BI 456906: Discovery and preclinical pharmacology of a novel GCGR/GLP-1R dual agonist with robust anti-obesity efficacy

Abstract

Objective: Obesity and its associated comorbidities represent a global health challenge with a need for well-tolerated, effective, and mechanistically diverse pharmaceutical interventions. Oxyntomodulin is a gut peptide that activates the glucagon receptor (GCGR) and glucagon-like peptide-1 receptor (GLP-1R) and reduces bodyweight by increasing energy expenditure and reducing energy intake in humans. Here we describe the pharmacological profile of the novel glucagon receptor (GCGR)/GLP-1 receptor (GLP-1R) dual agonist BI 456906.

Methods: BI 456906 was characterized using cell-based in vitro assays to determine functional agonism. In vivo pharmacological studies were performed using acute and subchronic dosing regimens to demonstrate target engagement for the GCGR and GLP-1R, and weight lowering efficacy.

Results: BI 456906 is a potent, acylated peptide containing a C18 fatty acid as a half-life extending principle to support once-weekly dosing in humans. Pharmacological doses of BI 456906 provided greater bodyweight reductions in mice compared with maximally effective doses of the GLP-1R agonist semaglutide. BI 456906's superior efficacy is the consequence of increased energy expenditure and reduced food intake. Engagement of both receptors in vivo was demonstrated via glucose tolerance, food intake, and gastric emptying tests for the GLP-1R, and liver nicotinamide N-methyltransferase mRNA expression and circulating biomarkers (amino acids, fibroblast growth factor-21) for the GCGR. The dual activity of BI 456906 at the GLP-1R and GCGR was supported using GLP-1R knockout and transgenic reporter mice, and an ex vivo bioactivity assay.

Conclusions: BI 456906 is a potent GCGR/GLP-1R dual agonist with robust anti-obesity efficacy achieved by increasing energy expenditure and decreasing food intake.

Keywords: G protein coupled receptor; Glucagon; Glucagon-like peptide-1; Obesity; Peptide.

Figure 1

Structural properties and in vitro profile of BI 456906. (a) BI 456906 was designed based on a modified glucagon sequence, incorporating an unnatural amino acid and a glycine–serine linker in position 24, containing a C18 di-acid. (b) BI 456906 potency at the human GCGR and GLP-1R to stimulate cAMP in cells with recombinant (CHO-K1) and endogenous (MIN6, hepatocytes) receptor expression. The functional potency at the GCGR was confirmed by inhibition of glycogen synthesis in rat hepatocytes. (c) Potency of BI 456906, glucagon, GLP-1, oxyntomodulin and semaglutide at the human GCGR and GLP-1R based on the luciferase induction in CRE-Luc cells in the presence of 0.5% and 100% mouse or human plasma. (d–f) Glucose-stimulated insulin secretion after treatment with GLP-1 (1 nM), BI 456906, or liraglutide (0.01, 0.1, 1, 3, 10, 30, or 100 nM) in (d) mouse, (e) rat, and (f) perifused human pancreatic islets. Results are shown as insulin secretion expressed as a percentage of total insulin ±SEM. CRE-Luc, cAMP response element-luciferase; EC50, half maximal effective concentration; GCGR, glucagon receptor; GLP-1, glucagon-like peptide-1; GLP-1R, glucagon-like peptide-1 receptor; IC50, half maximal inhibitory concentration; OXM, oxyntomodulin; pEC50, negative log EC50; pIC50, negative log IC50.

Figure 2

Glucagon-like peptide-1 receptor engagement of BI 456906 to reduce food intake and gastric emptying and to improve glucose tolerance in lean animals after single dosing. The effects of BI 456906 and semaglutide are shown on food intake (a–c), glucose tolerance (d–f), and gastric emptying (g and h). Food intake is shown as percentage of vehicle food intake 24 h after dosing with (a) BI 456906 or (b) semaglutide in wild-type mice, and (c) in GLP-1R knockout mice. Glucose tolerance tested by intraperitoneal glucose administration after dosing with (d) BI 456906 or (e) semaglutide in wild-type mice and (f) in GLP-1R knockout mice. Gastric emptying was tested by acetaminophen concentration after ingestion of an acetaminophen-glucose bolus in wild-type mice after dosing with (g) semaglutide and h) BI 456906. (i) Summary of acute potencies (ED₅₀) of BI 456906 compared with semaglutide. Data are shown as mean ± SEM. Statistical analysis was done using one-way ANOVA followed by Dunnett’s method for multiple comparisons versus vehicle with significance defined at ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. APAP, acetaminophen; GLP-1R, glucagon-like peptide-1 receptor; iGTT, intraperitoneal glucose tolerance test; oGTT, oral glucose tolerance test.

