Alessandro Boianelli (1), Jacqueline Naylor (2), David Hornigold (2), Pär Nordell (1), Monika Sundqvist (1)
(1) DMPK, Research and Early Development Cardiovascular, Renal and Metabolism, BioPharmaceuticals R&D, AstraZeneca, Gothenburg, Sweden, (2) Bioscience Metabolism, Research and Early Development, Cardiovascular, Renal and Metabolism (CVRM), BioPharmaceuticals R&D, AstraZeneca, Cambridge, UK
Objectives: GLP-1 receptor agonists (GLP-1RAs) represent a compound class that play an important role in type 2 diabetes treatment. As target-related adverse events such as nausea and vomiting typically limits the therapeutic window, optimal design of early proof of concept (POC) trials is critical for efficient clinical development of new GLP-1RAs. While efficacy related to treatment regimen across key endpoints has been studied for this class of molecules (1), an in vitro–in vivo potency correlation analysis is still lacking. Main objectives of our work are:
- Derive a general quantitative exposure response relationship for the clinically available GLP-1RAs
- Establish an empirical relationship between target cover from in vitro potency and incidence of nausea in long term studies
Methods: We collected human pharmacokinetics data or human pharmacokinetic parameters for six approved GLP-1RA molecules to be able to simulate the pharmacokinetic profiles at different dosing regimens. In addition, we collected clinical efficacious exposures and in vitro potency from different cell-based systems including endogenous GLP-1R expressing EndoC-βH1 in the presence of 0.1% bovine serum albumin (EndoC0.1%BSA), CHO cells recombinantly expressing GLP-1R in the presence of 0% albumin (CHO0%HSA), and 4.4% human serum albumin (CHO4.4%HSA). Based on this data, we evaluated the correlation between target coverage, expressed as a ratio between steady-state exposure and unbound potency, body weight or HbA1c reduction in T2D patients, using published dose-response relationships (1). Unbound in vitro potency was calculate by the method in (2). The total ratio was used for the CHO4.4%HSA ratio.
Results: Based on efficacy data from all evaluated compounds, general models of bodyweight and HbA1c reduction in relation to target cover over assay dependent in vitro EC50 were derived. The model based on CHO0%HSA data showed excellent agreement with observations, while residual variability was larger using EndoC0.1%BSA and CHO4.4%HSA derived potencies. The observed in vitro-in vivo correlation was highly variable and in vitro assay dependent. More specifically in vivo EC50s for HbA1c and body weight reduction were estimated to be 4.4-fold (69% CV) and 214-fold (13% CV) times the in vitro CHO0%HSA EC50, 84-fold (47% CV) and 1.54-fold (24.8% CV) times the in vitro CHO4.4%HSA, and 5.10-fold (22.5% CV) and 0.035-fold (37.6 % CV) times the in vitro EndoC0.1%BSA, respectively. This large shift indicates that an approximate 50 to 135 fold higher concentration is required for effect on body weight compared to HbA1c reduction. Furthermore, an empirical relationship showing a good trend between target cover from CHO0%HSA potency and incidence of nausea in long term studies (48-156 weeks) was established. The direct response model indicated that exposures below 50-fold in vitro CHO0%HSA EC50 result in less than a 20% incidence of nausea, which is comparable to placebo. However exposures over CHO0%HSA, efficacious on body weight, were in the range increasing the nausea incidence between 30% to 50% based on the model.
Conclusions: Overall, the established relationships to efficacy and safety help increase confidence in human dose predictions and trial design for new GLP-1RAs in discovery and early clinical phases.
References:
[1] Maloney et al. A model-based meta-analysis of 24 antihyperglycemic drugs for type 2 diabetes: comparison of treatment effects at therapeutic doses.” Clinical Pharmacology & Therapeutics 105.5 (2019): 1213-1223.
[2] Wan et al. High Throughput Screening of Drug-Protein Binding in Drug Discovery, Journal of Liquid Chromatography & Related Technologies,(2017), 30:5-7, 681-700,
Reference: PAGE 29 (2021) Abstr 9669 [www.page-meeting.org/?abstract=9669]
Poster: Drug/Disease Modelling - Endocrine