated octanol-water partition coefficient (cLogP) were similar, ranging from 1.80 for synthetic chalcone 6 to 3.33 for synthetic chalcone 14 (Table 1). Scheme I: General scheme for the synthesis of chalcones. Synthetic chalcones were synthesized through the coupling of the corresponding either MOM-protected or phenolic acetophenones and benzaldehydes. Hydrolysis of MOM groups or selective demethylation in the presence of AlCl3 were achieved. These procedures led to thirteen synthetic polyoxygenated chalcones whose vasorelaxant effect was evaluated.
Scheme 1. General scheme for the synthesis of chalcones. Synthetic chalcones were synthesized through the coupling of the corresponding either MOM-protected or phenolic acetophenones and benzaldehydes. Hydrolysis of MOM groups or selective demethylation in the presence of AlCl 3 were achieved. These procedures led to thirteen synthetic polyoxygenated chalcones whose vasorelaxant effect was evaluated.

Evaluation of Vasorelaxant Activity
Vasorelaxation induced by each synthetic chalcone was first evaluated on WT mice thoracic aorta rings. Compounds 3, 8, 13 and 15 with identical substituents at C-2' and C-4' positions, absence of substituent at the C-6' position and presence of OH and/or OCH 3 groups at C-3, C-4 and C-5 positions induced a high vasorelaxation reaching more than 80%. Compounds 6, 10, 11, 16 and 17 with OH or OCH 3 group at the C-6' position or unsubstituted at C-3 and C-5 positions induced an average vasorelaxation between 50% to 80%. Finally, compounds 12, 14, 18 and 20 with both OH and OCH 3 groups at C-2', C-4' and/or C-6' positions or a bulkier OCH 2 COOH group at C-3 position exerted a low vasorelaxation effect below 50% (Figure 1 and Table 2

Evaluation of Vasorelaxant Activity on ERα KO Mice Aorta
To evaluate the involvement of the ERα pathway in the chalcones-induced vasorelaxation, experiments have been conducted on ERα KO mice aortas. To focus on the role of OH or OCH 3 groups at C-6' position with synthetic chalcones exhibiting a high vasorelaxant effect, compounds 3, 6, 8, 10, 13, 15, 16 and 17 have been tested on ERα KO mice thoracic aorta rings. Interestingly, only the synthetic chalcone 3 induced a significant lower vasorelaxation in ERα KO mice thoracic aorta rings compared to WT mice thoracic aorta rings (56.23 ± 4.91% vs. 84.28 ± 3.24%, respectively) while no significant differences were observed with synthetic chalcones 6, 8, 10, 13, 15, 16 and 17 (Figures 2  and 3 and Table 2). . Mice were ovariectomized 7 days before the experiment. Vessels were pre-contracted with U46619, phenylephrine, and serotonin, at 80 % of their maximal response. The presence of functional endothelium was assessed by determining the ability of acetylcholine (10 µM) to induce more than 50% relaxation of pre-contracted rings. Results are expressed in percentage of precontraction (y-axis) following the concentration (x-axis). Only the synthetic chalcone 3 induced a significant lower vasorelaxation in ERα KO mice thoracic aorta rings compared to WT mice thoracic aorta rings (56.23 ± 4.91% vs 84.28 ± 3.24% respectively) while no significant differences were observed with chalcones 6, 8 and 10. N = 4-7, * p < 0.05 (Mann-Whitney). . Mice were ovariectomized 7 days before the experiment. Vessels were pre-contracted with U46619, phenylephrine, and serotonin, at 80 % of their maximal response. The presence of functional endothelium was assessed by determining the ability of acetylcholine (10 µM) to induce more than 50% relaxation of pre-contracted rings. Results are expressed in percentage of precontraction (y-axis) following the concentration (x-axis). Only the synthetic chalcone 3 induced a significant lower vasorelaxation in ERα KO mice thoracic aorta rings compared to WT mice thoracic aorta rings (56.23 ± 4.91% vs. 84.28 ± 3.24% respectively) while no significant differences were observed with chalcones 6, 8 and 10. N = 4-7, * p < 0.05 (Mann-Whitney). . Mice were ovariectomized 7 days before the experiment. Vessels were pre-contracted with U46619, phenylephrine, and serotonin, at 80 % of their maximal response. The presence of functional endothelium was assessed by determining the ability of acetylcholine (10 µM) to induce more than 50% relaxation of pre-contracted rings. Results are expressed in percentage of precontraction (y-axis) following the concentration (xaxis). No significant differences were observed with synthetic chalcones 13, 15, 16 and 17. N = 4-7 (Mann-Whitney).

