The role of vascular endothelium in nitroglycerin‐mediated vasodilation

Abstract

Aims

Nitroglycerin (or glyceryl trinitrate, GTN) has been long considered an endothelium‐independent vasodilator because GTN vasodilation is intact in the absence of the endothelium and in the presence of endothelial dysfunction. However, in animal and in vitro models, GTN has been shown to stimulate the release of certain endothelium‐derived vasodilators such as nitric oxide (NO) and prostacyclin (PGI₂). In addition, chronic GTN therapy leads to endothelial dysfunction. In this series of experiments, we explored how GTN might interact with the vascular endothelium in normal humans, without cardiovascular disease or risk factors associated with abnormalities in vascular function.

Methods

We examined the effect of inhibition of NO, PGI₂, and epoxyeicosatrienoic acids (EETs, a class of endothelium‐derived hyperpolarizing factor) on GTN‐mediated vasodilation. We measured arterial blood flow responses to brachial artery infusions of GTN in the absence and presence of L‐NMMA (n = 13), ketorolac (n = 14) and fluconazole (n = 16), which are inhibitors of endothelium‐derived NO, PGI₂ and EETs, respectively, in healthy volunteers.

Results

Our results demonstrate that inhibition of endothelium‐dependent vasodilator mechanisms does not alter forearm resistance vessel responses to GTN.

Conclusion

We conclude that GTN‐mediated dilation of forearm resistance vessels is largely independent of vascular endothelium.

What is Already Known about this Subject

• Nitroglycerin is considered to be an endothelium‐independent vasodilator.

• Nitroglycerin has been shown to increase production of endothelium‐derived vasodilators such as nitric oxide and prostacyclin in endothelial cells.

• No human, in vivo investigation has been conducted to examine whether vascular endothelium is important in nitroglycerin‐mediated vasodilation.

What this Study Adds

• Inhibition of the production of endothelium‐derived vasodilators nitric oxide, prostacyclin or epoxyeicosatrienoic acids had no impact on nitroglycerin‐mediated dilation of human resistance vessels in vivo.

• Nitroglycerin produces its vasodilatory effect on forearm resistance vessels in a way that is independent of vascular endothelium in humans.

Introduction

Since its therapeutic properties were initially described in 1879 1, glyceryl trinitrate (GTN), also known as nitroglycerin, has been used to treat angina, acute coronary syndromes and congestive heart failure 2, 3, 4. Although GTN has long been considered to be an endothelium‐independent vasodilator, with potent effects on venous capacitance vessels and conduit arteries 2, 3, a number of lines of evidence suggest that some of the vascular effects of GTN may, in part, be to effect mediated by the endothelium 4. Studies performed in animal models suggest that acute administration of GTN stimulates the basal activity of endothelial nitric oxide synthase (eNOS) 5 and inhibition of eNOS can significantly decrease GTN‐mediated vasodilation 6. In addition, studies in cultured human endothelial cells have demonstrated the ability of GTN to stimulate synthesis of prostacyclin (PGI₂) 7, a potent vasodilator derived from the vascular endothelium 8. However, these observations should be interpreted with caution as these endothelial cells were extracted from conduit arteries, which might not be applicable to other vascular distributions. GTN may also affect endothelium‐derived hyperpolarizing factors (EDHFs) since chronic GTN therapy is found to downregulate the EDHF pathway 9. Epoxyeicosatrienoic acids, which are metabolites of arachidonic acid, have been identified as EDHFs 10. In vivo studies in animals and humans have also suggested that epoxyeicosatrienoic acids are involved in the regulation of vascular tone 11, 12, 13, 14, 15, 16, 17, 18. In humans, miconazole, a cytochrome P₄₅₀ inhibitor, decreases bradykinin‐induced vasodilation in the presence of eNOS and cyclooxygenase (COX) inhibition 15. Another cytochrome P₄₅₀ antagonist, fluconazole, caused a small decrease in basal vascular tone in healthy volunteers 16, 17, 18. To date, no study has explored whether the acute effects of GTN might be mediated in part by EDHFs.

