Non-Insulin Injectable Treatments (Glucagon-Like Peptide-1 and Its Analogs) and Cardiovascular Disease

Introduction

I ncretin hormones serve a major physiologic role in postprandial nutrient metabolism. Their effects include glucose-dependent stimulation of insulin secretion and, in some incretins, inhibition of glucagon secretion. The two well-characterized incretin hormones, glucose-dependent insulinotropic polypeptide and glucagon-like peptide-1 (GLP-1), have quite different spectrums of activity, with GLP-1 being more potent and having more diverse effects. ¹ GLP-1 action is mediated both by systemic circulation with subsequent binding to a specific cell surface G-protein-coupled receptor (GLP-1 receptor) and by transmission via the vagus nerve and/or the nodose ganglia to the hind brain where it activates specific neuronal pathways. ² GLP-1's effects are concentration dependent. ^1,2 Low concentrations have effects on insulin and glucagon secretion. High concentrations, in addition, delay gastric emptying, increase satiety, reduce food intake, and have profound effects on organs such as the cardiovascular (CV) and immune systems.

Several aspects of the GLP-1 system are unusual. GLP-1 receptors are widely expressed in such diverse tissues as pancreatic islets, kidney, lung, heart, endothelium, and multiple regions of the peripheral and central nervous system. ² The existence of these GLP-1 receptors raises the likelihood that GLP-1 could have effects on these other tissue. Several recent reviews have summarized those published effects of GLP-1 receptor agonists that either directly or indirectly might have an impact on the CV system. ^3,4 The present review focuses on the complex direct effects of GLP-1 and GLP-1 receptor agonists on cardiac events with a perspective on potential mechanisms. A distinction has been made between such direct effects and alterations in cardiac risk factors, which secondarily might reduce future CV events.

Secreted GLP-1 is metabolized within a minute or two by a ubiquitous enzyme, dipeptidyl peptidase-4 (DPP-4), which cleaves peptides with a penultimate alanine residue (Fig. 1). ⁵ The metabolite generated is GLP-1 (9–36)-amide. Very little circulating GLP-1 is in the active form, with the majority being the metabolite. This very rapid degradation of GLP-1 (7–36)-amide and generation of GLP-1 (9–36)-amide raises the possibility that GLP-1 (9–36)-amide might have some significant physiologic function. ⁶ The rapid degradation of the secreted GLP-1 (7–36)-amide minimizes its value as a pharmacologic treatment and has led to the development of GLP-1 analogs resistant to DPP-4 action (Fig. 2) and to molecules that inhibit the DPP-4 enzyme as therapeutic agents. ^1,2,5 Because endogenous GLP-1 and glucose-dependent insulinotropic polypeptide are rapidly cleaved by DPP-4, it is obvious that differences are likely to exist in the effects of administered GLP-1 versus GLP-1 analogs that resist DPP-4 degradation (exenatide and liraglutide) and administered GLP-1 or GLP-1 analogs and administered drugs that inhibit the DPP-4 enzyme. ^1,2,5

OPEN IN VIEWER

FIG. 1.

Secretion and metabolism of glucose-dependent insulinotropic polypeptide (GIP) and glucagon-like peptide-1 (GLP-1). In response to meal ingestion, GIP is secreted from the K-cells in the upper small bowel, and GLP-1 (7–36)-amide is secreted from the L-cells of the distal small bowel. Both are very rapidly metabolized by the dipeptidyl peptidase-4 enzyme. The cleavage occurs between amino acids 2 and 3 and generates GIP (3–42) and GLP-1 (9–36)-amide.

OPEN IN VIEWER

FIG. 2.

Structures of glucagon-like peptide-1 (GLP-1) (7–36)-amide, its metabolite GLP-1 (9–36)-amide, its commercially available analogs, and exendin (9–39), an inhibitor of the GLP-1 receptor.

There are two tools that can be used to ascertain whether an effect observed with GLP-1 or a GLP-1 analog is brought about via the specific GLP-1 receptor or through a non–GLP-1 receptor-mediated action. An analog of exendin-4 (exenatide) with the N-terminal two amino acids deleted, exendin (9–39)-amide (Fig. 2), binds to the GLP-1 receptor but has no activity and therefore is an inhibitor of GLP-1 activity mediated through the GLP-1 receptor. ⁷ A strain of mice has been developed in which GLP-1 receptors were deleted. ⁸ An action of GLP-1 or its analogs observed in these GLP-1 receptor knockout mice should be mediated by a GLP-1 receptor-independent mechanism.

