Venkatesh Pilla Reddy1, Oleg Stepanov2, Asfiha Gebre3, Isabelle Pouliquen1, Kirill Zhudenkov2, and Rosalin Arends3
1Clinical Pharmacology and Quantitative Pharmacology, R&D, AstraZeneca, Cambridge, United Kingdom; 2M&S Decisions, Moscow, Russia; 3Clinical Pharmacology and Quantitative Pharmacology, R&D, AstraZeneca, Gaithersburg, United States
Introduction: AZD7442, a combination product of 2 mAbs (AZD8895 and AZD1061), is being studied for the prophylaxis and treatment of mild to moderate and severe COVID-19. AZD8895 and AZD1061 are SARS CoV-2-specific neutralizing mAbs that bind to distinct neutralizing epitopes on the RBD of the S protein of SARS CoV-2. AZD7442 has been engineered to provide half-life extension compared to conventional antibodies and the combination of two antibodies reduces the risk of resistance developed by the SARS-CoV-2 virus.
Objectives: To build a population-based PK model to describe the time course of AZD7442 exposure and identify covariates that can explain the inter-subject PK variability.
Methods: The population-based PK model was built using AZD7442 Phase I study (D8850C00001; NCT04507256) serum concentration profiles from different dose levels (300 mg to 3000 mg) after single dose administration either as an IV or IM injection [1,2]. A nonlinear mixed effects modeling approach (NONMEM version 7.4.4 or greater, ICON, Elicott City, MD) was used for the PK model development. R (www.r project.org, R 4.0.2) was used for data visualization. Exploratory graphical analysis of the concentration time data suggested a monophasic elimination profile and a 2 compartmental model with first order absorption and first order elimination for AZD8895, AZD1061, and AZD7442 PK following IV and IM administration described the data adequately. Inter-individual variability was modelled using an exponential form. Residual variability was modelled using an additive plus proportional error form. The predictive performance of the final base PK model for AZD8895, AZD1061 and AZD7442 was assessed via prediction corrected visual predictive checks. Covariates available from this study including age, body weight, sex, race and parameters related to hepatic and kidney function were explored as possible statistically significant variables by using automated stepwise covariate search. PK simulations were performed to predict and confirm the dose required for long-term protection and for treating COVID-19.
Results: AZD7442 concentration-time profiles were well described by a 2-compartment model with first order absorption and first order elimination. Overall, the population PK analysis supports that the pharmacokinetics of the two mAbs in AZD7442 are very similar and that these 2 mAbs follow linear kinetics with a long terminal t1/2. Bodyweight on CL, Q as fixed allometric coefficient of 0.75 and on V2 and V3 as fixed allometric coefficient of 1 resulted in the reduction of the objective function value by 90, 58 and 42 units for AZD8895, AZD1061 and AZD7422 respectively. Therefore, body weight was included as a structural covariate in the base PK model with these fixed exponents. Pop-PK model-based simulations suggest that AZD7442 drug concentrations remain above the minimal protective serum target concentration long-term.
Conclusion: Overall, the population PK analysis supports that the pharmacokinetics of AZD8895 and AZD1061 are very similar and that these 2 mAbs follow linear kinetics with a long AZD7442 terminal t1/2. The extended t1/2 should allow for prolonged maintenance of efficacious exposure
References:
[1] https://www.astrazeneca.com/media-centre/press-releases/2021/us-supply-agreement-for-additional-azd7442-doses.html
[2] Van Erp EA et al. Front. Immunol. 2019, https://doi.org/10.3389/fimmu.2019.00548.
Reference: PAGE 29 (2021) Abstr 9843 [www.page-meeting.org/?abstract=9843]
Poster: Drug/Disease Modelling - COVID-19