Blood flow effects of activating two related hormone receptors in healthy people

This was a single‐centre, randomized, placebo‐controlled, exploratory, single‐blinded, physiological, mechanistic study. The protocol received approval from the East Midlands Nottingham 1 Research Ethics Committee (18/EM/0417). The study was registered on Clinicaltrials.gov↗ (NCT03835013↗) and performed according to the principles of the Declaration of Helsinki. Recruitment commenced in February 2019 and completed in September 2021. Recruitment was interrupted in 2020 due to the impact of the COVID‐19 pandemic. Part A was a single‐blinded, dose‐comparison study to assess the tolerability of a 2‐h glucagon infusion in an escalating dose. Part B was a single‐blinded, experimental study to assess the safety, tolerability and non‐invasive haemodynamic effects of dual agonism at the GLP‐1 and glucagon receptor. To simulate the acute haemodynamic effects of GLP‐1 and glucagon receptor dual agonism, exenatide (a licensed GLP‐1 receptor agonist which shares 53% amino acid homology with native GLP‐116) and glucagon were intravenously infused. At the time of protocol development, dual GLP‐1:glucagon receptor agonists were in early‐stage clinical development and therefore the decision was made to use IV agents. Importantly, this has the advantage of being able to study drug effects with 100% bioavailability.

Using a Latin square block design, participants acted as their own control and received a different infusion at each study visit. Two IV cannulas were sited in the upper limbs on the opposite sides, one for the infusion and the other reserved for blood sampling. Two pairs of infusions consisting of combinations of 0.9% normal saline (Macopharma), glucagon (Novo Nordisk) at concentrations of 25 ng/kg/min (low dose) or 50 ng/kg/min (high dose) and exenatide (50 ng/min loading dose for 30 min then 25 ng/min) were simultaneously infused through a single cannula, via a double lumen extension set, for 120 min (Part A) or 60 min (Part B) at either 0.5 or 1 mL/min (Alaris GH syringe pump, Becton Dickinson) as per the experimental protocol in Figure 1. A number of adverse events, specifically nausea and vomiting, were reported in Part A and accordingly the decision was taken to only infuse the 25 ng/kg/min glucagon dose in Part B to reduce the risk of adverse events and better simulate dual agonist compounds. A low‐sorb infusion line (Alaris, P7000 extension set, Becton Dickinson) was used for all IV infusions. Dose selection was based on previous similar studies.17, 18, 19, 20

The study enrolled non‐smoking, adult male participants between the ages of 18 and 40 years with a body mass index (BMI) of 18‐30 kg/m². Participants attended an initial screening visit where written informed consent was obtained. Participants were determined to be healthy by means of a medical history, physical examination including anthropometric measurements, vital signs, 12‐lead electrocardiogram (ECG) and transthoracic echocardiogram. Those with hypertension, clinically significant heart disease, renal or liver impairment, or diabetes were excluded.

Participants were studied in the resting supine position following an 8‐h overnight fast. They were asked to refrain from alcohol and caffeine consumption for 12 h and to avoid non‐steroidal anti‐inflammatory drugs for 24 h before study visits. Strenuous exercise was avoided for 12 h prior to infusion visits. Visits were carried out in a quiet, temperature‐controlled room. Haemodynamic measurements were taken every 15 min and blood samples every 30 min (Accu‐Check glucose monitor, Roche) (Figure 1).

Cardiac output, stroke volume and heart rate were measured in duplicate over an average of 60‐s time intervals using the Cheetah Starling Stroke Volume Haemodynamic Monitoring System (Cheetah Medical, a validated, non‐invasive monitoring technique based on thoracic bioreactance.21 Peripheral vascular resistance was calculated as MAP divided by cardiac output, and then multiplied by 80 to convert arbitrary units to dynes/s/cm⁵ as per previously published research.22 Brachial systolic BP, brachial diastolic BP, central systolic BP, central diastolic BP, MAP, heart rate and augmentation index were measured using SphygmoCor XCEL (AtCor Medical). The rate pressure product, an indirect measure of cardiac work, was calculated by multiplying systolic BP by heart rate. Heart rate variability was measured using the SphygmoCor Device (AtCor Medical). Using a three‐lead ECG, a continuous, short‐term recording of heart rate was made over 5 min to provide an indication of autonomic nervous system activity. Fast Fourier transform of the ECG signal was used to derive the low frequency to high frequency power ratio (normalized for total power).23

Blood samples in EDTA tubes were collected for analysis of total GLP‐1, total active GLP‐1, glucagon, glucose‐dependent insulinotropic polypeptide (GIP) and peptide YY. Heparin lithium tubes analysed insulin, C‐peptide, free fatty acids and triglycerides. All venous blood was collected in chilled 5‐mL syringes containing 20 μL of dipeptidyl peptidase‐4 inhibitor (DPP4i) (Merck). Blood was then transferred into a chilled 2.6‐mL EDTA tube containing 100 μL of aprotinin 10 000 Kallikrein Inactivator Units/mL (Trasylol, Nordic Pharma). All syringes and blood tubes were stored in wet ice prior to sampling. Aprotinin is an enzyme inhibitor that prevents glucagon degradation and DPP4i inhibits degradation of GLP‐1. Samples were collected in chilled lithium heparin tubes. Blood tubes were placed immediately on wet ice, centrifuged at 4 °C, 3005 xg for 10 min, snap frozen on dry ice and stored at less than −40 °C until analysis.

