CD8+ T cells control SIV infection using both cytolytic effects and non-cytolytic suppression of virus production

Abstract

Whether CD8⁺ T lymphocytes control human immunodeficiency virus infection by cytopathic or non-cytopathic mechanisms is not fully understood. Multiple studies highlighted non-cytopathic effects, but one hypothesis is that cytopathic effects of CD8⁺ T cells occur before viral production. Here, to examine the role of CD8⁺ T cells prior to virus production, we treated SIVmac251-infected macaques with an integrase inhibitor combined with a CD8-depleting antibody, or with either reagent alone. We analyzed the ensuing viral dynamics using a mathematical model that included infected cells pre- and post- viral DNA integration to compare different immune effector mechanisms. Macaques receiving the integrase inhibitor alone experienced greater viral load decays, reaching lower nadirs on treatment, than those treated also with the CD8^–depleting antibody. Models including CD8⁺ cell-mediated reduction of viral production (non-cytolytic) were found to best explain the viral profiles across all macaques, in addition an effect in killing infected cells pre-integration (cytolytic) was supported in some of the best models. Our results suggest that CD8⁺ T cells have both a cytolytic effect on infected cells before viral integration, and a direct, non-cytolytic effect by suppressing viral production.

Introduction

Understanding the host immune response against HIV/SIV infection is essential for developing effective therapeutic and preventive strategies. Unfortunately, HIV continuously evades and subdues the host’s immune responses, muddling our attempts at elucidating the nature of the immune mechanisms needed to control infection. Examples of HIV evasion strategies include: (i) undergoing rapid mutation of its proteins due to host immune pressures to effectively evade host adaptive responses^1,2,3; (ii) inducing down regulation of MHC-I expression through the viral protein Nef to reduce host cytotoxic capabilities to target infected cells⁴; (iii) taking advantage of virus-specific adaptive responses that generate activated CD4⁺ T cells, the preferential target of HIV, to propagate the infection⁵; (iv) chronically stimulating the immune system, thus resulting in production of nonfunctional exhausted cytotoxic T lymphocytes (CTLs)^6,7. As such, there is uncertainty regarding which immune response should be emphasized in current research efforts.

Of the various immune responses against HIV, the response exerted by CD8⁺ T cells has been shown to be critical for control of HIV/SIV⁸. This is supported by: (i) the temporal association that exists between the increase in virus-specific CD8⁺ T cell responses and the post-peak decline in plasma viremia^9,10; (ii) the CD8⁺ CTLs ability to suppress new infections in vitro^11,12; (iii) the virus escape mutations that consistently arise in response to the host CD8⁺ T cell response during all stages of infection^1,2,13,14; (iv) the strong association between specific host MHC-I alleles and HIV/SIV disease progression¹⁵; and (v) the association of circulating escape mutants with the prevalence of specific HLAs in the population^16,17. Experimental in vivo CD8⁺ cell depletion studies in SIV-infected macaques have strengthened this argument and provided more direct evidence for the role of CD8⁺ cells in HIV infection^{18,19,20,21,22,23,24,25,26,27,28}. CD8⁺ cell depletion results in a rapid and sustained rebound of plasma viremia, which is controlled when CD8⁺ cells are restored. These results are consistent in many models of SIV infection: elite controller^18,29, nonpathogenic³⁰, rapid progressor^21,31, antiretroviral treated^20,22,23 and untreated models^19,24,26,32. Accordingly, understanding the mechanisms of action of CD8⁺ cells and identifying strategies to boost CD8⁺-specific immune responses is a key priority, both for HIV vaccine and cure research.

Although CD8⁺ cells hold strong potential for cure efforts, their specific mechanism(s) of action is not well understood⁸. CD8⁺ CTLs could exert a direct cytotoxic response against viral-infected cells via release of granzyme/perforin and/or stimulation of the Fas/FasL pathway^33,34. Alternatively, CD8⁺ cells could act by interfering with de novo infection or the release of new virions through soluble antiviral factors, including the CCR5-binding proteins MIP-1α, MIP-1β, RANTES, the cellular anti-HIV factor (CAF), α-defensins, and other factors^35,36,37. To help shed light on this question, two groups studied the lifespan of SIV-infected cells after nucleotide reverse transcriptase inhibitor (NRTI) treatment, either in the presence or absence of CD8⁺ cells. They showed that the average lifespan was not different with or without CD8⁺ cells, concluding that CD8⁺ CTLs do not exert a cytolytic effect on infected cells^20,23. Another study quantified the lifespan of infected cells after antiretroviral treatment in infected people with different HLA background, both favorable and unfavorable for HIV progression³⁸. These studies found no difference in the lifespan of infected cells and concluded that protective CD8⁺ T cells may exert their effect before onset of productive infection, or via noncytolytic mechanisms, but none of them directly demonstrated this^20,23,38. These data have both been corroborated and challenged^{39,40,41,42,43}, leaving the field to question what the true mechanism(s) of action of CD8⁺ T cells is against HIV.

To help settle this question, we interrogated whether CD8⁺ cells exert a cytolytic response against infected cells prior to viral production (i.e., before viral DNA integration into the host genome). The hypothesis is that once viral integration occurs and the cell starts producing virus, viral cytopathic effects dominate. Although this hypothesis has been proposed^20,42, it has never been tested experimentally. Here, to test this hypothesis, we administered the integrase inhibitor raltegravir (RAL) to SIV-infected rhesus macaques (RMs), in the presence or absence of CD8⁺ cells. To analyze this data, we developed new viral dynamic models to account for CD8 depletion, adapted from our previous model⁴⁴, and fitted them to the data to study the possible effector mechanisms contributing to the observed viral load profiles. We found that the half-life of infected cells before viral integration in the RAL-treated only group is significantly shorter than in the RAL-treated plus CD8⁺ depleted group, suggesting that CD8⁺ T cells have a cytolytic role prior to viral integration. Further, the best models also indicated that the viral production rates increased in the absence of CD8⁺ cells, indicating that CD8⁺ T cells also exert a non-cytolytic effect in suppressing viral production.

Results

Experimental design and mathematical modeling approach

Twenty (20) Indian-origin rhesus macaques (RM) were IV-infected with 300 TCID50 of SIVmac251 (Fig. 1). The initial dynamics of the virus were similar in all animals (Fig. S1). At fifty-six (56) days post-infection (dpi), after the virus reached quasi steady-state, CD8⁺ cells were depleted by administration of the monoclonal antibody M-T807R1 to 12 RMs. Three weeks later, these animals received an additional M-T807R1 infusion. In 8 of these RMs, 2 days following the first CD8⁺ cell depletion, RAL monotherapy was initiated for 23 days (CD8 depletion plus RAL Tx group, Fig. 1a). The remaining 4 RMs served as untreated controls (CD8 depletion group, Fig. 1a). An additional 8 RMs without CD8⁺ T cell depletion were also treated with RAL monotherapy under the same conditions (RAL Tx group, Fig. 1a).Fig. 1: Study approach...