Dual glucagon‐like peptide‐1 and glucagon receptor agonism reduces energy intake in type 2 diabetes with obesity

To establish which components of energy balance mediate the clinically significant weight loss demonstrated with use of cotadutide, a glucagon‐like peptide‐1 (GLP‐1)/glucagon receptor dual agonist, in early‐phase studies.


Conclusion:
Weight loss with cotadutide is primarily driven by reduced EI, with relatively small compensatory changes in EE.

K E Y W O R D S
clinical trial, drug development, energy regulation, incretin physiology, weight control

| INTRODUCTION
3][4] MASH and T2D can further contribute to the progression of kidney and cardiovascular disease. 5ight loss has been associated with improvements in MASH 6 and glycaemic control 7 ; therefore, weight loss is considered a key treatment target for patients with MASH and/or T2D. 8,9Glucagon-like peptide-1 (GLP-1) receptor agonism has been shown to provide effective glycaemic control in T2D.However, weight loss with GLP-1 receptor agonism is relatively modest at the doses at which maximum glucose reduction is achieved, ranging from 1.5 to 6.0 kg in clinical trials. 10,11This highlights the need for T2D treatments to complement glucose control with a greater reduction in body weight.A major challenge for achieving weight loss by reduced energy intake (EI; whether through diet or pharmacotherapy) is the physiological compensatory reduction in energy expenditure (EE) that tends to counteract reduced EI. 12,13Therapies that can simultaneously reduce appetite while preventing or limiting a reduction in EE might therefore allow greater weight loss and metabolic benefits.
One such approach is to combine the beneficial effects of GLP-1 receptor agonism with other therapies (glucose-dependent insulinotropic polypeptide [GIP] receptor agonism or antagonism, or glucagon receptor agonism).6][17] However, glucagon infusion longer than 72 h was not associated with increased EE. 18 In these studies, glucagon was used in low doses because high doses can induce nausea and vomiting. 19veral GLP-1/glucagon receptor agonists are in development, including BI 456906, LY3305677, LY4347943, JNJ-54729518, 20,21 HM15211, NNC9204-1706, Alt-801, and G3215. 22Two further agents have been discontinued owing to excessive gastrointestinal adverse events (AEs) during phase 2 studies (SAR425899) 22 and cardiovascular safety issues (NN117). 23Although the magnitude of glucose lowering and weight loss varies across these agents, the optimal activity ratio of the two agonists from the perspective of health outcomes is yet to be determined.
Cotadutide is a synthetic linear peptide, which functions as a dual GLP-1/glucagon receptor agonist 24 and is under development for treating MASH and chronic kidney disease with T2D. 25,26[28][29][30][31] The beneficial renal effects of glucagon include natriuresis, increased renal vasodilation and an increase in estimated glomerular filtration rate, 32 while favourable hepatic effects include increased lipid oxidation, reduced lipid synthesis, and increased bile acid production. 33With a 5:1 ratio of GLP-1 and glucagon signalling, 26 cotadutide has been optimized so that the antiglycaemic properties of GLP-1 balance the glucose-elevating properties of glucagon action, such that the overall effect of cotadutide results in a glucose decrease while retaining appropriate glucagon signalling.
In rodent models, cotadutide promoted weight loss through both increased EE and reduced EI. 34Based on comparisons with liraglutide in the same study, increased EE might account for up to 25% of body weight loss while the remainder was attributable to reduced EI. 34Clinical studies of cotadutide in obese or overweight individuals with T2D demonstrated decreased body weight and improvements in glycaemia, 5,25,35 with the latter being mediated by increased insulin secretion and delayed gastric emptying. 26However, the interplay among EE, EI and weight loss induced by cotadutide has not been studied in humans.
Here, we report a randomized, placebo-controlled, double-blind, phase 2a study performed in adults with T2D and obesity, which included a detailed examination of the mechanisms of cotadutide action.We show that cotadutide-associated weight loss is driven by markedly decreased EI, achieved without any major compensatory change in EE.We show that glycaemic effects are likely mediated by increased insulin secretion and reduced gastric emptying.