CV Effects of Administered GLP-1 in Humans

For many years clinicians had explored whether infusions of glucose, insulin, and potassium could improve clinical outcomes in patients with acute myocardial infarctions by enhancing myocardial glucose uptake and oxidation. Those studies generally were equivocal, with a few patients improving and others showing no benefit or adverse effects. ⁹ The demonstration of myocardial GLP-1 receptors and the availability of recombinant GLP-1 made it possible to explore whether continuous infusions of GLP-1 could improve myocardial function in ischemic heart disease. Ten patients (five with diabetes mellitus) with acute myocardial infarction and decreased left ventricular (LV) ejection fraction (LVEF) (<40%) were given a 72-h infusion of GLP-1 (1.5 pmol/kg/min) in addition to standard background therapy following successful primary angioplasty. ¹⁰ Compared with the 11 patients (four with diabetes mellitus) treated similarly but without GLP-1, the GLP-1-treated patients had improved LVEF, global wall motion score indices, and regional wall motion score indices. A subsequent study evaluating the effects of GLP-1 infusion (1.5 pmol/kg/min) on the outcome of patients with preserved baseline LV function undergoing coronary bypass grafting showed that GLP-1-treated patients required less postoperative use of inotropic and vasoactive infusions and had fewer postoperative arrhythmias compared with the control patients. ¹¹ Glycemic control was better, and insulin requirements were 45% less in the GLP-1-treated patients. Several studies have examined the effects of GLP-1 infusion on patients with chronic heart failure. Sokos et al. ¹² treated 12 patients (eight with diabetes mellitus) with New York Heart Association (NYHA) Class III/IV heart failure with a GLP-1 (2.5 pmol/kg/min) infusion for 5 weeks plus standard therapy and compared their results with those of nine patients (five with diabetes mellitus) with similar baseline demographics, background therapy, and degree of LV dysfunction treated only with standard therapy. GLP-1 treatment significantly improved LVEF, maximum O₂ consumption, 6-min walk distance, and the Minnesota Living with Heart Failure questionnaire score. Halbirk et al. ¹³ examined the effects of a 48-h infusion of GLP-1 (0.7 pmol/kg/min) in 20 patients without diabetes exhibiting heart failure (NYHA Class II and III) and ischemic heart disease (LVEF=30±2%) in a randomized, double- blind, placebo-controlled crossover design. They found that GLP-1 infusion increased circulating insulin and reduced plasma glucose but had no major CV effects. GLP-1 infusion caused a minor increase in heart rate and diastolic blood pressure. Mild to modest hypoglycemia occurred in eight patients during GLP-1 infusion. This 2-day study was unable to confirm the results of 5-week treatment of Sokos et al. ¹²

A cardioprotective effect of GLP-1 in humans has been demonstrated in a recent pilot study in which an infusion of GLP-1 (1.2 pmol/kg/min) given for 30 min after a 1-min low-pressure balloon occlusion at the site of a left anterior descending coronary artery stenosis improved recovery of LV systolic and diastolic function 30 min after the balloon occlusion compared with a saline control. ¹⁴ Coronary balloon occlusion caused LV stunning in the control group with cumulative LV dysfunction on subsequent occlusion. LV stunning did not occur in the GLP-1-treated patients. The clinical implications of these data are that GLP-1 may reduce the damage accompanying short-term coronary artery closure during the placement of coronary stents.

Mechanistic Studies of CV Effects of GLP-1 in Animal Models

Human studies do not permit the extensive manipulations and measurements necessary to characterize the mechanisms by which GLP-1 exerts its beneficial CV effects. Studies in animal models of CV disease (CVD) have been carried out to better define how GLP-1 exerts its beneficial CV effects. GLP-1 infusions have been shown to increase myocardial function and metabolism in both ischemic and non-ischemic models of myocardial disease.

