Results from three phase 1 trials of NNC9204-1177, a glucagon/GLP-1 receptor co-agonist: Effects on weight loss and safety in adults with overweight or obesity

Abstract

Objective

Glucagon/glucagon-like peptide-1 (GLP-1) receptor co-agonists may provide greater weight loss than agonists targeting the GLP-1 receptor alone. We report results from three phase 1 trials investigating the safety, tolerability, pharmacokinetics and pharmacodynamics of the glucagon/GLP-1 receptor co-agonist NNC9204-1177 (NN1177) for once-weekly subcutaneous use in adults with overweight or obesity.

Methods

Our focus was a 12-week, multiple ascending dose (MAD), placebo-controlled, double-blind trial in which adults (N = 99) received NN1177 (on an escalating dose regimen of 200, 600, 1300, 1900, 2800, 4200 and 6000 μg) or placebo. Two other trials also contributed to the findings reported in this article: a first human dose (FHD)/single ascending dose (SAD), placebo-controlled, double-blind trial in which adults (N = 49) received NN1177 (treatment doses of 10, 40, 120, 350, 700 and 1100 μg) or placebo, and a drug–drug interaction, open-label, single-sequence trial in which adults (N = 45) received a 4200-μg dose of NN1177, following administration of a Cooperstown 5 + 1 index cocktail. Safety, tolerability, pharmacokinetic and pharmacodynamic endpoints were assessed.

Results

For the FHD/SAD and MAD trials, baseline characteristics were generally balanced across treatment cohorts. The geometric mean half-life of NN1177 at steady state was estimated at between 77 and 111 h, and clinically relevant weight loss was achieved (up to 12.6% at week 12; 4200 μg in the MAD trial). Although NN1177 appeared tolerable across trials, several unexpected treatment-related safety signals were observed; increased heart rate, decreased reticulocyte count, increased markers of inflammation (fibrinogen and C-reactive protein), increased aspartate and alanine aminotransferase, impaired glucose tolerance and reduced blood levels of some amino acids.

Conclusion

Although treatment with NN1177 was associated with dose-dependent and clinically relevant weight loss, the observed safety signals precluded further clinical development.

Highlights

• •Glucagon/glucagon-like peptide-1 receptor co-agonism may enhance body weight loss.
• •NN1177 is a glucagon/GLP-1 receptor co-agonist investigated for weight management.
• •NN1177 administration resulted in dose-dependent, clinically relevant weight loss.
• •Safety concerns with NN1177 were considered too pronounced for continued clinical development.

Abbreviations

AE

adverse event

ALT

alanine aminotransferase

AST

aspartate aminotransferase

AUC

area under the curve

AUC_0–168h

AUC from time 0 to 168 hours

AUC_0–inf

from time 0 to infinity

BMI

body mass index

CL/F

apparent total serum clearance of NN1177

C_max

maximum concentration

DDI

drug–drug interaction

FHD

first human dose

GI

gastrointestinal

GLP-1 RA

GLP-1 receptor agonist

GLP-1

glucagon/glucagon-like peptide-1

HbA_1c

glycated haemoglobin

hsCRP

high-sensitivity C-reactive protein

MAD

multiple ascending dose

MMRM

mixed model for repeated measurements

MRT

mean residence time

OGTT

oral glucose tolerance test

PD

pharmacodynamics

PK

pharmacokinetics

QTc

corrected QT interval

QTcF

QT corrected for heart rate by Fridericia's cube root formula

QTcI

individual-specific corrected QT

s.c.

subcutaneous

SAD

single ascending dose

SAE

serious adverse event

SS

steady state

${\mathrm{t}}_{½}$

terminal half-life

T2D

type 2 diabetes

TEAE

treatment-emergent adverse event

t_max

time to maximum serum concentration of NN1177

V_z/F

apparent volume of distribution

1. Introduction

As a chronic disease, obesity increases cardiometabolic risk [1] and ultimately confers an increased risk of premature mortality [2,3]. While 5–10% body weight loss in individuals with overweight or obesity has been shown to reduce obesity-related complications, weight loss of ≥15% can further improve risk factors for cardiometabolic disease and improve health-related quality of life [4,5].

Glucagon is a peptide hormone secreted by the alpha cells of the pancreas in response to low blood glucose levels, stimulating glycogenolysis and gluconeogenesis in the acute phase, resulting in a subsequent increase in blood glucose [6]. Recent evidence suggests that the hepatic actions of glucagon on glycogenolysis and gluconeogenesis are dependent on a fasting state, with endogenous glucagon reducing glycaemia in an insulinotropic manner [7]. Beyond the role of glucagon in glucose homeostasis, there are additional metabolic effects; these include increasing satiety and reducing food intake, as well as effects on lipid homeostasis and the increase of energy expenditure [8,9]. These additional metabolic effects offer a potential opportunity in weight management, but the use of glucagon-based therapies has been limited due to the potential impact on glucose homeostasis.

Glucagon-like peptide-1 (GLP-1) is an incretin peptide hormone that, in a glucose-dependent manner, stimulates insulin secretion and suppresses glucagon secretion [10]. Moreover, through synergistic actions of GLP-1 in the gut and brain, GLP-1 reduces body weight by reducing appetite and increasing satiety [10]. GLP-1 receptor agonists (GLP-1 RAs) are widely used for the treatment of type 2 diabetes (T2D). The benefits of GLP-1 RAs on body weight and the complications associated with obesity are well documented and two members of the drug class, liraglutide and semaglutide, are approved for weight management (as an adjunct to diet and exercise) in individuals with overweight or obesity [[11], [12], [13], [14], [15]]. Furthermore, the drug class has a well-established cardiovascular safety profile, and dulaglutide, liraglutide and once weekly subcutaneous (s.c.) semaglutide are indicated to reduce cardiovascular risk in people with T2D [16].

Enhanced weight loss may be achieved through co-agonism; dual agonism at the GLP-1 and gastric inhibitory polypeptide (GIP) receptors with tirzepatide has demonstrated mean weight loss of up to 19% [17]. Combining both glucagon receptor and GLP-1 receptor agonism in a single molecule could target weight loss through multiple mechanisms of action to enhance the level of weight loss possible [18]. However, the question of appropriate receptor balance, i.e. the relative potency of the co-agonist at each individual receptor, remains to be clarified in order to achieve maximal benefit with an acceptable side-effect profile [19]. Animal models and early clinical development trials demonstrated that glucagon/GLP-1 co-agonism can result in greater weight loss compared with GLP-1 RAs alone, without detrimental effects on glycaemic control [[20], [21], [22], [23], [24]]; however, recent studies have highlighted challenges in achieving this balance [19,25].

The glucagon/GLP-1 receptor co-agonist NNC9204-1177 (hereafter NN1177) with 2- to 4-fold higher affinity for GLP-1 than glucagon receptors was developed as a novel pharmacotherapeutic option for weight management [25]. Here, we report data and findings from three phase 1 trials investigating the safety, tolerability, pharmacokinetics (PK) and pharmacodynamics (PD) of NN1177 for once-weekly s.c. use in adults with overweight or obesity. The focus of this report is a 12-week multiple ascending dose (MAD) trial, with additional, clinically relevant findings reported from the first human dose (FHD)/single ascending dose (SAD) trial and a drug–drug interaction (DDI) trial. The DDI trial is described in detail elsewhere [26]; only safety endpoints of specific interest are reported herein to complement findings of the MAD and FHD/SAD trials. The preclinical evaluation of NN1177, and the translational challenges it presented, are also reported separately [25]. Although clinical development of NN1177 has been discontinued, the accrued data and the associated interpretations as presented in this article highlight challenges to, and mechanisms of potential relevance for, the development and use of the glucagon/GLP-1 receptor co-agonist drug class.

2. Materials and methods

2.1. Phase 1 FHD/SAD study design and population

The FHD/SAD trial was a phase 1, single-centre, placebo-controlled, double-blind, randomised, SAD clinical trial (NCT02941042). Adult male participants (18–55 years old) with BMI 25.0–34.9 kg/m² were randomised 3:1 to receive ascending single doses of s.c. NN1177 or placebo. Exclusion criteria included: those aged ≥40 years with estimated 10-year risk of atherosclerotic cardiovascular disease ≥5% [27] or with clinically relevant conditions (including diabetes), prior obesity surgery, and clinically significant laboratory or electrocardiography results. Full inclusion and exclusion criteria are provided in Appendix S1.

The FHD/SAD study was originally planned with eight NN1177 dose cohorts and two placebo cohorts each comprising eight participants with study drug doses up to 2000 μg (Table S1A). However, only six doses (10, 40, 120, 350, 700 and 1100 μg, in ascending order) were investigated. Within each cohort, administration of the study drug began with two participants on the first day (one participant received the active treatment and one participant received placebo), followed by a 70-hour safety observation period; dosing of the remaining participants followed if no serious adverse event (SAE) nor other significant safety concerns with probable or possible relation to the study drug had been observed. Participants were followed for 39 days post-dose.

