Bilobalide increases SK_Ca current via 5-HT_2BR-PLC–PKC pathway, causing hyperpolarization and enhanced EET production in human coronary artery endothelial cells

Abstract

Background

Ginkgo biloba extract and its active component bilobalide induce endothelium-dependent vasodilation in coronary and other vascular beds. This effect is associated with increased Ca²⁺-activated K⁺ channel activity in endothelial cells. We aimed to identify ion channels targeted by bilobalide in human coronary artery endothelial cells (HCAECs) and to explore the underlying mechanisms.

Methods

Whole-cell patch-clamp recordings were performed on cultured HCAECs. Epoxyeicosatrienoic acids (EETs) and prostacyclin were quantified with enzyme-linked immunosorbent assays; nitric oxide and hydrogen peroxide were assessed with commercial kits.

Results

Bilobalide dose-dependently increased whole-cell current through activation of the small-conductance Ca²⁺-activated K⁺ (SK_Ca) channel, but not through Cl⁻, transient receptor potential, or other K⁺ channels. The current increase persisted when all channels except SK_Ca were blocked and was abolished by apamin, a selective SK_Ca blocker; additionally, apamin prevented bilobalide-stimulated current. Bilobalide also induced HCAEC hyperpolarization, which was inhibited by apamin. The bilobalide-induced current increase was blocked by antagonists of 5-hydroxytryptamine (5-HT)_1/2 and 5-HT_2B receptors, phospholipase C, and protein kinase C, but not by inhibitors of 5-HT_1B receptors, cytochrome P450, or G_i/o proteins. Furthermore, bilobalide stimulated EET production but not nitric oxide, prostacyclin, or hydrogen peroxide production; this effect was also apamin-sensitive.

Conclusions

Bilobalide increases HCAEC current and induces hyperpolarization by activating SK_Ca via the 5-HT_2B receptor–phospholipase C–protein kinase C signaling pathway, leading to EET generation. These findings provide insights into bilobalide’s endothelial mechanisms and highlight its potential as an alternative therapy for coronary artery disease.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12906-026-05418-y.

Background

Coronary artery endothelial cells regulate coronary vascular tone via vasodilator release, a process significantly modulated by endothelial ion channels [1]. In human coronary artery endothelial cells (HCAECs), these channels include the small-, intermediate-, and large-conductance Ca²⁺-activated K⁺ channels (SK_Ca, IK_Ca, and BK_Ca, respectively); the inward rectifier K⁺ channel (K_ir); the ATP-sensitive K⁺ channel (K_ATP); the voltage-gated K⁺ channel (Kv); the transient receptor potential (TRP) channel; and the Cl⁻ channel [2, 3]. In particular, SK_Ca and IK_Ca channels are essential for endothelial hyperpolarization and vasodilator production in human and rodent vessels [4, 5]. Moreover, apamin, an SK_Ca blocker, has been shown to decrease membrane hyperpolarization in coronary arteries [6] and to inhibit substance P–induced aortic vasodilation [7].

Ginkgo biloba extract and its constituent bilobalide cause vasodilation in some vascular beds [8]. A study in healthy elderly adults found that G. biloba extract improved endothelium-dependent vasodilatory capacity in coronary arteries [9], and G. biloba extract also activated endothelium-dependent vasorelaxation in rabbit aortic rings [10]. In aortic endothelial cells, G. biloba extract dose-dependently activated K_Ca channels [8, 11]; Li et al. [8, 11] identified bilobalide as the active ingredient in G. biloba extract that elicited vasoactive responses. In addition, studies have demonstrated that bilobalide exerts anti-ischemic effects in neonatal rat cardiomyocytes [12] and decreases infarct volume after middle cerebral artery occlusion [13]. Thus, bilobalide may affect K_Ca in various cell types and exert anti-ischemic effects.

To date, no specific bilobalide receptor has been reported. In Caenorhabditis elegans, bilobalide significantly inhibited 5-hydroxytryptamine (5-HT, or serotonin)–controlled egg-laying behavior by blocking 5-HT receptor (5-HTR) homologs [14]. A study in HCAECs demonstrated that 5-HT–induced nitric oxide (NO) production was mediated by both the 5-HT_1B receptor (5-HT_1BR) and the 5-HT_2B receptor (5-HT_2BR) [15]. Moreover, 5-HT–stimulated endothelium-dependent vasorelaxation was mediated by the 5-HT_2BR in rat jugular vein and in pig pulmonary and cerebral arteries [16, 17]. We therefore hypothesized that bilobalide may affect HCAEC K_Ca channels via a 5-HTR.

In this report, we identified the SK_Ca channel as the main contributor to bilobalide-induced current and showed that bilobalide acted via the 5-HT_2BR–phospholipase C (PLC)–protein kinase B pathway. Furthermore, we demonstrated that the SK_Ca channel was involved in bilobalide-induced hyperpolarization and the production of epoxyeicosatrienoic acids (EETs), but not NO, prostacyclin (PGI₂), or hydrogen peroxide (H₂O₂), in this cell type.

Methods

Cell culture

HCAECs were purchased from Lonza (Walkersville, MD, USA) and grown in endothelial cell growth medium (EGM-MV BulletKit, Lonza) at 37 °C in 5% CO₂ humidified air. Cells at passages 4–8 were used. Three to four hours before an experiment, cells were trypsinized and plated on petri dishes containing poly-l-lysine–coated coverslips to facilitate attachment.

Chemicals

Unless indicated otherwise, all chemicals were obtained from Sigma-Aldrich Corporation (St Louis, MO, USA). Bilobalide (C₁₅H₁₈O₈; molecular weight, 326.3) was dissolved in dimethyl sulfoxide (DMSO) to obtain a stock solution, which was stored at − 2 °C to − 5 °C until use. On an experiment day, various bilobalide concentrations were prepared using fresh external solutions. The maximum final DMSO concentration in the external solution was 0.33% (volume/volume). Inhibitors used in our experiments are listed in Table 1.

Table 1.

