Endothelium-Derived Hyperpolarizing Factor Mediates Bradykinin Stimulated Tissue Plasminogen Activator Release In Humans

Abstract

Aims

Bradykinin stimulates tissue plasminogen activator (t-PA) release from human endothelium. Although bradykinin stimulates both nitric oxide and endothelium-derived hyperpolarizing factor (EDHF) release, the role of EDHF in t-PA release remains unexplored. This study sought to determine the mechanisms of bradykinin-stimulated t-PA release in the forearm vasculature of healthy human subjects.

Methods

In 33 healthy subjects (age 40.3±1.9 years) forearm blood flow (FBF) and t-PA release were measured at rest, and after intra-arterial infusions of bradykinin (400ng/min) and sodium nitroprusside (3.2 mg/min). Measurements were repeated after intra-arterial infusion of TEA (1 μmol/min), fluconazole (0.4 μmol^.min^-1.L^-1), and N^G-monomethyl-L-arginine (L-NMMA, 8 μmol/min) to block nitric oxide, and their combination in separate studies.

Results

Bradykinin significantly increased net t-PA release across the forearm (P<0.0001). Fluconazole attenuated both bradykinin-mediated vasodilation (-23.3±2.7% FBF, P<0.0001) and t-PA release (from 50.9±9.0 to 21.3±8.9 ng/min/100ml, P=0.02). TEA attenuated FBF (-14.7±3.2%, P=0.002) and abolished bradykinin-stimulated t-PA release (from 22.9+5.7 to - 0.8±3.6 ng/min/100ml, P=0.0002). L-NMMA attenuated FBF (P<0.0001), but did not inhibit bradykinin-induced t-PA release (P=NS).

Conclusion

Bradykinin-stimulated t-PA release is partly due to cytochrome P450-derived epoxides, and is inhibited by K⁺_ca channel blockade. Thus, bradykinin stimulates both EDHF-dependent vasodilation and t-PA release.

INTRODUCTION

The endogenous fibrinolytic system is a crucial component of blood flow regulation and protects against intravascular thrombosis. Impairment of endogenous fibrinolysis characterized by inhibition of tissue plasminogen activator (tPA) release is thought to be an underlying factor in endothelial dysfunction and atherothrombosis.^1,2 The endothelium regulates vasodilator tone and fibrinolytic capacity by releasing vasoactive substances including nitric oxide (NO), prostacyclin, and endothelium-derived hyperpolarizing factor (EDHF).^{3, 4} Bradykinin is an endogenous cardioprotective vasoactive polypeptide that stimulates the release of these molecules and t-PA from the endothelium.^5-7 t-PA is synthesized in the endothelial cell, stored in small dense granules and released in a dose-dependent manner in response to bradykinin.^{6, 8} t- PA converts plasminogen to a proteolytic enzyme, plasmin, which digests fibrin-dependent thrombi.⁹ Although it is known that bradykinin causes vasodilation via activation of the B₂ receptor and subsequent release of NO, prostacyclin, EDHF, and t-PA, the downstream mechanisms governing t-PA release have not been fully elucidated.^{5, 10, 11}

There is mounting pre-clinical evidence to suggest that t-PA release is mediated by EDHF.^{12, 13} Persistent vasodilation after inhibition of NO and cyclooxygenase can be largely attributed to hyperpolarization by EDHF.^{3, 14} The hallmark of hyperpolarization is activation of endothelial and/or smooth muscle cell calcium-activated potassium (K⁺_ca) channels and their inhibition by apamin and charybdotoxin in pre-clinical studies.^{3, 15} In human studies, tetraethylammonium chloride (TEA) antagonizes K⁺_ca channels, thus facilitating characterization of EDHF-dependent responses.^{16, 17} Endothelium-dependent hyperpolarization may also be partly due to the release of epoxyeicosatrienoic acids (EETs) from cytochrome P450-dependent metabolism of arachidonic acid and subsequent stimulation of K⁺_ca channels on endothelial cells.¹⁸ The role of EETs as potential EDHFs has been investigated in humans using azole compounds such as fluconazole that selectively inhibit epoxidation (EET generation) of arachidonic acid.^{14, 18, 19}

Since bradykinin-mediated t-PA release may be crucial in maintaining endogenous thrombolytic potential, it is important to determine its underlying mechanisms. Our overall goal was to investigate whether EDHF release stimulates t-PA, with the hypothesis that bradykinin stimulates EDHF-mediated t-PA release in the human circulation. We investigated whether activation of the K⁺_ca channels and/or cytochrome P450 derived EETs mediates t-PA release.

