Selma El Messaoudi[1], Antonio Gonçalves[1], Annabelle Lemenuel-Diot[2], Jérémie Guedj [1]
[1] IAME, Université de Paris, INSERM, F-75018 Paris, France. [2] Roche Pharmaceutical Research and Early Development, Pharmaceutical Sciences, Roche Innovation Center Base
Objectives: Current treatments for chronic HBV (CHB) infection rely on nucleosid analogs (NUCs) and pegylated interferon (peg-IFNα). These treatments suppress viral replication but do not eradicate HBV and thus need to be taken lifelong[1]. Following the successes obtained against HCV infection, important efforts are being made to develop anti-HBV agents that can achieve a sustained cure. Because these new drugs will likely be used in combination with the current standard of care, it is essential to understand the effects of NUC and/or peg-IFNα on viral dynamics. In the past, models inspired from HCV dynamics have been developed to understand the determinants of early viral dynamics, but these models failed to predict the long term effects of NUC and/or Peg-IFN. In addition, given the complexity of HBV lifecycle[2], it is essential to model not only viral load, but also the other virus transcripts, such as Hepatitis B e antigen (HBeAg) and Hepatitis B surface antigen (HBsAg)[3][4].
This work aimed to fill this gap and to develop a modeling framework to characterize the long-term dynamic of the main HBV transcripts, relying on a very rich data set of >1,000 individuals treated with the nucleoside analog lamivudine and/or peg-IFNα.
Methods: A total of 1330 patients, HBeAg-positive and HBeAg-negative patients, from 2 randomized, multi-arms and multicentric clinical trials were included in the analysis[5][6]. Patients received either peg-IFNα monotherapy, lamivudine monotherapy or peg-IFNα plus lamivudine for 48 weeks. HBV DNA, HBsAg and for positive patients, HBeAg levels were sampled during the whole analysis. Sex, age, BMI, race and the level of ALTs at baseline were
informed.
Two viral dynamics models (one for each HBeAg status) were developed to fit the dynamics of the viral transcripts in patients treated with drug monotherapy. Our modeling framework included two populations of infected cells, that produced viral transcripts at different rates, representing population of cells that are either transcriptionally very active (called I1), or less transcriptionally active (called I2), but remain infected. A model selection based on a corrected LRT[7] was performed to include significant drug effects and baseline covariates were included. Once the best model based on monotherapies was selected, we used the parameters estimated to predict virologic outcomes with the combination based on 500 replicates of the dataset, using several drug-drug interactions models.
Results: At baseline, we predicted a majority of I2 for both HBeAg-positive (median=99.29%) and HBeAg-negative patients (median=93.53%). The race was included as a covariate on the production of HBsAg by I2 cells in both population. The levels of ALT at baseline could partly explain the variability and allowed to account for the natural history of the disease. The effect of lamivudine on blocking viral production was estimated at 99,4% and 99,9% in HBeAg negative and positive patients, respectively. In contrast the effect of peg-IFNα was lower, with values equal to 92.4% and 94.3% in HBeAg negative and positive patients, respectively. The immunomudulator role of peg-IFNα was highlighted in our model, by an enhancement of the death rates of I1 (×1.7) and I2 (×7.6) with HBeAg positive patients, and an enhancement of the loss of I2 (×7.9) in HBeAg-negative patients. In both populations, peg-IFNα had a significant effect on the transition rate from I1 to I2, δtr (×2.1 and ×4.6 respectively in HBeAg-negative and HBeAg-positive patients). The model was able to reproduce the proportion of undetectable HBV DNA at the end of treatment (observed = 75.5%, median predicted = 72.3% in HBeAg-positive patients, and observed = 94.8% and median predicted =93.6% in HBeAg-negative patients), as well as the decline of HBsAg and HBeAg.
Conclusions: This mechanistic model provides a new framework to improve the understanding of the complex long-term dynamic of chronic HBV infection under standard of care therapy. It will support the optimization of combination therapy with new anti-HBV agents in future analyses.
References:
[1] Anna S. Lok et al. Hepatitis B cure: From discovery to regulatory approval. Hepatology (Baltimore, Md.), 66(4):1296–1313, October 2017. [2] Antonio Gonçalves et al. What drives the dynamics of HBV RNA during treatment? Journal of Viral Hepatitis, 28(2):383–392, February 2021. [3] Margaret T. Chen et al. A function of the hepatitis B virus precore protein is to regulate the immune response to the core antigen. Proceedings of the National Academy of Sciences of the United States of America, 101(41):14913–14918, October 2004. [4] Michelle Martinot-Peignoux et al. HBsAg quantification: useful for monitoring natural history and treatment outcome. Liver International 34 Suppl 1:97–107, February 2014. [5] George K. K. Lau et al. Peginterferon Alfa-2a, lamivudine, and the combination for HBeAg-positive chronic hepatitis B. The New England Journal of Medicine, 352(26):2682–2695, June 2005. [6] Patrick Marcellin et al. Peginterferon alfa-2a alone, lamivudine alone, and the two in combination in patients with HBeAg-negative chronic hepatitis B. The New England Journal of Medicine, 351(12):1206–1217, September 2004. [7] Solène Desmée et al. Nonlinear Mixed-effect Models for Prostate-specific Antigen Kinetics and Link with Survival in the Context of Metastatic Prostate Cancer: A Comparison by Simulation of Two-stage and Joint Approaches. The AAPS journal, 17(3):691–699, May 2015.
Reference: PAGE 30 (2022) Abstr 10221 [www.page-meeting.org/?abstract=10221]
Poster: Drug/Disease Modelling - Infection