Inhibitors used in this study

Blockers	Other name	Concentration	Target	Ref
Apamin		500 nmol/L	SKCa	[6]
TEA	tetraethylammonium	1 mmol/L	BKCa	[18]
Iberiotoxin		100 nmol/L	BKCa	[19]
Clotrimazole		10 μmol/L	IKCa	[20]
La3+		10 μmol/L	TRP	[21]
BaCl2		100 μmol/L	Kir	[22]
DIDS	4,4′-diisothiocyanatostilbene-2,2′-disulfonic acid	250 μmol/L	Cl−	[23]
4-AP	4-aminopyridine	1 mmol/L	Kv	[24]
glibenclamide		10 μmol/L	KATP	[25]
methiothepin		1 µmol/L	5-HT1/2R	[26]
SB204741	N-(1-methyl-1H-5-indolyl)-N′-(3-methyl-5-isothiazolyl)urea	10 μmol/L	5-HT2BR	[27]
SB224289 *	1′-methyl-5-[[2′-methyl-4′-(5-methyl-1,2,4-oxadiazol-3-yl)biphenyl-4-yl]carbonyl]−2,3,6,7-tetrahydrospiro[furo[2,3-f]indole-3,4′-piperidine] hydrochloride	1 μmol/L	5-HT1BR	[28]
U73122	1-[6-[((17β)−3-methoxyestra-1,3,5[10] -trien-17-yl)amino]hexyl]−1H-pyrrole-2,5-dione	2 μmol/L	PLC	[29]
Gö−6983 *	3-[1-[3-(dimethylamino)propyl]−5-methoxy-1H-indol-3-yl]−4-(1H-indol-3-yl)−1H-pyrrole-2,5-dione	200 nmol/L	PKC	[30]
MS-PPOH *	N-methylsulfonyl-6-(2-propargyloxyphenyl)hexanamide	10 μmol/L	cytochrome P450	[31]
PTX	pertussis toxin	1 μg/mL	Gi/o protein	[32]

All were available through Sigma-Aldrich Corporation (St Louis, MO, USA), except those marked * (Santa Cruz Biotechnology, Dallas, TX, USA). Ref., references justifying the concentrations used

Whole-cell patch clamping

Whole-cell currents were recorded with the following solutions (modified from Ledoux et al. [33]). External solution contained (in mmol/L): 140 NaCl, 5.6 KCl, 1.2 MgCl₂, 2.6 CaCl₂, 10 d-glucose, and 10 HEPES–NaOH (pH 7.3, adjusted with NaOH; 320–325 mOsm/kg, adjusted with d-mannitol). Internal solution contained (in mmol/L): 140 KCl, 1 MgCl₂, 4 CaCl₂, 5 HEPES, 7 EGTA, and 3 ATP (pH 7.2, adjusted with KOH; 290–300 mOsm/kg); estimated free Ca²⁺ concentration was 223 nmol/L (calculated with the WEBMAXC Extended online program [34]), a low activating concentration for K_Ca channels. To study primarily the SK_Ca channel, we used “SK_Ca solution,” which contained external solution plus blockers of most known HCAEC channels except SK_Ca. Blockers included were 1 mmol/L TEA, 10 μmol/L clotrimazole, 10 μmol/L La³⁺, 100 μmol/L Ba²⁺, and 250 μmol/L DIDS, to inhibit BK_Ca, IK_Ca, TRP, K_ir, and Cl⁻ channels, respectively.

To examine the involvement of 5-HTR and intracellular signaling pathway, HCAECs were first preincubated for at least 30 min with a culture medium containing an antagonist of 5-HTR or pathway components, followed by electrophysiological recording in external solution containing the same preincubating antagonists.

All patch-clamp studies were performed at 25 °C to 26 °C. Cells were allowed to equilibrate for 3 min in a new solution before recording. In a voltage-clamp experiment, currents were recorded in response to a voltage ramp protocol consisting of a continuous depolarization from − 110 to + 100 mV within 200 ms (Fig. 1A, inset). The holding potential was − 40 mV, which is close to reported mean values for HCAEC resting membrane potential [35]. Cells with a total current of < 25 pA at + 80 mV and the seal resistance < 0.8 GΩ were excluded. For current clamping, cell membrane potential was recorded at zero current. All recording protocols were applied via the Axopatch 200B patch-clamp amplifier, Digidata 1440A analog-to-digital converter, and Clampex program (Axon Instruments, Molecular Devices, Sunnyvale, CA, USA).

Fig. 1.

Dose-dependent whole-cell current increase by bilobalide and possible ion channels involved. A Representative whole-cell current recorded from an HCAEC upon stimulation with a voltage ramp protocol (inset), comparing currents in 100 μmol/L bilobalide (BB) vs control (External Sol.). B Average current density–voltage relationship after exposure to bilobalide (filled circles) compared with that in external solution (open circles); ***P < 0.001, BB vs External Sol., Wilcoxon signed rank test. C Concentration–response relationship of bilobalide effects on HCAEC current at + 60 mV; continuous line represents a nonlinear fit; **P < 0.01 and ***P < 0.001 vs solvent (DMSO), one-way ANOVA with Tukey post hoc test. (See also supplementary Table S1). D Percentage change after adding the indicated channel blocker to the bilobalide-exposed current at + 60 or − 100 mV; negative percentages indicate inhibition. ^#P < 0.05 and ^##P < 0.01 vs zero (i.e., with respect to bilobalide alone), one-sample t test. (See also supplementary Table S2.) Numbers in parentheses indicate sample size. 4-AP; 1 mmol/L 4-aminopyridine; Apa, 500 nmol/L apamin; Ba²⁺, 100 μmol/L BaCl₂; CTZ, 10 μmol/L clotrimazole; DIDS, 250 μmol/L diisothiocyanato stilbene disulfonic acid; Gli, 10 μmol/L glibenclamide; La³⁺, 10 μmol/L LaCl₃; 1 mmol/L TEA, tetraethylammonium

From previous study, Current of each cell was depended on the area of cell membrane. For the compared the sum of current, current density was determined by membrane currents divided by cell capacitance that was proportional to cell membrane area; the result was showed in pA/pF.