It is now known that chronic GTN therapy is associated with the development of endothelial dysfunction 19, 20, which has been demonstrated to occur in both peripheral resistance vessels 21, 22, 23 as well as peripheral and coronary conduit arteries 24. Mechanistic studies demonstrate that chronic GTN therapy results in an increase in free radical production, which causes a decline in eNOS function 21, 22, 25, 26 and a decrease in the activity of PGI₂ synthase 27. In addition, chronic GTN therapy has also be shown to cause a downregulation on the EDHF pathway, although the mechanism and consequences of this effect remain unclear 9. These observations indicate that there is an interaction between GTN and endothelium‐derived vasodilatory mediators. In the current study, we examined the role of vascular endothelium in GTN‐mediated vasodilation in humans to determine whether, in part, it involves the endothelium‐derived vasodilators eNOS, PGI₂ and/or epoxyeicosatrienoic acids.

Methods

The protocols described herein were approved by the Mount Sinai Hospital Research Ethics Board and written informed consent was obtained from all participants.

Experimental protocol

Fifty‐eight healthy, non‐smoking males aged between 18 and 30 were recruited to participate in the study. Of this total, 48 were divided into three groups of 16 who were randomly assigned to sequential intra‐arterial infusions of GTN in the presence and absence of Levo‐N‐monomethylarginine (L‐NMMA), ketorolac or fluconazole, which are inhibitors of the production of endothelium‐derived NO, PGI₂ and epoxyeicosatrienoic acids, respectively. The remaining 10 subjects were assigned to receive two sequential infusions of GTN and serve as a control. All subjects received local anesthetic and had a 3 French arterial line inserted into the brachial artery of the non‐dominant arm using a Seldinger technique. Each subject first received normal saline infusion for 15 min and baseline forearm blood flow (FBF for baseline 1) was determined during the last 3 min. Immediately after baseline 1 was recorded, subjects received three infusions of GTN at increasing infusion rates of 10, 30 and 100 nmol min⁻¹. Each infusion lasted for 7 min and FBF measurements were performed continuously during the last 3 min followed by a 4 min washout period before the next infusion was initiated. The last infusion of GTN (100 nmol min⁻¹) was followed by a washout interval of 10 min. Depending on the assigned treatment, each subject then received a 15 min infusion of normal saline (control condition), L‐NMMA (a loading dose of 5 mg min⁻¹ for 5 min followed by a maintenance dose of 1.25 mg min⁻¹ and a co‐infusion of verapamil at 3.5 μg min⁻¹ for 10 min), ketorolac (a loading dose of 600 μg min⁻¹ for 5 min followed by a maintenance dose of 300 μg min⁻¹ for 10 min), or fluconazole (0.4 μg min⁻¹ for 15 min), immediately after which a new baseline FBF (baseline 2) was measured for 3 min in the presence of normal saline or one of these inhibitors. The infusion rates of L‐NMMA, ketorolac and fluconazole were derived from previous clinical studies that demonstrated effective inhibition of NO, PGI₂ and epoxyeicosatrienoic acids, respectively 16, 18, 28, 29, 30, 31, 32, 33, 34, 35.

The infusion rate of verapamil was determined in a pilot study (n = 10) where it was co‐infused with L‐NMMA to prevent the reduction in FBF which is associated with NOS inhibition 36, 37, 38, 39. In the pilot study, infusion of L‐NMMA was initiated with a loading dose of 5 mg min⁻¹ for 5 min followed by a maintenance dose of 1.25 mg min⁻¹ for 10 min. Co‐infusion of verapamil was then initiated starting at a rate 10 μg min⁻¹, based on prior reports 40, 41, and continued for 10 min until a steady‐state FBF was observed. This procedure was repeated in further volunteers until an infusion rate of verapamil (3.5 μg min⁻¹) that, in combination with L‐NMMA, resulted in FBF that was similar to the baseline value prior to the infusion of L‐NMMA.

After L‐NMMA (and verapamil), ketorolac and fluconazole had been infused for 15 min, those infusions were continued while intra‐arterial GTN was again infused using the same infusion rates (10, 30 and 100 nmol min⁻¹). All drugs were prepared in normal saline and all the infusions were kept at a constant rate of 0.8 ml min⁻¹ with a precision pump (Harvard apparatus, South Natick, Massachusetts). The experimental protocol is summarized in Figure 1. Intra‐arterial blood pressure and heart rate were recorded after each infusion (PowerLab 8/30) using the average of 15 cardiac cycles. The electrocardiogram was monitored continuously.

Figure 1.

Summary of experimental procedures: the procedure timeline for all four experimental conditions

GTN = glyceryl trinitrate (or nitroglycerin).

