Comparative analysis of within-host dynamics of acute infection and viral rebound dynamics in postnatally SHIV-infected ART-treated infant rhesus macaques

Abstract

Viral dynamics of acute HIV infection and HIV rebound following suspension of antiretroviral therapy may be qualitatively similar but must differ given, for one, development of adaptive immune responses. Understanding the differences of acute HIV infection and viral rebound dynamics in pediatric populations may provide insights into the mechanisms of viral control with potential implications for vaccine design and the development of effective targeted therapeutics for infants and children. Mathematical models have been a crucial tool to elucidate the complex processes driving viral infections within the host. Traditionally, acute HIV infection has been modeled with a standard model of viral dynamics initially developed to explore viral decay during treatment, while viral rebound has necessitated extensions of that standard model to incorporate explicit immune responses. Previous efforts to fit these models to viral load data have underscored differences between the two infection stages, such as increased viral clearance rate and increased death rate of infected cells during rebound. However, these findings have been predicated on viral load measurements from disparate adult individuals. In this study, we aim to bridge this gap, in infants, by comparing the dynamics of acute infection and viral rebound within the same individuals by leveraging an infant nonhuman primate Simian/Human Immunodeficiency Virus (SHIV) infection model. Ten infant Rhesus macaques (RMs) orally challenged with SHIV.C.CH505 375H dCT and given ART at 8 weeks post-infection. These infants were then monitored for up to 60 months post-infection with serial viral load and immune measurements. We use the HIV standard viral dynamics model fitted to viral load measurements in a nonlinear mixed effects framework. We find that the primary difference between acute infection and rebound is the increased death rate of infected cells during rebound. We use these findings to generate hypotheses on the effects of adaptive immune responses. We leverage these findings to formulate hypotheses to elucidate the observed results and provide arguments to support the notion that delayed viral rebound is characterized by a stronger CD8+ T cell response.

Figure 1:. Description of the experiment.

Study schematic showing ART timing (grey shaded area), as well as viral load trajectories for the 10 subjects in the experiment.

Figure 2:. Standard viral dynamics model schematic.

A) Schematic of the standard model. Target cells, $T$ are sourced at rate $s$ and die at rate $d$. Virus $V$ infects target cells at rate $\beta$. Infected cells, $I$ die at rate $\delta$ and produce virions at rate $p$. Free virus gets cleared at rate $c$.

Figure 3:. Model Fitting Schematic.

Schematic of the approach we follow for model fitting and types of models tested.

Figure 4:. Distribution of △AIC.

Histogram of AIC values relative to the lowest AIC of all the models tested. We reject models with $△AIC$ > 10.

Figure 5:. Parameters differing between acute infection and viral rebound.

Bar chart depicting the proportion times of a parameter is predicted to differ between acute infection and viral rebound in all selected models. Viral clearance rate $\left(c\right)$ is predicted to be different between acute infection and viral rebound in 1 out of 13 best models; death rate of infected cells $\left(\delta \right)$ and start of exponential viral growth $\left(t_{\mathit{start}}\right)$ in 3 out 13 models; mass-action infectivity $\left(\beta \right)$ in 5 out of 13 models and the product of viral production rate $\left(p\right)$ and start of initial number of target cells $\left(T_0\right)$ in all 13 models. Red slices signify a decrease in the value of the parameter for rebound, whereas blue-toned slices depict an increase in the rebound value of the parameter.

Figure 6:. Difference in mean parameter values between viral rebound and acute infection.

Mean value for the product of viral production rate and initial number of target cells, ${pT}_0$ (A) and mass action infectivity $\beta$ (B) for acute infection (red) and viral rebound (blue) for the 13 models selected. For both parameters, the difference is mean values is statistically significant (T test, p<0.001).