METHODS

Subjects

Thirty-three healthy subjects aged 21 to 60 years and free of smoking, hypercholesterolemia, hypertension, diabetes, cardiovascular disease, medication use, or any other systemic disorder participated in the study (ClincalTrials.gov Identifier: NCT00166166) (Table 1). Written informed consent was obtained and the study was approved by the Emory University Institutional Review Board and the Food and Drug Administration.

Table 1.

Subject characteristics

Male, n (%)	21 (63.6)
Age, years	40.3±1.9
Ethnicity, n (%)
White	8 (24.2)
Black	21 (64.3)
Hispanic	4 (12.1)
Systolic BP, mmHg	121.6±2
Diastolic BP, mmHg	73±1.4
Heart rate, bpm	68±1.5
Body mass index (kg/m2)	28±1
Fasting blood glucose (mg/dL)	86.4±1.4
Triglycerides (mg/dL)	102.4±9.5
Total Cholesterol (mg/dL)	170±6.4
High Density Lipoprotein Cholesterol (mg/dL)	54.3±2.3
Low Density Lipoprotein (mg/dL)	95.3±5.7
Tobacco smoking (%)	0
Hematocrit (%)	39.4±0.6

Experimental Protocols

Subjects were enrolled into 3 separate protocols described below, performed sequentially. Measurements were performed after an overnight fast in a quiet, temperature-controlled (22 to 24°C) room. Subjects received 975mg of oral aspirin 90 minutes prior to initiation of the study to inhibit prostanoid release.^{20, 21} An intravenous catheter was placed into a deep antecubital vein for venous sampling and an arterial cannula into the brachial artery for arterial pressure monitoring, drug delivery and blood sampling.

Forearm Blood Flow Measurements

Simultaneous measurements of FBF were obtained in both arms with the use of a dual-channel venous occlusion strain gauge plethysmograph (model EC6, DE Hokanson, Bellevue, WA) as previously described.¹⁴ The mercury-filled silastic strain gauge was placed around the forearm and connected to a plethysmograph calibrated to measure the percent change in volume.¹⁴ An upper arm cuff was inflated to 40 mmHg to occlude venous outflow and FBF measurements were recorded every 15 seconds up to eight times, and a mean FBF value in mL·min^-1·100 mL-¹was computed. Forearm vascular resistance was calculated as the mean arterial pressure ÷ FBF and expressed as mmHg/ mL·min^-1·100 mL^-1.

Protocol 1: Effect of K⁺_ca Channel Activation on BK-Stimulated t-PA Release

FBF was measured after a 30-minute rest during saline infusion (total infusion rate 2.5ml·min^-1) and after intra-arterial infusions of 100, 200, and 400 ng/min of bradykinin (Clinalfa, Switzerland) given for 8 minutes each (n=18). After 30 minutes, intra-arterial TEA (Sigma-Aldrich, Allentown, PA, sterilized and tested for pyrogenicity by the Emory Investigational Drug Pharmacy) was infused at 1 μmol/min for 8 minutes. When given at 0.25 to 1 mg/min, TEA selectively inhibits K⁺_Ca channels, but reduces FBF with bradykinin only at 1 mg/min (<0.6μmol/min).^{16, 22, 23, 24} This dose and timing was effective in attenuating FBF and agonist-stimulated vasodilation in previous studies.^{4, 16, 25} While continuing the infusion of TEA, bradykinin infusions were re-administered. Arterial blood pressure and FBF measurements were repeated in the last 2 minutes of each infusion. Arterial and venous blood sampling were performed at baseline prior to the infusion of drug and at peak doses of bradykinin.

