Nitroglycerin protects the endothelium from ischaemia and reperfusion: human mechanistic insight

Abstract

What is already known about this subject

• The production of oxygen free radicals upon administration of nitroglycerin (GTN) and other organic nitrates has been advocated as one of the causes of nitrate-induce tolerance and endothelial dysfunction.

• It has been shown that nitrates also cause a protective phenomenon that is similar to ischaemic preconditioning, but the mechanisms of this effect have never been investigated in humans.

What this study adds

• We show, in vivo in humans, that GTN causes endothelial protection against ischaemic damage.

• Furthermore, we show that this effect is mediated by release of oxygen free radicals and by opening of the mitochondrial permeability transition pore.

Aims

Nitroglycerin (GTN) modulates tissue damage induced by ischaemia and reperfusion (IR) in a mechanism that is similar to ischaemic preconditioning. We set out to study, using a human model of endothelial IR injury, whether GTN-induced endothelial preconditioning is mediated by reactive oxygen species (ROS) formation and/or opening of mitochondrial permeability transition pores (mPTP).

Methods

In two double-blind, randomized, parallel studies, a total of 66 volunteers underwent measurement of radial artery endothelium-dependent, flow-mediated dilation (FMD) before and after local IR. Subjects were treated, 24 h before IR, with different drugs in order to test the mechanism of GTN-induced endothelial protection.

Results

Transdermal GTN (0.6 mg h⁻¹ for 2 h, administered 24 h before IR) significantly reduced the impairment of FMD caused by IR (placebo group: FMD after IR, 1.3 ± 0.8%; GTN group: FMD after IR, 5.3 ± 0.9%, P < 0.01 compared with placebo). This protective effect was lost when vitamin C (2 g i.v. at the time of GTN administration) or ciclosporin (an inhibitor of mPTP, 100 mg 2 h prior to GTN administration) were coadministered (FMD after IR: vit C + GTN group, 2.1 ± 1.0%; ciclosporin + GTN group, 1.7 ± 0.8%; both P < 0.05 compared with GTN alone).

Conclusions

We demonstrate that GTN protects the endothelium against IR-induced endothelial dysfunction, in an effect similar to delayed ischaemic preconditioning. Using a human model, we provide evidence supporting the concept that this protective effect is mediated by ROS release and mPTP opening upon GTN administration.

Introduction

The expression ‘ischaemic preconditioning’ (i.e. exposure to brief periods of reversible ischaemia 24–72 h prior to an infarction) refers to a protective phenotype characterized by increased tissue resistance towards ischaemia and reperfusion (IR) injury [1]. Although this phenomenon has been more thoroughly investigated in vitro at the level of cardiomyocytes and whole organs, multiple studies, including one from our group, have demonstrated that a similar protective state can be observed in humans in vivo at the level of the vascular endothelium [2–5]. Given the role of the endothelium in modulating vasomotor, antithrombotic and anti-inflammatory mechanisms, such protective effects might have important clinical implications in limiting tissue damage associated with acute coronary syndromes. Stimuli that induce (or mimic) this protective endothelial phenotype include brief ischaemia (ischaemic preconditioning [2]) as well as specific pharmacological agents such as l-arginine [4], sildenafil [5] and the ATP-dependent potassium (K-ATP) channel opener diazoxide [3]. While the mechanism(s) of ischaemic and pharmacological preconditioning are still incompletely understood, reactive oxygen species (ROS) and mitochondrial channels such as the permeability transition pore (mPTP) are considered to be important components of this phenomenon [6, 7]. A recent hypothesis states that small bursts of ROS triggered by short ischaemia and/or pharmacological stimuli might cause reversible (low conductance) opening of the mPTP, leading to transient mitochondrial uncoupling. This subcritical uncoupling of the oxidative phosphorylation would subsequently limit cellular death upon IR, possibly by preventing prolonged (high conductance) opening of the mPTP and subsequent massive mitochondrial matrix swelling [8]. Confirming this view, inhibition of the mPTP, as well as incubation with antioxidants, blocked ischaemic preconditioning in an animal model of cardiac IR injury [7].