FBF was determined on both arms by venous occlusion strain‐gauge plethysmography (D.E. Hokanson Inc., Bellevue, Washington) with calibrated mercury‐in‐silastic strain gauges. Briefly, circulation of the hand was excluded by inflating a wrist cuff to 200 mm Hg during measurement periods. The upper arm cuff was inflated to 54 mm Hg and deflated at 10 s intervals (Hokanson rapid cuff inflator, D.E. Hokanson Inc. , Bellevue, Washington) for a period of 3 min, and FBF was recorded as the average of consecutive measurements. FBF was measured simultaneously in both arms and the responses. All FBF values were presented as the ratio of the infused vs. the non‐infused arm. This approach to the measurement of FBF normalizes the results obtained for the normal variability of FBF over time and is considered more repeatable and reliable than the absolute values of FBF in the infused arm alone. This is a standard approach used by our laboratory and others 22, 23, 41, 42 that allows for an assessment of changes in FBF associated with drug infusions while accounting for systemic influences on FBF which are reflected by changes in the non‐infused arm.

Statistical analysis

Two‐way analysis of variance (ANOVA), with appropriate contrast statements, was performed using SPSS Statistics (version 24) to analyse the changes in FBF in response to GTN at various infusion rates in the presence of one of the three endothelium inhibitors to determine the significance of vascular endothelium in GTN‐mediated vasodilation. A statistical significance level of <0.05 was taken as the threshold. For all results, post hoc comparisons between groups were performed with the Bonferroni correction. All results are expressed as mean ± SEM.

Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY 43, and are permanently archived in the Concise Guide to PHARMACOLOGY 2017/18 44.

Results

Effect of repeated infusions of GTN combined with saline

In the absence of any endothelium inhibitor, intra‐arterial GTN infusions had no effect on mean arterial blood pressure (MAP) and heart rate (HR) (Table 1). GTN caused a significant dose‐dependent increase in FBF which was not altered after repeated administration (Figure 2).

Table 1.

Mean arterial pressure (MAP) and heart rate (HR) at baseline and after GTN infusion at the highest rate (100 nmol min⁻¹)

MAP	Without endothelium inhibitor		With endothelium inhibitor (or saline)
	Baseline 1	GTN 100 nmol min −1	Baseline 2	GTN 100 nmol min −1
Saline	80 ± 2	81 ± 2	83 ± 2	81 ± 2
L‐NMMA	85 ± 3	82 ± 2	87 ± 3	88 ± 3
Ketorolac	76 ± 2	79 ± 2	80 ± 2	82 ± 2
Fluconazole	79 ± 1	80 ± 2	83 ± 2	84 ± 2

HR	Without endothelium inhibitor		With endothelium inhibitor (or saline)
	Baseline 1	GTN 100 nmol min −1	Baseline 2	GTN 100 nmol min −1
Saline	63 ± 2	68 ± 3	65 ± 2	69 ± 3
L‐NMMA	57 ± 2	59 ± 2	59 ± 2	64 ± 3a
Ketorolac	64 ± 2	67 ± 2	67 ± 2	71 ± 2
Fluconazole	67 ± 2	67 ± 2	71 ± 2	72 ± 2

Values expressed as mean ± SEM

^a P < 0.05 compared to the baseline HR value in the presence of L‐NMMA

Figure 2.

The effect of endothelium inhibition on GTN‐mediated vasodilation: forearm blood flow responses to intra‐arterial infusions of GTN in the absence and presence of endothelium inhibitor L‐NMMA, ketorolac, and fluconazole

FBF = forearm blood flow; GTN = glyceryl trinitrate (or nitroglycerin)

Pilot study: effect of L‐NMMA combined with verapamil on FBF

The pilot study showed that L‐NMMA alone significantly decreased FBF (0.85 ± 0.11 vs. 1.22 ± 0.12 at baseline, P < 0.05) and the co‐infusion of L‐NMMA and verapamil (3.5 μg min⁻¹) restored FBF to a level similar to baseline (1.45 ± 0.16 vs. 1.22 ± 0.12 at baseline, P = NS).

Effect of L‐NMMA (and verapamil) on GTN responses

Intra‐arterial GTN infusions had no effect on MAP; however, there was a small but significant increase in HR after GTN infusions in the presence of L‐NMMA and verapamil (Table 1). FBF at baseline (baseline 1) and in the presence of L‐NMMA and verapamil (baseline 2) was similar. GTN caused a significant dose‐dependent increase in FBF which was not altered in the presence of L‐NMMA and verapamil (Figure 2).