Data were analysed using SPSS (version 28, IBM). No formal power calculations were used as this was an exploratory study with a sample size based on similar experimental pilot studies.18, 19, 20 Cardiovascular data were analysed using a two‐way repeated measures analysis of variance (ANOVA) with post hoc testing (paired t‐test). Metabolic data were analysed using a two‐way repeated measures ANOVA with post hoc testing (paired t‐test with Bonferroni correction). Area under the curve (AUC) data were analysed using a one‐way ANOVA and post hoc testing (paired t‐test). Heart rate variability data (non‐parametric) were plotted using a forest plot and analysed using the Wilcoxon signed rank test. Data are expressed as mean ± standard error of the mean (SEM), or percentages. Line and bar graphs represent the mean and error bars represent SEMs. A P value <.05 was deemed significant for all statistical analyses.

Ten adverse events were recorded in Part A and 10 in Part B. In Part A, hypoglycaemia (defined as a blood glucose <4.0 mmol/L) occurred six times, all in the glucagon infusion arm. Four of these hypoglycaemic episodes were symptomatic (three with nausea and one with vomiting). Asymptomatic hypoglycaemia occurred 10 times in Part B (twice in the glucagon infusion arm and eight times in the exenatide:glucagon co‐infusion arm). Hypoglycaemia was treated with IV dextrose until blood glucose was >4.0 mmol/L. There were no serious adverse events.

Our data align with previous studies showing that IV exenatide increases heart rate through a number of established mechanisms, including direct activation of pacemaker cells of the sino‐atrial node.24 Activation of the autonomic system is reported to entail chronotropic effects,25 but data from large animal models24 and humans3 fail to support GLP‐induced activation of the sympathetic nervous system. In agreement, GLP‐1‐related changes in sympathetic activity were not observed in this study based on heart rate variability. The small increase in heart rate (~4 bpm) had no significant effect on cardiac output. It is possible that compensatory physiological changes in stroke volume and peripheral vascular resistance were too small to detect in this pilot study. Similar haemodynamic patterns in heart rate (significant increase), stroke volume (no significant change), cardiac output (no significant change) and peripheral vascular resistance (no significant change) were observed in the other infusion arms. The rate pressure product significantly increased in all three infusion arms and this is expanded on below. Intravenous exenatide did cause a small, but significant, 3 mmHg increase in MAP in our infusion study, which broadly aligns with other studies of IV exenatide (Supporting Information Table S1). By contrast, other GLP‐1 receptor agonists have been reported to reduce systolic BP by ~0.7‐4.6 mmHg.26 This may suggest that exenatide has differential effects on the cardiovascular system compared with other GLP‐1 receptor agonists, although we cannot determine if the chronic effects on BP may be different to our findings in this short‐term study. Mechanisms relating to the contrasting effects of acute infusions of GLP‐1 receptor agonists (either native GLP‐1 or exenatide) versus the long‐term effects of licensed GLP‐1 receptor agonists on BP are not fully understood. Blood pressure reductions occur early (after 2 weeks) with prolonged use of GLP‐1 receptor agonists, suggesting that weight loss alone is not the only driver for BP reductions. Additional mechanisms which decrease BP include arterial and smooth muscle activation of GLP‐1 receptors, improvements in vascular endothelial function, nitric oxide induced vasodilation and inhibition of the renin‐angiotensin‐aldosterone system (natriuresis).27 The timescale and the relative contribution of these mechanisms to overall BP reduction is somewhat unclear, although likely accounts for at least some of the observed differences in BP between acute infusions and chronic treatment with GLP‐1 receptor agonists.

Glucagon demonstrated an increase in heart rate that is consistent with previous data.17, 18, 28 No change in stroke volume, cardiac output or MAP was observed and this aligns with other low‐dose infusion studies.29 Mechanistically, it is unclear if these chronotropic effects result from any direct effects on the myocardium given that the glucagon receptor (and mRNA transcripts) is absent from the sino‐atrial node, atria and ventricles of multiorgan donors and explanted hearts.30 One mechanism for an increase in heart rate is via glucagon‐induced activation of myocardial GLP‐1 receptors.28 Glucagon and native GLP‐1 share 47% amino acid homology plus overlapping binding sites, enabling glucagon to bind to the GLP‐1 receptor with low affinity.31, 32 Glucagon‐induced activation of the sympathetic nervous system has been shown to potentiate the chronotropic effects by stimulating the hypophysis‐hypothalamus‐adrenal axis33 and through increased catecholamine release.34 However, based on heart rate variability, we found no evidence of increased sympathetic activity in this study and it is possible these effects are only seen with higher doses of glucagon.