Vasodilator assays

To assess vasodilator production, HCAECs were cultured in 6-well plates at a density of 1 × 10⁵ cells/well. Cells were preincubated for 30 min with 500 nM apamin, followed by a 10-min exposure to 100 μmol/L bilobalide. The culture medium was centrifuged at 14 000 × g for 15 min at 4 °C, and the supernatant was collected for NO, PGI₂, or H₂O₂ assays. For EET assay, cells were lysed on ice in RIPA buffer supplemented with protease inhibitors (Thermo Scientific, Waltham, MA, USA) for 5 min, followed by centrifugation at 14 000 × g for 15 min at 4 °C to collect the supernatant. Results were compared among HCAECs in basal conditions, HCAECs treated with bilobalide alone, and HCAECs pretreated with apamin before bilobalide exposure.

NO assay

NO was quantified by measuring its products, nitrate and nitrite, in the supernatant using a commercial NO assay kit (ab65328, Abcam). Briefly, nitrate was first converted to nitrite with nitrate reductase, and then nitrite reacted with the Griess reagent to form a deep purple azo compound. The optical density of the azochromophore was detected at 540 nm using a microplate reader (Bio-Rad Laboratories, Hercules, CA, USA). Total protein concentration was determined with the micro bicinchoninic acid assay kit (Pierce Biotechnology, Rockford, IL, USA). Values were normalized to protein concentration in the supernatant.

PGI₂ assay

Because PGI₂ undergoes rapid metabolism, its stable metabolite, 6-keto-PGF₁α, was detected instead in the supernatant using competitive enzyme-linked immunosorbent assay (ELISA; 6-keto-PGF₁α ELISA kit, 515211, Cayman Chemical, Ann Arbor, MI, USA). The optical density of 6-keto-PGF₁α was read at 405 nm. Values were normalized to protein concentration in the supernatant.

EET assay

EET levels in whole-cell lysates were quantified by measuring the stable metabolite 14,15-dihydroxyeicosatrienoic acid (14,15-DHET) using a competitive 14,15-EET/DHET ELISA kit (ab175812, Abcam). Briefly, 14,15-DHET in the lysate competed with horseradish peroxidase–conjugated 14,15-DHET for binding to a precoated antibody. Optical density was read at 450 nm. Data were normalized to whole-cell protein concentration to account for variations in cell number.

H₂O₂ assay

H₂O₂ concentration in the supernatant was measured using an H₂O₂ assay kit (ab102500, Abcam) based on a horseradish peroxidase–catalyzed reaction that produces a colorimetric product detectable at 570 nm. H₂O₂ levels were normalized to whole-cell protein concentration.

For comparison, all induced vasodilator generation was expressed as the percentage of control (i.e., production in the absence of an agonist).

Data analysis

Initial analyses were performed with the pClamp 10.6 program (Axon Instruments) and Microsoft Excel (Microsoft Corporation, Redmond, WA, USA). Data were discarded if seal resistance was less than 1 GΩ, maximum voltage error was greater than 3 mV, or initial cell current amplitude (+ 80 mV) in the external solution was less than 25 pA. Statistical analyses were performed using the GraphPad Prism program (GraphPad Software Inc, San Diego, CA, USA). All data were tested for normality using the Shapiro–Wilk test. Paired data were compared using paired t test, Wilcoxon signed rank test, or repeated measures ANOVA with Greenhouse–Geisser correction followed by Tukey post hoc test, as appropriate. %Control or %Change data were compared to 100 or zero, respectively, using one-sample t test. Between-group differences were analyzed according to data type using unpaired t test or Mann–Whitney U test. Multiple comparison was performed using one-way ANOVA with Tukey post hoc test, for if the assumption of equal variance was met (using Brown-Forsythe test); otherwise Kruskal–Wallis test with post hoc Dunn multiple comparison test was employed. All data are expressed as mean ± SEM. Statistical significance was accepted at P < 0.05.

The bilobalide concentration–response curve was fitted to a dose–response equation using GraphPad Prism to obtain the EC₅₀ and maximum response.

Results

Bilobalide-enhanced whole-cell current primarily goes through the SK_Ca channel

Figure 1A shows an increase in outward whole-cell current after an HCAEC was exposed to 100 μmol/L bilobalide. The change could be seen within five minutes after starting the bilobalide-containing external solution. The corresponding change in average current density–voltage (I–V) curves derived from 34 cells is shown in Fig. 1B. This bilobalide-induced current increase was concentration-dependent (Fig. 1C and supplementary Table S1). A statistically significant increase was observed with 10, 30, 100, and 300 µmol/L bilobalide (%increase [mean ± SEM] was 33.19% ± 7.53% [n = 6], 42.40% ± 4.97% [n = 6], 39.61% ± 4.55% [n = 17], and 42.76% ± 10.36% [n = 5], respectively; P < 0.01), whereas the changes with 1 and 3 µmol/L bilobalide were not statistically significant (− 2.14% ± 1.64% [n = 5] and 5.80% ± 4.44% [n = 6]). The solvent for bilobalide (0.33% DMSO maximum concentration used) by itself did not affect HCAEC current (− 2.26% ± 7.80% [n = 4]). The EC₅₀ and maximum response were 5.31 ± 1.51 μmol/L and 43.51% ± 9.43%, respectively. Bilobalide at 100 µmol/L was chosen for use in the remaining experiments because this concentration could elicit a readily distinguishable vasodilation in rat aortic ring experiments [8, 36].