Proof that GLP-1 limits myocardial stunning has been shown in studies with conscious dogs. A 24-h continuous GLP-1 infusion (1.5 pmol/kg/min) in dogs undergoing 10-min occlusion of the left circumflex coronary artery followed by 24-h reperfusion restored regional wall motion earlier (92±4% at 15 min) and was complete (99±4%) at 24 h, whereas it was later (57±5% at 15 min) and residual contractile dysfunction persisted (78±3% at 24 h) in control treated dogs. ¹⁵ This occurred despite identical recovery of coronary blood flow. The reduction of myocardial stunning was independent of changes in systemic hemodynamics or global systolic function. Isovolumic LV relaxation improved significantly in the GLP-1-treated dogs.

In a non-ischemic CVD study, recombinant GLP-1 infusions (1.5 pmol/kg/min) for 48 h increased LV dP/dt (98%), stroke volume (102%), and cardiac output (57%) in conscious dogs with pacing-induced dilated cardiomyopathy (DCM) compared with saline-treated controls with DCM. ¹⁶ Additionally, GLP-1 treatment decreased LV end-diastolic pressure, heart rate, and systemic vascular resistance in DCM dogs. ¹⁶ GLP-1 treatment increased basal myocardial glucose extraction and uptake, decreased plasma norepinephrine and glucagon, and increased insulin-stimulated myocardial glucose uptake in dogs with induced DCM. ¹⁶ Infusions of GLP-1 in conscious dogs prior to the induction of cardiomyopathy had no effect on LV systolic, end-diastolic, or mean arterial pressure and showed a trend toward decreased dP/dt and cardiac output. ¹⁶ GLP-1 infusion increased LV and systemic hemodynamics in dogs with DCM but not in normal dogs. GLP-1 increased myocardial glucose uptake but had no effect on plasma norepinephrine or glucagon levels in normal dogs.

Detailed analyses of the metabolic and biochemical effects of a 48-h GLP-1 infusion in dogs with DCM ¹⁷ showed the following: (1) GLP-1 infusion caused up-regulation and increases in the active form of myocardial GLP-1 receptors. (2) GLP-1 infusion increased myocardial glucose uptake that was not accompanied by increased cyclic AMP (cAMP) level, Akt phosphorylation, activation of AMP kinase, or translocation of Glut-4 glucose transporters. (3) Increased myocardial glucose uptake occurred in the absence of an increase in peripheral plasma insulin and was associated with increased myocardial p38 α mitogen-activated protein (MAP) kinase activity, induction of nitric oxide synthase 2, and increased Glut-1 glucose transporter translocation. The importance of the latter pathway in increasing GLP-1-mediated myocardial glucose uptake in this model of myocardial disease was confirmed by showing that blocking MAP kinase activity by a MAP kinase inhibitor or nitric oxide synthase by nitro-l-arginine eliminated the GLP-1-mediated increase in myocardial glucose uptake. These data suggest that the mechanism of action of GLP-1 on the myocardium differs from that of its effects on the beta cell.

In intact rats, GLP-1 infused during a 30-min cardiac ischemia followed by a 120-min reperfusion reduced the infarct size in the susceptible myocardial area to 20% compared with the infarct size of 44% in the susceptible area of control animals. ¹⁸ In vitro studies showed that the protection was abolished by inhibition with exendin 9-39 (an inhibitor of GLP-1 receptor), Rp-cAMP (an inhibitor of cAMP), LY294002 (an inhibitor of phosphoinositide 3-kinase), and UO126 (an inhibitor of p42/44 MAP kinase). Continuous GLP-1 infusions administered from Month 9 to Month 12 in spontaneously hypertensive heart failure rats increased survival at 12 months (72%) compared with saline-treated controls (44%). ¹⁹ Control animals after 3 months had decreased stroke volume and LVEF and increased LV end-diastolic pressure. GLP-1-treated animals after 3 months had increased cardiac output and no change in LVEF and LV end-diastolic pressure. GLP-1 increased myocardial glucose uptake, Akt phosphorylation, and Glut-4 translocation. GLP-1 treatment reduced body weight and body mass index. GLP-1 treatment decreased myocardial apoptosis as measured by terminal deoxynucleotidyl transferase dUTP nick end-labeling staining and capase-3 activation.

