Annabelle Lemenuel-diot1, Selma El Messaoudi2, Jeremie Guedj2, Sylvie Retout1, Franziska Schaedeli-Stark1, Amine Ait-Ali1, Nelson Guerreiro1, Sijie Lu3, Farouk Chughlay1, Claire McGeown1, Rémi Kazma1
1F. Hoffmann La roche, 2Université Paris Cité, INSERM, 3Roche R&D Center China
Introduction: Existing mechanistic models of chronic hepatitis B infection (CHB) focus on a subset of biomarkers that do not capture the full complexity of CHB processes. Therefore, these models cannot adequately describe the various modes of action of novel therapeutics aiming to achieve functional cure in patients with CHB. To address this gap, the complexity of the HBV viral life cycle has been incorporated into a novel mechanistic viral dynamics modeling (MVDM) framework.Using this model, the source of response variability due to different individual characteristics was evaluated and simulations showed their influence on model parameters and ultimately on efficacy estimation. Methods: Leveraging the Piranga Phase 2 platform study (NCT04225715), the MVDM framework (see figure) was developed to fit jointly the dynamics of the full pharmacokinetic and biomarker datasets. Data originated from 188 nucleos(t)ide analogue-treated participants randomized into one of 6 arms with different combination regimens of xalnesiran, a GalNAc-conjugated small interfering RNA targeting HBsAg transcripts, with or without an immunomodulator (ruzotolimod, pegylated interferon a, or PD-L1 LNA). This MVDM framework included a pool of infected hepatocytes producing viral transcripts at different rates, either from cccDNA (ps1), or from integrated HBV sequences (ps2). The death rate of infected hepatocytes (d) was modelled primarily using ALT data. A model selection process estimated the effect of each molecule and the influence of individual characteristics, such as ethnicity or the HBV genotype, on the observed differences in efficacy. Using the MVDM framework, simulations predicted the in silico performance of different combination regimens tested in Piranga in populations with different distributions of these individual characteristics. Results: The effect of xalnesiran was estimated as an inhibition of HBsAg production from both origins (ps1 and ps2) and an acceleration of hepatocytes death rate (d) by a factor 4.24. Furthermore, the model estimated the add-on effect of each immunomodulator mainly through an enhancement of the d parameter (increased by 70%, 5% and 58% with peg-IFN-a, ruzotolimod,and PDL-L1 respectively) and, specifically for pegylated interferon a, an increase in anti-HBs production rate.The investigation of individual characteristics influencing the effect of xalnesiran revealed that the HbsAg Baseline as well as the HBV genotype play a role in modulating its effect. Using the MVDM framework, simulations of different combination regimens tested in Piranga showed how variations of the HBV genotype impacted efficacy estimations depending on the potential underlying geographic distribution. In addition, using the model to challenge the treatment combination, simulations did not predict any substantial increase in rates of HBsAg loss (HBsAg lower than 0.05 IU/mL) when simulating changes to the dose, duration, or regimen of the different combinations of molecules tested in Piranga. Conclusion: This novel MVDM framework improves the characterisation of the complex long-term dynamics of viral parameters during and after treatment with different combination regimens. Furthermore, it allowed the detection and evaluation of the influence of the individual characteristics on the efficacy of the different combinations. Optimisation of treatment combinations can therefore be supported by the model, paving the way to potential in silico efficacy prediction of new anti-HBV agents.
Malinovsky JM. et al. Anesth. Analg. 2000;91:1457-602 Djerada Z.et al. Clin. Pharmacokinet. 2018;57(9) :1135-47 Olofsen E. et al. Anesthesiology.2008;109(4):664-74
Reference: PAGE 33 (2025) Abstr 11531 [www.page-meeting.org/?abstract=11531]
Poster: Drug/Disease Modelling - Infection