As a screening test for ion channels involved, bilobalide-induced current was exposed to a blocker of a suspected channel. The results are shown in Fig. 1D (details in supplementary Table S2). Exposure to 500 nmol/L apamin (an SK_Ca blocker), 1 mmol/L TEA (a BK_Ca blocker), or 250 µmol/L DIDS (a Cl⁻ channel blocker) significantly reduced 100 μmol/L bilobalide-induced HCAEC current by 23.26% ± 2.20% (n = 6), 18.68% ± 4.60% (n = 7), and 23.91% ± 6.12% (n = 6), respectively (all P < 0.05). In contrast, a statistically significant difference in current amplitude was not obtained with 10 µmol/L clotrimazole (an IK_Ca blocker), 1 mmol/L 4-aminopyridine (a Kv blocker), 10 µmol/L La³⁺ (a nonspecific TRP channel blocker), 10 μmol/L glibenclamide (a K_ATP blocker), or 100 µmol/L Ba²⁺ (a K_ir blocker) (Currents were reduced by 3.85% ± 1.18% [n = 6], 3.61% ± 1.17% [n = 5], 7.83% ± 2.65% [n = 5], 7.09% ± 6.93% [n = 6], and 3.23% ± 4.32% [n = 5], respectively). Therefore, the contribution of IK_Ca, K_v, TRP, K_ATP, or K_ir channels in mediating the bilobalide’s action could be ruled out. However, it was still uncertain whether SK_Ca, BK_Ca, and Cl⁻ channels were involved because the respective channel blocker could have inhibited channels already open before the bilobalide exposure.

To resolve this, we conducted experiments where a blocker of a candidate channel was applied to inhibit basal channel opening before adding bilobalide (Fig. 2). Whole-cell currents were recorded in the presence of 500 nmol/L apamin, 100 nmol/L iberiotoxin ([19], a more selective BK_Ca blocker than 1 mmol/L TEA), or 250 μmol/L DIDS, followed by 100 µmol/L bilobalide. Only apamin was able to prevent bilobalide-induced current, whereas iberiotoxin or DIDS could not. The percentages of HCAEC current increase by bilobalide without prior exposure to any blocker vs with exposure to apamin, iberiotoxin, and DIDS were 38.42% ± 5.01% (n = 7) vs − 0.47% ± 2.42% (n = 5), 26.32% ± 4.07% (n = 5), and 31.37 ± 4.06 (n = 6), respectively (Fig. 2A-G, details in supplementary Table S3). Therefore, it was SK_Ca, and not BK_Ca or Cl⁻ channel, which mediated the bilobalide-increased current.

Fig. 2.

SK_Ca, but not BK_Ca nor Cl⁻ channel, mediates bilobalide-enhanced current. Representative whole-cell current traces and average current density–voltage relationships are shown for cells in external solution treated with 500 nmol/L apamin (A, B), 100 nmol/L iberiotoxin (C, D), or 250 μmol/L DIDS (E, F) for 5–10 min, followed by 100 µmol/L bilobalide exposure (Apamin vs Apamin + BB; Iberiotoxin vs Iberiotoxin + BB; DIDS vs DIDS + BB). *P < 0.05 and **P < 0.01, paired t test. G % Increase of current at + 80 mV after adding bilobalide in cells treated with apamin, iberiotoxin, or DIDS, as indicated. ^####P < 0.0001 vs non-treated control, ^$$P < 0.01 and ^$$$P < 0.001 vs 500 nmol/L apamin; ANOVA with Tukey post hoc test. Numbers in parentheses indicate sample size. (See also supplementary Table S3)

To investigate further, we conducted another set of experiments in SK_Ca solution containing blockers of most channels opening at positive membrane potentials (i.e., BK_Ca, IK_Ca, K_ir, TRP, and Cl⁻), except SK_Ca (see Methods), so that mainly SK_Ca current could be recorded and the other candidate channels would be inhibited. Traces from a representative experiment are shown in Fig. 3A. Currents in SK_Ca solution were enhanced by 100 µmol/L bilobalide and inhibited after addition of 500 nmol/L apamin. The average I–V curves from 6 cells are shown in Fig. 3B. The current amplitude at + 80 mV, calculated as %control (percentage of current in SK_Ca solution alone), is summarized in Fig. 3C (details in supplementary Table S4). The results showed that after bilobalide exposure, SK_Ca current at + 80 mV increased to 163.62% ± 18.98% (n = 6; P < 0.05). In the continued presence of bilobalide, 500 nmol/L apamin inhibited bilobalide-induced current to 88.76% ± 17.75% compared with SK_Ca solution alone, P > 0.05), an inhibition of 41.23% ± 10.66% (n = 6) compared with bilobalide-induced current (P < 0.05). These results strongly suggested that SK_Ca current was indeed significantly increased by bilobalide.

Fig. 3.

SK_Ca channel is a bilobalide target: experiments in SK_Ca solution. Results from experiments performed in SK_Ca solution (see Methods) are shown. A Representative whole-cell current traces and B average current density–voltage relationships comparing primarily SK_Ca current (SK_Ca) vs after exposure to 100 μmol/L bilobalide (SK_Ca + BB) and after subsequent addition of 500 nmol/L apamin (SK_Ca + BB + Apamin). C Currents at + 80 mV were expressed as %Control, comparing among current in SKCa solution alone (control), after bilobalide exposure, and after additional apamin exposure, as indicated. The error bar of the first bar graph is the normalized SEM to indicate the variability of the control currents (details in supplementary Table S4). **P < 0.01 vs 100% (representing control or SK_Ca solution alone), one-sample t test; ^##P < 0.01 vs SK_Ca + BB, paired t test

Bilobalide acts via the 5-HT_2B receptor

Because a 5-HTR can bind to bilobalide [37, 38] or mediate bilobalide actions in some tissues [14], we studied bilobalide-induced current in HCAECs pretreated with a blocker of the 5-HTR subtypes found in this cell. Cells were preincubated for 30 min in culture medium containing 1 µmol/L methiothepin (nonselective 5-HT_1/2R antagonist), 10 μmol/L SB204741 (5-HT_2BR antagonist), or 1 μmol/L SB224289 (5-HT_1BR antagonist). Basal current (i.e., before bilobalide exposure) was not statistically significantly different between control cells (preincubated without antagonists) and antagonist-preincubated cells (supplementary Tale S5). Addition of 100 μmol/L bilobalide in the continued presence of the same antagonist concentration increased HCAEC current in SB224289-preincubated cells (Fig. 4C) but not in cells preincubated with methiothepin or SB204741 (Fig. 4B and D). The %increase of current by bilobalide after methiothepin, SB224289, and SB204741 preincubation was − 7.81% ± 9.88% (n = 5; P > 0.05), 20.77% ± 5.55% (n = 5; P < 0.05), and − 4.31% ± 3.35% (n = 6; P > 0.05), respectively (Fig. 4E; details in supplementary Tale S6.) These data indicate that the 5-HT_2BR, but not the 5-HT_1BR, mediated the bilobalide action..

