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

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.


Aggregation of peptide antibody fusions
Compounds in 25 mM histidine buffer, 205 mM sucrose (pH 6.0) were concentrated using Amicon Ultra-15 centrifugal filters to achieve a target concentration of 50 mg/mL. Samples were then spiked with 0.02% Polysorbate 80, filtered on 0.2 m membrane, and 1 mL aliquots placed at 5°C for 4 weeks in type I borosilicate glass vials. Aggregation rates were then determined by size exclusion chromatography on a TSK gel G3000 SWXL 7.8 X 300 mm column after diluting the samples to 10 mg/mL in PBS and filtering on a 0.45 m membrane.

Affinity parameters for PCSK9 binding
Kinetic and affinity parameters for the binding of antibody and peptide antibody fusions to recombinant PCSK9 were determined by surface plasmon resonance (SPR) using a Biacore 2000 biosensor (GE Healthcare). An anti-human IgG surface was first created using the Human Antibody Capture Kit and CM5 sensor chip (GE Healthcare). Tested compounds at 1 nM were then flowed at 10 µL/min for 3 min. Human, cynomolgus (both in house) or rat PCSK9 (SinoBiological) diluted in running buffer (10 mM sodium phosphate (pH 7.4), 150 mM sodium chloride, 1 mg/mL BSA and 0.05% Tween 20) were injected for 10 min followed by a 10 min dissociation time period. Global dissociation rates were first calculated followed by global on-rate calculations both using a 1:1 binding kinetics model.

Biochemical PCSK9/LDL receptor inhibition assay
Biotinylated human PCSK9 (in-house) was incubated for 2 h with serial dilution of tested compounds in MaxiSorb plate (NUNC) coated with human LDL receptor (R&D Systems). Bound PCSK9 was detected using cryptate labelled streptavidin (Perkin Elmer) diluted at 100 ng/mL in Delfia Buffer (Perkin Elmer). Fluorescence signal was read on a Perkin Elmer Envision machine using a 340 nm excitation and 620 nm emission. Percentage of specific binding was calculated by subtracting the background signal obtained with no LDL receptor coated onto the plate normalized with the maximum specific binding signal obtained with no competitor compound minus background level.

Activity at human receptors closely related to GLP-1R
Activity of MEDI4166 at glucagon, glucagon-like peptide-2 (GLP-2), secretin and gastric inhibitory polypeptide (GIP) receptors was assessed in the cAMP accumulation assay previously described using stable cell lines expressing the different receptors (CHO derived cells for glucagon, GLP-2 and secretin receptors and HEK derived cells for GIP receptor). Specific agonist peptides for each of the four receptors, all purchased from Bachem, were used as positive controls: human glucagon peptide, human GIP peptide, human GLP-2 peptide and human secretin peptide.

Fat and fat free mass
Body composition (fat mass, fat-free mass) was determined using an EchoMRI instrument (Echo Medical Systems, Houston, TX).

Clinical chemistry parameters
Triglycerides, total cholesterol, and ALT from terminal serum samples were determined using an Olympus bioanalyzer (Olympus America Diagnostics, Center Valley, PA). Insulin concentrations were assessed via ELISA (MesoScale Discovery, Rockville, MD).

PKPD modelling and simulation of GLP-1 analogue peptide α-PCSK9 fusion
A two compartments human PKPD model was built using Berkeley-Madonna (version 8.3.14) based on the diagram described below: The pharmacokinetic profile of the peptide antibody fusion was assumed to be similar to anti-PCSK9 human IgG molecules. Antibody PK and PCSK9 parameters were obtained from the literature More specifically, bioavailability after subcutaneous administration was set to 80%, rate of absorption to 0.3 day -1 , volumes of the central and peripheral compartments to 2 and 3.47 L respectively, flow rate between central and peripheral compartments to 0.33 L/day and central compartment clearance to 0.21 L/day. Plasma PCSK9 concentration was assumed to be 3.4 nM with a clearance rate of 2.76 day -1 . Affinity of the compound to human PCSK9 was set to 0.6 nM with an off-rate of 9.7 day -1 . As the clearance of human PCSK9 is reported to be faster than that of a typical IgG, target turnover likely drives the clearance of the mAb/PCSK9 complex which has therefore been set up to 0.276 day -1 .
The pharmacokinetic profile of the benchmark GLP-1-Fc(4) was determined using parameters  GLP-1 analogue peptide Linker Antibody light chain identify site specific degradants generated by cleavage within the T1 and T2 peptides. Increase in the relative abundance of T1* a , T1* b and T2* peptides post in vivo exposure revealed (d) cleavage after G2, S12 and W25.  Supplementary Table 1