Several in vitro heart perfusion studies confirmed that normally perfused hearts show either no effect or a small reduction in myocardial function when GLP-1 is added to the perfusate. However, in hearts exposed to low-flow ischemia followed by reperfusion, GLP-1 treatment improves LV function and LV pressure development and decreases LV end-diastolic pressure. ^{20 –22} Myocardial glucose uptake is increased. The biochemical effects of GLP-1 were not consistent through the studies. In some the increase in myocardial glucose uptake involved the Akt pathway, and in others it did not. Zhao et al. ²² found that GLP-1 in the ischemic reperfused heart increased myocardial nitric oxide production, activated p38 MAP kinase activity, and stimulated Glut-1 glucose transporter translocation to a greater degree than saline control or insulin at 100 μU/mL.

The in vivo and in vitro animal models indicate that GLP-1 infusion improves many aspects of myocardial function in models of both ischemic and non-ischemic myocardial disease. The mechanism appears to be distinct from the mechanisms involved in the classic GLP-1 activation of cAMP and insulin's activation of Akt phosphorylation and Glut-4 glucose transporter translocation in increasing glucose uptake.

GLP-1 (9–36)-Amide: A Biologically Active Peptide

GLP-1 (9–36)-amide is generated by the action of DPP-4 on GLP-1 (7–36)-amide. ^1,2,5 Most of the total plasma GLP-1 measured in assays is GLP-1 (9–36)-amide. While initially thought to be an inactive metabolite, GLP-1 (9–36)-amide has been shown in many studies to have biologic activity that is not mediated by the classic G-protein-associated GLP-1 receptor. ^{23 –27} This has been verified by showing that GLP-1 (9–36)-amide is active in mice with no GLP-1 receptors ^26,27 or that it exerts its effects in the presence of the GLP-1 receptor antagonist exendin (9–39). ^23,26,27

Elahi et al. ²³ demonstrated in obese subjects that GLP-1 (9–36)-amide reduced hepatic glucose production by 50% during a euglycemic hyperinsulinemic clamp. Simultaneous administration of GLP-1 (9–36) amide plus exendin (9–39)-amide inhibited hepatic glucose production and modestly increased insulin secretion in both lean and obese subjects during a euglycemic hyperinsulinemic clamp. Meier et al. ⁶ showed that GLP-1 (9–36)-amide lowered postprandial plasma glucose levels in normal volunteers and that this effect occurred independently of changes in insulin or glucagon secretion or in the rate of gastric emptying.

A study in conscious dogs with DCM compared the effects of 48-h infusions of 1.5 pmol/kg/min GLP-1 (7–36)-amide and its metabolite GLP-1 (9–36)-amide on LV function and transmyocardial substrate uptake under basal and insulin-stimulated conditions using euglycemic–hyperinsulinemic clamp conditions. ²⁴ GLP-1 and its metabolite reduced LV end-diastolic pressure and increased LV dP/dt and cardiac output equally. ²⁴ Both increased myocardial glucose uptake without a significant increase in plasma insulin.

In an isolated rat heart ischemia/reperfusion preparation infusion of either exendin-4 or GLP-1 (9–36)-amide for the first 15 min of the 120-min reperfusion increased LV performance during the last 60 min of perfusion. ²⁵ These effects were only partially blocked by the GLP-1 receptor antagonist exendin (9–39). Exendin-4 but not the GLP-1 metabolite limited infarction size of the ischemic area (from 33.2% to 14.5%), and this effect was abolished by exendin (9–39). ²⁵ Studies in isolated perfused mouse hearts with ischemia reperfusion injury have shown that both exendin-4 and the GLP-1 metabolite GLP-1 (9–36)-amide improved functional recovery and reduced infarction size. ²⁷ Metabolic studies in cultured neonatal mouse cardiomyocytes showed that both increased levels of cAMP and phosphorylation of extracellular signal-regulated kinase 1/2 and protein kinase B/Akt. ²⁷ The cardioprotective effects of GLP-1 (9–36) were blocked by exendin (9–39) but were preserved in cardiomyocytes without GLP-1 receptors. ²⁷ GLP-1 (9–36)-amide but not exendin-4 improved survival of aortic endothelial cells undergoing ischemia reperfusion injury. ²⁷ This effect of GLP-1 (9–36)-amide was blocked by the nitric oxide synthase inhibitor NG-nitro-l-arginine methyl ester. ²⁷

The available data indicate that there are GLP-1 receptor-dependent and -independent pathways that provide myocardial protection and functional improvement in damaged myocardial cells. GLP-1 (7–36)-amide, its metabolite, and its analogs can activate some aspects of these pathways.