Fig. 4.

The 5-HT_2B receptor mediates bilobalide-induced whole-cell HCAEC current increase. Average current–voltage relationships recorded from A control cells (preincubated in culture medium only), comparing currents in external solution (External sol.) vs 100 μmol/L bilobalide (BB); B methiothepin-preincubated cells (1 μmol/L), comparing currents in external solution + methiothepin vs the same solution with 100 μmol/L bilobalide added (Methiothepin vs Methiothepin + BB); C SB224289-preincubated cells (1 μmol/L), comparing currents in external solution + SB224289 vs the same solution with bilobalide (SB224289 vs SB224289 + BB); and D SB204741-preincubated cells (10 μmol/L), comparing currents in external solution + SB204741 vs the same solution with bilobalide (SB204741 vs SB204741 + BB). *P < 0.05 and **P < 0.01, paired t test. E %Current increase in bilobalide at + 80 mV in cells preincubated with the antagonist indicated. ^##P < 0.01 compared with cells preincubated in culture medium alone; ANOVA with Tukey post hoc test; ^$P < 0.05 and ^$$P < 0.01 vs zero (representing currents before adding bilobalide), one-sample t test. Numbers in parentheses indicate sample size. (See also supplementary Table S5)

Bilobalide effect is mediated by the PLC–protein kinase C pathway

Activated 5-HT_2BR mainly exerts its effects through either the PLC–protein kinase C (PLC–PKC) [39] or arachidonic acid pathways [40]. In addition, EETs, which are cytochrome P450 products and an endothelium-derived hyperpolarizing factor, can modulate K_Ca channels [41]. Moreover, the platelet-activating factor (PAF) receptor can mediate bilobalide or G. biloba extract effects in some tissues [12], and PAF effects are usually mediated via G_i/o protein [42]. G_i/o protein also mediates 5-HT_1BR activation in endothelial cells and others [43–45].

To evaluate the intracellular pathways involved in the bilobalide effect, groups of HCAECs were incubated for 30 min in culture medium containing 2 μmol/L U73122 (selective PLC antagonist), 200 nmol/L Gö−6983 (selective PKC antagonist), or 10 μmol/L MS-PPOH (selective cytochrome P450 antagonist), or for 90 min with 1 μg/mL pertussis toxin (PTX, a selective G_i/o protein inhibitor). As in the previous section, preincubated antagonists were correspondingly present throughout the experiment, and no difference in basal current was observed between control and antagonist-preincubated groups (supplementary Table S4).

Addition of 100 μmol/L bilobalide increased HCAEC current in MS-PPOH– or pertussis toxin–preincubated cells (Fig. 5D and E) but not in cells preincubated with U73122 or Gö−6983 (Figs. 5B and C). The percentage current increase in bilobalide at + 80 mV was − 23.16% ± 6.88% (n = 6; P < 0.01), − 38.78% ± 5.19% (n = 5; P < 0.01), 43.94% ± 11.17% (n = 6; P < 0.05), and 34.86% ± 10.29% (n = 6; P < 0.05) with U73122, Gö−6983, MS-PPOH, and PTX preincubation, respectively (Fig. 5F; details in supplementary Tale S6.) These results suggest that the PLC–PKC pathway may mediate the bilobalide-induced current increase. Our data did not support a role for EETs or the PAF receptor in bilobalide-enhanced HCAEC whole-cell current, because neither cytochrome P450 nor G_i/o protein inhibition affected the current increased by bilobalide.

Fig. 5.

The PLC–PKC pathway is involved in the bilobalide effect. Average current–voltage relationships recorded from A control cells (preincubated in culture medium alone), comparing currents in external solution (External sol.) vs 100 μmol/L bilobalide (BB); B U73122-preincubated cells (2 μmol/L), comparing currents in external solution + U73122 vs the same solution with bilobalide (U73122 vs U73122 + BB); C Gö−6983–preincubated cells (200 nmol/L), comparing currents in external solution + Gö−6983 vs the same solution with bilobalide (Gö−6983 vs Gö−6983 + BB); D MS-PPOH–preincubated cells (10 μmol/L), comparing currents in external solution + MS-PPOH vs the same solution with bilobalide (MS-PPOH vs MS-PPOH + BB); and E PTX-preincubated cells (1 μg/mL), comparing currents in external solution + PTX vs the same solution with bilobalide (PTX vs PTX + BB). *P < 0.05, **P < 0.01, and ***P < 0.001, comparing currents with vs without bilobalide, Wilcoxon signed rank test (A) or paired t test (B–E). F %Current increase in bilobalide at + 80 mV in cells preincubated with the antagonist indicated. ^####P < 0.0001 vs cells preincubated in culture medium without antagonists (first bar), one-way ANOVA with Tukey post hoc test; ^$P < 0.05, ^$$P < 0.01, and P < 0.0001 vs zero (representing currents before bilobalide exposure), one-sample t test. Numbers in parentheses indicate sample size. (See also supplementary Table S6)

Apamin reverses bilobalide-induced hyperpolarization

If bilobalide opens the SK_Ca channel, causing K⁺ efflux, hyperpolarization should be expected. However, the apparent reversal potential when bilobalide successfully induced a current increase (Figs. 1A, 2C & E) did not show a distinct hyperpolarization shift. This may be due to the small current in our cells causing low signal-to-noise ratio. Moreover, in a voltage ramp protocol the membrane current and voltage may not be in the steady state [46]. We therefore studied HCAEC membrane potential with current clamp experiments. The results (Fig. 6) demonstrated that bilobalide hyperpolarized HCAECs, with an average control membrane potential of − 22.59 ± 5.74 mV in external solution vs − 27.43 ± 5.73 mV during exposure to 100 µmol/L bilobalide (n = 7; P < 0.05). Addition of 500 nmol/L apamin virtually abolished bilobalide-induced hyperpolarization, reducing the membrane potential to − 19.35 ± 5.4 mV (n = 7; P < 0.01 compared with bilobalide; P > 0.05 compared with control). This supports our conclusion that bilobalide caused SK_Ca opening, which led to endothelial hyperpolarization.