2.2. Phase 1 MAD study design and population

The MAD trial was a phase 1, single-centre, placebo-controlled, double-blind, randomised, sequential MAD clinical study (NCT03308721). Seven cohorts of adult participants (age 18–55 years) with BMI 27.0–39.9 kg/m² were randomised (9:3) to receive ascending multiple doses of once-weekly s.c. NN1177 or placebo. Exclusion criteria included an estimated ≥5% risk of 10-year atherosclerotic disease in participants age ≥40 years [27]. Full inclusion and exclusion criteria are provided in Appendix S1.

Seven dose levels (200, 600, 1300, 1900, 2800, 4200 and 6000 μg) of NN1177 were investigated in ascending order. In each cohort, a total of 12 participants were treated for up to 12 weeks. To mitigate gastrointestinal (GI) adverse events (AEs) associated with GLP-1 RA treatment, stepwise escalation to the final dose level was used for cohorts two to seven (see Table S1B). In-house safety monitoring periods were planned approximately every second week and coincided with dose-escalation visits to the study site for cohorts two to seven. Before the initiation of each dose cohort, an internal Novo Nordisk trial safety group reviewed blinded safety, PK and PD parameters from the preceding dose cohort to decide whether to ascend to the next dose level. Protocol-defined participant-level stopping criteria were: plasma glucose ≥11.1 mmol/L (200 mg/dL) during a 2-hour oral glucose tolerance test (OGTT) or cardiovascular abnormalities measured after an at least 10-minute rest in the supine position. Specifically, the considered cardiovascular abnormalities were: QT interval prolongation (increase from baseline of >60 ms for the heart-rate-corrected interval [QT corrected for heart rate by Fridericia's cube root formula; QTcF]), absolute incident heart rate of >115 beats per minute (bpm) or sustained absolute heart rate of >100 bpm over 24 h. Participants started treatment on day 1 (baseline), continued treatment until day 85 (end of treatment), and were followed up for safety until day 110.

2.3. DDI trial study design and population

The design and results of the DDI trial have been reported in detail elsewhere (NCT04059367) [26]; only safety endpoints of specific interest that complement the FHD/SAD and MAD trials are reported herein. In brief, the DDI trial was a single centre, open-label, single-sequence trial. Adult participants (age 18–60 years) with BMI 23.0–29.9 kg/m² and glycated haemoglobin (HbA_1c) <6.5% received once-weekly s.c. NN1177 (dose escalated every 2 weeks until a final dose of 4200 μg for 3 weeks). NN1177 was initiated on day 1 (baseline) and the last dose was administered on day 78 (end of treatment). A Cooperstown 5 + 1 index cocktail, containing representative index substrates for multiple cytochrome P (CYP)450 enzymes (caffeine [CYP1A2, N-acetyltransferase-2 and xanthine oxidase]; dextromethorphan [CYP2D6]; omeprazole [CYP2C19]; intravenous midazolam [hepatic CYP3A]; and warfarin + vitamin K [CYP2C9]) [28], was administered before NN1177 administration on day 1 and on day 78 (prior to the final 4200-μg dose of NN1177).

2.4. Ethics

The clinical trials were all conducted in accordance with the Declaration of Helsinki, the Council for International Organizations of Medical Sciences International Ethical Guidelines, and the International Conference on Harmonisation Good Clinical Practices Guideline. The studies obtained approval from local institutional review boards (Aspire IRB 11491 Woodside Ave, Santee, CA 92071 and Midlands Independent Review Board 8417 Santa Fe Drive/Suite 100, Overland Park, KS 66212, US). All participants provided written informed consent before participating in the study.

2.5. Endpoints

2.5.1. Safety endpoints and assessments

For both the FHD/SAD and MAD trials, the primary endpoint was the number of treatment-emergent adverse events (TEAEs) recorded from start of dosing of NN1177 (day 1) until completion of the post-treatment follow-up visit (day 39 in the FHD/SAD trial and day 110 in the MAD trial). All TEAEs were coded using MedDRA v22.1. Potential relationship of a TEAE with treatment with study drug was assessed by the investigator.

In the DDI trial, TEAEs were recorded, and safety data of specific interest (TEAEs and clinical laboratory analyses of amino acids) are reported in this article as supportive information.

2.5.2. Heart rate, QT interval and blood pressure

In the FHD/SAD and MAD trials, vital signs (comprising heart rate, systolic/diastolic blood pressure, respiratory rate and body temperature) were taken at each study visit. The potential exposure–response relationship between NN1177 exposure and changes in the QTc interval (measured during Holter monitoring) were also evaluated in the MAD trial. Prior to lock of the MAD trial database, the pre-specified dependent variable QTcF was changed to individual-specific corrected QT interval (QTcI) to correct for the increased heart rate.

2.5.3. Clinical laboratory analyses

In the MAD trial, clinical laboratory analysis of haematology, biochemistry and coagulation parameters was performed at each study visit.

2.5.4. PK endpoints

PK characteristics of NN1177 were assessed as supportive secondary endpoints in both the FHD/SAD and MAD trials. In the FHD/SAD trial, the following secondary PK endpoints were assessed after a single dose: area under the NN1177 serum concentration–time curve (AUC) from time 0 to infinity (AUC_0–inf), AUC from time 0 to 168 hours (AUC_0–168h), maximum concentration (C_max), time to maximum serum concentration of NN1177 (t_max), terminal half-life (t_1/2), apparent total serum clearance of NN1177 (CL/F), apparent volume of distribution (V_z/F) and mean residence time (MRT). In the MAD trial, the following secondary PK endpoints were assessed at steady state (SS): AUC_{0–168h, SS}, C_{max, SS}, t_{max, SS}, and t_1/2, _SS. Exploratory PK endpoints comprised CL/F, apparent volume of distribution (at steady state and during elimination) (V_SS/F and V_z/F) and MRT. PK parameters assessed in the DDI trial are not reported herein.

2.5.5. PD endpoints

Change in body weight was assessed from baseline to end of follow-up as a supportive secondary PD endpoint in the FHD/SAD trial and from baseline to end of treatment as an exploratory PD endpoint in the MAD trial. Glucose metabolism parameters were also assessed as exploratory PD endpoints in both the FHD/SAD and MAD studies, comprising change from baseline to end of follow-up in fasting glucose, fasting insulin, HbA_1c and fasting glucagon.

In the MAD trial, an OGTT was conducted on day 1, day 29 and day 85, with blood samples taken for assessment of glucose (measured in plasma) and insulin at 0, 10, 20, 30, 60, 90 and 120 min after glucose intake. For each OGTT, the following endpoints were assessed: AUC from time 0 to 2 h (AUC_0–2h) for insulin and glucose, and insulin secretion ratio (the ratio of the AUCs calculated from time 0 to 2 h during OGTT for insulin and glucose).

Fasting leptin and fasting soluble leptin receptor concentrations were also assessed from baseline to end of treatment in the MAD trial.

2.6. Statistical analysis

For both the FHD/SAD and MAD trials, no formal statistical sample size calculations were performed; the selected sample size was considered sufficient and necessary to evaluate the primary objectives for both trials. Participants were replaced after screening as needed, in order to have the minimum number of participants with sufficient evaluable safety data to complete the safety evaluation for each cohort. In both trials, all safety endpoints and assessments, including the primary endpoint (number of TEAEs), were summarised using the safety analysis set consisting of all participants who were exposed to at least one dose of NN1177. No methods for imputing missing data were applied.

Analyses of PK and PD endpoints in both the FHD/SAD and MAD trials were based on the full analysis set, consisting of all participants who were randomised and received at least one dose of NN1177. In general, PK endpoints were summarised using descriptive statistics and presented as geometric means (coefficient of variation [%]), medians (minimum, maximum) or as arithmetic means (minimum, maximum). AUC_0–168h and C_max (PK endpoints) were compared across using an analysis of variance (ANOVA) model based on log-transformed data with dose as a factor. Exploratory analysis for dose proportionality was performed on AUC_0–inf and C_max (PK endpoints) using a linear regression model based on logarithmically transformed data, log(dose) as a covariate, with β = 1 meaning that the measured outcome (AUC_0–inf or C_max) increases in a dose-proportional manner with increasing dose (estimated quantity 2β with 95% confidence reported). Change in body weight, fasting glucose and HbA_1c (PD endpoints) were analysed using a mixed model for repeated measures (MMRM) with post-baseline measurements and treatment as variables. Fasting insulin (PD endpoint) was logarithmically transformed and analysed using an MMRM with post-baseline measurements and treatment as variables. For the OGTT-related PD endpoints, glucose and insulin AUC_0–2h were logarithmically transformed and analysed using an ANOVA model with dose as a factor and logarithmically transformed baseline value as a covariate. Glucose concentration at 2 h was analysed using an ANOVA model with dose as a factor and baseline value as a covariate. Change in fasting leptin and fasting soluble leptin receptor concentrations (PD endpoints) were analysed using an MMRM with logarithmically-transformed relative changes from baseline in fasting leptin or fasting soluble leptin receptor as the dependent variables and treatment, visit and logarithmically-transformed baseline fasting leptin or fasting soluble leptin receptor as fixed effects; treatment and baseline effect were nested within visit. QTcI exposure–response (PD endpoint) was analysed using an MMRM on change from baseline in QTcI, with fixed effects for treatment (active/placebo), time point, baseline QTcI and NN1177 concentration and random effects for participant and NN1177 concentration [29]. No methods for imputing missing data were applied.