CV Effects of GLP-1 Receptor Agonists

Two GLP-1 receptor agonists, exenatide and liraglutide, are clinically approved for the treatment of type 2 diabetes. Exenatide is a synthetic analog of exendin-4, a natural product found in the salivary secretions of the Gila monster, a reptile indigenous to the Southwest United States. ^1,2,28 This 39-residue peptide shares 53% homology with human GLP-1. It has equal binding to and activation of the human GLP-1 receptor as human GLP-1. ²⁸ It is resistant to DPP-4 cleavage, and 47% of its amino acid structure that differs from that of human GLP-1. ²⁸ Liraglutide is a synthetic derivative of human GLP-1 with a long-chain fatty acid linked to the peptide chain that prevents exposure of the alanine in position 2 so that it is not cleaved by the DPP-4 enzyme. ^1,2,28 Treatment with either of these GLP-1 receptor agonists does not generate a GLP-1 (9–36) amide-like metabolite.

Limited experimental animal data are available concerning their CV effects. Rats 2 weeks after coronary artery ligation and myocardial infarction were treated for 11 weeks with GLP-1, exendin-4, or vehicle. ²¹ Compared with controls, animals treated with either GLP-1 or exenatide had improved cardiac function (LVEF, LV end-diastolic pressure), normalized fasting plasma insulin and glucose, reduced body fat and fluid mass, and increased survival. ²¹ Exendin-4 in an in vitro rat heart ischemia/reperfusion model reduced the area of infarct in the ischemic region from 33% to 14% as well as augmenting LV performance. ²⁵ An in vivo study of myocardial ischemic injury followed by reperfusion in an in vivo pig study has provided the most dramatic demonstration of the cardioprotective effect of exenatide in ischemic cardiovascular disease. Dall and Landrace pigs had their left circumflex coronary artery occluded for 75 min and then reperfused for 3 days. ²⁹ Animals were given exenatide 10 μg subcutaneously and 10 μg intravenously 5 min before the onset of reperfusion and 10 μg twice daily subcutaneously on the following 2 days. The control animals received saline injections. Exenatide reduced myocardial infarction size (32.7±6.4%) compared with the control treatment (53.6±3.9%) and prevented deterioration of systolic and diastolic cardiac function and myocardial stiffness. ^29,30 Biochemical studies of the myocardial tissues showed that exenatide reduced oxidative stress and apoptosis. ²⁹ A recently published study ³¹ comparing an acute 6-h infusion of exenatide (0.12 pmol/kg/min) with saline in a randomized blinded crossover study in 20 male patients with type 2 diabetes and congestive heart failure showed that extenatide increases cardiac index (from 20% to 30%) and decreases pulmonary capillary wedge pressure (from 8% to 15%). The short exenatide infusion increased heart rate by 20%.

Studies with liraglutide showed improved outcomes and survival after experimental infarction in mice. ³² In an in vivo study of a 40-min occlusion of the left anterior descending coronary artery followed by a 2.5-h reperfusion in Landrace pigs, administration of liraglutide (10 μg/kg subcutaneously once daily) for 3 days prior to and during the ischemia reperfusion had no beneficial effect on infarction size, nor did it improve hemodynamic parameters compared with the control treated animals. ³³

Effects of GLP-1 Receptor Agonist Treatment of Patients with Diabetes on CV Risk Factors and Clinical CV Events

A large clinical base is available to evaluate the effects of exenatide and liraglutide on CV risk factors and CV safety. Both exenatide and liraglutide cause a statistically significant decrease in systemic blood pressure that appears to be more than the decrease that could be explained by their effects in reducing weight. An analysis of pooled data from six clinical trials examined the effect of 6 months or greater treatment of exenatide twice daily compared with placebo or insulin in 2,171 patients. ³⁴ Six months of exenatide treatment was associated with a greater reduction in systolic blood pressure than treatment with placebo (−2.8±0.75 mm Hg, P=0.0002) or insulin (−3.7±0.85 mm Hg, P=0.0001). No difference was noted in diastolic blood pressure among the groups. The group differences in systolic blood pressure were greater depending on the magnitude of the baseline systolic blood pressure. A weak correlation existed between the amount of weight loss and the reduction in systolic blood pressure in the exenatide-treated patients (r=0.09, P=0.002). In the LEAD 1–5 studies, liraglutide consistently showed a reduction in systolic blood pressure of 2.5–5 mm Hg versus the comparator treatment. ^4,35 The extent of blood pressure reduction in the LEAD clinical trials did not appear to be explained by the degree of weight loss.

