Calycosin and Formononetin Induce Endothelium-Dependent Vasodilation by the Activation of Large-Conductance Ca2+-Activated K+ Channels (BKCa)

Calycosin and formononetin are two structurally similar isoflavonoids that have been shown to induce vasodilation in aorta and conduit arteries, but study of their actions on endothelial functions is lacking. Here, we demonstrated that both isoflavonoids relaxed rat mesenteric resistance arteries in a concentration-dependent manner, which was reduced by endothelial disruption and nitric oxide synthase (NOS) inhibition, indicating the involvement of both endothelium and vascular smooth muscle. In addition, the endothelium-dependent vasodilation, but not the endothelium-independent vasodilation, was blocked by BKCa inhibitor iberiotoxin (IbTX). Using human umbilical vein endothelial cells (HUVECs) as a model, we showed calycosin and formononetin induced dose-dependent outwardly rectifying K+ currents using whole cell patch clamp. These currents were blocked by tetraethylammonium chloride (TEACl), charybdotoxin (ChTX), or IbTX, but not apamin. We further demonstrated that both isoflavonoids significantly increased nitric oxide (NO) production and upregulated the activities and expressions of endothelial NOS (eNOS) and neuronal NOS (nNOS). These results suggested that calycosin and formononetin act as endothelial BKCa activators for mediating endothelium-dependent vasodilation through enhancing endothelium hyperpolarization and NO production. Since activation of BKCa plays a role in improving behavioral and cognitive disorders, we suggested that these two isoflavonoids could provide beneficial effects to cognitive disorders through vascular regulation.


Introduction
Calycosin and formononetin ( Figure 1) are two structurally similar isoflavonoids that are present abundantly in traditional Chinese medicine (TCM) such as Radix Astragali (Huang Qi) and phytoestrogenic herb including Trifolium pretense L. (red clover), and they have a long clinical history in treating various cardiovascular diseases [1,2]. Previous studies have shown that calycosin and formononetin produced antihypertensive effects and improved endothelial and cardiovascular functions [3][4][5][6]. They have been shown to display vasoactive effects in various vascular beds [3,[5][6][7]. In rat aorta, calycosin induced vasodilation mainly by inhibiting voltage-dependent Ca 2+ channel (VDCC) in vascular smooth muscle [7], while formononetin caused vasodilation by releasing nitric oxide (NO) in endothelial cells, as well as by the activation of large-conductance Ca 2+activated K + (BK Ca ) and ATP-sensitive potassium (K ATP ) channels in aortic smooth muscle cells [6]. In addition, these two isoflavonoids were reported to ameliorate cerebral ischemia and reperfusion injury by improving endothelial dysfunction [8,9]. These observation led us to investigate the pharmacological underlying mechanisms of calycosin and formononetin in small resistance arteries (internal diameter ≤ 300 m) and vascular endothelial cells.
Small resistance arteries are major sites of peripheral vascular resistance and are closely related to endothelial dysfunction and the pathogenesis of hypertension [10,11]. Interestingly, it has been shown that endothelium-dependent hyperpolarization (EDH) was more pronounced in small resistance arteries than large conduit arteries such as aorta [10,12]. In the vascular walls, calcium-activated potassium (K Ca ) channel is the main contributor for endothelium-derived hyperpolarizing factor-(EDHF-) mediated responses, which plays a crucial role in the regulation of vascular tone and the maintenance of systemic blood pressure [13,14]. Recently, endothelial K Ca channels have been used as new drug targets for cardiovascular diseases such as hypertension to stimulate EDHF and NO production to improve endothelial dysfunction [13,[15][16][17]. There are three types of K Ca channels based on their conductances, including IK Ca (intermediate conductance), SK Ca (small conductance), and BK Ca (large conductance) [18]. Although IK Ca and SK Ca are the major K Ca channels present in the endothelial cells of arteries, BK Ca channels have been identified in the endothelium of rat pulmonary and mesenteric arteries [19,20] and cultured endothelial cells [21,22]. It has been suggested that BK Ca channel has a compensatory role for improving vasoreactivity in environment such as hypertension and cardiovascular diseases [23,24]. In addition, it was shown that BK Ca channel acts as a negative feedback mechanism for vascular dysfunction impaired by reactive oxygen species (ROS) and is overexpressed in diseases associated with endothelial dysfunction [25][26][27].
In the present study, we investigated the effects of calycosin and formononetin in rat mesenteric resistance arteries and their underlying mechanisms with a focus on endothelial K + channel. We demonstrated that calycosin and formononetin induced endothelium-dependent vasodilation through NO production and BK Ca channel activation. We also showed that calycosin and formononetin increased NO production through endothelial nitric oxide synthase (eNOS) and neuronal nitric oxide synthase (nNOS) pathway and activated endothelial BK Ca channels in human umbilical endothelial cells (HUVEC). Taken together, our study demonstrated that calycosin and formononetin are endothelial BK Ca channel activators, suggesting a novel mechanism for vasodilation by these isoflavones, and they might be potential for treating vascular and cerebrovascular diseases associated with endothelial dysfunction.