Fig. 6.

Apamin inhibits bilobalide-induced hyperpolarization. A A membrane potential trace recorded from an HCAEC in a current clamp experiment; BB, 100 μmol/L bilobalide; apamin, 500 nmol/L apamin. B Membrane potential changes of individual cells after bilobalide exposure (BB) and apamin addition (BB + apamin), demonstrating that hyperpolarization with bilobalide and reversal by apamin occurred in all but one cell tested. Notice the variation in the baseline resting membrane potential among cells. Values are average levels of the corresponding tracing. C Bar graphs comparing membrane potential changes before and after adding bilobalide and apamin, respectively, as indicated below each bar graph. *P < 0.05 vs control (external solution, first bar), ^##P < 0.01 vs bilobalide (second bar); repeated measures ANOVA with Greenhouse–Geisser correction followed by Tukey post hoc test. (See also supplementary Table S7)

SK_Ca channel mediates bilobalide-induced EET production

We further investigated whether bilobalide induced the generation of vasodilators—NO, PGI₂, EETs, and H₂O₂—and whether this effect involved SK_Ca channels. Figure 7 shows that 100 μmol/L bilobalide increased the average %control for EETs to 182.3% ± 23.14% (n = 5; P < 0.05 compared to control), and bilobalide-induced EET production was prevented by preincubation with apamin (79.48% ± 10.39% [n = 5]; P < 0.01 compared with bilobalide alone). In contrast, bilobalide had no statistically significant effect on NO, PGI₂, or H₂O₂ production (Fig. 7). These results suggest that the SK_Ca channel is involved in bilobalide-induced EET production in HCAECs.

Fig. 7.

The SK_Ca channel is involved in bilobalide-induced EET production. Percentage of control for bilobalide-induced vasodilator production from HCAECs, comparing cells pretreated with 500 nmol/L apamin followed by 100 μmol/L bilobalide addition (in the continued presence of apamin; Apa + BB) vs cells exposed to bilobalide alone (BB). Bradykinin (BK) served as a positive control. The error bars of the first bar graphs (Control, gray bars) are the normalized SEM to indicate the variability of the control currents. *P < 0.05 and ***P < 0.001 vs Control, ^††P < 0.01 vs Bilobalide, ^‡P < 0.05 vs Apamin + Bilobalide; one-way ANOVA with Tukey post hoc test (EET and H₂O₂) or Kruskal–Wallis test with post hoc Dunn’s multiple comparison test (NO and PGI₂). (Details in supplementary Table S8)

Discussion

This study showed that bilobalide dose-dependently increased HCAEC whole-cell current with a maximum response of 43.51% ± 9.43%. The EC₅₀ of 5.31 ± 1.51 µmol/L is in the same range as the bilobalide concentrations eliciting vasodilation in rat aorta [8]. The increased current primarily consisted of SK_Ca current, whereas other K⁺ channels (BK_Ca, IK_Ca, K_ir, K_ATP, and Kv), Cl⁻ channels, and Ca²⁺-conducting TRP channels were excluded (Figs. 1 and 2). Moreover, bilobalide significantly increased HCAEC current when most channels except SK_Ca were blocked (Fig. 3), and apamin virtually eliminated or prevented the bilobalide-induced current (Figs. 1, 2 and 3). These data clearly demonstrate that the SK_Ca channel is a target of bilobalide. Our findings are consistent with studies in aortic endothelial cells reporting that G. biloba extract or bilobalide can activate a K_Ca channel [8, 11]. Of note, endothelial SK_Ca is also a target of other vasodilatory substances and hormones [47, 48].

Although the in vivo bilobalide effect was expected to occur at the resting membrane potential (RMP), we could only monitor currents at high positive potentials because the net membrane current at rest was zero and our HCAEC currents were often tiny. However, for the non-voltage-gated channels (SK_Ca, IK_Ca, Cl⁻, and TRP), currents at positive potentials actually reflected channel opening at any potentials. Therefore, if bilobalide’s target was a non-voltage-gated channel (as was the case in our investigation), its effects at + 60 or + 80 mV would indicate what occurred at the RMP and were physiologically relevant. (Voltage-sensitive channels could still have physiological role at rest, depending on the actual mechanism and conditions; currents at high positive or negative potentials just signified whether bilobalide affected them.)

In our screening experiments (Fig. 1D), bilobalide was applied first, followed by a channel blocker. The results that apamin, 1 mmol/L TEA, or DIDS decreased the current exposed to bilobalide could be interpreted as bilobalide raising the SK_Ca, BK_Ca, or Cl⁻ current, or the blockers reducing existing currents (not caused by bilobalide). Therefore, in Fig. 2, apamin, iberiotoxin, or DIDS was applied first to block existing open channels followed by bilobalide. Only apamin could prevent the bilobalide’s action, and thus SK_Ca was the most probable bilobalide target. Note that although 1 mmol/L TEA was reported to block mainly BK_Ca [18], higher TEA concentrations may block other K⁺ channels. Therefore, the more specific BK_Ca channel blocker iberiotoxin was employed in Fig. 2, proving the non-participation of BK_Ca. Besides, some studies have suggested that BK_Ca is present only in culture and not in native cells [49]. Notice also that HCAECs seemed to have a high Cl⁻ conductance (Fig. 1 D, DIDS) which may serve to resist changes in resting membrane potential (E_Cl was ~ −0.5 mV). Nonetheless, the fact that bilobalide could still hyperpolarize the cells in Fig. 6 is another indication that Cl⁻ channel was not affected by bilobalide.