Exenatide:glucagon co‐infusion significantly increased heart rate as well as the rate pressure product, an indirect measure of cardiac work and myocardial oxygen consumption. Rate pressure product is an independent predictor of both in‐hospital cardiac mortality and long‐term all‐cause mortality in patients with acute coronary syndrome. The increase in the rate pressure product in our study was driven by an increase in heart rate rather than systolic BP. Increases in heart rate due to native GLP‐1 and glucagon co‐infusion18 (Supporting Information Table S2) and dual GLP‐1:glucagon receptor agonists are widely published but less is known about their short‐term effects on the rate pressure product. This increase in the rate pressure product seen with co‐infusion may be consistent with long‐acting GLP‐1:glucagon dual receptor agonists in the initial period after starting therapy. Reassuringly, any possible short‐term increase in the rate pressure product is shown to be ameliorated over time due to the beneficial effects on BP.35 Previous concerns were raised in this field over whether the increase in heart rate may negatively affect those with diastolic dysfunction, cardiomyopathy and heart failure. Clinical trial data examining the effect of GLP‐1:glucagon dual receptor agonists on heart failure related outcomes are awaited.

Haemodynamic data on GLP‐1:glucagon dual receptor agonists in clinical development are available for several compounds, including cotadutide, survodutide and pemvidutide. Cotadutide (MEDI0382), a GLP:glucagon receptor dual agonist with a ratio of approximately 5:1 GLP‐1 to glucagon activity, increases heart rate by 5 bpm (P < .001) in patients with type 2 diabetes and chronic kidney disease.36 Similar findings are reported in patients with type 2 diabetes and obesity (mean heart rate +12 bpm, systolic BP −10 mmHg).37 Survodutide (BI 456906) increases heart rate by 1‐6 bpm in patients with type 2 diabetes and is currently being tested in a phase 3 cardiovascular outcome trial in adults with obesity and increased cardiovascular risk.38, 39 Pemvidutide (ALT‐801) reduces systolic and diastolic BP without leading to any clinically significant increase in heart rate.35

From a metabolic perspective, the metabolic changes in all infusion arms align with previous published studies confirming GLP‐1 and glucagon receptor engagement. It was not possible to measure exenatide concentrations but note is made of supraphysiological glucagon levels in Part A (mean at 120 min 1945 ± 328 pg/mL) and Part B (mean at 30 min 899 ± 59 pg/mL). It is evident that glucagon offsets the glucose lowering efficacy of exenatide in this study. This highlights the importance of the respective ratios of GLP‐1:glucagon dual agonist compounds for future treatments in metabolic populations to avoid any deleterious effects.

This was a small exploratory pilot study meaning it is not possible to draw firm conclusions. Only healthy adult males under 40 years old were eligible to participate. Females, older adults and those with chronic conditions, including type 2 diabetes, were excluded, meaning the current findings may not necessarily apply in other clinical settings. Stroke volume and cardiac output were measured using non‐invasive devices. Although these devices are validated, catheter‐based testing methods (ie thermodilution method, Fick's method) remain the gold standard. Adverse events, in particular nausea and vomiting, may have affected haemodynamic parameters. Observed haemodynamic effects may have also been affected by other hormones, including insulin. The infusion duration was short and therefore it is not possible to extrapolate the cardiovascular and metabolic outcomes over a longer duration, and even less so in the case of a purported future chronic oral treatment. GLP‐1 and glucagon receptor occupancy and activation differ compared to molecules in clinical development, meaning caution should be applied when making direct comparisons. Finally, although participants were blinded to the infusion, the study team was unblinded.

J.G., V.P., C.M.M., L.J., J.C., P.A. and I.B.W. conceived and designed the study. I.B.W. acted as the chief investigator for the study. J.G., G.D.S., A.H., E.V., J.H. and F.K. performed all study visits. N.J. performed all echocardiograms. P.B. analysed bloods. J.G., V.P., C.M.M., L.J., J.C., P.A. and I.B.W. analysed and interpreted the data. J.G. wrote the first draft of the manuscript. All authors critically revised the manuscript and approved the final version. The authors confirm that the Principal Investigator for this paper is I.B.W. and that he had direct clinical responsibility for patients.

V.P., L.J. and P.A. are employees and shareholders of AstraZeneca.

The data that support the findings of this study are available from the corresponding author upon reasonable request.