Abstract

​

Serum hepatitis B virus (HBV) kinetics in urokinase-type plasminogen activator/severe combined immunodeficient (uPA-SCID) mice reconstituted with humanized livers from inoculation to steady state is highly dynamic despite the absence of an adaptive immune response. We developed a stochastic agent-based model that includes virion production cycles in individual infected human hepatocytes. The model was calibrated using a genetic algorithm approach with the serum HBV kinetics observed in mice inoculated with 108 HBV genome equivalents and fit the data well when the following viral production parameters were assumed: (1) An eclipse phase lasting 5-50 hours and (2) a post-eclipse phase production rate that is based on increasing production cycles initially starting with a long production cycle of 1 virion per 20 hours that gradually reaches 1 virion per hour after approximately 3-4 days before virion production increases dramatically to reach to a steady state production rate of 4 virions per hour per cell. The model was then validated by showing it could accurately simulate the viral kinetics observed with lower HBV inoculation doses (104-107 genome equivalents) in which similar, but delayed patterns were observed. Together, modeling suggests that it is the cyclic nature of the virus lifecycle combined with an initial slow but increasing rate of HBV production from each cell that plays a role in generating the observed multiphasic HBV kinetic patterns in humanized mice.