Protocol 2: Effect of Cytochrome P450 Metabolites to BK-Stimulated t-PA Release

FBF was measured at rest during saline infusion and after intra-arterial infusion of escalating doses of bradykinin (n=11). Thirty minutes later, fluconazole (Pfizer, New York, NY) was infused intra-arterially at 0.4 μmol·min^-1.L^-1 for 5 minutes. This dose has been effective in attenuating vasodilation during sustained flow conditions.^{4, 17} While continuing the infusion of fluconazole, escalating doses of bradykinin were administered. Finally, combined intra-arterial infusions of fluconazole and TEA were given for 8 minutes and bradykinin infusions repeated. This protocol allowed assessment of the contribution of cytochrome P450 metabolites alone and in combination with K⁺_ca channel activation to BK-stimulated t-PA release.

Protocol 3: Effect of NO to BK-Stimulated t-PA release

The contribution of NO to bradykinin stimulated t-PA release was also determined in the presence of K⁺_ca channel activation (n=13). FBF was measured at rest and after intra-arterial infusion of bradykinin. Thirty minutes later, TEA (1 μmol/min) was infused for 8 minutes, and followed by repeat infusions of bradykinin. After 30 minutes, TEA and N^G-monomethyl-L-arginine (L-NMMA) (Clinalfa, Switzerland) at 8 μmol/min were infused together for 8 minutes to inhibit EDHF and NO synthesis, followed by repeat administration of bradykinin in escalating doses while continuing infusions of the two antagonists. This dose and timing of L-NMMA was effective in attenuating FBF and agonist-stimulated vasodilation in previous studies.^{16, 25}

Effect of sodium nitroprusside on t-PA release

Finally, in 30 subjects, the comparative contribution of the endothelium-independent vasodilator, sodium nitroprusside (SNP) and bradykinin was investigated. FBF was measured during rest and after intra-arterial infusion of escalating doses of bradykinin. Thirty minutes later, SNP (Hospira, Lakeforest, IL) was infused intra-arterially at 1.6, and 3.2 μg/min.

Blood Sampling and Biochemical Assays

Simultaneous arterial and venous blood samples were obtained from the infused arm before and after peak doses of vasodilator drugs, cooled, centrifuged, and stored in tubes containing 0.5 M citrate buffer (Diapharma, West Chester, OH) at -70°C. t-PA antigen levels were determined in duplicate using a two-site enzyme linked immunosorbent assay (Diapharma, West Chester, OH) as described previously.²⁶ Net release or uptake rates at each time point were calculated: net release = (C_V-C_A) X¹ , where C_V and C_A represent the concentration of t-PA in the brachial vein and artery, respectively.

Statistical Analysis

To determine the effects of the interventions on FBF and forearm vascular resistance, changes in the flow and resistance responses between the control, single blockade, and combined blockade were compared using repeated measures ANOVA. To determine the effects of treatments on net t-PA release, changes in net t-PA release of control, single blockade, and combined blockade were compared using one-way ANOVA with the Tukey post hoc test. Changes of the response were converted to percentage changes of least squares means when presented in the text. Data are presented as mean ± SEM in the text. The least squares means and standard errors are presented in the figures. A P value less than 0.05 was considered statistically significant.

RESULTS

Baseline Subject Characteristics

Thirty-three healthy (age 40.3±1.9 years, 64% male) normotensive, non-diabetic, non-obese, nonsmoking subjects with normal serum cholesterol were studied (Table 1). During intrabrachial drug infusions, no changes in blood pressure or heart rate were observed.

Effect of K⁺_ca Channel Activation on FBF and BK-Stimulated t-PA Release

Bradykinin produced dose-dependent increase in FBF (P=0.002) and decrease in vascular resistance (P=0.03, Figure 1A, 1B). Bradykinin significantly increased the arterio-venous t-PA concentration gradient and net t-PA release across the forearm (from -0.3±0.5 to 36.9±5.5 ng/min/100ml, P=0.002, Table 2, Figure 1C). There was no association between net t-PA release and age, sex, BMI, total cholesterol, and LDL levels.

Figure 1. Contribution of K⁺_ca channel activation to bradykinin-stimulated vasodilation, and t-PA release.

Forearm blood flow (A) and forearm vascular resistance (B) changes in response to bradykinin (100, 200, and 400 ng/min) alone, and after infusion of TEA (n=18). Net t-PA release (C) in response to peak dose of bradykinin (400 ng/min) and after infusion of TEA. Data shown as mean ± SEM.