2.7. Role of the funding source

The study sponsor was involved in the study design, data collection, data review, data analysis and drafting of the report. All authors had full access to the data related to these studies and approved the report for publication.

3. Results

3.1. Participant disposition, baseline characteristics and dose levels

In the FHD/SAD trial, 44 (89.8%) of the 49 participants randomly assigned to treatment completed the trial (Figure 1A; Appendix S2). Six of the eight planned dose levels were investigated; the trial safety group raised GI tolerability concerns (based on a review of the frequency and severity of events [preferred terms] within the ‘GI disorder’ system organ class) following administration of 1100 μg and it was decided not to initiate the 1600-μg and 2000-μg cohorts (Table S1A).

Figure 1.

Participant flow. A. First in human dose/single ascending dose trial. B. Multiple ascending dose trial. The number of participants exposed to the NN1177 1900 μg group (11 participants) and 4200 μg group (20 participants) was higher than in the other NN1177 groups. In cohort 4 (1900 μg) and cohort 6 (4200 μg), a blinded assessment of the timing and reasons for participants being prematurely withdrawn after randomisation led to replacements of participants to aim for at least eight participants for safety evaluation. ∗One participant in the NN1177 4200 μg cohort reported a non-TEAE leading to premature withdrawal from the trial, the remaining adverse events leading to withdrawal were all TEAEs. FAS, full analysis set; SAS, safety analysis set; TEAE, treatment-emergent adverse event.

In the MAD trial, a total of 99 participants were randomly assigned to treatment (200 μg, 600 μg, 1300 μg [n = 9 each]; 1900 μg [n = 11]; 2800 μg [n = 9]; 4200 μg [n = 20]; and 6000 μg [n = 9]) or placebo (n = 23) (Figure 1B). More participants were randomised and exposed to the 1900-μg and 4200-μg dose levels than to the other dose levels. This was due to a larger number of prematurely withdrawn participants assigned to these dose levels (three participants in the 1900-μg cohort and 13 participants in the 4200-μg cohort; Appendix S2); consequently, more participants were entered into the 1900 μg and 4200 μg cohorts in order to achieve at least eight participants for safety evaluation. A total of 68 participants (68.7%) completed the trial. The maximum dosing level was 6000 μg; the dose-escalation regimen to this dose level in cohort seven was modified based on data from cohorts one to six and the regimen was modified to incorporate a lower initial dose level (Table S1B).

In the DDI trial, 45 participants were enrolled, of whom 37 completed the trial; details are reported elsewhere [26].

Overall baseline characteristics for the FHD/SAD, MAD and DDI trials are shown in Table 1. For the FHD/SAD and MAD trials, demographics and baseline characteristics were generally similar across cohorts (Table S2A and B). There were no clinically relevant imbalances across groups with respect to medical history at screening.

Table 1.

Baseline characteristics.

	FHD/SAD trial N = 49	MAD trial N = 99	DDI trial N = 45
Mean age, years (range)	33.9 (21–55)	36.0 (18–55)	33.9 (21–55)
Sex, n (%)
Male	49 (100)	45 (45.5)	49 (100)
Female	0	54 (54.5)	0
Race, n (%)
American Indian or Alaska Native	1 (2.0)	0	0
Asian	0	2 (2.0)	0
Black or African American	33 (67.3)	57 (57.6)	11 (24.4)
White	15 (30.6)	40 (40.4)	32 (71.1)
Other	0	0	2 (4.4)
Ethnicity, n (%)
Hispanic or Latino	4 (8.2)	7 (7.1)	23 (51.1)
Not Hispanic or Latino	45 (91.8)	92 (92.9)	22 (48.9)
Mean body weight, kg (SD)	92.5 (10.0)	94.6 (14.1)	75 (60.0–95.4)
Mean BMI, kg/m2(SD)	29.5 (2.5)	32.8 (3.3)	26.5 (1.9)
Mean blood pressure, mmHg (SD)
SBP	117 (10)	113 (11.8)	N/A
DBP	69 (8)	73 (8.2)	N/A
Heart rate, bpm (SD)	61 (10)	65.5 (10.6)	N/A
Mean HbA1c, % (SD)	5.3 (0.3)	5.3 (0.4)	N/A
Fasting glucose, mmol/L (SD)	4.41 (0.39)	4.7 (0.4)	N/A
Mean triglycerides, mmol/L (SD)	1.07 (0.58)	1.03 (0.45)	N/A

Data from safety analysis set, which consists of all participants who were exposed to at least one dose of study drug.

%, percentage of subjects with characteristic; BMI, body mass index; bpm, beats per minute; DBP, diastolic blood pressure; DDI, drug–drug interaction; FHD, first human dose; HbA_1c, glycated haemoglobin; MAD, multiple ascending dose; n, number of participants with characteristic; N, number of participants with available data; N/A, not available; SAD, single ascending dose; SBP, systolic blood pressure; SD, standard deviation.

3.2. Safety and tolerability

3.2.1. TEAEs

Summary statistics for TEAEs in the FHD/SAD trial are shown in Table 2A. In brief, across NN1177 doses the proportion of participants reporting TEAEs ranged from 33.3% (40-μg dose) to 100% (350-μg dose), and across all doses 25 participants experienced 48 events, compared with 84.6% (11 participants; 19 events) with placebo. The majority of TEAEs were general disorders and administration-site conditions (Table 2A). There were no safety issues that precluded initiation of the MAD trial.

Table 2.

Treatment-emergent adverse events.

A. First in human dose/single ascending dose trial
	NN1177 10 μg		NN1177 40 μg		NN1177 120 μg		NN1177 350 μg		NN1177 700 μg		NN1177 1100 μg		Placebo
	n (%)	E	n (%)	E	n (%)	E	n (%)	E	n (%)	E	n (%)	E	n (%)	E
Number of participants	6		6		6		6		6		6		13
Adverse events	5 (83.3)	8	2 (33.3)	3	3 (50.0)	3	6 (100)	6	5 (83.3)	14	4 (66.7)	14	11 (84.6)	19
Serious adverse events	0	0	0	0	0	0	0	0	0	0	0	0	0	0
Relation to trial productb
Probable	0	0	0	0	0	0	0	0	0	0	3 (50.0)	10	1 (7.7)	5
Possible	1 (16.7)	1	0	0	0	0	0	0	2 (33.3)	4	0	0	0	0
Unlikely	5 (83.3)	7	2 (33.3)	3	3 (50.0)	3	6 (100)	6	5 (83.3)	10	3 (50.0)	4	11 (84.6)	14
Severity
Severe	0	0	0	0	0	0	0	0	0	0	1 (16.7)	1	0	0
Moderate	2 (33.3)	2	1 (16.7)	1	2 (33.3)	2	3 (50.0)	3	4 (66.7)	4	2 (33.3)	6	5 (38.5)	5
Mild	4 (66.7)	6	1 (16.7)	2	1 (16.7)	1	3 (50.0)	3	5 (83.3)	10	4 (66.7)	7	8 (61.5)	14
Outcome
Not recovered	0	0	0	0	0	0	0	0	0	0	0	0	0	0
Recovered	5 (83.3)	8	2 (33.3)	3	3 (50.0)	3	5 (83.3)	5	5 (83.3)	14	4 (66.7)	14	11 (84.6)	19
Unknown	0	0		0	0	0	1 (16.7)	1	0	0	0	0	0	0
General disorders and administration-site conditions	4 (66.7)	4	2 (33.3)	2	2 (33.3)	2	6 (100)	6	5 (83.3)	6	2 (33.3)	2	8 (61.5)	8
Medical device-site dermatitis	4 (66.7)	4	1 (16.7)	1	2 (33.3)	2	5 (83.3)	5	4 (66.7)	4	2 (33.3)	2	6 (46.2)	6
Vessel puncture site bruise	0	0	1 (16.7)	1	0	0	0	0	0	0	0	0	1 (7.7)	1
Vessel puncture site haematoma	0	0	0	0	0	0	0	0	2 (33.3)	2	0	0	0	0
Catheter site bruise	0	0	0	0	0	0	0	0	0	0	0	0	1 (7.7)	1
Medical device site pruritus	0	0	0	0	0	0	1 (16.7)	1	0	0	0	0	0	0
Injection site reactions	0	0	0	0	0	0	0	0	0	0	0	0	0	0
Gastrointestinal disorders	0	0	0	0	0	0	0	0	2 (33.3)	2	3 (50.0)	6	3 (23.1)	5
Nausea	0	0	0	0	0	0	0	0	1 (16.7)	1	3 (50.0)	3	1 (7.7)	1
Vomiting	0	0	0	0	0	0	0	0	0	0	3 (50.0)	3	1 (7.7)	1
Abdominal pain	0	0	0	0	0	0	0	0	0	0	0	0	2 (15.4)	2
Diarrhoea	0	0	0	0	0	0	0	0	0	0	0	0	1 (7.7)	1
Dyspepsia	0	0	0	0	0	0	0	0	1 (16.7)	1	0	0	0	0

