Alberto Vegas Rodriguez1, Nieves Velez de mendizabal2, Iñaki F. Troconiz1,4,5, Justin Feigelman3
1Department of Pharmaceutical Sciences, School of Pharmacy and Nutrition, University of Navarra, Pamplona, Spain. , 2Eli Lilly and Company. Formerly Gilead Sciences., 3Gilead Sciences, , 4Navarra Institute for Health Research (IdiSNA), 5Institute of Data Science and Artificial Intelligence (DATAI)
Introduction Human immunodeficiency virus (HIV) remains a global health challenge, with up to 75% of infected individuals undergoing antiretroviral therapy (ART) [1,2]. Understanding the complex interplay between virus and the host immune system represents an ideal niche to establish predictive models that integrate the dynamics of both with drug pharmacokinetics (PK) and pharmacodynamics (PD). Such quantitative models are instrumental for the successful development of new therapeutics, including establishing optimal dosing strategies. However, many available models require some parameters to be fixed to ensure identifiability which may limit the assessment of patient variability and dose individualization [3,4]. Objectives We sought to develop a fully identifiable, population semi-mechanistic pharmacokinetic/pharmacodynamic (popPKPD) model integrating drug PK, and CD4+ T cell and viral RNA longitudinal data obtained from clinical trials of the antiretroviral drugs Lenacapavir, Bictegravir, Tenofovir Alafenamide, Emtricitabine or Elvitegravir, given as monotherapy to people living with HIV. Methods Data was collected from six phase 1 clinical trials and combined for model development using NONMEM 7.5.1 through a sequential approach. First a PK model was developed, and then empirical Bayes estimates of the PK model parameters were used to describe the dynamics of viral load and CD4+ T cells. Viral dynamics were captured by expanding the “basic” viral model to include two additional infected cell populations: chronically and latently infected cells. The model assumes that, at treatment initiation, viral and immune cells are in steady state with CD4+ T cells comprising T, I, C, and L cell types. Drug effects were incorporated following known mechanisms of action of each antiretroviral agent. Model exploration was performed through deterministic simulations in mrgsolve to predict the antiviral efficacy of different dosing strategies. Results Baseline levels of biomarkers varied significantly among individuals but were consistent across studies. No notable placebo effects were observed. Most patients showed a marked decline in viral titers following treatment initiation, a trend that continued through the treatment period. Weakly-informative priors applied to fixed-effect parameters resulted in precise model parameter estimates, with the death rates of infected cells (dI=1.14d?¹, dC=0.03d?¹, dL=2.3·10?4d?¹) and virus (dV=26.9d?¹) agreeing with established data. The estimated mean baseline HIV-RNA concentration was 43,460 copies/mL, with a 141% inter-individual variability (IIV). The baseline CD4+ T cell count was 410 cells/µL, roughly 20 times higher than the concentration of infected cells. Simulations indicated that time to reach an undetectable state (50 copies/mL) ranged from three to six months, while the duration of viral rebound varied between two and four weeks. No clear correlation was observed between in vivo and in vitro EC50 values. Conclusions The popPKPD model effectively captures key aspects of viral dynamics, drug behaviour, and treatment response without the need for fixed parameters. Estimated parameters align with HIV pathogenesis [5,6]. EC50 estimates varied among drugs and exceeded in vitro predictions, a pattern observed in other studies [7]. Our study is the first to provide in vivo EC50 estimates for five antiretroviral drugs based on clinical trial data, highlighting its pioneering approach and contribution to understanding HIV treatment dynamics.
1. Global HIV Programme [Internet]. [cited 2023 Dec 15]. Available from: https://www.who.int/teams/global-hiv-hepatitis-and-stis-programmes/hiv/strategic-information/hiv-data-and-statistics 2. HIV Treatment: The Basics | NIH [Internet]. [cited 2023 Dec 15]. Available from: https://hivinfo.nih.gov/understanding-hiv/fact-sheets/hiv-treatment-basics 3. Nowak MA, May RM. Virus Dynamics: Mathematical Principles of Immunology and Virology. Oxford: Oxford University Press; 2000. 16–43 p. 4. Gonçalves A, Mentré F, Lemenuel-Diot A, Guedj J. Model Averaging in Viral Dynamic Models. AAPS J. 2020 Mar 13;22(2):48. 5. Perelson AS, Kirschner DE, De Boer R. Dynamics of HIV infection of CD4+ T cells. Math Biosci. 1993;114(1):81-125. doi:10.1016/0025-5564(93)90043-A 6. Mendoza P, Jackson JR, Oliveira TY, et al. Antigen-responsive CD4+ T cell clones contribute to the HIV-1 latent reservoir. Journal of Experimental Medicine. 2020;217(7). doi:10.1084/jem.20200051 7. Fang J, Jadhav PR. From in vitro EC50 to in vivo dose–response for antiretrovirals using an HIV disease model. Part I: A framework. J Pharmacokinet Pharmacodyn. 2012;39(4):357-368. doi:10.1007/s10928-012-9255-3
Reference: PAGE 33 (2025) Abstr 11730 [www.page-meeting.org/?abstract=11730]
Poster: Drug/Disease Modelling - Infection