Figure 3

BI 456906 demonstrates significantly greater bodyweight-lowering efficacy compared with maximally effective doses of semaglutide in diet-induced obese mice by engaging the glucagon receptor. Mice received repeated SC dosing of vehicle, semaglutide, or BI 456906 for 30 days. (a) Percentage change in bodyweight from baseline to Day 28. (b) Effects of BI 456906 and semaglutide vs vehicle on food intake to Day 28. Concentrations of (c) liver triglycerides, d) liver cholesterol, (e) plasma cholesterol, (f) plasma triglycerides, (g) plasma ALT, and (h) plasma AST at Day 28. Plasma levels of (i) ghrelin, (j) insulin, (k) leptin, (l) glucagon, and m) FGF-21 at Day 28. (n) The nM exposures and ex vivo bioactivity of BI 456906 and semaglutide displayed as relative nM activity based on (o) the EC₅₀ for the human GLP-1R and GCGR determined in 100% mouse plasma. Data are shown as mean ± SEM, with individual data shown. Statistical analysis was done using one-way ANOVA followed by Dunnett’s method for multiple comparisons versus vehicle with significance defined at ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 vs vehicle; ^#p < 0.05 vs 20 nmol/kg semaglutide. ALT, alanine aminotransferase; AST, aspartate aminotransferase; FGF-21, fibroblast growth factor-21; GCGR, glucagon receptor; GLP-1, glucagon-like peptide-1; GLP-1R, glucagon-like peptide-1 receptor; hGCGR, human glucagon receptor; hGLP-1R, human glucagon-like peptide-1 receptor; SC, subcutaneous.

Figure 4

Effects of BI 456906 on glycemia and energy expenditure in diet-induced obese mice. DIO mice received chronic repeated dosing of vehicle (BID), BI 456906 (QD), semaglutide (QD), or LA-GCG (BID) for 10 days. (a) Change in bodyweight from baseline. Plasma levels of (b) glucose, (c) glucagon, and (d) active ghrelin over 10 days of treatment. Energy expenditure over 10 days in DIO mice treated with daily dosing of (e) 5 or 10 nmol/kg BI 456906 or (f) 20 nmol/kg BI 456906. (g) Data in panel e have been analyzed separately per day using a general linear model with energy expenditure as a dependent variable, a baseline observation (Day 1 measurement prior to dosing), treatment as a fixed factor, and the covariate bodyweight to compare different dose levels of the test item with negative control (vehicle). Degrees of freedom were calculated according to Kenward and Roger, and one-sided pairwise comparisons with vehicle were performed using a 5% level of significance. Significance level at Day 8 represents the treatment effect of the BI 456906 10 nmol/kg group from the ANCOVA analysis. BID, twice daily; DIO, diet-induced obese; EE, energy expenditure; LA-GCG, long-acting glucagon; QD, once daily.

Figure 5

Modulation of gene expression after dosing with BI 456906 and semaglutide. Dose-dependent gene expression modulation by BI 456906. (a) Heat map of differential gene expression (93 gene set; log2 fold change >1 or <−1, and FDR <0.05) based on different levels of BI 456906 (3, 10, 20, and 30 nmol/kg) vs the control vehicle, semaglutide (20 and 100 nmol/kg) vs BI 456906 30 nmol/kg, and the control vehicle alone. Each row represents a gene, each column represents a sample, and the color code shows the dose-dependent Z-score of TPM (deviation from a gene’s mean expression in standard deviation units). (b–n) BI 456906 compound-specific dose-dependent effect on genes involved in (b and c) the methionine (GCGR biomarkers), (c–f) GPCR signaling, (g and h) oxidative phosphorylation, (i–k) cholesterol metabolism, and (l–n) cell cycle and differentiation. The middle line denotes median TPM. FDR, false discovery rate; GCGR, glucagon receptor; GPCR, G protein-coupled receptor; TPM, transcripts per million.