Nitroglycerin (GTN) is a nitric oxide donor commonly prescribed for the therapy of myocardial ischaemia and congestive heart failure. Recent studies have shown that short-term (1–4 h) administration of GTN modulates cardiac sensitivity to IR injury through mechanisms that mimic the molecular pathways of ischaemic preconditioning [9]. This effect is manifest, in animal models, as a reduction in infarct area after coronary artery ligation [10] and, in humans, as a reduction in measures of myocardial ischaemia in models of low-flow and demand-induced ischaemia [11, 12]. Of importance, it has been recently documented that exposure to GTN causes a rapid increase in mitochondrial ROS production [13]. Because the role of ROS and mPTP in (ischaemic or pharmacological) preconditioning has never been investigated in humans, we set out to test, in humans in vivo, whether ROS production and/or mPTP opening play a role in the preconditioning-mimetic effect of GTN.

Methods

Studies were approved by the local Ethics Committee. Written informed consent was obtained in all cases.

ROS-dependent endothelial protection – protocol 1

The protocol and model used for the human IR injury studies have been described in detail in a previous publication from our group [5].

Twenty-two healthy male nonsmoking volunteers (22–46 years old) were enrolled in this double-blind, randomized, placebo-controlled parallel trial. Subjects were randomized to receive i.v. placebo (normal saline) or 2 g of vitamin C. Immediately after administration of vitamin C or placebo, all subjects received transdermal GTN (0.6 mg h⁻¹, Minitran; 3M, St Paul, MN, USA) for 2 h. Twenty-four hours after removing the transdermal patch, subjects returned to the laboratory and radial artery endothelium-dependent flow-mediated dilation (FMD) was measured in the nondominant arm before and after local IR injury. In order to induce radial artery IR, a pneumatic cuff placed at the level of the brachial artery was inflated to 200 mmHg for 15 min, after which 15 min of reperfusion was allowed. We and others have previously showed that this cycle of IR causes specific impairment of endothelium-dependent vasodilation while leaving smooth muscle responsiveness unaltered, as shown by sublingual GTN administration [2–5]. Two separate groups of volunteers underwent the same FMD–IR–FMD protocol to serve as controls. These subjects received no therapy (n = 10) or i.v. vitamin C alone (n = 7). The interventions received by each of the groups are presented in Figure 1.

Figure 1.

Subdivision in groups for the human protocols. A total of 66 subjects was enrolled. All groups underwent measurement of radial artery flow-mediated dilation (FMD, arrows) before and after local ischaemia/reperfusion (IR)

MPTP-dependent endothelial protection – protocol 2

Twenty healthy male nonsmoking volunteers (25–30 years old) were enrolled in a second double-blind, randomized, controlled parallel trial. All subjects received ciclosporin (Sandimmune Neoral, 100 mg per os; Novartis, Basel Switzerland). This dosage yielded an average plasma concentration of 450 ± 92 µm. Two hours later, subjects were randomized to receive placebo treatment or transdermal GTN (0.6 mg h⁻¹) for 2 h. Twenty-four hours after removing the transdermal patch, the same protocol described for endothelial IR injury above was applied. An additional control group (n = 7 volunteers) received GTN alone. There were no side-effects in any of the groups.

Analysis

Recently developed (Department of Informatics Engineering, University of Siena), fully automated edge detection software was used for the measurement of radial artery FMD (http://www.heartworks.it). Using this software, consecutive measurements of radial artery diameter yield a coefficient of variation of <1%. The methods employed to study IR-induced endothelial dysfunction are described in detail in a previous report from our group [5]. Data are presented as mean ± SE. The effect of IR on radial artery diameters, reactive hyperaemia and FMD within each group was tested using a paired t-test and differences among groups were analysed using unpaired t-test or analysis of variance (anova), as appropriate, with P < 0.05 set as the threshold for significance. Statview version 5 (SAS Institute Inc., Cary, NC, USA) was employed for statistical analysis.

