Immune-viral dynamics modeling for SARS-CoV-2 drug development

Abstract

Coronavirus disease 2019 (COVID-19) global pandemic is caused by severe acute respiratory syndrome-coronavirus 2 (SARS-CoV-2) viral infection, which can lead to pneumonia, lung injury, and death in susceptible populations. Understanding viral dynamics of SARS-CoV-2 is critical for development of effective treatments. An Immune-Viral Dynamics Model (IVDM) is developed to describe SARS-CoV-2 viral dynamics and COVID-19 disease progression. A dataset of 60 individual patients with COVID-19 with clinical viral load (VL) and reported disease severity were assembled from literature. Viral infection and replication mechanisms of SARS-CoV-2, viral-induced cell death, and time-dependent immune response are incorporated in the model to describe the dynamics of viruses and immune response. Disease severity are tested as a covariate to model parameters. The IVDM was fitted to the data and parameters were estimated using the nonlinear mixed-effect model. The model can adequately describe individual viral dynamics profiles, with disease severity identified as a covariate on infected cell death rate. The modeling suggested that it takes about 32.6 days to reach 50% of maximum cell-based immunity. Simulations based on virtual populations suggested a typical mild case reaches VL limit of detection (LOD) by 13 days with no treatment, a moderate case by 17 days, and a severe case by 41 days. Simulations were used to explore hypothetical treatments with different initiation time, disease severity, and drug effects to demonstrate the usefulness of such modeling in informing decisions. Overall, the IVDM modeling and simulation platform enables simulations for viral dynamics and treatment efficacy and can be used to aid in clinical pharmacokinetic/pharmacodynamic (PK/PD) and dose-efficacy response analysis for COVID-19 drug development.

FIGURE 1

Target‐cell limited viral dynamics model for SARS‐CoV‐2 viral infection and disease severity. I_max, maximum unbound systemic concentration; PK, pharmacokinetic; SARS‐CoV‐2, severe acute respiratory syndrome‐coronavirus 2

FIGURE 2

Viral dynamics profiles in different disease severity levels. SARS‐CoV‐2, severe acute respiratory syndrome‐coronavirus 2; VL, viral load

FIGURE 3

VL data and fitting. Black dots are VL data. Horizontal red dashed lines are the LLOQ of VL (15.49 copies/ml). Solid lines are fits of VL for each subject based on their individual parameters. Subjects with green, blue, and red solid lines are mild, moderate, and severe cases, respectively. LLOQ, lower limit of quantification; SARS‐CoV‐2, severe acute respiratory syndrome‐coronavirus 2; VL, viral load

FIGURE 4

Correlations between predicted +VL days, VL AUC and severity. Positive VL duration and VL AUC are calculated from individually simulated VL profiles from the dataset. Duration of positive VL calculated from the dataset are overlaid to show the consistence between data and simulations. AUC, area under the curve; SARS‐CoV‐2, severe acute respiratory syndrome‐coronavirus 2; VL, viral load

FIGURE 5

Predicted responses in hypothetical scenarios of VL AUC versus start time of treatment from virtual population simulations. Solid lines and dots are the mean VL AUC, and shaded areas are 95% confidence intervals. AUC, area under the curve; SARS‐CoV‐2, severe acute respiratory syndrome‐coronavirus 2; VL, viral load

FIGURE 6

Predicted positive in hypothetical scenarios of VL duration versus start time of treatment from virtual population simulations. Solid lines and dots are the mean +VL duration, and shaded areas are 95% confidence intervals. SARS‐CoV‐2, severe acute respiratory syndrome‐coronavirus 2; VL, viral load