B. Multiple ascending dose trial
	NN1177 200 μg		NN1177 600 μg		NN1177 1300 μg		NN1177 1900 μg		NN1177 2800 μg		NN1177 4200 μg		NN1177 6000 μg		Placebo
	n (%)	E	n (%)	E	n (%)	E	n (%)	E	n (%)	E	n (%)	E	n (%)	E	n (%)	E
Number of participants	9		9		9		11		9		20		9		23
Adverse events	8 (88.9)	29	6 (66.7)	37	9 (100)	31	11 (100)	90	9 (100)	95	20 (100)	219	9 (100)	93	16 (69.6)	62
Serious adverse events	0	0	0	0	0	0	0	0		0	0	0	1 (11.1)	6	0	0
Events leading to premature withdrawal from triala	1 (11.1)	1	1 (11.1)	1	0	0	1 (9.1)	1	1 (11.1)	1	6 (30.0)	6	3 (33.3)	3	0	0
Relation to study drugb
Probable	6 (66.7)	16	3 (33.3)	15	6 (66.7)	21	11 (100)	60	9 (100)	65	19 (95.0)	156	9 (100)	68	12 (52.5)	31
Possible	2 (22.2)	3	4 (44.4)	8	0		7 (63.6)	19	9 (100)	22	16 (80.0)	43	8 (88.9)	15	9 (39.1)	13
Unlikely	6 (66.7)	10	6 (66.7)	14	8 (88.9)	10	7 (63.6)	11	4 (44.4)	8	13 (65.0)	20	5 (55.6)	10	13 (56.5)	18
Severity
Severe	0	0	0	0	0	0	0	0	0	0	0	0	1 (11.1)	5	0	0
Moderate	1 (11.1)	2	4 (44.4)	5	1 (11.1)	1	3 (27.3)	10	5 (55.6)	9	8 (40.0)	18	4 (44.4)	16	3 (13.0)	3
Mild	8 (88.9)	27	6 (66.7)	32	9 (100)	30	11 (100)	80	9 (100)	86	20 (100)	201	9 (100)	72	15 (65.2)	59
Treatment-emergent adverse events by system organ class and preferred termc
Gastrointestinal disorders	6 (66.7)	12	2 (22.2)	5	4 (44.4)	14	7 (63.6)	21	8 (88.9)	34	17 (85.0)	69	7 (77.8)	36	7 (30.4)	11
Nausea	3 (33.3)	3	0	0	3 (33.3)	8	4 (36.4)	9	7 (77.8)	12	13 (65.0)	14	4 (44.4)	9	3 (13.0)	3
Vomiting	1 (11.1)	1	0	0	2 (22.2)	5	4 (36.4)	7	4 (44.4)	5	10 (50.0)	21	5 (55.6)	13	2 (8.7)	2
Diarrhoea	2 (22.2)	2	1 (11.1)	1	1 (11.1)	1	1 (9.1)	1	3 (33.3)	4	7 (35.0)	8	3 (33.3)	3	3 (13.0)	3
Dyspepsia	3 (33.3)	4	1 (11.1)	1	0	0	1 (9.1)	1	1 (11.1)	2	5 (25.0)	5	4 (44.4)	6	1 (4.3)	1
Abdominal pain	1 (11.1)	1	0	0	0	0	1 (9.1)	1	1 (11.1)	1	6 (30.0)	9	2 (22.2)	2	1 (4.3)	1
Cardiac disorders	2 (22.2)	3	1 (11.1)	2	3 (33.3)	5	7 (63.6)	20	6 (66.7)	16	14 (70.0)	34	5 (55.6)	7	8 (34.8)	12
Sinus tachycardia	1 (11.1)	2	1 (11.1)	2	3 (33.3)	5	5 (45.5)	11	4 (44.4)	13	12 (60.0)	30	5 (55.6)	7	4 (17.4)	5
Ventricular tachycardia	0	0	0	0	0	0	3 (27.3)	3	2 (22.2)	2	2 (10.0)	2	0	0	3 (13.0)	3
Metabolism and nutrition disorders	4 (44.4)	4	1 (11.1)	1	2 (22.2)	2	8 (72.7)	9	6 (66.7)	6	14 (70.0)	15	7 (77.8)	12	4 (17.4)	5
Decreased appetite	4 (44.4)	4	1 (11.1)	1	1 (11.1)	1	7 (63.6)	7	6 (66.7)	6	14 (70.0)	14	7 (77.8)	7	4 (17.4)	4
General disorders and administration-site conditions	0	0	2 (22.2)	10	1 (11.1)	2	6 (54.5)	22	6 (66.7)	15	11 (55.0)	37	7 (77.8)	21	6 (26.1)	11
Injection site reaction	0	0	2 (22.2)	10	1 (11.1)	2	5 (45.5)	19	3 (33.3)	8	6 (30.0)	23	5 (55.6)	16	6 (26.1)	11
Fatigue	0	0	0	0	0	0	0	0	3 (33.3)	3	4 (20.0)	4	3 (33.3)	4	0	0
Nervous system disorders	2 (22.2)	2	3 (33.3)	5	1 (11.1)	1	2 (18.2)	6	5 (55.6)	12	14 (70.0)	22	2 (22.2)	3	3 (13.0)	4
Headache	2 (22.2)	2	2 (22.2)	3	0	0	2 (18.2)	5	4 (44.4)	9	9 (45.0)	13	1 (11.1)	1	2 (8.7)	2

Data from the safety analysis set consisting of all participants who were exposed to at least one dose of study drug. TEAEs recorded from time of dosing of NN1177 (day 1) until completion of the post-treatment follow-up visit (day 110). AEs were coded using the most recent version of MedDRA (version 22.1). Participants who received placebo were pooled across cohorts.

%, percentage of participants with an AE; AE, adverse events; E, number of events; MedDRA, Medical Dictionary for Regulatory Activities; n, number of participants; SAE, serious adverse event; TEAE, treatment-emergent adverse event.

^aThe TEAEs leading to withdrawal were (doses in bold refer to the assigned dose group whereas doses not in bold refer to the dose level achieved [in escalation] when the event occurred): 13 TEAEs leading to withdrawal; 200 μg: blood glucose increased (n = 1) at dose 200 μg, 600 μg; upper respiratory tract infection (n = 1) at dose 600 μg, 1900 μg; vomiting (n = 1) at dose 1300 μg, 2800 μg; arrhythmia (n = 1) at dose 1300 μg, 4200 μg: abdominal pain (n = 2) at doses 1300 μg and 1900 μg, asthenia (n = 1) at dose 2800 μg, sinus tachycardia (n = 2) at doses 1300 μg and 4200 μg, nausea (n = 1) at dose 4200 μg, 6000 μg: vomiting (n = 1) at dose 3200 μg, fatigue (n = 1) at dose 3200 μg, hyperglycaemia (n = 1) at dose 1600 μg.

^bAs assessed by investigator.

^cTEAEs are included if the cumulative proportion of participants with at least one event across cohorts including the pooled placebo cohort was ≥10%.

In the MAD trial, 94.7% (72 participants; 594 events) of participants reported TEAEs with any dose of NN1177 compared with 69.6% (16 participants; 62 events) with placebo (Table 2B). Fewer participants in the 200-μg and 600-μg cohorts reported TEAEs compared with those receiving the higher doses, but there was no clear dose relationship (Table 2B). There was a tendency towards a dose relationship for NN1177 (≥1300-μg dose groups) with TEAEs of moderate severity. For relation to study drug, most participants had TEAEs that were judged by the investigator as probably or possibly related to NN1177. Amongst participants receiving NN1177, 13 participants (17.1%) prematurely withdrew from the trial due to a TEAE, compared with no participants in the placebo group (Table 2B).

