Xuanlin Liu1, Maisam Jasinga1, Lalitya M. Sudarsono1, J. G. Coen van Hasselt1, Anne-Grete Märtson1
1Leiden Academic Centre for Drug Research, Leiden University
Introduction Cytomegalovirus (CMV) is a herpesvirus which is one of the major complications in solid organ and stem cell transplant patients. Ganciclovir and its prodrug valganciclovir are the standard treatments for CMV infection. Despite their efficacy, these antiviral agents have a very narrow therapeutic window due to high myelotoxicity. Mechanistic modeling is a powerful tool to gain thorough understanding of the CMV viral dynamics, which builds the framework for the identification of PK-PD relationship and further clinical translation. Although the in vivo viral dynamics of CMV have been studied, previous published models have relied solely on viral DNA profiles, which indicates the amount of both infectious and non-infectious virus instead of infectious virus only. So far, no in vitro CMV models have been built and no viral titer (i.e. infectious virus) data has been reported from either in vitro or in vivo studies. In this study, both viral DNA (i.e., both infectious and non-infectious virus) and viral titer were characterized, and their relationship was elucidated through the development of a mechanistic viral dynamic model of in vitro CMV infection. Methods Experimental design: MRC5 cells and HCMV Towne strain were used for the viral infection. Experiment 1: 1.5 x 104 cells were seeded with (MOI = 0.1) and without the presence of virus. Cell counts were measured at day 1, 3, 5, 10, 14 and 17. Experiment 2: 1.5 x 106 cells were infected with HCMV (MOI = 0.1) with no treatment and inoculated for 2 hours. After the incubation the cells were washed and replenished with tissue culture medium. Viral supernatants were collected serially for 25 days over the course of infection. Viral load burden was measured using qPCR assay to determine the viral DNA (copies/mL) and using 50% tissue culture infectious dose (TCID50) assay to determine the infectious viral titer (PFU/mL). Model development: The viral dynamic model development was carried out using NONMEM (v.7.5). First, natural cell growth rate of the susceptible cell counts (S) and death rate (d) of infected cells (I) were derived from the cell counts data of the uninfected and infected group. Next, a target cell-limited (TCL) model [1] was applied to characterize the viral dynamics by fitting both the viral DNA and viral titer data. d were fixed to the final estimates from the previous fitting. Key viral kinetic parameters estimated included viral infection rate (ß), cell growth rate (?), production rate of viral DNA (p) and viral clearance (?). A linear model, a power-law model and a saturation model [2] were tested to describe the relationship between total (V) and infectious virus (Vi). The estimation was performed using the first-order conditional estimation with interaction (FOCEi) method. The final model was evaluated by comparing the predictions and observations. Results CMV viral load burden was quantified by two methods in this study: total viral DNA (copies/mL) from qPCR and the amount of infectious virus (PFU/mL) from TCID50 assay. The fitting results showed that the natural cell growth curve can be best described by an exponential saturated function of which the natural growth rate was 0.77 day-1. When the cells were infected, the proliferation of susceptible cells drastically reduced by 100-fold, decreasing from 0.77 to 0.0077 day-1 with a death rate of infected cells of 0.74 day-1. A TCL model well captured the viral load profile with ß = 4.97*10^(-8) copy-1/day, ? = 0.0077 day-1, p = 1.81*105 copy/cell/day, ? = 0.03 day-1. The relationship between V and Vi can be best characterized using a power-law model: Vi = f(V) = B*V^hill, where B = 0.037 and hill = 0.871, indicating each new DNA copy came with approximately 0.03 infectious virion. Conclusions This is the first known viral kinetic model of in vitro CMV infection that characterizes the dynamics of CMV as well as the relationship between viral DNA and viral titer. This model gives a better understanding of the infectivity of CMV. The final estimates showed good agreement with the published in vivo CMV infection model, providing an ideal basis to study the in vitro PK-PD of CMV and ganciclovir with the potential to translate to a clinical setting afterwards.
[1] Duke, Elizabeth R et al. “Mathematical Modeling of Within-Host, Untreated, Cytomegalovirus Infection Dynamics after Allogeneic Transplantation.” Viruses vol. 13,11 2292. 16 Nov. 2021, doi:10.3390/v13112292 [2] Iyaniwura, Sarafa A et al. “The kinetics of SARS-CoV-2 infection based on a human challenge study.” Proceedings of the National Academy of Sciences of the United States of America vol. 121,46 (2024): e2406303121. doi:10.1073/pnas.2406303121
Reference: PAGE 33 (2025) Abstr 11547 [www.page-meeting.org/?abstract=11547]
Poster: Drug/Disease Modelling - Infection