(M-002) Population Pharmacokinetic Modeling of MEDI7352, a Bispecific Fusion Protein for Dual Targeting of NGF and TNFα, in Patients with Painful Diabetic Neuropathy and Painful Osteoarthritis of the Knee

Ahmed Salem – BioPharmaceuticals R&D, Clinical Pharmacology and Quantitative Pharmacology, AstraZeneca, Gaithersburg, MD, USA – AsztraZeneca; Kirsten Scott – BioPharmaceuticals R&D, Neuroscience, AstraZeneca, Cambridge, UK – AstraZeneca; Edward Emery – BioPharmaceuticals R&D, Neuroscience, AstraZeneca, Cambridge, UK – AstraZeneca; Fraser Welsh – BioPharmaceuticals R&D, Neuroscience, AstraZeneca, Cambridge, UK – AstraZeneca; Iain Chessell – BioPharmaceuticals R&D, Neuroscience, AstraZeneca, Cambridge, UK – AstraZeneca; Keith Tan – BioPharmaceuticals R&D, Neuroscience, AstraZeneca, Cambridge, UK – AstraZeneca; Lindsay Clegg – BioPharmaceuticals R&D, Clinical Pharmacology and Quantitative Pharmacology, AstraZeneca, Gaithersburg, MD, USA – AstraZeneca; Weifeng Tang – BioPharmaceuticals R&D, Clinical Pharmacology and Quantitative Pharmacology, AstraZeneca, Gaithersburg, MD, USA – AstraZeneca; Daniel Kaschek – IntiQuan AG - A ProductLifeGroup Company, Basel, Switzerland. – IntiQuan AG; Isabelle Pouliquen – BioPharmaceuticals R&D, Clinical Pharmacology and Quantitative Pharmacology, AstraZeneca, Cambridge, UK – AstraZeneca; Meng Li – BioPharmaceuticals R&D, Clinical Pharmacology and Quantitative Pharmacology, AstraZeneca, Gaithersburg, MD, USA – AstraZeneca

Objectives: Nerve growth factor (NGF) and tumor necrosis factor (TNF)-α are key mediators in nervous system sensitization and pain pathophysiology [1,2]. MEDI7352, a bispecific fusion protein targeting both NGF and TNFα, has shown synergistic efficacy in preclinical pain models and is under development for chronic pain management. We aimed to develop a population pharmacokinetic (PK) model of MEDI7352 in painful diabetic neuropathy (PDN) and osteoarthritis patients, identifying covariates affecting its exposure with emphasis on immunogenicity due to high incidence associated with MEDI7352.

Methods: Longitudinal MEDI7352 serum concentrations were collected from three placebo-controlled studies: a Phase I study (NCT02508155) of single ascending doses (0.3–1000 μg/kg intravenous, 50 μg/kg subcutaneous) and multiple ascending doses (1–450 μg/kg intravenous biweekly) in osteoarthritis population, a Phase IIa study (NCT03755934) (5–450 μg/kg intravenous biweekly) in PDN population, and a Phase IIb study (NCT04675034) (7.5–150 mg subcutaneous biweekly) in osteoarthritis population. PK model development and validation were conducted using non-linear mixed-effect modeling. Baseline demographic characteristics (weight, age, sex, race, albumin), anti-drug antibody (ADA) titers and status, biomarker (C-reactive protein), and disease type were tested as covariates. Covariate selection was done based on biological plausibility, graphical inspection, and statistical significance (95% confidence interval not crossing the null value). Simulations with the final model assessed covariate effects on PK parameters and predicted steady-state exposure for future dosing regimens.

Results: 4470 PK samples from 413 subjects were included in PK modeling. ADA incidence ranged from 64% to 81%, with onset around day 14 post-first dose, and persistence over time. Exploratory analysis showed that ADA formation significantly impacted MEDI7352 exposure without clear effect on efficacy and safety. MEDI7352 PK was best characterized by a one-compartment model with first-order absorption and elimination. Covariates included body weight on clearance and volume of distribution (Vd), and ADA titer on clearance. The ADA impact on clearance was modeled through a combination of a titer-dependent Emax function to describe the magnitude of ADA impact on exposure and a time-dependent sigmoidal function to describe the time course of ADA impact. Disease type and other covariates had no significant effect on MEDI7352 PK. Parameter estimates for a typical subject of 84 kg and negative ADA status were 1.1 L/day for clearance, 5 L for Vd, and 54% for subcutaneous bioavailability. Simulations showed maximum ADA effect on exposure by week 12 post first dose.

Conclusions: A PK model of MEDI7352 was developed using pooled data from phase I & II studies, including both intravenous and subcutaneous administrations. Body weight and ADA status (titer) were significant covariates impacting PK parameters. An innovative approach was used to model immunogenicity impact on MEDI7352 PK. The model was used to guide MEDI7352 dose selection and derisk immunogenicity's effect on exposure for future studies.

Citations: [1] Clarke, J. Going for gold to improve anti-NGF therapy in OA. Nat Rev Rheumatol 17, 508 (2021). https://doi.org/10.1038/s41584-021-00670-7
[2] Miller RE, Block JA, Malfait AM. What is new in pain modification in osteoarthritis? Rheumatology (Oxford). 2018 May 1;57(suppl_4):iv99-iv107. doi: 10.1093/rheumatology/kex522. PMID: 29361112; PMCID: PMC5905627.

Keywords: Bispecific, Pain, Immunogenicity