Engineering of a GLP-1 analogue peptide/anti-PCSK9 antibody fusion for type 2 diabetes treatment

Abstract

Type 2 diabetes (T2D) is a complex and progressive disease requiring polypharmacy to manage hyperglycaemia and cardiovascular risk factors. However, most patients do not achieve combined treatment goals. To address this therapeutic gap, we have developed MEDI4166, a novel glucagon-like peptide-1 (GLP-1) receptor agonist peptide fused to a proprotein convertase subtilisin/kexin type 9 (PCSK9) neutralising antibody that allows for glycaemic control and low-density lipoprotein cholesterol (LDL-C) lowering in a single molecule. The fusion has been engineered to deliver sustained peptide activity in vivo in combination with reduced potency, to manage GLP-1 driven adverse effects at high dose, and a favourable manufacturability profile. MEDI4166 showed robust and sustained LDL-C lowering in cynomolgus monkeys and exhibited the anticipated GLP-1 effects in T2D mouse models. We believe MEDI4166 is a novel molecule combining long acting agonist peptide and neutralising antibody activities to deliver a unique pharmacology profile for the management of T2D.

Figure 1

Dual activity of GLP-1 receptor agonist peptide and anti-PCSK9 antibody fusions. (a) Schematic representation of GLP-1 analogue peptide anti-PCSK9 antibody fusion. (b) Potency at human GLP-1R in CHO transfected cells of the 27 tested peptide antibody fusions in the cAMP accumulation assay using anti-PCSK9 Ab#1 (pink), Ab#2 (orange), Ab#3 (khaki), Ab#4 (green), Ab#5 (blue), Ab#6 (purple) and Ab#7 (grey). Potency of the reference compound GLP-1-Fc(γ4) is shown with a dashed red line. The red arrow shows the fusion chosen for in-depth characterisation. (c) Relative PCSK9 binding of the 27 tested peptide antibody fusions in the biochemical epitope competition assays compared to the parental anti-PCSK9 mAbs. A relative binding of 1, shown with a dashed red line, is corresponding to the parent antibody EC₅₀. (d) Surface plasmon resonance (SPR) sensorgrams for binding of recombinant human PCSK9 to Ab#2 (black) and peptide antibody fusion Ab#2_GLP1 (orange). The graph illustrates data of a typical experiment done 8 times independently. (e) Uptake of fluorescently labelled LDL-cholesterol by HepG2 cells treated with 45 nM of recombinant human PCSK9 and Ab#2 (black), peptide antibody fusion Ab#2_GLP1 (orange) or isotype control mAb (grey), (n = 2) (f) Activation of human GLP-1R as % of maximum GLP-1(7–36) amide peptide response in CHO transfected cells using cAMP accumulation assay following treatment with Ab#2_GLP1 (orange), GLP-1-Fc(γ4) (pink), GLP-1(7–36) amide peptide (green) or isotype control mAb (grey), (n = 2). For (b) and (c), values are shown as mean of n = 2. For (e) and (f), values are presented as mean (±SD) of a typical experiment performed independently at least in duplicate.

Figure 2

Stability engineering of GLP-1 analogue peptides in fusion with anti-PCSK9 Ab#2. Human Fc concentration (pink) and GLP-1R active compound concentration (orange) following single intravenous administration in C57BI/6 mice (n = 3 per time-point) of (a) Ab#2_GLP1 at 5 mg/kg; (b) Ab#2_EX4 at 1 mg/kg; (c) N-glycosylated GLP-1 analogue W25N/V27S Ab#2 fusion at 40 mg/kg; (d) disulphide bridge stabilised variant Ab#2_DSB#1 at 10.8 mg/kg. Values are presented as mean (±SEM). Concentration of the active compound at 168 h post-dose for Ab#2_EX4 was not reported as it was below the lower limit of quantification of the assay.

Figure 3

Optimising aggregation profile of disulphide bridge exendin-4 in anti-PCSK9 fusion. (a) Chromatogram profile of preparative size exclusion for Ab#2_DSB#1 compound following protein-A purification. (b) Structure (PBD 3C5T) of exendin-4 (cyan for the α-helix and orange for the tryptophan cage) in complex with the N-terminal domain of human GLP-1R (green) showing the targeted residues (pink) for rational design of peptide antibody fusion with low aggregation propensity. (c) Percentage aggregate in Protein-A purified samples of disulphide bridge variants in fusion with Ab#2 compared to exendin-4 fusion (Ab#2_EX4) − n = 1. Colour code corresponds to the cysteine mutation in the α-helix of exendin-4: D9C (pink), A18C (orange), L21C (khaki) and W25C (green). Fusions chosen for further analysis are shown with a blue arrow. (d) Human Fc concentration (pink) and GLP-1R active compound concentration (orange) following single intravenous administration in CD rats (n = 3) of Ab#2_DSB#7 at 10 mg/kg. For (d), values are presented as mean (±SEM).