Exenatide and liraglutide treatment both result in modest weight loss (2.5–5.0%) ^{35 –39} and a reduction in body fat (11%) and waist circumference (5%). ^{39 –41} In limited studies, exenatide treatment reduced fasting plasma triglyceride 12% and low-density lipoprotein-cholesterol 6% and increased high-density lipoprotein cholesterol 24%. ^38,42,43 Serum C-reactive peptide decreased 61%, whereas interleukin-6, resistin, endothelin-1, and monocyte chemoattract protein-1 were unchanged. ^38,42,43 Adiponectin increased 14%. Liraglutide decreased fasting plasma triglycerides 22%, plasminogen activator inhibitor-1 25%, and brain naturetic peptide 38% and had no effect on serum adiponectin, leptin, interleukin-6, or tumor necrosis factor-α. ^35,41,44

The improvements in weight and CV risk profile might be expected to predict a decrease in long-term CV events. Such data are being pursued in ongoing long-term clinical CV end-point trials. Analyses of the available clinical trial data for both exenatide and liraglutide indicate that there is no evidence to indicate any CV safety issues with either. A retrospective analysis of pooled data from 12 clinical trials (12–52 weeks in duration) of exenatide twice daily versus either placebo or insulin treatment of patients with type 2 diabetes examined the reported adverse events for major CV events. ⁴⁵ The events included in the analysis were stroke, myocardial infarction, cardiac mortality, and acute coronary syndrome and revascularization procedures. The population consisted of 2,316 exenatide-treated patients and 1,629 randomized placebo- or comparator-treated patients. The primary major CV events risk rate for exenatide was 0.7 with 95% confidence limits of 0.38 to 1.31. A retrospective analysis of the Life Link database of medical and pharmaceutical insurance claims for June 2005 through March 2009 analyzed incident CV events in 39,275 patients with type 2 diabetes treated with exenatide twice daily and 381,218 patients treated with other glucose-lowering therapies. ⁴⁶ Exenatide patients were less likely to have a CV event than the non–exenatide-treated patients (hazard ratio 0.81; 95% confidence levels 0.68–0.95) and had lower rates of CVD-related hospitalizations (0.88; 95% confidence interval 0.79–0.98). A recent meta-analysis of randomized clinical trials of GLP-1 receptor antagonists and CV events calculated the Mantel–Haenszel odds ratio with 95% confidence levels for major CV events for all GLP-1 receptor agonists to be 0.74 (0.50–1.08), for exenatide 0.85 (0.50–1.45), and for liraglutide 0.69 (0.40–1.22). ⁴⁷

A recently approved slow-release formulation of exenatide has been approved by the U.S. Food and Drug Administration. This preparation is administered once weekly, has a greater effect in improving glycemic control than exenatide twice daily, and appears to be equally as effective as exenatide twice daily in reducing CV risk factors. ⁴⁸

Summary

Human and animal data show that infusion of GLP-1 or its metabolite GLP-1 (9–36)-amide improves CV outcomes in both ischemic and non-ischemic myocardial disease. The extent to which these improved outcomes are direct effects of GLP-1 receptor activation or GLP-1 receptor-independent activation is not clear. Data concerning the effects of the GLP-1 receptor analogs exenatide and liraglutide are limited to animal data, where exenatide has shown reduction in ischemia/reperfusion injury whereas the data with liraglutide are equivocal. Because the structures of exenatide and liraglutide are quite different and neither is cleaved by DPP-4, it has yet to be determined whether one or the other will improve outcomes in patients with CVD. A separate question to be addressed by ongoing current long-term clinical trials is whether the improvement in CV risk factors reported in patients with type 2 diabetes treated with GLP-1 receptor agonists will lead to a reduction in new cardiovascular events.