figure a

Introduction

Methionine aminopeptidase 2 (MetAP2) inhibitors are a novel drug class being investigated for the treatment of type 2 diabetes, obesity and metabolic disease. In preclinical models of obesity and diabetes, MetAP2 inhibitors produce weight loss characterised by markedly reduced adiposity and increased glycaemic control, as well as transiently reduced food intake [1, 2]. The MetAP2 inhibitor beloranib has demonstrated consistent and substantial weight loss and glucose-lowering effects in clinical studies of general obesity, hypothalamic-injury-associated obesity and Prader–Willi syndrome (PWS) [3,4,5,6]. This phase 2 clinical trial is the first to study the effects of MetAP2 inhibition with beloranib compared with placebo on glycaemic control and body weight in individuals with type 2 diabetes and obesity.

Methods

This multicentre randomised double-blind placebo-controlled phase 2 trial was conducted at 16 investigational sites in Australia from December 2014 to January 2016. The trial was terminated early because of an imbalance of venous thromboembolism events in beloranib-treated vs placebo-treated individuals in beloranib clinical trials. The institutional review boards at all clinical sites approved the protocol before study initiation. All participants provided written informed consent. A sample size of 50 people per group was to provide 91% power to detect a mean difference of 5.0 kg assuming a standard deviation of 7.5 kg.

Eligible individuals aged 18–65 years with BMI ≥30 kg/m2 had type 2 diabetes with HbA1c 53–97 mmol/mol (7%–11%) and fasting plasma glucose <15.6 mmol/l and were treated with diet/exercise or a stable dose of non-insulin glucose-lowering medication. Participants were randomised in a 1:1:1 ratio via a centralised interactive web response system to placebo, 1.2 mg beloranib or 1.8 mg beloranib administered subcutaneously twice weekly (electronic supplementary material [ESM] Fig. 1). Randomisation was stratified based on sulfonylurea use (yes/no) and screening HbA1c (< or ≥75 mmol/mol [< or ≥9%]). Lifestyle (diet/exercise) counselling was provided to all participants.

The primary endpoint was the change in weight from baseline to week 26. Key secondary endpoints included the change from baseline to week 26 in glycaemic variables, the proportion of participants achieving weight loss targets and HbA1c targets and the change in cardiometabolic variables. Safety evaluations throughout the study included assessment of incidence and severity of adverse events (AEs) as well as changes in laboratory measures, physical examination, vital signs and other safety measures.

The primary efficacy analysis was conducted on the per protocol population (n = 66), which included all randomised participants who received ≥75% of planned doses, had an end-of-treatment efficacy assessment and did not have any notable protocol violations. Data were analysed using an ANCOVA model with change in weight as the dependent variable, treatment group, sulfonylurea use and baseline HbA1c as factors and baseline weight as a covariate without imputation of missing data. Analyses performed on the per protocol population used observed data. Safety analyses were conducted on the safety population. The safety and intent-to-treat populations (n = 152 each) included all randomised participants who received ≥1 dose of trial drug.

Data management and randomisation services were performed by BioClinica (Audubon, PA, USA) and statistical analyses were conducted by Agility Clinical (Carlsbad, CA, USA) and Pharmapace (San Diego, CA, USA). Data are available from the corresponding author on reasonable request.

Results

Participant disposition during the trial is shown in ESM Fig. 1. At the time of early study closure, 70 individuals had completed the study, 65 had not completed 26 weeks of treatment and 18 had discontinued the study (11 withdrew because of AEs, five withdrew consent, one was lost to follow up and one discontinued because of investigator decision).

Individuals in the intent-to-treat population were primarily white (89%), non-Hispanic (98%) and male (57%), with mean ± SD baseline age 54 ± 7 years, BMI 39.9 ± 7.3 kg/m2, HbA1c 67 ± 11 mmol/mol (8.3 ± 1.0%) and duration of diabetes of 7.1 ± 4.9 years. The majority (93%) of participants were treated with ≥1 glucose-lowering medication, most commonly biguanides (72%) and sulfonylureas (42%). There were no unexpected differences in demographics and baseline characteristics between the intent-to-treat and per protocol populations (ESM Table 1). Mean ± SD exposures to placebo, 1.2 mg and 1.8 mg beloranib were 20.7 ± 6.8 (median 23.6 [min 1.6, max 27.4]) weeks, 19.5 ± 7.7 (23.9 [1.6, 26.6]) weeks and 18.5 ± 8.3 (22.1 [0.4, 26.7]) weeks, respectively.

Weight loss with beloranib was observed as early as week 4 and progressed through to week 26 (Fig. 1a). At week 26, the least squares mean weight change (± SE) in the placebo group was −3.1 ± 1.2% compared with −13.5 ± 1.1% in the 1.2 mg beloranib group (placebo-corrected difference [95% CI]: −10.4% [−13.6, −7.2]) and −12.7 ± 1.3% in the 1.8 mg beloranib group (−9.6% [−12.9, −6.4]), both p < 0.0001) (ESM Table 2). Additionally, a substantially greater proportion of beloranib-treated participants achieved week 26 weight loss targets ≥5%, ≥10% and ≥15% compared with placebo (Fig. 1b).

