Whack-a-virus: HIV-specific T cells play an exhausting game

Abstract

T cell responses are important for the control of acute HIV infection but become progressively dysfunctional. In this issue of Cell Host & Microbe, Dubé et al. and Takata et al. provide insights into their ongoing interplay with persistent HIV reservoirs, with implications for harnessing functional, durable responses to eliminate HIV.

HIV infection initiates a complex and well-studied interplay between the virus and T cell responses, which strive to control viral replication. But what happens to this game of “cat and mouse” once antiretroviral therapy (ART) eliminates viral replication, leaving only a sparse “reservoir” of infected cells? The answer may inform efforts to harness HIV-specific T cells to either eradicate reservoirs or to control viral rebound on ART interruption. Two studies in this issue of Cell Host & Microbe, by Dubé et al. and Takata et al., reveal and begin to unravel an intricate ongoing relationship.^1,2

Traditionally, infected cells in ART-treated individuals are bundled under the term “latent reservoirs,” presumed to be undetectable to HIV-specific T cells. However, this popular notion, overlooks underlying complexities. While “latency” often denotes transcriptional silence, Yuki et al., report that HIV transcripts, though predominantly short and non-productive, remain abundant during ART.³ If defined instead as a lack of virion production, then the large majority of infected cells are “latent” but may yet be visible to T cells, which can recognize miniscule amounts of antigen.⁴ Relatedly, proviruses harboring genetic defects, which account for most infected cells on ART, cannot produce infectious virions, yet a subset can present HIV epitopes to T cells.⁵ Thus, while antigenic silence of some cells harboring intact proviruses is a key mechanism of persistence, there is a gap in our knowledge as to how various forms of spontaneous HIV expression, from both intact and defective proviruses, stimulate and shape HIV-specific T cell responses on ART.

Using a multiplexed single-cell RNA-flow-FISH assay to both discriminate between cells producing short/abortive versus elongated transcripts and assess concomitant HIV p24 protein expression, Dubé and colleagues demonstrate that, although a median of 4% of cells with integrated HIV DNA spontaneously transcribe viral RNA (vRNA+), this was predominately short/abortive with only very rare cells expressing p24 protein. Stimulation with PMA/ionomycin (“maximal” latency reversal) induced only a 2.5× increase in %vRNA+ cells but substantially increased representation of elongated transcripts and p24 expression. These findings support that latency is predominately attributable to blocks to the elongation and processing of transcripts, rather than to transcriptional initiation. Each of the vRNA+ cells sampled harbored defective HIV proviruses. While this highlights defective proviruses as a key driver of spontaneous expression, representation of rare intact proviruses was unlikely at this sampling depth. Leveraging a unique strength of the RNAflow-FISH assay, the group showed that vRNA+ cells were phenotypically diverse but enriched in clusters expressing markers of central memory T cells and activation, as well as chemokine receptors delineating T-helper differentiation, especially CCR6 (Th17/22), CCR4 (Th2), and CXCR5 (Tfh). These latter enrichments potentially reflect the egress of vRNA+ cells from anatomical compartments serving as sanctuaries for their maintenance. Interestingly, the frequencies of vRNA+ cells correlated directly with duration of pre-ART infection, a potential clue to their genesis and maintenance for further study.

Dubé et al. next assessed CD4⁺ and CD8⁺ HIV-specific T cell responses in relation to the varying forms of spontaneous and inducible HIV expression described above. While there were no significant correlations between T cell responses and HIV DNA, frequencies of spontaneous vRNA+ cells correlated directly with HIV-specific CD4⁺ T cell responses, most strongly for those targeting HIV-Nef. Weaker parallel trends were observed for HIV-specific CD8⁺ T cell responses. The strongest relationships, however, were between HIV-specific T cell responses and spontaneous HIV-p24 protein production, which was further associated with HIV-specific CD4⁺ T cells exhibiting a specific phenotypic profile (namely, CXCR3+, CXCR3+/CCR6+, and CXCR3/CXCR5+ superclusters). Correlations with HIV expression were lost when this was measured following PMA/Ionomycin induction, highlighting the importance of spontaneous p24 expression. Overall, these results highlight how distinctive forms of spontaneous HIV expression may be responsible for the maintenance of different T cell responses under ART.