Figure 6

Plasma and transcriptional markers of glucagon receptor activation by BI 456906. Amino acid regulation by BI 456906. (a–g) Plasma amino acid levels at Day 28 after subchronic dosing with vehicle, BI 456906, or semaglutide in DIO mice. Data are shown as mean ± SEM, with individual data shown. (h–m) The dose-dependent regulation of genes encoding amino acid-metabolizing enzymes by BI 456906 versus semaglutide. The middle line denotes median transcripts per million (TPM) (n–p) Correlation of transcriptional changes in Sds (n), Gls2 (o), and Got1 (p) mRNA with plasma levels of serine (Sds) and glutamine (Gls2 and Got1) after treatment with BI 456906, vehicle, or semaglutide. Significant negative correlation (p < 0.0001, Spearman) was only observed in BI 456906 treatment, but not in semaglutide. R2 values are represented on the plots. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. DIO, diet-induced obese; TPM, transcripts per million.

Figure 7

Effects of BI 456906 on glucagon receptor markers in lean mice and on luciferase activity in target tissues of glucagon-like peptide-1 receptors and glucagon receptors in CRE-Luc mice. Activation of GCGRs and GLP-1Rs by BI 456906 in lean and in CRE-Luc mice. (a) Liver Nnmt mRNA expression after single dosing with vehicle, semaglutide or BI 456906. Plasma levels of (b) glutamine and (c) serine after single dosing with vehicle, semaglutide or BI 456906. Data are shown as mean ± SEM, with individual data shown. (d–m) CRE-Luc mice were dosed with BI 456906, semaglutide and LA-GCG and imaged over 12 h. (d–f) Left lateral imaging of mice pre-dose, 4 h, 8 h, and 12 h after dosing with 100 nmol/kg, (d) semaglutide, (e) BI 456906, or (f) LA-GCG. (g) Ex vivo tissue imaging of the liver (top row) and pancreas (bottom row). (h) Bioluminescence signal for ex vivo tissue imaging. (i–k) Ex vivo determined x-fold induction of luciferase activity in the pancreas and liver after treatment with (i) semaglutide, (j) BI 456906, or (k) LA-GCG versus vehicle and isoproterenol. (l and (m) Liver Nnmt mRNA expression in CRE-Luc mice after treatment with (l) BI 456906 or (m) semaglutide versus vehicle and isoproterenol. Data are shown as mean ± SEM. Statistical analysis was done using one-way ANOVA followed by Dunnett’s method for multiple comparisons versus vehicle with significance defined at ∗∗p < 0.01, ∗∗∗p < 0.001. CRE-Luc, cAMP response element-luciferase; GCGR, glucagon receptor; GLP-1R, glucagon-like peptide-1 receptor; Iso, isoproterenol; LA-GCG, long-acting glucagon; NNMT, nicotinamide N-methyltransferase.

Figure 8

Gene expression patterns of differentially expressed genes in diet-induced obese mice and humans with non-alcoholic fatty liver disease. To find BI 456906 dose-dependent genes associated with NAFLD, we compared current study clusters with previously published transcriptomic data on NAFLD by Pantano et al. [46]. (a) Gene clusters that were upregulated over NAFLD fibrosis stages in the Pantano study. (b) Similar gene clusters as in (a) were downregulated with increasing BI 456906 treatment in diet-induced obese mice. (c) The overlap between the clusters was significant (p < 0.0001, hypergeometric test); the Z-score represents the scaled transformation of normalized counts. (d) Gene clusters that were downregulated over NAFLD fibrosis stages in the Pantano study. (e) Similar gene clusters as in (d) were upregulated with increasing BI 456906 treatment. (f) The overlap between the clusters was significant (p < 0.0001, hypergeometric test); the Z-score represents the scaled transformation of normalized counts. (g) Schematic for the working hypothesis of the actions of BI 456906 at GCGRs in the liver, based on transcriptional changes and cluster analyses. GCGR, glucagon receptor; NAFLD, non-alcoholic fatty liver disease; NASH, non-alcoholic steatohepatitis.