Evaluation of Vasorelaxant Activity in the Presence of Nω-nitro-L-arginine methyl ester (L-NAME)
In order to evaluate the involvement of the NO pathway, experiments have then been performed after incubation for 20 min with NOS inhibitor (10 -4 M) on both WT and ERα KO mice thoracic aorta rings. To focus on the role of groups at C-2', C-4' and C-4 positions, synthetic chalcones 3, 8, 13 and 15 have been tested. Interestingly, L-NAME prevented the chalcone-induced vasorelaxation only for chalcone 13 in WT mice thoracic aorta rings at the highest concentration (10 -2 g/L, Figure 4 and Table  2). It is noteworthy that L-NAME seemed to prevent the chalcone-induced vasorelaxation for compounds 3 and 8 at low concentrations and this effect is not recovered at 10 -2 g/L. No difference was observed for the synthetic chalcone 15 ( Figure 4A and 4B). Similarly, the presence of L-NAME on ERα KO mice thoracic aorta rings prevented the vasorelaxation only with the synthetic chalcone 13 (81.96 ± 2.46% vs 26.96 ± 6.72%) while having no effect with synthetic chalcones 3, 8 and 15 ( Figure  5 and Table 2). . Mice were ovariectomized 7 days before the experiment. Vessels were pre-contracted with U46619, phenylephrine, and serotonin, at 80 % of their maximal response. The presence of functional endothelium was assessed by determining the ability of acetylcholine (10 µM) to induce more than 50% relaxation of pre-contracted rings. Results are expressed in percentage of precontraction (y-axis) following the concentration (x-axis). No significant differences were observed with synthetic chalcones 13, 15, 16 and 17. N = 4-7 (Mann-Whitney).

Evaluation of Vasorelaxant Activity in the Presence of Nω-nitro-L-arginine methyl ester (L-NAME)
In order to evaluate the involvement of the NO pathway, experiments have then been performed after incubation for 20 min with NOS inhibitor (10 −4 M) on both WT and ERα KO mice thoracic aorta rings. To focus on the role of groups at C-2', C-4' and C-4 positions, synthetic chalcones 3, 8, 13 and 15 have been tested. Interestingly, L-NAME prevented the chalcone-induced vasorelaxation only for chalcone 13 in WT mice thoracic aorta rings at the highest concentration (10 −2 g/L, Figure 4 and Table 2). It is noteworthy that L-NAME seemed to prevent the chalcone-induced vasorelaxation for compounds 3 and 8 at low concentrations and this effect is not recovered at 10 −2 g/L. No difference was observed for the synthetic chalcone 15 ( Figure 4A,B). Similarly, the presence of L-NAME on ERα KO mice thoracic aorta rings prevented the vasorelaxation only with the synthetic chalcone 13 (81.96 ± 2.46% vs. 26.96 ± 6.72%) while having no effect with synthetic chalcones 3, 8 and 15 ( Figure 5 and Table 2). Mice were ovariectomized 7 days before the experiment. Vessels were pre-contracted with U46619, phenylephrine, and serotonin, at 80 % of their maximal response. The presence of functional endothelium was assessed by determining the ability of acetylcholine (10 µM) to induce more than 50% relaxation of pre-contracted rings. Results are expressed in percentage of precontraction (y-axis) following the concentration (x-axis). L-NAME only prevented the chalconeinduced vasorelaxation for the synthetic chalcone 13 at the highest concentration (10 -2 g/L) and seemed to prevent the chalcone-induced vasorelaxation for compounds 3 and 8 at low concentrations.