Fig. 1
figure 1

(a) Least squares mean ± SE for the change from baseline in weight. (b) Percentage of participants with weight loss of ≥5%, ≥10% or ≥15% from baseline to week 26. (c) Least squares mean ± SE for the change from baseline in HbA1c. (d) Percentage of participants who achieved an HbA1c level of <53 mmol/mol (<7%) or ≤48 mmol/mol (≤6.5%) at week 26. (e) Least squares mean HbA1c by visit. Data reflect observed data for the per protocol population (n = 66). For (a, c, e), white circles, placebo (n = 22); white squares, 1.2 mg beloranib (n = 25); black squares, 1.8 mg beloranib (n = 19). For (b, d), white bars, placebo (n = 22); grey bars, 1.2 mg beloranib (n = 25); black bars, 1.8 mg beloranib (n = 19). The differences between the 1.2 mg and 1.8 mg beloranib groups vs placebo for the body weight and HbA1c efficacy endpoints were consistent and statistically significant (p < 0.05) regardless of the analysis population (per protocol or intent to treat) or imputation strategy (last observation carried forward [LOCF] or mixed-model repeated measures [MMRM]) used to explore the sensitivity of the analysis (data not shown). *p < 0.05, **p < 0.01 and ***p < 0.0001 for beloranib vs placebo

Full size image

At week 26, the least squares mean change in HbA1c (± SE) in the placebo group was −6.6 ± 2.2 mmol/mol (−0.6 ± 0.2%) compared with −21.9 ± 2.2 mmol/mol (−2.0 ± 0.2%) in the 1.2 mg beloranib group (placebo-corrected difference [95% CI]: −15.3 mmol/mol [−21.9, −9.8] [−1.4% (−2.0, −0.9)]) and −21.9 ± 3.3 mmol/mol (−2.0 ± 0.3%) in the 1.8 mg beloranib group (−15.3 mmol/mol [−21.9, −8.7] [−1.4% (−2.0, −0.8)]); both p < 0.0001 (Fig. 1c, ESM Table 2). Additionally, a greater proportion of individuals in the beloranib treatment groups achieved HbA1c levels <53 mmol/mol (<7%) and ≤48 mmol/mol (≤6.5%) at week 26 compared with placebo (Fig. 1d). The week 26 least squares mean HbA1c in both beloranib groups was 45 mmol/mol (6.3%) and the mean HbA1c in the placebo group was 61 mmol/mol (7.7%) (Fig. 1e).

Fasting plasma glucose was significantly reduced vs placebo in both beloranib groups by week 8 and continued to decline through week 26 in beloranib-treated individuals (ESM Fig. 2a). Beloranib also produced improvements in other glycaemic variables compared with placebo (postprandial glucose, insulin sensitivity and beta cell function) and cardiometabolic markers (high-sensitivity C-reactive protein [hsCRP], adiponectin and leptin; ESM Fig. 2b, c and ESM Table 2).

More participants in the beloranib groups withdrew from the study because of an AE (n = 5 per beloranib dose group) than in the placebo group (n = 2). The most common reasons for withdrawal of beloranib were sleep related (e.g. insomnia, abnormal dreams, sleep disorder). There were no deaths. The most common AEs with greater frequency in the beloranib groups compared with placebo were sleep related (e.g. abnormal dreams, sleep disorder), diarrhoea, injection-site bruising, injection-site erythema, lower respiratory tract infection and cough (Table 1). Most AEs were mild to moderate in severity and self-limiting. The incidence of hypoglycaemia AEs was low and similar across treatment groups (4–6%); all were mild and resolved. There were three AEs of thromboembolic nature (superficial thrombophlebitis, deep vein thrombosis, non-fatal pulmonary embolism); all occurred in beloranib-treated participants.

Table 1 AEs occurring from baseline to week 26
Full size table

Discussion

Here, we report that treatment with a MetAP2 inhibitor for 26 weeks produced placebo-corrected clinically meaningful weight loss (−10%) and reductions in HbA1c (−15.3 mmol/mol [−1.4%]) in individuals with obesity and type 2 diabetes, and most participants achieved clinical HbA1c and obesity treatment goals. Improvements in weight and HbA1c were similar for both beloranib doses, despite a higher baseline weight in the lower (1.2 mg) dose group, suggesting that the lower dose likely produced maximum efficacy.

The occurrence of unexpected serious venous thromboembolism AEs across beloranib clinical trials led to termination of beloranib development and investigation of the mechanism. Recently published data show that beloranib has effects on endothelial cells that influence pro- and anticoagulant factors [7]. Furthermore, prolonged (>24 h) exposure, resulting from unique pharmacological properties of beloranib [7] as well as the suspension formulation, likely exacerbates these effects. Importantly, other MetAP2 inhibitors were recently shown to have similar beneficial effects on body weight and glycaemic control, but with a substantially improved safety profile compared with beloranib [7].

In addition to reductions in body weight and HbA1c, improvements in waist and hip circumference, fat mass, lipids, hsCRP, leptin and adiponectin (ESM Table 2) are consistent with previous observations of beloranib [3,4,5,6] and likely result from rapid weight loss as well as other weight-independent effects of MetAP2 inhibition. MetAP2 inhibitors are hypothesised to reduce body weight by increasing fat mobilisation and oxidation [2] and reducing food intake—beloranib produces a marked but transient reduction in food intake in preclinical studies [2] and improves measures of hunger and prospective food intake in obese individuals [3,4,5] and hyperphagia in PWS [6]. However, preclinical studies of obese pair-fed mice demonstrate that the decrease in food intake does not completely account for the reduction in body weight produced by MetAP2 inhibition [1]. Although weight loss of the magnitude documented in this study is sufficient to substantially improve glycaemic control [8], weight-independent effects of MetAP2 inhibition also likely play a key role. For example, MetAP2 inhibition attenuates activity of extracellular signal regulated kinase (ERK1/2) [9] and downstream factors such as retinoic acid-related orphan receptor α (RORα) that influence the development of insulin resistance, inflammation and other metabolic disorders [10].

The current observations with MetAP2 inhibition demonstrate a potential new pathway to impact obesity and type 2 diabetes and support further investigations of this pathway in metabolic disease. Another MetAP2 inhibitor with an improved safety profile is in clinical development [7].