Results

ROS-dependent endothelial protection – protocol 1

Radial artery diameter measurements are presented in Table 1. Resting blood pressure, radial artery diameter as well as FMD before IR were similar among groups. FMD data are presented in Figure 2. Also, radial artery diameters were not modified by IR. After IR, as previously described [2, 5], FMD was significantly blunted compared with pre-IR values in the control group of subjects that received no therapy (FMD after IR, 1.3 ± 0.8%) and in those who received vitamin C alone (FMD after IR, 1.3 ± 1.3%), P < 0.01 compared with before IR for both groups. Providing evidence of the endothelial pharmacological preconditioning effect of GTN, administration of GTN + i.v. placebo almost entirely prevented IR-induced blunting of FMD (FMD after IR, 5.3 ± 0.9%, P < 0.01 compared with no therapy, after IR; Figure 2). In the GTN + vitamin C group, IR caused a reduction in FMD that was essentially identical to that observed in the subjects who received no therapy (FMD after IR, 2.1 ± 1.0%, P < 0.01 compared with before IR; P < 0.05 compared with GTN + placebo and P = NS compared with no therapy, after IR). There was no effect of IR or of any drug on radial artery blood flow at any time point; in particular, there was no difference in reactive hyperaemia between the GTN + placebo (before IR, from 34 ± 10 to 282 ± 74 ml min⁻¹; after IR, from 24 ± 5 to 254 ± 52 ml min⁻¹) and the GTN + vitamin C group (before IR, from 33 ± 8 to 514 ± 56 ml min⁻¹; after IR, from 37 ± 12 to 353 ± 100 ml min⁻¹).

Table 1.

The effect of ischaemia/reperfusion (IR) on flow-mediated dilation (FMD) in the seven groups

	Before IR Resting	FMD, diameter change	After IR Resting	FMD, diameter change
Protocol 1
Vitamin C alone	2.60 ± 0.1	0.20 ± 0.08	2.65 ± 0.1	0.03 ± 0.03*‡
No therapy	2.38 ± 0.1	0.18 ± 0.03	2.41 ± 0.1	0.03 ± 0.02*‡
GTN + placebo	2.55 ± 0.1	0.22 ± 0.02	2.52 ± 0.1	0.13 ± 0.03†‡
GTN + vitamin C	2.48 ± 0.1	0.18 ± 0.02	2.57 ± 0.1	0.05 ± 0.02*‡
Protocol 2
GTN alone	2.65 ± 0.1	0.19 ± 0.03	2.68 ± 0.1	0.15 ± 0.03§
GTN + ciclosporin	2.48 ± 0.1	0.15 ± 0.01	2.49 ± 0.1	0.04 ± 0.02*§
Ciclosporin	2.33 ± 0.1	0.17 ± 0.02	2.42 ± 0.1	0.08 ± 0.02*§

Radial artery diameter data (in mm). ‘Baseline’ data refer to radial artery diameter before wrist cuff occlusion. ‘FMD, diameter change’ data refer to maximal dilation after wrist cuff release.

^*P < 0.01 compared with before IR

^†P < 0.05 compared with before IR, within-group

^‡P < 0.01 anova

^§P < 0.05 anova.

Figure 2.

Reactive oxygen species-mediated preconditioning (protocol 1). Percent increase of radial artery diameter from baseline during reactive hyperaemia before and after ischaemia/reperfusion (IR). There was no difference between groups at baseline. After IR, endothelium-dependent vasodilation was significantly blunted in all groups (*P < 0.01 compared with corresponding group before IR; †P < 0.05 compared with corresponding group before IR), although it was significantly higher in the subjects who received nitroglycerin (GTN) + placebo (P< 0.05 compared with all other groups, after IR). Abbreviations as in Table 1

MPTP-dependent endothelial protection – human protocol 2

Radial artery diameter measurements are presented in Table 1. Resting blood pressure, radial artery diameters and blood flows were similar (both before and after IR) among groups. FMD data are reported in Figure 3. Administration of GTN alone caused similar protection against ischaemic injury to that seen in protocol 1 (FMD after IR, 5.5 ± 1.3%). After IR, FMD was blunted in subjects who received ciclosporin alone (FMD after IR, 3.5 ± 1.0%, P = 0.01 compared with before IR) and in those that received GTN + ciclosporin (FMD after IR, 1.7 ± 0.8%, P < 0.01 compared with before IR; P < 0.05 compared with GTN alone).

Figure 3.