Our findings provide direct evidence that bilobalide causes endothelial hyperpolarization and increases EETs, which are known endothelium-derived hyperpolarizing factors (EDHF). Importantly, both effects were completely neutralized by apamin (Figs. 6 and 7). Our results strongly indicate that the SK_Ca channel is principally responsible for hyperpolarization and EET production. Previously, it was reported that G. biloba extracts, including bilobalide and ginkgolides, can induce cytochrome P450 enzyme activity [50]. By extension, this may in turn enhance EET production, as found in our study. Interestingly, we found that bilobalide did not affect endothelial NO or PGI₂ production. This is consistent with a previous study in human umbilical vein endothelial cells [51]. However, it has been reported that bilobalide can induce relaxation through NO production in rat thoracic aorta [8]. Moreover, Li et al. reported that G. biloba extract activated a K_Ca channel and nitric oxide synthase (NOS) activity in porcine aortic endothelial cells [11]. The discrepancy may be explained by the difference in vessel type: it is known that NO is dominant in large vessels, whereas EDHFs predominate in small vessels [52]. We demonstrated in HCAECs that bilobalide induces only the generation of EETs, not other vasodilators (Fig. 7), and apamin virtually eliminated bilobalide-induced EET production.

Our data may support the SK_Ca channel as a mediator of bilobalide's known cardioprotective effects on coronary flow [53] and infarct size [12]. Interestingly, down-regulation of endothelial SK_Ca channels was found in pathological circumstances like hypertension [54], diabetes [55], and cardioplegic-hypoxia [56, 57], causing impaired vasodilatory response. Moreover, treatment with SK_Ca activator improved SK_Ca function and restored the vasorelaxation in the cardioplegic-hypoxia scenario [57]. These findings call for further investigation into bilobalide's potential to alleviate these conditions.

This research also shows that the effect of bilobalide in endothelial cells is mediated by the 5-HT_2BR. Both the 5-HT_1BR and the 5-HT_2BR are expressed in endothelial cells, including HCAECs [17]. Moreover, activation of both 5-HTR subtypes causes endothelium-dependent vasodilation in these cells [15]. There has been no previous report associating bilobalide with the 5-HT_2BR. Only Brown and Luo [14] found that bilobalide could modulate C. elegans metabotropic SER-4 receptors, a mammalian 5-HT_1/2R homolog. On the other hand, bilobalide can inhibit 5-HT₃R, a neurotransmitter-gated ion channel not found in endothelial cells [58, 59], with an IC₅₀ of 470 μmol/L or higher [38]. We did not test other 5-HTR subtypes because the selective 5-HT_2BR antagonist SB204741 completely inhibited the action of bilobalide (Figs. 4D and E).

The 5-HT_2BR is a G_q/11 protein–coupled receptor that typically sends a downstream signal via PLC and PKC activation [39]. Our data are consistent with this conventional signal transduction pathway. The PKC isoforms found in HCAECs are the Ca²⁺-dependent α and β1 and the Ca²⁺-independent δ, ɛ, and ι isoforms [60]. It should be noted that the intracellular (pipette) Ca²⁺ concentration in our whole-cell patch-clamp experiments (223 nmol/L, see Methods) was sufficient for a low PKC activation [61, 62]. Nevertheless, the available literature does not provide evidence that PKC can directly phosphorylate SK_Ca channel subunits. On the other hand, a recent study convincingly showed that inhibition of HCAEC SK_Ca currents by NAD was mediated by PKCβ, with experiments using the PKC activator phorbol 12-myristate 13-acetate or PKC inhibition (both subtype-specific and nonspecific) [63]. This suggests the possibility that bilobalide-induced PKC activation could be compartmentalized, probably involving one of the other PKC subtypes in these cells [60]. Finally, it is intriguing that currents in cells preincubation with and in the presence of U73122 or Gö−6983 was actually reduced by bilobalide (Fig. 5). This may suggest a parallel inhibitory effect of bilobalide that was unmasked when the stimulatory pathway (PLC and PKC) was blocked.

SK_Ca channel, G_q/11 protein, and PLC are known to colocalize in endothelial caveolae [64], the same location as NO synthase, which catalyzes NO synthesis [65]. Hence, the finding that bilobalide elicited the generation of EETs and not NO may suggest the presence of SK_Ca in another compartment. Therefore, our results imply that the 5-HT_2BR and the associated SK_Ca channel are probably not localized in caveolae. In addition, because the intracellular Ca²⁺ concentration was kept constant in our experiments, the mechanism of channel modulation may involve increased Ca²⁺ sensitivity of the channels.

Bilobalide elicits vasodilation in various tissues, acting on both vascular smooth muscle and endothelial cells [8, 66]. Our data provide a plausible explanation for the responsible mechanism in endothelial cells: Bilobalide binds to the 5-HT_2BR and activates endothelial SK_Ca channels via the PLC–PKC pathway, causing hyperpolarization and thus increasing the electrochemical force for Ca²⁺ entry [1]. In intact cells (where Ca²⁺ was not kept constant as in our whole-cell patch clamping experiments), intracellular Ca²⁺ concentration will become higher, which can then stimulate phospholipase A₂ to convert phospholipids to arachidonic acid [67], a substrate of cytochrome P450 enzyme with EETs as products, leading to smooth muscle relaxation and vasodilation [68]. Concurrently, membrane hyperpolarization induced by SK_Ca channel opening can spread passively via myoendothelial gap junctions to hyperpolarize smooth muscle cells [1]. Studies have shown that SK_Ca channel activation is involved in EDHF-like hyperpolarization in endothelial cells and the underlying smooth muscles [69]. Furthermore, in intact cells the increased Ca²⁺ entry can activate more SK_Ca channels, providing a positive feedback loop for sustaining intracellular Ca²⁺ increase [4]. This is a likely chain of events that collectively leads to bilobalide-induced vasodilation in human coronary arteries (Fig. 8). However, the whole-cell patch clamp technique can elucidate only membrane-delimited components, and other pathways may exist in vivo, likely acting in concert with that described here to elicit vasodilation.