| Study design and approval
This randomized, placebo-controlled, double-blind, phase 2a trial was conducted in a single centre (Cambridge, UK).The study design and patient disposition are detailed in Figures S1 and 1, respectively.
The study was approved by the East of England-Essex Research Ethics Committee (reference 18/EE/0204) and was performed according to the International Council for Harmonization Guidance for Good Clinical Practice, the ethical principles outlined in the Declaration of Helsinki, and applicable regulatory requirements.The study was registered at ClinicalTrials.gov(NCT03596177).All patients provided written informed consent before enrolment.

| Participants
Participants were recruited via a radio advertisement and through local primary care practices between October 2018 and December 2019.The full list of eligibility criteria is shown in S1.

| Randomization and follow-up
Eligible participants underwent 16-day single-blind treatment with placebo to exclude non-compliers or placebo-responders, thus increasing study power. 36Subsequently, participants were randomly assigned 2:1 to receive subcutaneous cotadutide or placebo for 42 days with a fixed-dose escalation of 100 μg daily for the first 4 days, 200 μg daily for the following 4 days and 300 μg daily for the remaining 34 days.Cotadutide and placebo were identically labelled and indistinguishable in appearance.Therefore, the pharmacists, participants, and study site personnel were masked to treatment allocation.Randomization was performed using an interactive web response system.All participants entered a 28-day follow-up period of safety monitoring, with the final visit occurring 28 days post last dose.

| Statistical analysis
A sample size of 16 participants in the cotadutide group and eight participants in the placebo group was planned to provide >90% power to detect a 2.5% weight change from baseline difference between treatment groups at Day 59, assuming SD = 2.0% and two-sided α = 0.1.
Descriptive statistics were used to summarize data at specified time points.The primary efficacy outcome of percentage change in body weight between Days 17 and 59 was compared between cotadutide and placebo groups using analysis of covariance (ANCOVA), adjusting for treatment group and baseline body weight.The analyses for EE variables was planned to be conducted using kJ/kg of fat mass.
Additional sensitivity analyses were performed using kJ/kg of lean body mass and kJ/kg of total body mass.The aim of these analyses was to compare mean EE between treatment groups while controlling for the effect of covariates.
A modified intention-to-treat population was used for the analysis of primary and secondary efficacy outcomes, and ANCOVA models adjusting for treatment group and baseline value of the outcome of interest were fitted.The least squares (LS) mean and the corresponding 90% confidence interval (CI) are presented for the ANCOVA models.Selected secondary and exploratory endpoints were analysed using the per-protocol population.Safety outcomes were analysed in the as-treated population.The last observation carried forward method was used to impute post-baseline missing values when applicable.
Statistical analyses were performed using SAS ® , version 9.

| Study population
Between October 2018 and December 2019, 44 participants were screened for eligibility, of whom 28 were randomly assigned to receive cotadutide (n = 19) or placebo (n = 9).All 28 participants received placebo in the single-blind run-in period and 25 progressed to receive double-blind treatment (Figure 1) with cotadutide (n = 18) or placebo (n = 7).Of n = 28 enrolled participants, 19 (68%) completed the study.Of the nine participants who discontinued, two withdrew and seven were discontinued for other reasons (Figure 1).Demographic and baseline characteristics were balanced between the treatment groups (Table 1).Participants were on average 60 years old and included predominantly White men, with obesity and well-controlled T2D.Two participants in the cotadutide arm were overweight (BMI ≥25 kg/m 2 and <30 kg/m 2 ) while the remaining participants in both groups were obese (BMI ≥30 kg/m 2 ).

| Body weight and composition
Cotadutide therapy resulted in 4% weight loss over 42 days of treatment, with significantly greater weight loss than placebo T A B L E 1 Participant demographics and baseline characteristics, as-treated population.

