Relationship between Measures of HIV Reactivation and Decline of the Latent Reservoir under Latency-Reversing Agents

Abstract

Antiretroviral-free HIV remission requires substantial reduction of the number of latently infected cells and enhanced immune control of viremia. Latency-reversing agents (LRAs) aim to eliminate latently infected cells by increasing the rate of reactivation of HIV transcription, which exposes these cells to killing by the immune system. As LRAs are explored in clinical trials, it becomes increasingly important to assess the effect of an increased HIV reactivation rate on the decline of latently infected cells and to estimate LRA efficacy in increasing virus reactivation. However, whether the extent of HIV reactivation is a good predictor of the rate of decline of the number of latently infected cells is dependent on a number of factors. Our modeling shows that the mechanisms of maintenance and clearance of the reservoir, the life span of cells with reactivated HIV, and other factors may significantly impact the relationship between measures of HIV reactivation and the decline in the number of latently infected cells. The usual measures of HIV reactivation are the increase in cell-associated HIV RNA (CA RNA) and/or plasma HIV RNA soon after administration. We analyze two recent studies where CA RNA was used to estimate the impact of two novel LRAs, panobinostat and romidepsin. Both drugs increased the CA RNA level 3- to 4-fold in clinical trials. However, cells with panobinostat-reactivated HIV appeared long-lived (half-life > 1 month), suggesting that the HIV reactivation rate increased by approximately 8%. With romidepsin, the life span of cells that reactivated HIV was short (2 days), suggesting that the HIV reactivation rate may have doubled under treatment.IMPORTANCE Long-lived latently infected cells that persist on antiretroviral treatment (ART) are thought to be the source of viral rebound soon after ART interruption. The elimination of latently infected cells is an important step in achieving antiretroviral-free HIV remission. Latency-reversing agents (LRAs) aim to activate HIV expression in latently infected cells, which could lead to their death. Here, we discuss the possible impact of the LRAs on the reduction of the number of latently infected cells, depending on the mechanisms of their loss and self-renewal and on the life span of the cells that have HIV transcription activated by the LRAs.

Keywords: human immunodeficiency virus; latency; reactivation.

FIG 1

Effects of underlying latent-reservoir dynamics. Depending on the mechanisms driving latent-cell decay under ART, we would see a different relationship between additional virus reactivation and the change in reservoir decay under LRAs. We do not know all the processes that occur in the reservoir, but we show the effects of three scenarios, schematically depicted on the left. (A) HIV reactivation is the only mechanism of loss of latently infected cells (activated cells are represented by a different symbol, because we did not follow them in the model). (B) Latently infected cells can reactivate HIV transcription or simply die without HIV reactivation. (C) Latently infected cells can reactivate HIV, die, and renew by division or new infections. L, latently infected cells; A, cells with reactivated HIV transcription; α, natural HIV reactivation rate (solid red arrow); 3α, 3-fold increase in the HIV reactivation rate in the presence of an LRA (dashed red arrow); δ_L, death rate of latently infected cells; ρ, renewal rate of the latent reservoir; Δ_n, natural decay rate of the latent reservoir; Δ_d, decay rate of the latent reservoir in the presence of an LRA. (Middle) The bars represent the decay rates of the reservoir under ART and with an LRA, with the contribution of HIV reactivation in red. (Right) Reservoir decay under LRA that triples the HIV reactivation rate (red line) compared to the natural decay (solid black line) and triple natural decay (dashed line).

FIG 2

Simplest dynamics of infected cells with reactivated HIV transcription under ART and under LRAs. (A) Model 1. All cells with reactivated HIV (A) die at the same rate, δ_n, irrespective of whether virus was reactivated naturally (α_n) or by an LRA (α_d); here, the model does not follow the processes in the latent reservoir, and thus, a different symbol is used. (B) Virus-transcribing cells under ART (dashed line) and with an LRA (red line) when the LRA triples the natural HIV reactivation rate.

FIG 3

Life span of cells after HIV reactivation by an LRA. Cells with HIV transcription reactivated by an LRA may live longer than cells with naturally reactivated virus. (A) Model 2, unlike model 1, has two types of HIV-transcribing cells: short-lived cells with naturally reactivated HIV, A_n, and potentially long-lived cells with LRA drug-reactivated HIV, A_d; again, the model does not depend on the processes within the latent reservoir. (B) Increase in the number of all virus-transcribing cells (relative to baseline with no LRA) when the half-life of the cells with LRA-reactivated HIV is 0.87 days (δ_d = 0.8 day⁻¹; red line), 8.7 days (δ_d = 0.08 day⁻¹; green line), or 34.7 days (δ_d = 0.02 day⁻¹; blue line) and the LRA-induced HIV reactivation rate, α_d, is 2α_n, so that the total HIV reactivation rate, α, is 3α_n. (C) LRA HIV reactivation rate (as a percentage of the natural virus activation rate) that would be necessary to measure a tripling of the number of HIV-transcribing cells over 50 days, depending on the half-life of the cells with LRA-reactivated virus (half-life: 0.87 days, red bar; 8.7 days, green bar; 34.5 days, blue bar). (D) Dynamics of tripling of HIV-transcribing cells with different life spans under LRA (half-life: 0.87 days, red line; 8.7 days, green line; 34.5 days, blue line).

FIG 4

Steps from measuring the increase in CA HIV RNA under LRA treatment to estimating the reduction of the latent HIV reservoir. LRAs increase the rate of reactivation of HIV transcription in latently infected cells. By measuring the maximal fold increase in the quantity of CA HIV RNA in blood under LRA treatment and the rate of decay after LRA treatment is stopped, one can estimate the greatest possible increase in the HIV reactivation rate. HIV reactivation is only one of a number of mechanisms of loss and renewal that regulate the number of cells latently infected by HIV in patients on ART. Other factors involved in the maintenance of the latent reservoir can also significantly affect our estimates of reservoir reduction. The final reduction of the latent reservoir caused by increasing HIV reactivation can be estimated only when the relative contributions of these mechanisms are better understood.