Table 2.

Effect of interventions on arterial-venous (AV) gradient

Venous-arterial t-PA difference (ng/ml)	Control (n=33)	+TEA (n=18)	+Fluconazole (n=11)	+Fluconazole +TEA (n=10)	+TEA +L-NMMA (n=13)
Baseline	−0.2 ± 0.3	−0.3 ± 0.3	−0.2 ± 0.8	−0.2 ± 0.8	−0.2 ± 0.4
Bradykinin (400 ng/min)	5.6 ± 0.8*	0.03 ± 0.7†	4.4 ± 1.4†	1.6 ± 0.4	2.6 ± 0.9

^*P < 0.05 versus baseline

^†P < 0.01 versus control

Infusion of TEA decreased resting FBF by 14%, P=0.0001 and increased vascular resistance by 18%, P=0.03. Co-administration of TEA with bradykinin attenuated the overall response with a 15%, P=0.002 lower FBF and a 21%, P=0.03 higher vascular resistance (Figure 1A, 1B). TEA infusion effectively abolished t-PA release with bradykinin (from 22.9±5.7 to - 0.8±3.6 ng/min/100ml, P=0.0002) and the arterio-venous t-PA concentration gradient (Table 2), indicating that the effect of bradykinin on endothelial t-PA release is mediated entirely through K⁺_ca channel activation (Figure 1C).

Effect of Cytochrome P450 Metabolites and K⁺_ca Channel Activation on FBF and BK-Stimulated t-PA Release

Infusion of fluconazole decreased resting FBF by 19%, P=0.001 and increased vascular resistance by 27%, P=0.001. Fluconazole also blunted the overall vasodilator response to bradykinin with a 23%, P<0.0001 reduction in FBF and a 35%, P<0.0001 higher vascular resistance (Figure 2A, 2B). Fluconazole infusion attenuated the t-PA release with BK (from 50.9+9 to 21.3+8.9 ng/min/100ml, P=0.02) and the arterio-venous t-PA concentration gradient (Table 2), indicating significant contribution of cytochrome P450 metabolites to BK-stimulated t-PA release (Figure 2C). Infusion of TEA after fluconazole reduced FBF by an additional 20% (P<0.0001, n=10) and increased vascular resistance by a further 24%, P<0.0001 (Figure 2A, 2B). In the presence of fluconazole, infusion of TEA trended to attenuate the arterio-venous t-PA gradient (Table 2), but the further reduction in net t-PA release did not reach statistical significance (Figure 2C). However, TEA infusion attenuated the t-PA release with bradykinin to a greater extent than fluconazole (85±4% versus 61±11% reduction, respectively, P=0.05), indicating that of the potential hyperpolarization mechanisms, K⁺_ca channel activation is the predominant EDHF signaling pathway mediating bradykinin-stimulated t-PA release.

Figure 2. Contribution of cytochrome P450 metabolites and K⁺_Ca channel activation to bradykinin-stimulated vasodilation and t-PA release.

Forearm blood flow (A) and forearm vascular resistance (B) changes in response to bradykinin (100, 200, and 400 ng/min) alone, after infusion of fluconazole, and combined infusions of fluconazole and TEA (n=11). Net t-PA release (C) in response to peak dose of bradykinin (400 ng/min) and after initial infusion of fluconazole, followed by combined fluconazole and TEA infusions. Data shown as mean ± SEM.

Effect of NO and K⁺_ca Channel Activation on FBF and BK-Stimulated t-PA release

To assess the comparative contribution of K⁺_ca channel activation and NO to bradykinin-stimulated t-PA release, individual and combined blockade with TEA and L-NMMA were studied. As seen previously, infusion of TEA attenuated the vasodilator response to bradykinin with a 17%, P<0.0001 reduction in FBF (Figure 3A). Similarly, t-PA release with bradykinin was reduced from 20.4±6 to 3.5±3 ng/min/100ml, P=0.02, indicating a significant contribution of K⁺_ca channel activation to bradykinin-stimulated t-PA release (Figure 3C). Infusion of LNMMA after TEA reduced FBF by an additional 21%, P<0.0001 and increased vascular resistance by a further 42%, P=0.04 (Figure 3A, 3B). In the presence of TEA, infusion of LNMMA did not further affect net t-PA release (P=0.25, Figure 3C, Table 2), indicating that bradykinin-stimulated t-PA release is through a K⁺_ca channel-dependent and NO-synthase-independent pathway.