| Self-reported hunger
There was no significant difference between treatment groups in any of the four items used to assess hunger measured by visual analogue scale (Table S1).However, the sensation of feeling sick became numerically greater in the cotadutide group versus the placebo group with no statistical significance, while a reduction in the desire to eat was observed in both groups (Table S1).
T A B L E 2 Body composition assessed by dual energy absorptiometry, modified intention-to-treat population.Note: p value is for the difference in LS mean versus placebo from analysis of covariance model adjusted for treatment arm and baseline value of the outcome.In the cotadutide group, complete data were available for N = 12 participants on Day 59.In the placebo group, the number of participants with complete data stayed at N = 7 throughout the study.In the total body weight analysis, Day 17 was considered baseline, while for the body composition analysis Day À1 (start of placebo run-in period) was considered baseline.Abbreviations: CI, confidence interval; LS, least squares; SD, standard deviation.

| Energy intake and expenditure
We then examined in more detail how changes in EI and EE might underpin the cotadutide-induced changes in body weight and composition.The placebo group had an approximately 600-kJ higher EI at baseline, which is a small difference in absolute terms.S3, Table S3).

| Glucose homeostasis
Fasting plasma glucose decreased in the cotadutide group (change from baseline -2.S4]).C-peptide change followed the same pattern as that observed for insulin, with increased responses during the MMTT with cotadutide (Table S4).
During the IVGTT, there was a decrease in AUC glucose accompanied by marked increases in AUC insulin and AUC C-peptide with cotadutide (Figure S4a-c not in the placebo group (113 [22, 323] to 122 [19, 216]; p = 0.003 T A B L E 3 Energy intake from ad libitum lunch, modified intention-to-treat population.Note: p value is for the difference in LS mean versus placebo from analysis of covariance model adjusted for treatment arm and baseline value of the outcome.In the cotadutide group, complete data were available for N = 14 participants on Day 32 and for N = 14 participants on Day 59.In the placebo group, the number of participants with complete data stayed at N = 7 throughout the study.In this analysis, Day 16 was considered the baseline.Abbreviations: CI, confidence interval; LS, least squares; SD, standard deviation.
for the difference in change from baseline for cotadutide vs. placebo [-

| Gastric, small bowel and colonic emptying time
There was a significant increase in gastric emptying time following treatment with cotadutide, with a change from baseline of 6.7 h (90% CI 5.0, 8.4) versus 2.6 h (90% CI 0.4, 4.7) in the placebo group ( p = 0.019).There were no changes in small bowel and colonic emptying times (Table S5).

| Muscle biopsy
Using RNA sequencing we found eight genes that were differentially expressed in response to cotadutide treatment (Figure S5, Table S6).
Among them, AC109583.4(ENSG00000284672) was a novel uncharacterized long non-coding RNA.Targeted analysis indicated potential improvement in indices of insulin sensitivity (Figure S6; enrichment p = 7*10 À12 ).Suggestive results were also obtained for lipid oxidation including increased expression of Peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC1alpha) in three out of four participants (Figure S7; enrichment p = 0.005).No dysregulation was observed for the other selected marker sets.
3.9 | Safety, pharmacokinetics and immunogenicity Most AEs were mild or moderate.Two participants in the cotadutide group (11.1%) had at least one event of Grade ≥3 severity.
Four participants receiving cotadutide (22.2%) had five AEs leading to permanent discontinuation of cotadutide (nausea in two participants and eructation, injection-site erythema, and a markedly decreased appetite in one participant each), and all were considered related to cotadutide.There was a low number of injection-site reactions (two participants in the cotadutide group [11.1%] reported three injection-site reactions), and an increase from baseline in heart rate was detected in both groups, which was in the order of 9 bpm.A trend toward a reduction in blood pressure was observed in the cotadutide group, and there were no clinically meaningful changes in laboratory values or ECGs.