Mitochondrial permeability transition pore-mediated preconditioning (protocol 2). Percent increase of radial artery diameter from baseline during reactive hyperaemia before and after ischaemia/reperfusion (IR). There was no difference between groups at baseline. After IR, endothelium-dependent vasodilation was significantly higher in the subjects who received nitroglycerin (GTN) alone ▪, while ciclosporin blocked GTN protection. In subjects who received ciclosporin alone , IR caused a reduction of flow-mediated dilation that, although blunted, was not significantly different compared with the GTN + ciclosporin group □. Abbreviations as in Table 1

Discussion

In a recent report, Kharbanda et al. demonstrated that the same IR stimulus employed here specifically impairs FMD of forearm conductance vessels while leaving smooth muscle responses unchanged [2]. Further studies have shown that exposure to brief episodes of ischaemia (ischaemic preconditioning) [2], L-arginine [4], as well as sildenafil [5], limit this impairment in endothelium-dependent relaxation induced by IR. The data presented here demonstrate that GTN also causes this protective effect. Furthermore, we provide evidence supporting the concept that this phenomenon is mediated by ROS production and mPTP opening.

There is now clear evidence that ROS production is an important component of both ischaemic and pharmacological (except for adenosine-induced) preconditioning [14]: in animal studies, antioxidants prevented this protective phenomenon [15, 16] and intracoronary administration of a ROS-generating solution, in the absence of ischaemia, was sufficient to induce an infarct-sparing effect equivalent to that observed after ischaemic preconditioning [17]. Within this mechanistic construct, the role of the mPTP in IR injury appears to be complex: upon reperfusion, changes in cytosolic Ca²⁺, pH or ROS induce prolonged (high-conductance) opening of this large channel positioned in the inner mitochondrial membrane, causing severe energy depletion, mitochondrial matrix swelling and release of the cytochrome C from the mitochondrial matrix [18]. These mPTP-dependent phenomena are considered terminal cellular events contributing to tissue injury via necrosis and apoptosis [19, 20]. In contrast, both ischaemic and pharmacological preconditioning stimuli cause transient (low conductance; possibly ROS-dependent) opening of the mPTP [7, 16]: this subcritical mitochondrial transition stimulates a molecular cascade which, among other protective effects, might ultimately limit lethal mPTP opening during IR [21]. Therefore, inhibitors of the mPTP might have a dual effect: when administered during IR, they prevent mPTP opening, therefore limiting infarct size; when administered before preconditioning, they block transient mPTP opening, inhibiting this protective phenotype [7]. Of importance, while it remains to be demonstrated whether ciclosporin inhibits ischaemic preconditioning (e.g. preinfarctual angina) in humans, this is the first study to provide evidence supporting a role of ROS and mPTP in human pharmacological preconditioning.

It has to be acknowledged that the present data were acquired in healthy human volunteers, and future studies will have to investigate whether the phenomena described here can translate into a clinical advantage. It also has to be acknowledged that ciclosporin has several other effects besides mPTP inhibition: by binding to cytosolic proteins of the ciclophilin family, ciclosporin has a high affinity for several key substrates, including calcineurin, in a multitude of cell types. We therefore cannot conclude with certainty that the observed effects of ciclosporin were mediated via the mPTP alone. While this limitation needs to be recognized, concentrations similar to those employed here were previously used in animal studies to achieve specific inhibition of mPTP opening [7].

In sum, in the last 10 years, evidence has been reported that sustained ROS production during prolonged continuous nitrate therapy [22] causes a variety of vascular abnormalities which have the potential to influence negatively the prognosis of cardiovascular patients [23]. By contrast, this study and others confirm that short-term exposure to the GTN can provide protection from ischaemic insult. In the current study, we provide evidence supporting a role of ROS production and mPTP opening, triggered by brief exposure to GTN, in the induction of a potent endothelial protection that is similar to that observed after ischaemic preconditioning. Taken together, these phenomena provide a novel mechanistic explanation for GTN-induced (ROS-mediated) activation of protein kinase C [24], which has been previously implicated in the development of both GTN-triggered and ischaemic preconditioning. Of note, providing further evidence in support of a biphasic effect of the molecular cascade induced by GTN administration, activation of protein kinase C has also been shown to be an important player in the vascular toxicity which follows prolonged high-dose nitrate therapy (reviewed in [25]).

In conclusion, we believe that the present findings provide a new view of the mechanisms of action of organic nitrates, and emphasize the need for a critical reappraisal of the effects of this class of drug. Finally, our data provide the first human in vivo evidence of a (paradoxical) beneficial role of ROS. Whether this effect can be exploited clinically, and whether our observations provide a mechanistic background for the inconsistent results of trials testing the effect of antioxidant therapy in patients with cardiovascular disease, remain to be studied.