Takata and colleagues compared post-ART dynamics and phenotypes of HIV-specific CD8⁺ T cell responses relative to HIV DNA and cell-associated RNA in cohorts of Thai participants that initiated ART at acute (AHI) versus chronic (CHI) stages of infection. This comparison uniquely allowed the group to assess the relative contribution of reservoir seeding dynamics to the T cell response. They observed lower total HIV DNA levels, as well as lower cell-associated RNA, in the AHI versus CHI participants. However, there was not a corresponding difference in HIV-specific CD8⁺ T cell responses. When these groups were combined, direct correlations were observed between: (1) CD8⁺ T cells targeting the HIV-Pol gene product and HIV DNA; (2) CD8⁺ T cells targeting HIV Gag, Pol, Nef, or Vif and cell-associated HIV RNA; and (3) CD8⁺ T cells targeting Pol, Nef, or Vif and HIV RNA/DNA ratio (a measure of relative transcription). These results suggest that transcriptionally active, HIV-infected cells on ART produce antigen to maintain an active CD8⁺ T cell response. Of note, lower levels of decay in total HIV-specific CD8⁺ T cells as well as the subsets targeting Gag, Nef, Vif, or Rev/Tat, were each associated with greater reductions in integrated HIV DNA in CHI participants, suggesting that stable CD8⁺ T cell responses drive gradual reductions in infected cells.

How might ongoing recognition of infected cells shape the functionality of HIV-specific CD8⁺ T cells on ART? TCF-1 and TOX transcription factors are largely responsible for regulating the differentiation state of CD8⁺ T-cells between “self-renewing” and “short-lived” or “exhausted.” Takata et al. showed that, pre-ART, almost all HIV-specific CD8⁺ T cells were TCF-1^low TOX^high short-lived effector cells. These remained predominantly TCF-1^low TOX^high dysfunctional short-lived cells in people who initiated ART during CHI but differentiated into TCF-1^high TOX^low stem-like, self-renewing cells in people initiating ART during AHI. Critically, total HIV DNA levels correlated inversely with the TCF-1^high TOX^low self-renewing population, implicating interactions with the HIV reservoir as a driving force in persistent T cell exhaustion.

Overall, this pair of independent studies, along with recent publications,^6–8 serve to firmly establish that HIV-specific T cells continue to interact with some infected cells following ART (Figure 1). Efforts to harness these responses, along with latency reversal, to purge reservoirs, may be informed by understanding why they fail to eliminate cells spontaneously expressing HIV. We identify three non-mutually exclusive possibilities: first, as alluded to by the results of Dubé et al., HIV-expressing cells in the periphery may comprise recent emigrants from CTL-deficient anatomical sanctuaries, such as lymph node follicles. Second, HIV-expressing cells may be killed but partially replenished through clonal expansion, with ongoing stochastic expression from progeny cells. Indeed, Takata et al. report an inverse relationship between T cell response durability and decay in HIV DNA. Third, recognition may not lead to killing—either due to limitations in effector function, or to the selection of infected cells that are resistant to CTL.^9,10

Figure 1. Snapshot of the interplay between HIV-specific T-cells and persistent infected cells during ART.

A subset of infected CD4⁺ T-cells during ART is able to spontaneously express vRNA. While most of these vRNA+ cells express short/abortive transcripts, an even more rare subset of them can spontaneously express viral antigens. The present studies propose that these cells expressing viral antigens engage in an exhaustive whack-a-mole game with HIV-specific T-cells, maintaining an active T cell response. This encounter, however, takes its toll on the responders, leaving behind immunogenic marks of chronic activation and exhaustion. Created with BioRender.com.

It is currently unclear whether there is a net benefit or detriment to this ongoing stimulation of T cells on ART. On one hand, the findings of Takata et al. suggest that this prevents HIV-specific T cells from acquiring the functional profiles likely needed to control viral rebound on ART interruption. On the other hand, the boosting of effector HIV-specific CD8⁺ T cell responses by defective proviral antigens may poise them to eliminate cells with intact proviruses, during either spontaneous or therapeutically induced latency reversal. Given that HIV-specific CD4⁺ T cells can both harbor HIV and, as shown by Dubé et al., be expanded by spontaneous HIV expression, this also raises the possibility of a detrimental chain reaction of clonal expansion of HIV-specific HIV-infected CD4⁺ T cells. Persistent activation of HIV-specific T cells might also contribute to the inflammation induced comorbidities seen on ART. While a previous study did not turn up evidence of this,¹¹ the question remains open to deeper examination. By supporting a paradigm shift, the current papers thus lay the foundations for these and other lines of inquiry with relevance to both HIV pathogenesis and efforts to cure infection.