Discussion
By inhibiting the angiotensin-converting enzyme (ACE) and inducing high vasodilation, the anti-hypertensive effect of amino-chalcones has been demonstrated in vitro and in vivo on rats, rabbits, and dogs [25,26]. Several molecular targets such as cyclooxygenase [27] and potassium channels [28], metabolism pathways such as triacylglycerol synthesis pathway [29][30][31] and pro-angiogenic properties [32] have also been identified to explain the potential role of chalcones in the prevention and/or treatment of cardiovascular diseases. However, all these effects have been mainly highlighted with functionalized chalcones bearing amino, prenyl or sulfonyloxy groups and little data are available for polyoxygenated chalcones [25].
The present study described the synthesis of thirteen synthetic chalcones bearing oxygenated substituents (OH and OCH 3 ). The pharmacological characterization of their vascular activity including the involvement of NO and ERα-dependent pathways allowed the identification of two promising compounds (3 and 13) inducing an endothelium-dependent vasorelaxation. Precisely, the synthetic chalcone 3 is preferable for an ERα-dependent vasorelaxation while the synthetic chalcone 13 is preferable for an ERα-independent and NO-dependent vasorelaxation.
ERα has already been identified as a target for chalcones, especially in the field of cancer [17,33] and it is now well accepted that estrogens can protect against cardiovascular diseases [11,34,35]. At the vascular level, ERα pathway can, through a non-genomic mechanism, lead to an endothelium-dependent vasodilation [34,36]. The identification in this work of compounds 3 and 13 with an ERα-or NO-dependent mechanism, respectively, clearly confirms that these two chalcones exert an endothelium-dependent vasorelaxation in mice thoracic aorta.
Other vascular mechanisms as phosphodiesterase (PDE) inhibition or calcium channel blocking have also been described to disclose chalcones-induced vasorelaxation [9,15]. These targets, mainly located on vascular smooth muscle cells, would explain the NO-and ERα-independent vasorelaxation of the other synthetic chalcones tested. In addition, these mechanisms may also be involved to explain the remaining vasorelaxation observed on ERαKO mice in the presence of L-NAME. Further studies are however required to confirm the involvement of PDE or calcium channel in the NO-and ERα-independent vasorelaxation of the present synthetic chalcones.
Modifications of the substitution pattern contributed to the study of structure-activity relationships. The synthetic chalcone 14, bearing an OH group at C-2' position and an OCH 3 group at C-4' position exerted a low vasorelaxation effect (less than 50%). Synthetic chalcones 6, 10, 16, 17 and 18, bearing an OH group or an OCH 3 group at the C-6' position, exerted an average or a low vasorelaxation effect (less than 80%). Altogether, the data collected here demonstrated that both C-2' and C-4' positions must bear the same group, either OH or OCH 3 , and an unsubstituted C-6' position is required for a better activity. Accordingly, in a study evaluating the relaxant potency and efficacy in rat thoracic aorta of prenylated and allylated chalcones, an increased vasorelaxant efficacy was also identified for chalcones without any OH group at the C-6' position [37]. Additionally, the number of OH groups bearing by the chalcone backbone was also described as an important factor for a higher activity [37]. The replacement of OH groups by OCH 3 groups revealed that the corresponding compounds did not exhibit a higher relaxant efficacy as demonstrated with compound 3 and 8 compared to compound 13 and 15. These data confirmed the importance of the number of OH groups but they also highlighted that OCH 3 groups must be considered. More precisely, we identified that the 3,5-dimethoxy-4-hydroxybenzene moiety is required for both an ERα-or NO-dependent vasorelaxation in both 2',4'-dihydroxychalcone or 2',4'-dimethoxychalcone series.
Recently, isoliquiritigenin has been described to induce an endothelium-independent vasodilation through a calcium-activated potassium channel-dependent (BK Ca ) pathway in mice [15]. By substituting OH groups by OCH 3 groups at the C-2' and C-4' positions, isoliquiritigenin derivatives such as compound 11 have been synthsized. This synthetic chalcone exerted a comparable vasorelaxation than isoliquiritigenin through an endothelium-independent pathway. The involvement of the BK Ca in compound 11-induced vasorelaxation now needs to be explored in order to further understand the role of the methoxylation at C-2' and C-4' positions in the BK Ca -dependent vasorelaxation.
Similar cLogP, ranging from 1.80 to 3.33 were observed for the thirteen synthesized chalcones. We and others demonstrated that a higher cLogP value of flavonoids was associated with a higher vasorelaxation potential [9,38]. However, this association has been observed for flavonoids with very low lipophilicity (cLogP < 1) as anthocyanidins [9]. In the present study, no association between cLogP and the vasorelaxant potential has been observed, likely because of cLogP mainly ranging around 2 and 3.
In conclusion, coupling reactions between acetophenones and benzaldehydes led to thirteen synthetic polyoxygenated chalcones bearing OH groups and/or OCH 3 groups and/or a bulkier OCH 2 COOH group. The vascular effects of chalcones were assessed on WT and ERα KO mice thoracic aorta and revealed four compounds with a very high vasorelaxation potential (3, 8, 13 and 15). More precisely, the synthetic chalcone 3 exerted an ERα-dependent and NO-independent vasorelaxation while the synthetic chalcone 13 exerted an ERα-independent and NO-dependent vasorelaxation. Structure-activity relationships (SAR) study allowed us to identify the 3,5-dimethoxy-4-hydroxybenzene moiety as a pharmacophore for both an ERα-or NO-dependent vasorelaxation. Finally, the current study provides new promising data for the preventive and therapeutic use of chalcones in the field of CVDs.