Figure 3. Contribution of NO and K⁺_ca channel activation to bradykinin-stimulated vasodilation and t-PA release.

Forearm blood flow (A) and forearm vascular resistance (B) in response to bradykinin (100, 200, and 400 ng/min) alone, after initial infusion of TEA, and combined infusion of TEA and L-NMMA in healthy (n=13) subjects. Data shown as mean ± SEM. Box plot of net t-PA release (C) in response to peak dose of bradykinin (400 ng/min) and after infusion of TEA, followed by combined TEA and L-NMMA infusions. Minimum and maximum values depicted by whiskers, the box signifies lower and upper quartiles, the median is represented by the short black line within the box, and the mean by the symbol marker.

Effect of sodium nitroprusside on t-PA release

Infusions of bradykinin and SNP resulted in similar increases in FBF (P=0.7) and decreases in FVR (Figure 4A). However, in contrast to bradykinin, the t-PA response (-1.8±2 versus -0.6±0.4 ng/min/100ml, p=NS) and the arterial-venous t-PA concentration gradient (-0.4±0.3 versus -0.2±0.3, p=0.5) remained unchanged with SNP, indicating that SNP does not stimulate endothelial t-PA release Figure 4B).

Figure 4. Contribution of bradykinin and sodium nitroprusside to vasodilation and t-PA release.

Forearm blood flow (A), vascular resistance (B), and net t-PA release (C) in response to peak doses of bradykinin (400 ng/min) and sodium nitroprusside (3.2 μg/min). Data shown as mean ± SEM.

DISCUSSION

Herein, we demonstrate that although both NO and EDHF contribute to bradykinin-induced vasodilation, t-PA release is solely mediated by EDHF signaling pathways. We confirmed that cytochrome P450-derived EETs partly contribute to bradykinin-stimulated t-PA release, and now show that it is completely abolished by inhibition of K⁺_ca channels, the final site of action for several putative EDHFs. ²⁷ Finally, we demonstrate a lack of contribution of NO to bradykinin-stimulated t-PA release.¹¹

Endothelial fibrinolytic activity is an endogenous protective mechanism against the development and propagation of arterial thrombi.²⁸ In the fibrinolytic system, plasminogen is activated to plasmin that degrades fibrin into soluble fibrin degradation products.²⁹ The endothelium releases t-PA through a calcium and G protein-dependent pathway in response to a variety of endogenous mediators including catecholamines, thrombin, and bradykinin.^{30, 31} Endogenous bradykinin is produced by activation of the plasma kallikrein-kinin system on endothelial cells and bradykinin B₂ receptor blockade inhibits t-PA release.^{10, 11}

In the intact human vasculature, endothelium-dependent hyperpolarization can be mediated by agonists such as bradykinin and by physical stimuli including increases in shear stress that all increase intracellular calcium, causing opening of endothelial K⁺_ca channels and initiating downstream processes that explain the EDHF phenomena.^{3, 4} Candidate EDHFs include EETs synthesized from the cytochrome P450-dependent metabolism of arachidonic acid, and hydrogen peroxide generated from the degradation of superoxides through various endothelial oxidases such as NADPH-oxidase (NOX) or through superoxide dismutase-dependent dismutation.^{32, 33} The availability of TEA, a specific K⁺_ca channel antagonist, for use in humans has enabled investigations of EDHF activity in the forearm microcirculation.^{16, 17, 22} We now demonstrate that in addition to stimulating vasodilation, bradykinin stimulates t-PA release that is completely abolished by K⁺_ca channel inhibition.^{6, 34}

Endothelium-derived cytochrome P450 metabolites of arachidonic acid can hyperpolarize membranes primarily by activating the K⁺_ca channels, although the identity of the specific EETs remains controversial.^35-37 Herein we demonstrate that fluconazole, an inhibitor of cytochrome 2C9, attenuated bradykinin-mediated t-PA release, suggesting cytochrome P450 metabolites contribute significantly to t-PA release,¹⁹ observations that have been confirmed in both pre-clinical,^{14, 38} and clinical studies in both healthy and hypertensive subjects.²⁷