| DISCUSSION
Our data show that dual GLP-1/glucagon receptor agonism with cotadutide leads to weight loss, primarily via a marked reduction in EI.We also provide evidence that the glucose-lowering actions of cotadutide are mediated by increased insulin secretion and prolonged gastric emptying.The metabolic benefits of GLP-1 receptor agonists are well established.In humans, weight loss with GLP-1 receptor agonist therapy is predominantly driven by reduced EI, ranging from 15% to 20% [37][38][39] with associated weight loss of 1.5-6 kg over 30 weeks. 10creased EE is seen with GLP-1 receptor agonists in rodent models and attributed to brown adipose tissue (BAT) activation, 40 but this is unlikely to affect weight loss in humans. 38,39Glucagon signalling plays a more dominant role in energy homeostasis through a variety of mechanisms, including upregulation of energy-expensive catabolic processes, especially in the liver, and activation of BAT via fibroblast growth factor 21 in rodent models. 41[43] Clinical trials with oxyntomodulin 44,45 showed 20%-35% reductions in EI with single 46 or repeated doses. 47In this study, we found a reduction in EI with cotadutide of >50% from baseline.Although we did not compare cotadutide therapy with GLP-1 receptor agonist treatment directly, the reduction in EI was markedly greater than that previously reported with GLP-1 receptor agonist treatment alone, 48 suggesting that glucagon receptor agonism may act synergistically with GLP-1 signalling to reduce appetite.A recent phase 1b trial in healthy adults demonstrated that treatment with SAR425899 (GLP-1/ glucagon co-agonist) over 19 days led to reduced selective metabolic adaptation (smaller decrease in sleeping metabolic rate than expected for weight loss) but there was no change in total EE measured by indirect calorimetry. 49 performed detailed measures of EE.In preclinical studies with cotadutide, increases in EE were attributed to the preferential hepatic fat oxidation mediated via PGC1alpha and enhanced mitochondrial maintenance and function, with these effects being differentiated from a mono-GLP-1 receptor agonist. 24A study of oxyntomodulin in overweight and obese humans demonstrated that weight loss was achieved through both reduced EI and increased AEE, but not REE. 50In a double-blind, randomized, crossover trial using infusions of GLP-1, glucagon, GLP-1 and glucagon co-infusion, and placebo in obese adults without diabetes, Tan et al. 16 demonstrated a significant increase in REE after glucagon and after GLP-1 and glucagon co-infusion, but not after GLP-1 alone.In contrast, Bagger et al. 51 assessed both oxyntomodulin and separate and combined infusions of GLP-1 and glucagon in young healthy adults, reporting reduced EI for all compounds tested, but no significant changes in REE.A further study in humans infusing glucagon demonstrated increased REE, which was independent of BAT activation.
These studies primarily evaluated acute glucagon infusions, while the effects of more sustained glucagon receptor agonism are less well characterized.Indeed, a study of a single patient with a glucagonoma observed minimal changes in EE. 52 Furthermore, it has been demonstrated that insulin infusion offsets the effect of glucagon on REE, which may explain the inability to detect significant differences with combined GLP-1/glucagon receptor agonism. 28Changes in EE in this study were in keeping with the anticipated compensatory adaptations to weight loss, except for a small numerical increase in REE observed in early dosing. 13,53,54However, the observation of increased expression of markers of lipid oxidation (e.g., PGC1alpha) in skeletal muscle is intriguing, although the interpretation is limited by the very small sample size in this substudy.Despite the reduction in physical activity, there was no significant difference between treatment groups in TEE measured using the doubly labelled water technique over 6 weeks.
We found a reduction in physical activity and daily step count in the cotadutide group, while AEE during exercise was unchanged and TEE (kJ/kg lean mass) from indirect calorimetry decreased significantly in the cotadutide group versus placebo.This suggests that weight loss was not explained by greater physical activity in the cotadutide group.
It may be explained by the commonly reported AEs in the cotadutide group, which included lethargy, nausea and loss of appetite; however, reduced physical activity has been observed in other studies 55 of glucose-lowering therapies that do not share the same AE profile (sodium-glucose cotransporter-2 inhibitors and dipeptidyl peptidase-4 inhibitors).A structured exercise programme could be considered in future studies with cotadutide to achieve greater weight loss and more metabolic benefits.