In the MAD trial, events related to GI tolerability (‘GI disorders’ system organ class) were the most frequently reported TEAEs, particularly nausea and vomiting, and were experienced by 67.1% of participants (191 events across 51 participants) with NN1177 and 30.4% of participants (11 events across seven participants) with placebo (Table 2B). The proportion of GI TEAEs possibly or probably related to NN1177 treatment was highest for participants treated with dose levels ≥1900 μg (data not shown). Injection-site reactions (pain, itching and redness) were mild, and reported by 28.9% of participants (78 events across 22 participants) with NN1177 and 26.1% of participants (11 events across six participants) with placebo.

Five participants receiving NN1177 prematurely withdrew from the trial because they met protocol-defined stopping criterion: three participants met the heart rate criterion (one participant in the 1900-μg dose cohort [sustained absolute heart rate >100 bpm over 24 h] and two participants in the 4200-μg dose cohort [one with sustained absolute heart rate >100 bpm over 24 h and one with incident absolute heart rate >115 bpm]); two participants met the OGTT criterion of plasma glucose ≥11.1 mmol/L (200 mg/dL) (one participant in 200-μg dose cohort and one participant in 6000-μg dose cohort).

The AE profiles observed in the DDI trial are shown in Table S3 and generally reflected the TEAE profile observed in the FHD/SAD and MAD trials.

3.2.2. Heart rate, QT interval and blood pressure

Increases in heart rate were seen in all three trials. In the FHD/SAD trial, heart rate increased by up to 22 bpm at 24 h post-dosing (Figure 2A). At follow-up, the heart rate increases had reversed or diminished, and QTcF exposure–response analysis found no relationship between blood concentrations of NN1177 and changes in QTcF. These results did not preclude initiation of the MAD trial.

Figure 2.

Heart rate and QTcI interval. A. Heart rate in the single ascending dose trial. B. Heart rate in the multiple ascending dose trial. C. Predicted placebo-adjusted QTcI interval change from baseline vs concentration. D. Heart rate by time of day in the drug–drug interaction trial. Data from safety analysis set. A. Participants were treated on day 1 (baseline). Mean heart rate from safety ECG assessments at Visit 2 (predose day 1; postdose, day 2, 3, 4, 5), Visit 3 (day 8), Visit 5 (day 15 to day 16), Visit 8 (day 39 to day 41) B. Participants were started treatment on day 1 (baseline), continued treatment until day 85 (end of treatment) and were followed up for safety until day 110; vertical reference lines represent first and last dosing of NN1177. Mean of Holter heart rate during 48 h for Visit 2 (day-2 to day 1), Visit 2 (day 8 to day 10), Visit 4 (day 22 to day 24), Visit 6 (day 36 to day 38), Visit 8 (day 50 to day 52) and Visit 10 (day 64 to day 66) are plotted on 0, 9, 23, 37, 51 and 65 days, respectively. C. Reference line represents 10 ms prolongation of the QTcI interval. ETDs were derived by analysing change from baseline in the QTcI interval at all time points using a mixed effect model with subject and corresponding NN1177 concentration as random effects, and baseline QTcI interval and NN1177 concentration as fixed covariates and treatment (active/placebo) and time point as fixed factors. For the two random effects, an unstructured 2 × 2 covariance matrix was applied. The model also includes independent and identically distributed residual error terms. The solid black line with shaded area denotes the model-predicted mean placebo-adjusted delta QTcI interval with 90% CI. D. Participants were started treatment on day 1 (baseline), continued treatment until day 78 (end of treatment). AUC, area under the curve; CI, confidence interval; DDI, drug–drug interaction; QTcF, QT interval corrected for heart rate using Fridericia's formula; QTcI, QT interval corrected using individualised formula.

In the MAD trial, heart rate increased by 1–15 bpm from baseline to end of treatment with NN1177 at day 85 (168 h after last dosing) across dose levels; the increases appeared to be more pronounced at the highest dose levels (Figure 2B). At follow-up, the heart rate increases had reversed or diminished. In the MAD trial, treatment with NN1177 appeared to prolong the QTcI interval by up to 18 ms in a concentration-dependent manner (Figure 2C). Figures for QTcI as a function of time and model diagnostic plots are shown in Figs. S1 and S2, respectively. While there seemed to be a concentration-dependent increase in the QTcI interval, the concentration–QTc slope estimate should be interpreted with caution as the assumptions for the pre-specified linear effect model were not fulfilled due to the observed effects of NN1177 on heart rate, and as the hysteresis plot demonstrated a delayed effect where the QTcI interval did not return to baseline values for all cohorts with declining concentration (Fig. S2B).

In the DDI trial, heart rate was measured continuously over 48 h. The observed increase in heart rate with NN1177 (12 bpm over a 24-hour period) was greatest at night (mean of 15 bpm); thus, the difference in average heart rate between night and day seen at baseline (day 1) diminished following treatment with NN1177 (day 66; Figure 2D).

No clinically relevant changes in systolic or diastolic blood pressure were observed in the FHD/SAD or MAD trials (Fig. S3A and B).

3.2.3. Clinical laboratory analyses

Clinical laboratory parameters for the FHD/SAD trial are shown in Table S4; no participant had any abnormal laboratory assessment that was deemed clinically significant.

Clinical laboratory parameters for the MAD trial are shown in Table 3. In the NN1177 cohorts in the MAD trial, there was a reduction in the reticulocyte count at end of treatment relative to baseline values, which appeared to be dose-dependent in participants treated with higher NN1177 doses (Table 3). Reticulocytes recovered to baseline levels by the end of the follow-up period. There were no clinically relevant changes in other haematological parameters (e.g. haemoglobin, or counts of erythrocytes, thrombocytes and leucocytes).

Table 3.

Clinical laboratory parameters in the multiple ascending dose trial.

