Modulation of the CD4 Ectodomain of HIV-Specific CAR T Cells to Enhance Efficacy

Abstract

Human immunodeficiency virus (HIV) infection persists despite effective antiretroviral therapy (ART) because upon treatment interruption, the virus rapidly rebounds from reservoirs of long-lived, latently infected cells. The development of therapeutic strategies capable of recognizing and eliminating infected cells is therefore central to achieving a functional cure of HIV infection. Chimeric antigen receptor (CAR) T cell immunotherapy offers a promising approach to generating a cytotoxic T cell response that can recognize and kill HIV-infected cells. The HIV envelope (Env) glycoprotein (gp) gp120 is the most highly expressed viral protein on the surface of infected cells and thus makes an optimal target, but the capacity of gp120 to adapt to, and escape from, immune responses is well described and poses a significant challenge. CD4- based CAR T cells represent a solution to the problem because the CD4 receptor binds a highly conserved region of gp120. As binding to CD4 is required for viral entry, HIV gp120 adaptation and escape is fundamentally constrained under selective pressure from CD4. However, although Env is the most highly expressed viral protein, its surface expression on HIV-infected cells is heterogeneous and often low, necessitating careful engineering of CD4-based CARs to optimize antigen sensitivity while preserving CAR T cell function. In this dissertation, we investigated the impact of modulations of the CD4 ectodomain (ECD) of CARs on the efficacy of HIV-specific CAR T cells. To do so, we first engineered a panel of CD4-based CARs in which the CD4 ECD was mutated to create a gradient of binding affinity to HIV Env. Biophysical and functional analyses of this panel of CD4 affinity variant CARs across multiple HIV Env variants demonstrated that CD4-to-Env binding affinity is positively associated with intermolecular binding kinetics, intercellular avidity between CD4 CAR T cells and Env- expressing target cells, CAR T cell responses to antigen, cytokine production and polyfunctionality, and Env-expressing target cell killing. We next examined whether structural properties of the CD4 ectodomain independently influence CAR function by modulating receptor dimerization and ex vivo culture conditions. First, we assessed the impact of CD4-based CAR dimerization on T cell function by introducing mutations designed to disrupt CD4–CD4 interactions. Disrupting CD4 CAR dimerization did not reveal enhanced cytotoxic potential, though activation state was notably reduced. We then evaluated the influence of ex vivo culture conditions, which can alter cell surface electrostatic interactions and cellular metabolism. These changes may shift activation states and affect CAR clustering, an effect reported to reduce the efficacy of other CAR T cells. To test this, we generated CAR T cells under high-salt conditions and observed enhanced in vivo proliferation in response to antigen. Together, our findings suggest the potential to improve the functionality of CD4-based CAR T cells by increasing CD4 affinity for HIV Env, thereby enhancing the potential to target and kill infected cells even at low densities of surface Env presentation. Importantly, these improvements occur without detectable increases in baseline activation or exhaustion markers, indicating that CAR activation in the context of HIV infection is limited by antigen sensitivity under conditions of sparse envelope expression. Our work highlights that there are multiple avenues of optimization and tuning of CD4-based CARs to generate an effective HIV-specific CAR T cell design and provide guiding principles for cellular immunotherapies targeting chronic viral infection.