Mechanistic PK/PD modeling of the simultaneous effects of bepirovirsen on hepatitis B surface antigen (HBsAg) and ALT in participants with chronic hepatitis B infection to support Phase 3 study design

Objectives: Bepirovirsen, an antisense oligonucleotide that targets all hepatitis B virus (HBV) RNAs, reduced hepatitis B surface antigen (HBsAg) levels in Phase 2 studies in participants with chronic HBV infection. Transient increases in alanine aminotransferase (ALT) were often observed concurrent with HBsAg reduction.[1] The objectives of this work were to develop a mechanistic pharmacokinetic/pharmacodynamic (PK/PD) model to describe the effects of bepirovirsen on HBsAg and ALT, identify predictors of response, and to perform model-based simulations to inform Phase 3 study design. Methods: The models for PK, HBsAg, and ALT were developed sequentially using data from Phase 2a (GSK 205695/NCT02981602; n=31) and 2b (GSK B-Clear 209668/NCT04449029; n=455) studies in participants with chronic HBV infection. Population modeling was performed using NONMEM interfaced with the Finch Studio NONMEM workbench.[2] Model simulations were performed in R using mrgsolve. A three-compartment PK model was used to describe the pharmacokinetics of bepirovirsen. An indirect-response model with transit compartments and inhibition on production was used to develop the PK-HBsAg model. The PK-HBsAg-ALT model was adapted from a general model for drug induced ALT release [3], and altered to incorporate mechanisms of ALT release driven by direct and indirect (driven by reduction in HBsAg) effects of bepirovirsen on hepatocyte cell death. Results: The model accurately predicted the time-course of changes in bepirovirsen PK, HBsAg, and ALT levels. Baseline HBsAg was a significant predictor of response to bepirovirsen treatment (covariate on IC50). In simulations of predicted response versus baseline HBsAg, lower baseline HBsAg increased the estimated percentage of participants achieving a sustained HBsAg suppression off-treatment. Age (as a continuous covariate) and concomitant nucleos(t)ide therapy were statistically significant covariates on baseline HBsAg level. Model results indicated that ALT increases were primarily driven by HBsAg reduction rather than a direct drug effect. Conclusion: In conclusion, this semi-mechanistic PK/PD model describing the relationships between bepirovirsen PK, HBsAg, and ALT provides key insights into bepirovirsen efficacy in treatment of chronic HBV infection and key predictors of response. Overall, simulations using this model support the use of bepirovirsen 300 mg per week for 24 weeks and the enrolment of participants with baseline HBsAg ≤3000 IU/mL in Phase 3 clinical trials to maximize the benefit of bepirovirsen treatment in Phase 3 studies. Funding: GSK (studies 205695, 209668)Citations: [1] Yuen MF, et al. Efficacy and Safety of Bepirovirsen in Chronic Hepatitis B Infection. N Engl J Med. [2] Ismail MHI, et. al. Finch studio: a next generation NONMEM modeling workbench. [3] Li J, et. al. A general model for cell death and biomarker release from injured tissues. J Pharmacokinet Pharmacodyn.