Chemicals and Reagents.
Calycosin and formononetin were purchased from Shanghai Forever Biotech (Shanghai, China), and they were dissolved in dimethyl sulphoxide (DMSO). All the chemicals were purchased from Sigma-Aldrich (St. Louis, MO, USA), while the cell culture reagents were from Gibco (Carlsbad, CA, USA). All the antibodies used for immunoblotting were purchased from Cell Signaling Technology (Danvers, MA, USA).

Animals.
All the procedures were carried out according to the ethical guidelines of the Institute of Chinese Medical Sciences (ICMS), University of Macau, and NIH guide for the Care and Use of Laboratory Animals. Male Sprague-Dawley rats (250 ± 350 g) were killed by cervical dislocation.

Preparation of Resistance Mesenteric
Arteries. The mesenteric arterial beds were dissected from rats immediately after cervical dislocation and were kept in cold Krebs-Henseleit buffer with the following compositions (mM): NaCl, 118; KCl, 4.7; MgSO 4 , 1.2; KH 2 PO 4 1.2; NaHCO 3 , 25; CaCl 2 , 2.5; D-glucose, 5.5. The Krebs-Henseleit buffer was continuously gassed with a mixture of 95% O 2 /5% CO 2 to maintain a pH of 7.4. Small (the third branches of the mesenteric artery) mesenteric arteries were placed in a 4-chamber wire myograph (DMT, Tissue Bath System 700MO, Aarhus, Denmark) and were maintained at 37 ∘ C in Krebs-Henseleit buffer. The normalized protocol was used as described previously [28]. The tension was recorded by the PowerLab recording system (ADInstruments, Hastings, UK). In endothelium-denuded experiments, the endothelium was removed by rubbing the inner surface of the segment with human hair. In endothelium-intact experiments, the presence of a functional endothelium was examined by precontracting the arteries with methoxamine (10 M) and then was relaxed by carbachol (10 M), and this vasodilation was greater than 80%.

Cell
Culture. Human umbilical vein endothelial cells (HUVECs) were purchased from Life Technologies (Carlsbad, CA, USA) and were cultured in Hams F-12k nutrient medium supplemented with 15% fetal bovine serum, 1% penicillin-streptomycin, 100 mg/mL heparin, and 10 mg/mL endothelial cell growth supplement (Sigma-Aldrich, St. Louis, MO, USA) at 37 ∘ C under an atmosphere of 5% CO 2 in air. The cells were used in passages 2-6. Plated cells were allowed to adhere for 2 h before patch clamp experiments. For the experiments measuring NO production, and the expressions of different nitric oxide synthases (NOS), calycosin and formononetin were treated for 1 h in HUVEC.