	NN1177 200 μg	NN1177 600 μg	NN1177 1300 μg	NN1177 1900 μg	NN1177 2800 μg	NN1177 4200 μg	NN1177 6000 μg	Placebo
Haematology
Haemoglobin (mmol/L)
Baseline	8.5 (12.8)	8.9 (13.6)	8.3 (12.5)	8.9 (8.3)	8.6 (9.0)	8.2 (10.8)	8.3 (12.2)	8.4 (13.1)
EOT	8.2 (12.7)	8.7 (15.0)	7.9 (10.2)	8.3 (10.9)	8.0 (10.2)	8.2 (14.2)	7.5 (5.4)	7.9 (13.7)
Follow-up	8.5 (12.8)	8.4 (14.6)	7.8 (11.5)	8.3 (10.8)	7.7 (11.5)	7.8 (13.4)	7.4 (11.8)	8.0 (16.2)
Ratio to baseline at EOT	0.95 (3.5)	0.97 (3.0)	0.95 (6.5)	0.94 (8.4)	0.93 (4.7)	0.97 (7.7)	0.93 (2.6)	0.96 (5.1)
Ratio to baseline at follow-up	1.00 (5.0)	0.94 (4.6)	0.94 (2.9)	0.94 (6.6)	0.90 (6.0)	0.94 (6.7)	0.92 (6.4)	0.97 (6.1)
Erythrocytes, 1012/L
Baseline	4.88 (10.6)	5.08 (13.5)	4.80 (14.6)	4.86 (10.3)	4.62 (9.0)	4.56 (9.7)	4.57 (9.5)	4.77 (11.1)
EOT	4.70 (10.7)	4.92 (11.0)	4.61 (13.3)	4.61 (12.2)	4.38 (9.1)	4.50 (15.0)	4.12 (7.4)	4.54 (9.9)
Follow-up	4.90 (11.4)	4.78 (12.7)	4.53 (14.5)	4.53 (11.2)	4.22 (10.3)	4.34 (12.5)	4.18 (11.7)	4.60 (13.6)
Ratio to baseline at EOT	0.96 (3.1)	0.96 (3.4)	0.96 (7.1)	0.95 (8.7)	0.95 (5.5)	0.98 (8.0)	0.96 (1.2)	0.97 (5.2)
Ratio to baseline at follow-up	1.00 (5.4)	0.94 (5.6)	0.94 (3.4)	0.93 (6.0)	0.91 (5.8)	0.95 (6.6)	0.93 (5.3)	0.97 (6.5)
Reticulocytes, 109/L
Baseline	64.35 (32.0)	67.97 (33.9)	59.10 (26.2)	62.78 (43.5)	69.94 (19.7)	61.69 (24.4)	68.96 (19.5)	63.50 (20.2)
EOT	60.66 (38.6)	64.88 (23.0)	45.11 (18.3)	40.63 (41.4)	42.73 (39.7)	38.19 (41.1)	38.32 (30.6)	58.10 (21.1)
Follow-up	68.54 (28.3)	75.23 (24.0)	60.78 (22.0)	67.39 (29.1)	63.43 (41.9)	70.32 (31.6)	70.62 (26.6)	65.00 (27.2)
Ratio to baseline at EOT	0.89 (22.7)	0.95 (24.7)	0.76 (24.0)	0.70 (17.5)	0.63 (27.2)	0.64 (21.2)	0.62 (20.3)	0.90 (18.2)
Ratio to baseline at follow-up	1.07 (14.4)	1.11 (15.6)	1.03 (15.5)	1.07 (24.6)	0.91 (33.1)	1.12 (17.8)	1.05 (38.9)	1.02 (24.5)
Thrombocytes, 109/L
Baseline	268 (17.1)	243 (14.9)	252 (23.8)	262 (24.7)	248 (9.5)	251 (21.5)	248 (23.1)	264 (28.5)
EOT	296 (13.4)	254 (18.3)	242 (24.5)	244 (26.3)	233 (11.2)	242 (15.5)	256 (15.6)	288 (23.4)
Follow-up	280 (19.8)	236 (20.2)	264 (25.9)	247 (20.7)	252 (13.3)	269 (22.8)	275 (22.6)	269 (22.4)
Ratio to baseline at EOT	1.08 (13.1)	1.03 (16.6)	0.96 (14.8)	0.96 (9.9)	0.94 (9.2)	0.94 (18.0)	1.04 (14.5)	1.08 (12.6)
Ratio to baseline at follow-up	1.04 (9.8)	0.97 (15.9)	1.05 (7.7)	0.94 (12.5)	1.02 (9.1)	1.07 (8.8)	1.10 (11.6)	0.99 (13.9)
Leucocytes, 109/L
Baseline	6.1 (36.4)	6.6 (23.9)	4.9 (22.1)	5.8 (19.3)	5.4 (13.2)	5.8 (33.1)	6.8 (22.1)	6.4 (26.6)
EOT	6.6 (21.5)	7.5 (23.2)	5.5 (26.9)	5.5 (24.7)	4.7 (27.3)	4.7 (26.8)	5.6 (26.9)	6.6 (23.0)
Follow-up	6.3 (31.6)	6.5 (20.8)	5.7 (24.4)	5.9 (27.0)	5.2 (23.2)	6.0 (27.4)	6.5 (19.5)	6.6 (24.7)
Ratio to baseline at EOT	1.04 (19.8)	1.14 (34.1)	1.11 (21.9)	0.97 (16.1)	0.87 (15.4)	0.89 (12.9)	0.85 (17.8)	1.06 (18.3)
Ratio to baseline at follow-up	1.02 (21.3)	0.98 (21.2)	1.15 (17.4)	1.01 (23.7)	0.97 (17.8)	1.03 (10.1)	0.94 (17.6)	1.03 (24.2)
Inflammatory biomarkers
Fibrinogen, g/L
Baseline	2.86 (26.2)	3.09 (19.7)	3.37 (15.8)	2.91 (20.3)	2.92 (18.5)	3.20 (15.4)	3.72 (15.8)	3.21 (24.9)
EOT	2.96 (23.2)	3.25 (20.1)	3.88 (22.6)	3.34 (17.5)	3.95 (11.9)	4.30 (15.5)	4.94 (22.7)	3.11 (24.6)
Follow-up	2.81 (18.9)	2.97 (19.0)	3.00 (17.8)	2.50 (24.7)	2.58 (24.7)	3.15 (26.5)	3.58 (22.5)	3.05 (29.9)
Ratio to baseline at EOT	1.01 (16.7)	1.05 (11.7)	1.15 (13.9)	1.14 (9.8)	1.28 (12.6)	1.33 (16.0)	1.36 (12.9)	0.94 (15.2)
Ratio to baseline at follow-up	0.98 (16.6)	0.95 (13.6)	0.89 (10.6)	0.86 (14.2)	0.88 (13.0)	0.98 (24.6)	0.96 (9.0)	0.93 (19.1)
hsCRP, mg/L
Baseline	1.1 (218.7)	1.7 (247.7)	2.9 (109.4)	1.4 (132.1)	1.3 (146.1)	2.0 (146.3)	2.6 (155.7)	2.3 (172.0)
EOT	1.3 (258.1)	1.1 (281.9)	2.2 (160.1)	1.5 (196.6)	2.6 (110.4)	3.7 (98.5)	4.7 (88.1)	2.1 (112.6)
Follow-up	1.6 (245.1)	1.5 (204.1)	2.6 (99.3)	1.1 (135.6)	0.7 (201.5)	2.6 (134.2)	3.5 (111.4)	2.1 (188.8)
Ratio to baseline at EOT	1.22 (93.4)	0.84 (64.0)	0.76 (87.4)	1.11 (82.1)	1.64 (79.5)	1.65 (35.6)	2.14 (99.4)	1.01 (116.4)
Ratio to baseline at follow-up	1.45 (56.0)	0.89 (32.4)	0.90 (38.5)	0.80 (106.5)	0.55 (94.8)	1.18 (106.1)	1.17 (76.3)	0.95 (224.2)
Liver parameters
Alanine aminotransferase (U/L)
Baseline	22 (45.1)	15 (34.8)	19 (76.4)	15 (52.9)	19 (54.2)	15 (54.1)	14 (63.7)	18 (64.1)
EOT	23 (71.3)	15 (29.1)	20 (61.2)	21 (49.0)	18 (48.6)	24 (66.3)	22 (100.0)	15 (62.1)
Follow-up	18 (40.7)	21 (71.6)	17 (67.0)	18 (63.5)	23 (85.8)	18 (67.9)	20 (68.0)	14 (47.8)
Ratio to baseline at EOT	1.02 (51.8)	0.97 (11.0)	1.07 (39.7)	1.25 (42.3)	0.93 (65.7)	1.40 (22.0)	1.62 (41.2)	1.01 (39.6)
Ratio to baseline at follow-up	0.82 (55.5)	1.38 (37.7)	0.91 (26.0)	1.19 (46.5)	1.20 (62.8)	1.20 (45.1)	1.56 (56.0)	0.89 (41.4)
Aspartate aminotransferase (U/L)
Baseline	17 (23.7)	16 (16.7)	18 (28.6)	16 (32.4)	17 (27.6)	15 (26.8)	15 (21.3)	17 (36.7)
EOT	17 (25.5)	16 (23.6)	18 (31.8)	17 (19.2)	15 (19.9)	18 (25.2)	19 (36.8)	15 (29.5)
Follow-up	16 (22.2)	20 (43.4)	16 (32.5)	17 (37.9)	18 (58.5)	16 (28.8)	16 (30.4)	16 (39.5)
Ratio to baseline at EOT	0.94 (28.6)	1.01 (20.3)	1.02 (23.3)	1.02 (29.1)	0.88 (24.0)	1.10 (19.0)	1.25 (19.5)	1.00 (22.8)
Ratio to baseline at follow-up	0.91 (28.9)	1.27 (33.2)	0.88 (20.7)	1.11 (22.1)	1.06 (43.5)	1.05 (27.7)	1.11 (24.9)	1.05 (36.0)
Other biochemistry
Potassium (mmol/L)
Baseline	4.2 (3.9)	4.5 (7.4)	4.4 (3.4)	4.5 (6.4)	4.3 (6.6)	4.4 (6.7)	4.4 (4.7)	4.4 (6.7)
EOT	4.2 (6.9)	4.1 (3.3)	4.2 (6.9)	3.9 (9.6)	3.8 (10.4)	3.8 (10.9)	3.6 (10.2)	4.1 (6.1)
Follow-up	4.4 (7.1)	4.4 (7.7)	4.4 (6.2)	4.2 (10.1)	4.2 (9.2)	4.2 (8.1)	4.3 (8.1)	4.4 (5.5)
Ratio to baseline at EOT	0.99 (6.9)	0.91 (9.8)	0.95 (6.9)	0.87 (7.0)	0.87 (12.1)	0.83 (11.5)	0.82 (11.1)	0.94 (7.2)
Ratio to baseline at follow-up	1.05 (4.9)	0.99 (9.2)	1.01 (6.1)	0.94 (6.2)	0.97 (10.7)	0.97 (10.3)	0.99 (10.1)	0.99 (5.4)

Data from safety analysis set consisting of all participants who were exposed to at least one dose of study drug. Baseline information is defined as the measurement at the latest assessment before first dosing. EOT was measured on day 85 (visit 15) and follow-up on day 110 (visit 19). All values are geometric mean (CV).

CV, coefficient of variation; EOT, end of treatment; hsCRP, high-sensitivity C-reactive protein.