Effect of ketorolac on GTN responses

Intra‐arterial GTN infusions had no effect on MAP or HR in the absence or presence of ketorolac (Table 1). FBF at baseline (baseline 1) and in the presence of ketorolac (baseline 2) was similar. GTN caused a significant dose‐dependent increase in FBF which was not altered in the presence of ketorolac (Figure 2).

Effect of fluconazole on GTN responses

Intra‐arterial GTN infusions had no effect on MAP or HR in the absence or presence of fluconazole. FBF at baseline (baseline 1) and in the presence of fluconazole (baseline 2) was similar. GTN caused a significant dose‐dependent increase in FBF which was not altered in the presence of fluconazole (Figure 2).

Discussion

The current study examined the role of vascular endothelium in in vivo GTN‐mediated vasodilation in humans. We used the brachial artery infusion forearm blood flow model in order to avoid systemic haemodynamic effects and consequent reflex neurohormonal responses. Indeed, our results demonstrate that the systemic effects of intra‐arterial GTN infusions on blood pressure and heart rate were largely neutral (Table 1), suggesting minimal impact from the peripheral nervous system in response to the local vasodilatory effect of GTN.

Feelisch et al. demonstrated the production of NO when GTN was incubated with human endothelial cells 45. In other in vitro endothelial cell models, GTN was shown to increase the update of L‐arginine, a precursor of NO 5, and stimulate eNOS 5, 6 possibly via activation of the phosphatidylinositol 3‐kinase and protein kinase B pathways 46. These prior findings suggest that GTN could stimulate the production of endothelial NO contributing to its vasodilatory effect. Interestingly, a clinical study found that intra‐arterial administration of GTN at doses that did not change resting blood flow increased FBF responses to subsequent intra‐arterial infusions of acetylcholine, a known endothelium‐dependent vasodilator, in patients with congestive heart failure (CHF) but not healthy volunteers 47, indicating that there was no interaction between GTN and acetylcholine in healthy volunteers thereby suggesting there was no interaction between GTN and the endothelium in this population. We compared the vasodilatory effect of GTN in the resistance vascular bed in the forearm in the presence and absence of L‐NMMA and found that the effect was not altered by the pharmacological inhibition of endothelial NO production (Figure 2). Consistent with previous observations 47, our findings suggest that either endothelial NO is not crucial to GTN‐mediated dilation of resistance vessels or inhibition of endothelial NO production alone is not sufficient to alter the vasodilatory effect of GTN in healthy volunteers. However, our observation might not be applicable to predicting the role of endothelium in the GTN‐mediated dilation of other vessel types. Similar to our observations, the vasodilatory effect of sodium nitroprusside, a spontaneous NO donor, on resistance vessels is not altered by NO synthase inhibition 16, 48. However, NO inhibition has been shown to enhance the in vivo vasodilatory effect of this NO donor in human conduit arteries 17 and veins 49. Therefore, it is possible that eNOS might play a different role in the effect of GTN in other types of blood vessels. It is also possible that other commonly used organic nitrates including isosorbide dinitrate and isosorbide‐5‐mononitrate might have different interactions with eNOS depending on the specific vascular bed.