NO Production Assay. NO was measured by a Nitric
Oxide Assay Kit (Abcam, Cambridge, UK) following the manufacturer's instruction. Briefly, 75 L of cell supernatants was mixed with 5 L enzyme cofactor solution and 5 L nitrate reductase. Following 2 h of incubation for converting nitrates to nitrites, 5 L enhancer was added to each sample and incubated for 30 min. 5 L DAN Probe was then added and incubated for 10 min. After that, 5 L NaOH was added in the mixture for 10 min. The fluorescence intensity was detected by a microplate reader (SpectaMax M5, Molecular Devices, USA) using excitation at 360 nm and emission at 450 nm wavelengths.

Western Blot Analysis.
After indicated treatment, the protein was extracted with ice-cold lysis buffer, and the concentrations of lysates were measured by the bicinchoninic acid kit (Pierce, US). 30 g proteins were used and separated by 10% SDS-PAGE gels and were transferred onto the nitrocellulose membranes. Membranes were incubated with primary antibodies (eNOS, p-eNOS, nNOS, and iNOS antibody using 1/1000 dilution, whereas GAPDH antibody using 1/4000 dilution) overnight at 4 ∘ C, and secondary antibodies (anti-rabbit with 1/1000 dilution) for 1 hr, and blots were developed by enhanced chemiluminescence (GE Healthcare Life Sciences, UK) with an imaging system (Bio-Rad Laboratories, USA). GAPDH were used as housekeeping controls.

Data and Statistical
Analysis. Data were expressed as mean ± SEM. vasodilation responses in each segment were expressed as a percentage of relaxation. The dose-response curves were fitted to a logistic equation described previously [30]. The maximum percentage of relaxation ( max ) and the concentration required to produce 50% of maximal response tone (EC 50 ) were calculated from the fitted curves. Currents curves were fitted by Boltzmann equations. The curve fitting and statistical analyzes were determined using GraphPad Prism 5 (San Diego, CA, USA). Significant differences were analyzed by -test, one-way ANOVA followed by a Dunnett's test, or two-way ANOVA followed by a Bonferroni post hoc test. < 0.05 was considered as significant.

Effect of Calycosin and Formononetin on NO Production and eNOS, iNOS, and nNOS Expression in HUVEC.
In order to further determine whether these two isoflavonoids could regulate the production of NO and the expression of NOS in endothelial cells, HUVEC was employed as a cellular model. HUVEC is a widely in vitro cell model for the study of the regulation of endothelial function and the vascular diseases [31]. Figure 4(a) showed that calycosin increased NO production in a dose-dependent manner in HUVEC, and similar effect was also observed with formononetin ( Figure 4(b)). There are three isoforms of NOS, including eNOS, inducible nitric oxide synthase (iNOS), and nNOS, which are responsible for the generation of NO in the vascular endothelium [32]. It has been reported that eNOS and nNOS are coexpressed in endothelial cells while iNOS is not [33]. Our results demonstrated that both calycosin (100 M) and formononetin (100 M) significantly induced the activation of eNOS (Figures 4(c) and 4(d)). In addition, both isoflavonoids also upregulated nNOS expression (Figures 4(c) and 4(d)). In contrast, iNOS expression was unaffected by neither calycosin nor formononetin (Figures 4(c) and 4(d)). These data suggested that calycosin and formononetin induced endothelium-dependent vasorelaxation via NO production through eNOS and nNOS.