In the MAD trial, treatment with NN1177 was associated with increased levels of the inflammatory biomarkers fibrinogen and high-sensitivity C-reactive protein (hsCRP), with the greatest increases observed with the higher dose levels (Table 3). An increase in fibrinogen levels was also seen in the DDI trial [26]. In the measures of liver function, there were increases in alanine aminotransferase (ALT) and aspartate aminotransferase (AST) observed in the 4200-μg and 6000-μg NN1177 cohorts vs placebo (Table 3); there were no noteworthy changes in bilirubin or alkaline phosphatase levels (data not shown). In the MAD trial, no effect of NN1177 was observed on plasma activated partial thromboplastin time or prothrombin time (data not shown); while in the DDI trial, the mean warfarin international normalised ratio increased following administration of the Cooperstown 5 + 1 index cocktail [26]. Although blood potassium levels decreased in participants treated with the four highest NN1177 dose levels in the MAD trial (Table 3), no participants had potassium levels requiring therapeutic intervention. There was no change seen in the MAD trial for blood sodium levels (data not shown); however, blood sodium levels decreased in the DDI trial [26]. An evaluation of outliers was performed as part of the standard evaluation of the study results, and no outliers of clinically meaningful relevance were identified.

Circulating levels of most amino acids decreased over the course of the DDI trial (Table S5).

3.3. Pharmacokinetic characterisation of NN1177

In the FHD/SAD trial, NN1177 was measurable in serum for a minimum of 168 h at all NN1177 dose levels and could be measured for more than 672 h after dosing at 700 and 1100 μg (Fig. S4A). Geometric mean AUC_0–inf increases were dose proportional across the dose range of 10 μg to 1100 μg, ranging from 27 to 2746 nmol∗h/L, and a similar increase in AUC_0–168h was observed (Table S6A). Geometric mean C_max increased from 0.2 to 13.7 nmol/L across the dose range of 10 μg to 700 μg, with no increase in C_max from the 700-μg to the 1100-μg dose level (Table 4A). Median t_max ranged from 33 to 45 h, and geometric mean of t_1/2 ranged from 76 to 87 h, with no apparent dose relationship in either (Table S6A). Estimated dose proportionality relationships for AUC_0–inf and C_max are shown in Fig. S5A; the estimates of 2β were 1.99 and 1.94 for AUC_0–inf and C_max, respectively (where a value of 2 corresponds to dose proportionality).

Table 4.

Oral glucose tolerance test in the multiple ascending dose trial.

Treatment contrasts from day 1 to day 85 (N1177 vs placebo)
N1177	Estimate [95% CI]	p-value
Glucose AUC0–2h(treatment ratioa)
200 μg	0.96 [0.82;1.11]	0.5479
600 μg	1.08 [0.93;1.27]	0.3063
1300 μg	1.03 [0.90;1.18]	0.6471
1900 μg	1.01 [0.87;1.16]	0.9066
2800 μg	1.14 [0.99;1.31]	0.0760
4200 μg	1.15 [0.99;1.33]	0.0617
6000 μg	1.41 [1.19;1.67]	0.0001
Glucose concentration at 2h(treatment differenceb), mmol/L
200 μg	−0.8 [–2.5;0.8]	0.3156
600 μg	−0.2 [–1.9;1.5]	0.8340
1300 μg	0.1 [–1.4;1.6]	0.9376
1900 μg	0.1 [–1.5;1.6]	0.9348
2800 μg	1.5 [–0.1;3.0]	0.0697
4200 μg	0.6 [–1.0;2.2]	0.4645
6000 μg	2.5 [0.7;4.4]	0.0077
Insulin AUC0–2h(treatment ratioa)
200 μg	1.30 [0.86;1.96]	0.2139
600 μg	1.60 [1.04;2.48]	0.0343
1300 μg	1.84 [1.24;2.72]	0.0030
1900 μg	1.87 [1.26;2.77]	0.0024
2800 μg	2.46 [1.62;3.75]	<0.0001
4200 μg	2.27 [1.53;3.36]	0.0001
6000 μg	2.69 [1.68;4.29]	<0.0001

Data from full analysis set, consisting of all participants who were randomised and received at least one dose of study drug.

AUC_0–2h, area under the NN1177 serum concentration–time curve from time 0 to 2 h; CI, confidence interval.

^aTreatment ratio: mean change in AUC_0–2h from day 1 to day 85 with each NN1177 dose/mean change in AUC_0–2h from day 1 to day 85 with placebo.

^bTreatment difference: mean change in glucose concentration from day 1 to day 85 with each NN1177 dose–mean change in glucose concentration from day 1 to day 85 with placebo. For glucose and insulin AUC_0–2h, the endpoint data were logarithmically transformed and analysed using an analysis of variance model with dose as a factor and logarithmically transformed baseline value as a covariate. For glucose concentration at 2 h, the endpoint was analysed using an analysis of variance model with dose as a factor and baseline value as a covariate.

In the MAD trial, geometric mean serum concentrations of NN1177 increased with increasing dose level (Fig. S4B). Geometric mean AUC_0–168 increased with increasing dose level, ranging from 559 to 27,959 nmol∗h/L (Table S6B). Geometric mean C_max at steady state (C_max,SS) increased from 5 to 239 nmol/L across the dose range of 200 μg to 6000 μg (Table 4B). The median t_max at steady state (t_max,SS) ranged from 30 to 48 h, and the geometric mean t_1/2 at steady state (t_1/2,SS) ranged from 77 to 111 h (Table S6B). Estimated dose proportionality relationships for AUC_0–168,SS and C_max,SS are shown in Fig. S5B; a minor deviance from dose proportionality was observed for AUC_0–168h,SS (estimate for 2β: 2.23; 95% confidence interval [CI] 2.10; 2.36) and C_max,SS (estimate for 2β: 2.21; 95% CI 2.08; 2.35).

3.4. Pharmacodynamics

3.4.1. Body weight

In the FHD/SAD trial, body weight decreased from day 1 to day 8 (7 days after dosing) in a dose-dependent manner in the three highest NN1177 dose groups (Fig. S6A); the maximum weight loss of 2.9% was seen at day 5 in the 1100-μg dose cohort.

Body weight in the MAD trial decreased from day 1 through end of treatment (day 85) in a dose-dependent manner with NN1177 (Fig. S7A). The placebo-adjusted estimated reductions in body weight ranged from 0.4% (200-μg dose) to 12.6% (4200-μg dose) and were statistically significant for the dose levels higher than 200 μg (p < 0.05). During the follow-up on cessation of NN1177, body weight increased in all NN1177 groups.

In the DDI trial, body weight decreased by 10.1% during treatment with NN1177 from day 1 to end of treatment at day 78 [26].

3.4.2. Glucose metabolism

There was no clear dose dependence in the change from baseline in fasting glucose or fasting insulin in the FHD/SAD trial (Fig. S6B and C).

In the MAD trial, fluctuations in fasting plasma glucose were observed throughout the treatment period with no dose dependency; statistical analysis of the treatment difference (NN1177 vs placebo) at end of treatment also showed no clear dose dependency (Fig. S7B). Similarly, fluctuations in fasting insulin were observed throughout the treatment period with no clear dose dependency; however, there was a statistically significant two-fold difference in the measured change in fasting insulin from baseline to end of treatment in the 6000-μg dose cohort compared with placebo (day 85) (Fig. S7C). There were no clinically relevant changes in HbA_1c at end of treatment across the cohorts (Fig. S7D). Fasting glucagon decreased from baseline to the follow-up visit for all NN1177 groups (Fig. S7E).

Mean glucose profiles from the OGTT conducted at day 1, day 29 and day 85 of the MAD trial are shown in Fig. S8A–C. A greater increase in the glucose AUC for 0–2 h (AUC_0–2h) from baseline to end of treatment was observed with the three highest N1177 dose levels, compared with the change in AUC_0–2h with placebo; however, statistically significance vs. placebo was shown only for the 6000 μg dose level (Table 4). Compared with the OGTT performed at day 1, there was an increase in glucose concentrations at 2 h in the OGTT at day 85 with the highest N1177 dose levels (data not shown); the changes were greater with NN1177 ≥1300 μg compared with placebo but statistically significant (vs placebo) for the 6000 μg dose level only (Table 4). Together, these indicate a deterioration in glucose tolerance, especially with the higher NN1177 dose levels. Furthermore, there were statistically significantly greater increases in the insulin AUC_0–2h from baseline to end of treatment with NN1177 ≥600 μg compared with placebo, with more than two-fold differences compared with placebo observed for the three highest NN1177 dose levels (Table 4), indicating an increase in insulin resistance.

3.4.3. Leptin

Fasting leptin levels in the MAD trial decreased with NN1177 and placebo between baseline and day 79 (data not shown); the reductions were statistically significantly greater with NN1177 ≥1300 μg than with placebo (Table S7). Between baseline and day 79, fasting soluble leptin receptors levels had increased with NN1177 ≥1300 μg and slightly decreased with placebo (data not shown); the increases with NN1177 ≥1900 μg were statistically significantly greater than the change with placebo (Table S7).

