HIV-1 INFECTED HUMAN LYMPHOID TISSUE REMAINS IMMUNE-ACTIVATED DESPITE ART: SOLUBLE AND EXTRACELLULAR VESICLE-ASSOCIATED CYTOKINES

Background In the age of efficient antiretroviral therapy (ART), premature aging (that includes the development of various pathologies) of HIV-1-infected patients with suppressed viral replication becomes an important health problem. These pathologies are associated with residual immune activation after successful ART. Cytokines, which sustain immune activation, can be either in a soluble form or can be associated with extracellular vesicles (EVs). Together with classical soluble cytokines, these EV-associated cytokines may play a key role in HIV-1 infection and pathogenesis. These hypotheses should be tested in a biologically-relevant system under controlled laboratory conditions. Human ex vivo lymphoid tissues constitute such a system. It was previously used to study HIV-1 pathogenesis, and similar to the situation in vivo, soluble cytokines were shown to be upregulated in this system upon HIV-1 infection. Here we investigated the modulation of both soluble and EV-associated cytokines following HIV-1 infection and upon application of ART. We investigated whether the HIV-1-triggered immune activation is decreased when viral replication is suppressed. We considered various hypotheses on the mechanisms of HIV-1-triggered immune activation: persistent immunological and inflammatory response after an initial cytokine storm; presence of defective virions sustaining immune activation; latent herpesvirus infection which may be reactivated during HIV-1 infection and leading to an inflammatory response; or proinflammatory effects of antiretroviral drugs themselves. Methods Ex vivo lymphoid tissues were infected with two HIV-1 strains: X4LAI04 or R5SF162. HIV-1 was either allowed to replicate for 16 days, or tissues were treated with ritonavir (RTV) or AZT-3TC at day 3 post-infection (n=8). Ex vivo cultures were also treated with cytokines (n=4), inactivated HIV-1 virions (n=4), or ART in the absence of HIV-1 infection (n=8). HIV-1 replication was analyzed in tissue culture supernatants by measurement of p24gag antigen. EV presence was confirmed by Western Blot and Nanoparticle Tracking Analysis (NTA). 33 cytokines in soluble and EV-associated forms were measured with multiplexed bead-based assays. Herpesvirus DNA copies in tonsillar tissue (n=4) were quantified with droplet digital PCR (ddPCR). Results We found that both strains of HIV-1 replicated well in tissues and triggered an upregulation of numerous soluble cytokines as early as day 3 post infection. Many cytokines that were upregulated upon HIV-1 infection in soluble form were increased in EV-associated form as well. However, some cytokines were uniquely upregulated only in the EV form, mainly in response to X4LAI04 infection. In early HIV-1 infection, there was a significant increase in the percentage of soluble RANTES and TNF-α compared to EV-associated; additionally, RANTES significantly increased in the percentage of EV surface associated compared to EV internal. Cytokines which were significantly upregulated throughout culture as evaluated by cumulative totals included: IL-2, IFN-γ, MIP-1α, MIP-1β and RANTES for both virus infections. X4LAI04 infection also led to increases in IL-7, IL-18, M-CSF and TNF-α. An even greater number of cytokines were upregulated in EV-associated form; including many of the same cytokines that were increased in soluble form, but additionally IL-1α, IL-6, IL-13, IL-21, IL-33, GM-CSF and TGF-β were upregulated following X4LAI04 infection. RTV and AZT-3TC treatment of tissues efficiently suppressed viral replication (>99% suppression for both treatments). Despite control of viral replication, cytokines remained upregulated after 13 days of ART treatment, and EV-associated cytokines were less likely to decrease than soluble ones. X4LAI04 induced stronger immune responses as measured by increased soluble and EV-associated cytokines, particularly pro-inflammatory cytokines and the -chemokines MIP-1α, MIP-1β and RANTES, compared to R5SF162 strain. These X4LAI04 cytokine upregulations, especially the β-chemokines, were also less likely to be restored after both antiretroviral treatments. Additional experiments demonstrate that this persistent immune activation was not due to the initial cytokine storm. The emulation of an initial cytokine stimulation was able to boost the production of other cytokines early in culture, but these increases were not maintained over time. Inactivated X4LAI04 was able to trigger a similar, but slightly weaker cytokine release compared to the infectious virus. A single inoculation of inactivated virus elevated many of the same cytokines as live virus; 7 of 13 upregulated cytokines with live virus were increased with inactivated virus. Repeated exposure to inactivated virus generated an even more similar response compared to real infection; almost all the same cytokines were upregulated and some of them were more elevated with inactivated virus (IL-21, MIP-1α, MIP-1β, RANTES and TNF-α. Herpesvirus (HSV-2, EBV, CMV, HHV-6, and HHV-7) DNA copies were detected at different time points at variable concentrations. However, no obvious pattern of herpesvirus reactivation was observed that would account for the immune activation. Finally, the antiretroviral treatments themselves were not responsible for immune activation since only small but significant decreases were observed for a few cytokines. Conclusions HIV-1 infection of ex vivo human lymphoid tissues led to upregulation of various cytokines, in both soluble and EV-associated forms. HIV infection altered distribution of certain cytokines between soluble and EV-associated forms, as well as between EV-surface and EV-encapsulation. Despite viral suppression by ART, the majority of the upregulated cytokines, especially β-chemokines, remained upregulated, similar to the in vivo situation. EV-associated cytokines were more likely to remain elevated than soluble ones. The ex vivo human tissue model was further employed to test various hypotheses on HIV-triggered immune activation. It was determined that an initial cytokine storm did not change the basal setpoint, there was no correlation with reactivation of latent herpesviruses, and ART by itself did not trigger immune activation; however, the presence of noninfectious viral particles can trigger a response similar to infectious virus.