Effect of Calycosin on Outward Currents in HUVEC.
The activation of K Ca is the major mechanism for EDH. We observed that exposure of calycosin induced dose-dependent outward currents in HUVEC recorded by whole cell patch clamp. The whole cell currents of HUVEC were recorded with 250 ms voltage steps between −100 mV and +100 mV from a holding potential of −60 mV. As shown in Figures 5(a) and 5(b), calycosin (10 nM-100 M) increased dose-dependent outward currents in HUVEC ( = 9). At 100 M calycosin, the current at +100 mV was significantly increased ( Figures  5(a) and 5(c), 31.2 ± 1.7 pA/pF; = 7) compared to control (9.9 ± 1.0 pA/pF), and this current was abolished by TEACl, a nonspecific inhibitor of K + channels (1 mM, Figures 5(c) and 5(e); 11.6 ± 1.5 pA/pF; = 6).
Although IK Ca and SK Ca are the major K Ca channels present in the endothelial cells, BK Ca is also expressed moderately, and it was identified in HUVEC [22,23]. We observed that K Ca channels in calycosin induced outward currents in HUVEC. An IK Ca channel inhibitor, apamin (100 nM), had no effect on the outward currents induced by calycosin ( Figure 5(e), = 6). However, it was abolished by BK Ca inhibitors, IbTX (200 nM, Figures 5(d) and 5(e); 13.8 ± 2.8 pA/pF; = 6) and ChTX (200 nM, Figure 5(e); 14.9 ± 0.9 pA/pF; = 7). In addition, by maintaining [Ca 2+ ] i at 250 nM, 500 nM, or 1000 nM, calycosin (100 M) significantly increased the outward currents stimulated by single +100 mV step, compared with Ca 2+ -free solution containing 10 mM EGTA ( Figure 5(f), = 7-9). These data strongly suggested that calycosin mainly activated endothelial BK Ca channels but has minimal effects on IK Ca or SK Ca .

Effect of Calycosin and Formononetin on [Ca 2+ ] i in HUVEC. Endothelial [Ca 2+
] i elevation has been implicated in endothelium-mediated vasodilation [18] and is also needed for K Ca activation [34]. Since calycosin and formononetin increased outward currents via endothelial K Ca channel, we next examined their effects on [Ca 2+ ] i in HUVEC. As expected, we observed a rapid increase in [Ca 2+ ] i in response to calycosin (Figures 7(a) and 7(b)). Similar results also showed that formononetin evoked Ca 2+ influx (Figures 7(a)  and 7(b)). Both of these two drugs induced approximately 45% increases in [Ca 2+ ] i compared to untreated cells. Therefore, these data suggested that calycosin and formononetin activated endothelial BK Ca channels probably by increasing