Figure 4

Optimising GLP-1R potency of disulphide bridge DSB#7 peptide in antibody fusion. (a) Split view model of DSB#7 analogue peptide (cyan) in complex with the N-terminal domain of human GLP-1R (green) showing in orange the targeted residues for rational design of peptide antibody fusion with reduced potency at human GLP-1R. Receptor residues predicted to interact with the targeted peptide amino acids are shown in yellow. The disulphide bridge between the peptide α-helix and C-terminus is shown in red. (b) Activity at human GLP-1R expressed in CHO cells using the cAMP accumulation assay for DSB#7 peptide variants with single point amino acid mutation at G2 (pink), E15 (orange), V19 (green), I23 (khaki) and L26 (blue) in fusion with the optimised antibody Ab#2.1. Reference compound GLP-1-Fc(γ4) is shown as a purple cross. Values are shown as mean of n = 2. Fusion molecules at target human GLP-1R potency are shown in a grey square. (c) Human Fc concentration following single intravenous administration in CD rats (n = 3 per compound) of Ab#2.1 in fusion with DSB#7 peptide variant G2V (pink) at 60 mg/kg, E15A (orange) at 53 mg/kg, V19A (green) at 58.5 mg/kg and L26I (blue) at 60 mg/kg. (d) Human Fc concentration (pink) and GLP−1R active compound concentration (orange) following single subcutaneous administration at 60 mg/kg in CD rats (n = 3) of the fusion Ab#2.1_DSB#7_V19A. For (c) and (d), values are presented as mean (±SEM).

Figure 5

Glucose control and weight loss induced by MEDI4166 in mouse models of diabetes and obesity. (a) Fasted blood glucose levels in male DIO mice (n = 10/group) following a single subcutaneous administration of anti-PCSK9 control antibody Ab#2.1 (10 mg/kg), MEDI4166 (10 mg/kg) or Ab#2.1_EX4 (10 or 30 mg/kg). Values are presented as mean (±SEM). In all cases, *p < 0.05; ***p < 0.001; ****p < 0.0001 compared to control mAb; ^$$p < 0.01; ^$$$$p < 0.0001 MEDI4166 compared to 10 mg/kg Ab#2.1_EX4; ^p < 0.05; ^^^p < 0.001; ^^^^p < 0.0001 MEDI4166 compared to 30 mg/kg Ab#2.1_EX4; ^##p < 0.01 30 mg/kg Ab#2.1_EX4 compared to 10 mg/kg Ab#2.1_EX4. (b) Fasted blood glucose levels in male diabetic db/db mice (n = 12/group) following weekly subcutaneous administration of vehicle (PBS) or MEDI4166 (3, 10 or 30 mg/kg) for 4 weeks. Values are presented as mean (±SEM);. **p < 0.05; ***p < 0.001; ****p < 0.0001 compared to vehicle. ^$p < 0.05; ^$$p < 0.01; ^$$$p < 0.001 3 mg/kg MEDI4166 compared to 10 mg/kg MEDI4166. ^p < 0.05; ^^p < 0.01; ^^^p < 0.001; ^^^^p < 0.0001 3 mg/kg MEDI4166 compared to 30 mg/kg MEDI4166. (c) Body weight change of DIO mice (n = 8/group) following once weekly subcutaneous administration of vehicle (PBS) or MEDI4166 (3, 10 or 30 mg/kg) for 26 days. Values are presented as mean ( ± SEM); ****p < 0.0001 for MEDI4166 (30 mg/kg) vs. vehicle; *p < 0.0001 for MEDI4166 (10 mg/kg) vs. vehicle.

Figure 6

MEDI4166 anti-PCSK9 activity and exposure profiles in lean cynomolgus monkey following single injection. (a) Change in plasma LDL-C and (b) free PCSK9 concentration following MEDI4166 intravenous administration at 1 (khaki) or 5 (green) mg/kg or subcutaneous administration at 10 (orange) or 100 (pink) mg/kg. (c) Human Fc concentration (dark blue) and GLP-1R active compound concentration (light blue) following intravenous administration of MEDI4166 at 5 mg/kg or (d) subcutaneous administration at 100 mg/kg. n = 3 animals per group. Values are presented as mean (±SEM). Concentration of GLP-1R active compound after 9 days and Fc concentration after 24 days for MEDI4166 at the 5 mg/kg (iv) dose were not reported as levels were below the lower limit of quantification of the assays.