Fig. 8.

Proposed mechanism of bilobalide-induced SK_Ca current and EET production in HCAECs. As inferred from our data, bilobalide acts via the endothelial 5-HT_2BR, PLC, and PKC pathway, leading to SK_Ca channel activation and increased outward current. The resulting K⁺ efflux causes membrane hyperpolarization, which is known to enhance the electrochemical Ca²⁺ driving force and promote Ca²⁺ influx through the TRP channel [5]. The rise in intracellular Ca²⁺ can lead to cytochrome P450–catalyzed EET production, probably via increased free arachidonic acid release by Ca²⁺-activated PLA₂ [67]. Gray letters and arrows indicate known interactions not addressed in this study

G. biloba extract and its constituents have long been used in traditional medicine and are widely available today as dietary supplements, particularly for patients with cardiovascular and neurovascular conditions [70, 71]. G. biloba extract contains primary active compounds—bilobalide, flavonoids, and ginkgolides—that have demonstrated antihypertensive, antiplatelet, anti-inflammatory, and antioxidant properties [72–74]. Our findings further support bilobalide as a promising therapeutic candidate for cardiovascular disorders, with the potential to transition from complementary use into mainstream clinical practice.

Limitations

There were certain limitations to the whole-cell patch clamp method. In this configuration, much of the intracellular content had been dialyzed, so only the membrane-delimited signal cascade could be studied. Additionally, our experiments could only reveal a Ca²⁺-independent pathway because our internal solution contained 7 mmol/L EGTA Ca²⁺ buffering. However, this does not rule out the existence of parallel cytosolic pathways in intact cells. Nonetheless, our findings showed that the SK_Ca-mediated route may be a principal process for bilobalide-induced EET production: in experiments with intact cells, apamin virtually eliminated EET production (Fig. 7C). Whole-cell configuration also allows a relatively short experimental time window. To minimize rundowns, we included ATP in the intracellular solution and discarded cells when the seal resistance was < 0.8 GΩ. Further, we limited our experiments within 10–15 min, during which HCAEC whole-cell currents were usually steady with little decay. However, we were not able to obtain wash-out currents for bilobalide within the experimental time frame, probably attributed to its dependence on intracellular signaling cascade.

Another limitation was that we did not unravel the mechanism leading from PKC activation to increased SK_Ca currents, nor did we investigate the specific PKC isoforms involved. This gap warrants further investigation. In addition, although SB204741 is a well-validated selective 5-HT_2BR antagonist [27], off-target effects cannot be fully excluded. Future studies demonstrating bilobalide binding on the 5-HT_2BR site and/or using receptor-specific knockdown approaches would strengthen these findings. Note also that in the experiments of Fig. 4 and 5, preincubation with vehicle was not tested. However, as shown in supplementary Table S5, the basal current of control group (preincubated with culture media without a blocker) was not different from basal current of other groups. Lastly, demonstrating the bilobalide action on coronary artery functions and SK_Ca involvement would be essential for the applicability of our findings, however this was beyond the scope of this study.

Conclusions

Our study demonstrates that bilobalide acts via the 5-HT_2BR, PLC, and PKC to enhance SK_Ca currents in HCAECs, resulting in increased whole-cell current, hyperpolarization, and EET generation. Bilobalide did not affect NO or PGI₂ production. The mechanism described herein may explain, at least partially, bilobalide-induced coronary vasodilation.

Abbreviations

14,15-DHET

14,15-Dihydroxyeicosatrienoic acid

5-HT

5-Hydroxytryptamine (or serotonin)

5-HTR

5-Hydroxytryptamine receptor

BK_Ca

Large-conductance Ca²⁺-activated K⁺ channel

DIDS

4,4′-Diisothiocyanatostilbene-2,2′-disulfonic acid

EDHF

Endothelium-derived hyperpolarizing factor

EET

Epoxyeicosatrienoic acid

EGTA

Ethylene glycol-bis (β-aminoethyl ether)-N, N, N ′, N ′-tetraacetic acid

ELISA

Enzyme-linked immunosorbent assay

Erev

Reversal potential

H₂O₂

Hydrogen peroxide

HCAEC

Human coronary artery endothelial cell

IK_Ca

Intermediate-conductance Ca²⁺-activated K⁺ channel

K_ATP

ATP-sensitive K⁺ channel

K_ir

Inward rectifier K⁺ channel

Kv

Voltage-gated K⁺ channel

NAD

Reduced nicotinamide adenine dinucleotide

NO

Nitric oxide

PAF

Platelet-activating factor

PGI₂

Prostacyclin

PKC

Protein kinase C

PLA₂

Phospholipase A₂

PLC

Phospholipase C

PTX

Pertussis toxin

RMP

Resting membrane potential

SK_Ca

Small-conductance Ca²⁺-activated K⁺

TEA

Tetraethylammonium

TRP

Transient receptor potential

Authors’ contributions

Conception and design of the experiments: KR, PT, NW, ST, and WBW. Collection of data: KR, PT, NW. Analysis and interpretation of data: KR, PT, NW, ST, and WBW. Drafting the article and revising it critically for intellectual content: KR, PT, WBW. All authors read and approved the final manuscript.

Funding

This work was partially supported by the Siriraj Graduate Thesis Scholarship (PT) and a Thesis Grant for Master’s Degree Student, National Research Council of Thailand (PT, WBW). We are also grateful to the Faculty of Medicine Siriraj Hospital, Mahidol University, for financial supports to KR, ST and WBW. None of the funders had any role in the conceptualization, design, data collection, analysis, decision to publish, or preparation of the manuscript.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

According to the Siriraj Institutional Review Board, Faculty of Medicine Siriraj Hospital, Mahidol University, this research was excluded from ethical approval requirement because all experiments were conducted on commercially available cells in culture. However, we made every effort to follow, where applicable, the ethical research conduct guidelines issued by the National Research Council of Thailand and the Siriraj Institutional Review Board. Consent to participate was not applicable for this research.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.