The evidence of a possible interaction between GTN and vasoactive prostaglandins has been inconsistent. GTN, at clinically relevant concentrations, has been shown to increase the production of prostacyclin (PGI₂) in cultured human endothelial cells 7 as well as saphenous 50, 51 and umbilical 50 veins. The mechanism via which GTN might increase PGI₂ production in these studies was unclear. Since sodium nitroprusside, a spontaneous NO donor, did not have an effect on PGI₂ production 51, it is possible that GTN itself (without biotransformation into NO or NO‐like moiety) might be able to stimulate one of the enzymes in the PGI₂ synthesis pathway. Since indomethacin, a non‐selective cyclooxygenase inhibitor, and a specific thromboxane A₂ synthase inhibitor were shown to decrease and enhance GTN‐stimulated release of PGI₂, respectively 50, 51, it seems likely that GTN might stimulate PGI₂ synthase and increase its activity. In contrast, another study found that GTN, at therapeutic concentrations, does not alter basal level of PGI₂ in human saphenous veins and mesenteric arteries 52. Clinical observations concerning the role of endothelium‐derived PGI₂ on the pharmacodynamic effects of GTN failed to find an important interaction. Studies using aspirin, which inhibits the production of PGI₂, potentiated the early hypotensive effect of sublingual GTN although this enhancement appeared to be mediated by changes in GTN pharmacokinetics 53, 54, 55. We elected to examine whether PGI₂ is involved in mediating the vasodilatory effect of GTN using a forearm infusion model. Our results show that the vasodilatory effect of GTN on resistance vessels in the forearm was not changed by pharmacologic inhibition of PGI₂ production with ketorolac (Figure 2). Our findings do not support the hypothesis that a component of vascular responses to GTN is mediated by PGI₂ in the human peripheral resistance vascular bed. Alternatively, it is possible that the concurrent inhibition of endothelium‐derived vasoconstrictor thromboxane A₂ production by ketorolac, an inhibitor of cyclooxygenase which synthesizes the common precursor of both PGI₂ and thromboxane A₂, might have counteracted the potential change resulting from the inhibition of PGI₂ production. This is unlikely since GTN has been shown to have no effect on endothelial thromboxane A₂ production 50.

Epoxyeicosatrienoic acids (EETs) have been identified as EDHFs that might act as a compensatory mechanism to regulate blood flow when NO and PGI₂ production is inhibited in human vessels 56, a role that might be particularly important in resistance vessels 57. To date, there have been no studies examining the potential role of EDHFs as mediators of the vasodilator responses to GTN. In light of this, we hypothesized that GTN would stimulate the release of EETs or have a synergistic interaction with these EDHFs, enhancing the overall vasodilatory effect of GTN. We examined whether inhibition of the production of EETs with fluconazole, an inhibitor of EETs producing enzymes (cytochorome P₄₅₀ 2C and 2J), would inhibit GTN‐mediated dilation of resistance vessels. We found that fluconazole did not cause any change in the vasodilatory effect of GTN. This suggests that EETs do not play a significant role in GTN‐mediated dilation of peripheral resistance vasculature.

Conclusion

Our current observations suggest GTN‐mediated dilation of forearm resistance vessels is largely independent of vascular endothelium. Over the past decade, GTN and acetylcholine have been used as endothelium‐independent and ‐dependent vasodilators of forearm resistance vessels, respectively, in human studies examining endothelium dysfunction 22, 23, 42, 58. The current study provides further evidence which supports the use of GTN as an endothelium‐independent vasodilator of resistance vessels.

Study limitations

Although not examined in the current study, it is possible that GTN could stimulate the production of multiple endothelial vasodilator species which may have compensatory capacity. In this case, inhibition of only individual vasodilator species might not be sufficient to cause a significant overall change in the vasodilatory effect of GTN. Although this possibility is acknowledged, the current findings suggest that GTN‐mediated dilation of the resistance vessels is largely independent of the vascular endothelium. However, it needs to be emphasized that our findings are restricted to the interaction between endothelium‐dependent vasodilators and the response of arterial resistance vessels to GTN. We have only explored the effect of the endothelium on GTN‐mediated dilation of forearm resistance vessels. We are unable to comment on whether our observations could be applied to other blood vessel types such as venous capacitance vessels and conduit arteries. It is possible that the endothelium may play a more important role as a mediator of GTN effects in veins and/or the conductance arterial system. In addition, we would like to acknowledge that the experimental conditions of previous in vitro studies are not directly comparable to the in vivo nature of our study. It is possible that the experimental conditions of prior in vitro studies and, in some cases, species differences explain how our results differ from those prior reports.

Competing Interests

There are no competing interests to declare.

The authors would like to thank their nurse manager Susan Kelly and research nurse Joyce Hill for their technical and medical assistance for the execution of all experiments. This work was supported, in part, by a grant in aid from the Canadian Institute for Health Research.

Contributors

K.Z. contributed to study design, execution of all the experiments, preparation and writing of the manuscript. J.D.P. provided consultation to K.Z. in study design, performed all intra‐arterial catheterizations, supervised all the experiments, writing and finalization of the manuscript. The principal investigator for this study is J.D.P. and he had direct clinical responsibility for patients.

Zhou, K. , and Parker, J. D. (2019) The role of vascular endothelium in nitroglycerin‐mediated vasodilation. Br J Clin Pharmacol, 85: 377–384. 10.1111/bcp.13804.