Because inhibition of EETs had a lesser effect than inhibition of K⁺_ca channels on bradykinin-stimulated t-PA release, it appears that other factors that stimulate K⁺_ca channels also contribute to EDHF activity, which in pre-clinical studies has been attributed to hydrogen peroxide, gap junctions or elevations in K⁺ release from endothelial cells.^{3, 18, 39-41} Hydrogen peroxide predominantly contributes to bradykinin-induced human arteriolar vasodilation⁴² and to release of t-PA from rat hearts,⁴³ but its role in the human circulation in vivo remains to be established. Thus, although hydrogen peroxide activates K⁺_ca channels and can function as an EDHF, its contribution to vascular homeostasis is complex.⁴⁴ It is not only a hyperpolarizing factor, but depending on the species, blood vessel, and concentration, hydrogen peroxide can be a signaling molecule and may also cause direct smooth muscle relaxation or depolarization of the smooth muscle and vasoconstriction.³

It may be argued that the bradykinin-mediated t-PA release is at least partly flow-mediated. However, with similar increases in blood flow with SNP, we and previous investigators found no increase in t-PA.^{6, 11, 27, 45, 46} Moreover, infusion of L-NMMA selectively attenuated FBF without affecting t-PA response to bradykinin.

Limitations

L-NMMA, fluconazole, and TEA are competitive inhibitors and may not completely inhibit NO, cytochrome P450 pathways, and K⁺_ca channels, respectively, and thus likely underestimate the physiologic importance of these mediators in vivo.^{47, 48} TEA also acts as a competitive inhibitor of the nicotinic receptor and reduces the water permeability of human AQP1 channels; however, in the doses administered, it has been shown to selectively inhibit K⁺_ca channels.^{49, 50} It is assumed that TEA effects are reflecting endothelium-dependent hyperpolarization as we cannot measure membrane potentials in this in vivo study. Fluconazole is not a specific inhibitor of cytochrome 2C9; however, we have shown significant inhibition of resting and bradykinin-mediated vasodilation with this inhibitor.⁴ Previous studies with a specific 2C9 inhibitor, sulphaphenazole, or with miconazole have yielded contradictory data.^{19, 51-53} We did not find any ethnic differences in bradykinin-mediated t-PA release even though ethnicity affects the vasodilator response to bradykinin; however, our study was not powered to explore this issue.⁵⁴ Our experiments were conducted in the setting of cyclooxygenase inhibition and therefore are unable to explore possible interactions with prostaglandins. Nevertheless, previous studies demonstrated no alteration in bradykinin-mediated t-PA release with cyclooxygenase inhibition.^{11, 55, 56}

CONCLUSIONS

We demonstrate that bradykinin-stimulated t-PA release is via endothelium-dependent hyperpolarization, partly through cytochrome P450-derived epoxides and ultimately via activation of K⁺_ca channels. Although bradykinin stimulates both NO and EDHF-dependent vasodilation, bradykinin-mediated t-PA release is purely EDHF-dependent. Both endothelium-dependent vasodilation and endothelial fibrinolytic capacity, measured as t-PA release, appear to predict the risk of future cardiovascular events.^{57, 58} Since abnormalities in the EDHF signaling pathway contribute to both abnormal vasodilation and increased thrombotic risk, agonists that enhance EDHF activity, such as epoxide hydrolase inhibitors, need to be investigated for their thrombolytic actions

Abbreviations

ASA

Acetylsalicylic acid

BK

Bradykinin

COX

Cyclooxygenase

EDHF

Endothelium-Derived Hyperpolarizing Factor

EETS

Epoxyeicosatrienoic acid

eNOS

Endothelial Nitric Oxide Synthase

FBF

Forearm Blood Flow

FLU

Fluconazole

H₂O₂

Hydrogen peroxide

K⁺_Ca

Calcium-activated potassium channels

L-NMMA

N^G-monomethyl-L-arginine

NO

Nitric Oxide

PGI₂

Prostacyclin

SNP

Sodium Nitroprusside

TEA

Tetraethylammonium chloride

tPA

Tissue Plasminogen Activator