Discussion
In the present study, we investigated the endothelial beneficial effects of calycosin and formononetin, two isoflavonoids isolated from the well-known antihypertensive herb, Radix Astragali, in rat small resistance arteries and HUVEC. The  the B-ring in calycosin [35]. Both isoflavonoids were shown to provide beneficial effects in vascular tone regulation and improvement in endothelial and cardiovascular dysfunction [3][4][5][6].
Here, we demonstrated that calycosin and formononetin produced very similar effects in rat small mesenteric arteries, that is, inducing vasorelaxation through endotheliumdependent and endothelium-independent pathways. We observed that relaxations elicited by both isoflavonoids could be reduced by endothelial disruption and were sensitive to L-NAME (inhibitor of NOS), apamin plus charybdotoxin (inhibitors of SK Ca and IK Ca ), and iberiotoxin (IbTX) (inhibitor of BK Ca ). Notably, the sensitivity to IbTX was only observed in endothelium-intact, but not in endotheliumdenuded vessels, indicating the involvement of BK Ca channels present in the endothelium. The relaxation elicited by both isoflavonoids was also reduced (to a lesser extent) when the arteries were contracted with high KCl (60 mM) depolarizing solution, indicating the inhibition of VDCC was also partly involved. Pretreatment with indomethacin to inhibit prostaglandin production did not affect the vasodilation, indicating that PGI 2 was not involved. Although calycosin and formononetin are very similar in structure, differential properties of their effects were also observed. Our data showed that relaxation induced by formononetin was sensitive to glibenclamide (K ATP channel inhibitor), but not calycosin-induced relaxation. Besides, we also observed some discrepancies in the effects of calycosin and formononetin in blood vessels from various vascular beds. Previous study intact aortic rings reported that the relaxation elicited by calycosin was endothelium-independent by acting as Ca 2+ channel blocker [7]. However, here in rat mesenteric arteries we observed that calycosin-elicited relaxation is both endothelium-dependent and endothelium-independent. For formononetin, results from previous studies in rat aortic rings showed that formononetin elicited relaxation through endothelium-dependent pathway involving NO synthesis, and endothelium-independent involving iberiotoxin-(IbTX-) sensitive BK Ca channel, glibenclamide-sensitive K ATP channel, and the inhibition of VDCC [5,6,36]. Here in rat mesenteric arteries, we observed very similar effects of formononetin, but in our preparation the sensitivity to IbTX was only observed in endothelium-intact vessels, indicating a more important role of endothelial BK Ca channels in the mesenteric resistance arteries. The discrepancies observed for calycosin and formononetin in rat aorta and mesenteric arteries may be explained by the differential physiological characteristics exhibited by conduit and resistance arteries for vascular homeostasis. In fact, it has been recognized that endothelium-dependent vasoactivities are more pronounced in small resistance arteries than in large conduit arteries such as aorta, and small resistance arteries are closely related to endothelial dysfunction [10,12]. Thus, our results suggested that calycosin and formononetin could effectively promote endothelial functions in the small resistance arteries, and endothelial BK Ca channels might have an important role. BK Ca channels are mainly expressed in vascular smooth muscle cells, but recent studies demonstrated that BK Ca channels are also present in vascular endothelium or freshly isolated endothelial cells from the small resistance arteries [19,20]. When expressed in endothelial cells, BK Ca channels are observed to regulate NO and EDH production, and their important roles have been implicated in disease conditions such as ischemic hypoxia [19,23,37,38]. In line with this, we demonstrated that both calycosin and formononetin significantly increased NO production and upregulated the activities and expressions of eNOS and nNOS, without affecting iNOS in HUVEC. It has been suggested that endothelial BK Ca channel might serve as a compensatory mechanism for improving vasoreactivity in disease environments such as hypertension and could be a potential therapeutic target for the regulation of blood pressure and flow through increased NO production vascular hyperpolarization [20,39].
By using whole cell patch clamp, we further showed that IbTX-sensitive outward-rectifying currents were induced by exposure to calycosin or formononetin in a dose-dependent manner in HUVEC, an endothelial cell model commonly used for endothelial functions. Moreover, these currents were also sensitive to charybdotoxin (also an inhibitor at BK Ca channel), but not apamin (selective SK Ca channel inhibitor). Buffering intracellular Ca 2+ by EGTA significantly reduced calycosin-and formononetin-induced outward currents, indicating that the activation of the K + channels was dependent on intracellular calcium which further indicated the characteristic of BK Ca channel [40]. We further demonstrated that calycosin and formononetin could induced an [Ca 2+ ] i increase in HUVEC, providing more evidences that calycosin and formononetin activate endothelial BK Ca channel by stimulating [Ca 2+ ] i increase.
Until recently, the emerging view of cerebrovascular dysregulation was implicated not only in cerebrovascular diseases, such as stroke, but also in neurodegeneration, like Alzheimer's disease (AD) [41]. Particularly, the inhibition of BK Ca activity was observed in 3xTg AD model mice and might be involved in early dysfunction in the AD brain [42]. Several studies suggested that the activation of BK Ca channels might be new therapeutic target for improving behavioral and cognitive disorders [43,44]. Since our results demonstrated that calycosin and formononetin act as novel BK Ca activators and also played a role in the regulation in small resistance arteries, we postulated these two isoflavonoids might have potential effects on cognitive disorders through the regulation of cerebral microcirculation by activating BK Ca channels.

Conclusions
In summary, our findings demonstrated that calycosin and formononetin induced vasodilation in rat small mesenteric arteries involving both the endothelium and the vascular smooth muscle. The endothelium-dependent responses were associated with eNOS/nNOS-dependent NO production and endothelium hyperpolarization, possibly by directly activating endothelial BK Ca channels. Therefore, we suggested that these isoflavonoids might provide potential therapeutic regiments to improve endothelial functions for treating