We observed lower fasting plasma glucose, reduced glucose and increased insulin and C-peptide excursions following a liquid meal in the cotadutide group.During the IVGTT, we found a marked increase in FPIR in the cotadutide group.Therefore, the glucose-lowering effects and improvements in postprandial glucose excursions associated with cotadutide are mediated by enhanced insulin secretion, rather than a change in responsiveness to insulin.Also likely contributing to improved glycaemia, we found delayed gastric emptying consistent with previous studies with cotadutide 26 and short-acting GLP-1 receptor agonists. 56key strength of our study was the comprehensive metabolic profiling of participants including 24-h indirect calorimetry and doubly The study was conducted in the National Institute for Health and Care Research (NIHR)/Wellcome Clinical Research Facility within the Cambridge Biomedical Campus.Overweight and obese (body mass index [BMI] >28 and ≤40 kg/m 2 ) adults aged 30-75 years with wellcontrolled T2D (glycated haemoglobin [HbA1c] level ≤ 64 mmol/mol [8.0%]) were eligible.Participants receiving metformin monotherapy were eligible if there was no significant change (increase or decrease >50%) in the dose 3 months before screening.Individuals treated with adjuncts to metformin (dipeptidyl peptidase-4 inhibitors, sulphonylureas, glinides, or sodium-glucose cotransporter-2 inhibitors) were eligible after a 4-week washout period.The key exclusion criteria were use of GLP-1 receptor agonists, insulin, or appetite-suppressing medications, acutely decompensated glycaemia, and regular high-intensity exercise.
The primary efficacy outcome was percentage change in body weight from the beginning to the end of treatment (Days 17-59).Secondary efficacy outcomes included absolute and percentage change in the following variables: EI (kJ) from ad libitum lunch from Days 16 to 32 and from Days 16 to 59, total energy expenditure (TEE), and activity-related energy expenditure (AEE) from Days 15 to 58, resting energy expenditure (REE) from Days 16 to 32, body fat mass from Days À1 to 59, fasting glucose, and area under the curve (AUC) glucose during mixed meal tolerance test (MMTT) from Days À1 to 59.All EE variables were expressed in kJ/kg of fat mass and additional analyses were performed using lean body mass instead of fat mass in separate models.Exploratory outcomes included: change in glucose metabolism parameters (insulin sensitivity, glucose effectiveness, first-phase insulin response [FPIR], disposition index) derived from an intravenous F I G U R E 1 CONSORT diagram showing participant disposition.*Informed consent was signed.glucose tolerance test (IVGTT; detailed in S2) between Days 1 and 14 and Days 41 or 42, gastric emptying time from Days 15 to 58, mean daily time (min) spent in activity, mean daily step count between Days 1-16 and Days 44-59.EE was measured using indirect calorimetry and the doubly labelled water technique.While the former is based on oxygen consumption and carbon dioxide production in a respiration chamber, the latter measures carbon dioxide production from isotope elimination in the urine.We measured and report both insulin and C-peptide responses during IVGTT and MMTT as the latter are more closely linked to insulin secretion.Procedures are detailed in S3 and S4 (study protocol).
4 or higher.Analyses of IVGTT data were conducted using minimal model and parameter estimation (S2) in SAAM II software (version 2.3) and STATA (version 16).GraphPad Prism (version 9) was used for figure generation.

Table 4
). REE showed the same pattern as TEE, although with a small numerical increase in REE noted after 16 days of therapy in the cotadutide group (+8.5% vs. +1.2%;LSmean difference 7.4% [90% CI À1.7, 16.5]; p = 0.177).There were no differences between placebo and cotadutide for AEE.There was a small reduction in respiratory quotient in the cotadutide group between baseline and the end of treatment (LS mean for % change: À0.73 [90% CI À2.28, 0.82] vs. 2.30 [90% CI 0.26, 4.33], cotadutide vs. placebo; p = 0.056) indicating a possible shift towards fat metabolism.3.5 | Physical activity Energy expenditure from indirect calorimetry, modified intention-to-treat population.Note: p value is for the difference in LS mean versus placebo from analysis of covariance model adjusted for treatment arm and baseline value of the outcome.In the cotadutide group, complete data were available for N = 12 participants on Day 58.In the placebo group, the number of participants with complete data stayed at N = 7 throughout the study.In this analysis, Day 15 was considered the baseline and Day 58 the final time point.N.B.Equilibration was not achieved in four participants in the cotadutide group and four participants in the placebo group.Abbreviations: CI, confidence interval; EE, energy expenditure; LBM, lean body mass; LS, least squares; SD, standard deviation.