Chemistry
All chemicals were purchased from Acros Organics (Geel, Belgium). Chromatographic separations for purification processes were achieved with flash chromatography IntelliFlash 310 (Analogix) using pre-packed C18 (Interchim) or silica gel column Chromabond®flash RS column (Macherey-Nagel). 1 H and 13 C NMR were recorded on Jeol GSX WB 270MHz. Chemical shifts are reported in parts per million (δ ppm).
Different procedures (A, B or C) have been used to achieve the Claisen Schmidt condensation (Scheme 2).
In conclusion, coupling reactions between acetophenones and benzaldehydes led to thirteen synthetic polyoxygenated chalcones bearing OH groups and/or OCH3 groups and/or a bulkier OCH2COOH group. The vascular effects of chalcones were assessed on WT and ERα KO mice thoracic aorta and revealed four compounds with a very high vasorelaxation potential (3, 8, 13 and  15). More precisely, the synthetic chalcone 3 exerted an ERα-dependent and NO-independent vasorelaxation while the synthetic chalcone 13 exerted an ERα-independent and NO-dependent vasorelaxation. Structure-activity relationships (SAR) study allowed us to identify the 3,5-dimethoxy-4-hydroxybenzene moiety as a pharmacophore for both an ERα-or NO-dependent vasorelaxation. Finally, the current study provides new promising data for the preventive and therapeutic use of chalcones in the field of CVDs.

Chemistry
All chemicals were purchased from Acros Organics (Geel, Belgium). Chromatographic separations for purification processes were achieved with flash chromatography IntelliFlash 310 (Analogix) using pre-packed C18 (Interchim) or silica gel column Chromabond® flash RS column (Macherey-Nagel). 1 H and 13 C NMR were recorded on Jeol GSX WB 270MHz. Chemical shifts are reported in parts per million (δ ppm).
Different procedures (A, B or C) have been used to achieve the Claisen Schmidt condensation. Procedure A: To a solution of acetophenone (0.5 mmol) in ethanol are added benzaldehyde (0.5 mmol) and a 60% solution of potassium hydroxide (KOH). The resulting reaction mixture is stirred at room temperature for 24 H. Then 10 mL of water is added. The corresponding chalcone is collected by vacuum filtration or purified by silica gel flash chromatography. When MOM protecting groups are used, their hydrolysis is undertaken with 2N HCl (2 eq / MOM group) at reflux in a methanol Scheme 2. Strategy for Procedure A.
Procedure A: To a solution of acetophenone (0.5 mmol) in ethanol are added benzaldehyde (0.5 mmol) and a 60% solution of potassium hydroxide (KOH). The resulting reaction mixture is stirred at room temperature for 24 H. Then 10 mL of water is added. The corresponding chalcone is collected by vacuum filtration or purified by silica gel flash chromatography. When MOM protecting groups are used, their hydrolysis is undertaken with 2N HCl (2 eq/MOM group) at reflux in a methanol solution. Completion of the reaction is determined by thin-layer chromatography (TLC). After cooling at room temperature (RT), water is added to the reaction mixture and the chalcone is extracted three times with ethyl acetate (EtOAc). The final product is purified by column chromatography.  194.6, 168.6, 167.4, 150.6, 147.2, 141.0, 134.6, 128.4, 120.1, 115.8, 110.1, 108.6, 104.7, 57.9. Procedure B: Thionyl chloride (0.5 eq) is added dropwise to a solution of acetophenone (1 eq) and benzaldehyde (1 eq) in 3 mL of dichloromethane (DCM). After being stirred at room temperature for 24 h, the reaction mixture is quenched with 10 mL of water. Crude chalcone is collected through filtration and could be purified by silica gel column chromatography when it is required.

Statistical Analysis
The statistical significance of the differences between samples was determined using the Mann-Whitney test. Differences were considered significant for p < 0.05. All statistical tests were carried out with GraphPad Prism 5 (GraphPad Software, Inc, San Diego, CA, USA).