4. Discussion

Although treatment with NN1177 in the trials reported here led to dose-dependent significant weight loss compared with placebo, this was not accompanied by the expected improvements in glucose metabolism or blood pressure. Furthermore, several cardiovascular and general safety concerns were raised for the higher NN1177 doses in the MAD trial, which prompted a decision to discontinue clinical development of this glucagon/GLP-1 receptor co-agonist. Preclinical evaluation of NN1177 revealed a complex picture of variability of compound exposure and study length when determining the optimal receptor balance [25], and this may partly explain the impeded clinical success.

The beneficial weight loss with NN1177 was accompanied by adverse tolerability and safety findings, including an increased prevalence of GI TEAEs, which appeared to be dose-related from 600 μg. This is consistent with other pipeline glucagon/GLP-1 co-agonists which, despite beneficial weight loss, have also demonstrated increased GI AEs [19,30,31]. In the MAD trial, there was some deviance from dose-proportionality of weight loss with the higher NN1177 dose levels – a lower-than-expected estimated placebo-adjusted weight loss was observed with the maximum 6000-μg dose (9.1%) compared with the 4200-μg dose (12.6%); this observation may have partly been due to the modified dose escalation implemented to mitigate observed tolerability concerns.

These studies identified an increased heart rate (including blunted night-time dip) with NN1177. This was not entirely unexpected as GLP-1 RAs are known to have chronotropic effects [32,33]; however, the increase in heart rate for NN1177 exceeds that observed for GLP-1 RAs. It is unknown if the observed heart rate increase would be sustained for periods longer than those of the present trials or if NN1177 tachyphylaxis would lead to heart rate normalisation over time. However, it was concluded that the magnitude of the heart rate increase observed with NN1177 at clinically relevant dose levels was not acceptable. Furthermore, the blunting of the nocturnal dip in heart rate may be detrimental, as data have shown that such blunting is associated with preclinical cardiac damage (left atrial enlargement) and is predictive of cardiovascular morbidity and mortality in the general population [34].

Analysis of electrocardiograms in the MAD trial showed a statistically significant increase in QTcI interval with clinically relevant doses of NN1177. The observed QTcI interval prolongation may be an effect of glucagon agonism, because QT interval prolongation has not been observed for GLP-1 RAs [35,36]. However, the concentration–QTc slope estimate for NN1177 in the MAD trial should be interpreted with caution as the assumptions for the pre-specified linear effect model were not fulfilled.

A physiological role for glucagon in the regulation of amino acid metabolism has been suggested. In the liver, glucagon stimulates enzymes that promote ureagenesis and clearance of amino acids, reducing their plasma levels [37]. Glucagon receptor agonism may thereby influence amino acid metabolism via this pathway, and indeed, the DDI trial identified marked reductions in levels of most amino acids. Reductions in blood amino acid levels have previously been observed with glucagon treatment [38], and may suggest that amino acid homeostasis is disturbed, potentially reflecting increased hepatic gluconeogenesis (through glucagon stimulation) with amino acids as substrate. Such amino acid reductions are not seen at similar magnitude following bariatric surgery [39], but are observed in patients with glucagonoma, where there is excessive production of glucagon [40,41]. Studies have also shown that people with elevated circulating concentrations of glucagon over seven-fold higher than normal are at risk of protein wasting, hypoaminoacidaemia and weight loss [42]. Furthermore, in patients with critical illness, elevated blood glucagon promotes gluconeogenesis, which requires skeletal muscle amino acid utilisation, ultimately leading to muscle wasting and depressed blood amino acid levels [43]. The consequences of the reduction of amino acids in the DDI trial are not clear, and the degree of lean mass loss following chronic administration of glucagon/GLP-1 RA co-agonists is not yet fully known. Furthermore, the results from the DDI trial do not provide information on whether the decrease in amino acids seen with NN1177 is dose dependent, or if it is sustained with prolonged administration. Body composition was also not assessed in these trials, so any loss of lean mass with NN1177 is also unknown.

The clinical laboratory analyses also revealed some potentially clinically relevant adverse effects of NN1177 treatment. In the MAD trial, decreases in reticulocyte counts were observed, which could indicate bone marrow suppression [44]; however, although a direct inhibitory effect of glucagon on erythropoiesis has been observed [45], evidence is sparse. Increases in INR were seen when NN1177 and warfarin were co-administered, despite administration of vitamin K, and as noted in the full report of the DDI trial [26], the increase in mean INR is not explained by warfarin PK. Further, although the interaction between glucagon and warfarin has been described previously, it is not fully understood [46]. Fibrinogen and hsCRP levels also increased in the trials, suggesting proinflammatory effects of NN1177 treatment. With respect to fibrinogen, the clinical significance of the observed increases is unknown because there is conflicting evidence for causality between hyperfibrinogenaemia and, for example, cardiovascular disease [47]. In addition, increased levels of liver enzymes (AST and ALT) were noted; however, there were no concerning cases in specific participants and no clear evidence of hepatocellular injury. There is evidence to suggest that some liver enzymes may transiently increase during a diet-induced weight loss with the effect considered due to a transient deterioration in hepatic steatosis prior to improvement, and therefore a benign effect [48,49]. In contrast, however, a previous phase 2 trial with semaglutide in participants with obesity showed decreases in liver enzymes (ALT) and hsCRP [50]; indeed, GLP-1 RAs such as liraglutide and semaglutide are known to reduce inflammation as measured by hsCRP [51].

While the trial populations did not have T2D, NN1177 had no beneficial effect on glycaemic parameters in the MAD and DDI studies, either directly, or through the indirect benefits of weight loss. However, the parameters from the OGTT in the MAD study indicated impaired glucose tolerance at doses ≥2800 μg. This result is in contrast to observations with GLP-1 RAs, for which increased glucose tolerance and weight loss are observed in adults with obesity following treatment [15]. Detrimental effects, or even lack of effect, on glycaemic parameters would not be well received for weight-loss compounds that would be used in a population where the prevalence of pre-diabetes and T2D is high. In people with T2D and obesity, results from studies of other GLP-1/glucagon co-agonists have been inconsistent; JNJ-64565111 failed to demonstrate improvements in HbA_1c and was associated with increases in fasting plasma glucose and fasting insulin [52], whereas cotadutide improved glycaemic control [31]. Early phase studies using GLP-1/GIP/glucagon receptor triagonists in healthy participants (SAR441255), as well as those with T2D (NN0090-2746), have observed a reduction in blood glucose to hypoglycaemic levels but in the absence of clinical signs or symptoms of hypoglycaemia in some patients [53,54]. Meanwhile, phase 2 studies of retatrutide (LY3437943), a once-weekly GLP-1/GIP/glucagon receptor triagonist, have demonstrated substantial weight loss in people with overweight or obesity with ≥1 weight-related condition but without T2D, and both weight loss and improvements in glyceamic control in people with T2D and BMI 25–50 kg/m² [55,56]. Furthermore, alternate targets to the glucagon receptor are also being explored in combination therapies with GLP-1, as seen with the dual GLP-1/GIP agonist tirzepatide [17], and with amylin agonists which are under investigation for weight management, including CagriSema (a combination of the GLP-1 RA semaglutide and the amylin analogue, cagrilintide) in the phase 3 REDEFINE trials [57,58].

Although the potential for enhancing weight-loss efficacy of the established GLP-1 RA class through co-agonism of related proglucacon family peptides holds promise for the management of obesity, the studies presented here highlight the importance of ensuring that safety is preserved. The key to unlocking the optimal combination of agonism may lie in understanding the implications of relative pharmacological activity at each receptor, and the optimal balance may depend on the indication, i.e. weight loss or glycaemic control [59]. For instance, GLP-1 receptor agonism may need to be sufficiently high to mitigate hepatic glucose production that is stimulated by glucagon agonism. Alternatively, given that glycaemic benefits of glucagon have been observed in the fed, but not fasting, state [7], further investigation of dosing schedules in relation to mealtimes may yield further insights into optimising co-agonists for weight management and related outcomes.

5. Conclusions

NN1177, a glucagon/GLP-1 receptor co-agonist, was investigated for weight management and demonstrated dose-dependent, clinically relevant weight loss. However, the dose-dependent safety concerns observed with up to 12 weeks of treatment with NN1177 were considered too pronounced, and the decision was made to end clinical development of NN1177.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

Martin Friedrichsen, Lars Endahl, Frederik Kreiner, Søren Toubro and Sune Nygård are employees of Novo Nordisk A/S. Martin Friedrichsen, Lars Endahl and Frederik Kreiner are shareholders in Novo Nordisk A/S. Martin Kankam has received funding from Diffusion Pharmaceutical Inc., Grifols, Urovant Sciences, ViroDefense, Merck, PhaseBio Pharmaceuticals, Inc., Idorsia Pharmaceuticals Ltd, DynPort Vaccine company/FDA/NIH, and Aerovate Therapeutics. Ronald Goldwater is an employee of Parexel International.

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Data availability

Data will be made available on request.