Harnessing immune cells to eliminate HIV reservoirs

Abstract

Purpose of the review:

Despite decades of insights about how CD8⁺ T cells and NK cells contribute to natural control of infection, additional hurdles (mutational escape from cellular immunity, sequence diversity, and hard-to-access tissue reservoirs) will need to be overcome to develop a cure. In this review, we highlight recent findings of novel mechanisms of antiviral cellular immunity and discuss current strategies for therapeutic deisgn.

Recent findings:

Of note are the apparent converging roles of viral antigen-specific MHC-E-restricted CD8⁺ T cells and NK cells, IL-15 biologics to boost cytotoxicity, and broadly neutralizing antibodies in their native form or as anitbody fragments to neutralize virus and engage cellular immunity, respectively. Finally, renewed interest in myeloid cells as relevant viral reservoirs is an encouraging sign for designing inclusive therapeutic strategies.

Summary:

Several studies have shown promise in many pre-clinical models of disease, including SIV/SHIV infection in non-human primates and HIV infection in humanized mice. However, each model comes with its own limitations and may not fully predict human responses. We eagerly await the results of clinical trails assessing the efficacy of these strategies to achieve reductions in viral reservoirs, delay viral rebound, or ultimately elicit immune based control of infection without cART.

INTRODUCTION

Defining the HIV reservoir – what can the immune system detect?

Understanding the nature of the reservoir is essential for designing strategies aimed at viral eradication to achieve immune-based control. The HIV reservoir is defined as any cell that can harbor virus, persist despite cART, and seed new infections. While CD4⁺ T cells constitute the majority of the HIV reservoir, HIV-infected macrophages and dendritic cells (DCs) are detected in tissues from virally suppressed PLWH (1–4)(* for 3 and 4), and macrophage-tropic viral sequences have been observed in PLWH following analytical treatment interruption (ATI) (5), emphasizing the importance of considering other cell subsets in addition to CD4⁺ T cells for HIV cure strategies. Latency (which has recently been described in human monocytes, the blood precursors for tissue-resident macrophages (6))* remains a major barrier for viral eradication. However, while there is no evidence for ongoing viral evolution during cART (7), residual viral RNA and protein expression during cART (3,4,8–10)(* for 8 and 10) may provide an opportunity for sustained immune detection and control of infection, even without the need for latency reversal. This review focuses on the challenges of eliminating productively infected cells and new strategies to control the virus.

CD8⁺ T cell-mediated control of infection

Many studies have shown the importance of CD8⁺ cytotoxic T lymphocytes (CTLs) in controlling HIV, initially by observing a drop in plasma viral load during acute stages of infection correlating with the expansion of HIV-specific CTLs (11,12). However, the selective pressure exerted by viral peptide-specific CTLs coupled with the downmodulation of MHC-I by HIV nef and vpu leads to the selection of cells infected with CTL-escaped virus, a major contributing factor to viral persistence (13). Additional evidence supporting the role of CTLs in controlling infection stems from individuals who naturally maintain a low (50–2000 HIV RNA copies/mL) or undetectable (defined by clinical assays <50 RNA copies/mL) viral load in the absence of cART, referred to as viremic controllers (VC) and elite controllers (EC), respectively. Numerous factors correlate with viral control in VC/ECs including protective HLA alleles (HLA-B*57 and HLA-B*27; reviewed in (11)), targeting of highly “networked” epitopes exhibiting little mutational tolerance (14), stable proliferative and cytolytic function of CTLs (15), and the ability of CTLs to respond to both lower concentrations of antigen, and a greater range of epitope variants (16). While loss of CD8⁺ T cell function in controllers has been shown to precede resurgence of viremia (17), no single factor likely determines initial protection, suggesting that control might require an optimal combination of both a highly functional immune response and targeting of “networked” epitopes. Complicating the role of CD8⁺ T cells in HIV control, it was recently observed that HIV-infected CD4⁺ T cells surviving co-culture with autologous CD8⁺ T cells show enrichment of genes associated with quiescence, suggesting that CD8⁺ T cells may play a role in promoting latency (18)*. With these challenges in mind, we will discuss strategies to improve CTL-mediated control of HIV.

NK cell contributions to control of acute and chronic infection

NK cells express a variety of germline-encoded activating and inhibitory receptors, whose ligands are differentially modulated on HIV-infected cells (19–25). Without the need for priming to elicit antiviral effects, they play a role in early stages of pathogenesis and potentially protection from viral acquisition. Cohort studies of highly exposed seronegative (HESN) individuals have identified particular NK cell phenotypes and enhanced effector functions associated with resistance to HIV acquisition (26–30). Furthermore, analysis of human samples prior to and then during acute HIV infection showed that increased NK cell activity was a correlate of future viremic control (31), and NK cell depletion in non-human primates (NHPs) enhanced acute viremia and inflammation (32), emphasizing the role of NK cells in control of early infection.

For chronic stages of viral infection, co-expression of KIR3DS1 and specific KIR3DL1 subtypes with HLA-B Bw4–80Ile is associated with lower plasma viremia and delayed progression to AIDS (33–35). For cART treated infection, Astorga-Gamaza et al recently showed that expression of the inhibitory receptor KLRG1 is associated with NK cell dysfunction, and that blocking KLRG1 enhances cellular function, promoting a reduction in intact HIV genomes (21)**. Humanized mouse models have yielded additional insights into the role of NK cells in chronic infection. Kim et al., showed that treatment of humanized mice with a latency reversal agent in combination with administration of human NK cells delayed the time to rebound following cART interruption (36). Furthemore, in elegant studies by Huot and colleagues, NK cells in african green monkeys (AGM), natural hosts for SIV that exhibit better viral control than rhesus macaques (RM), showed infiltration into the secondary lymphoid organs where they help control the infection in an MHC-E-restricted manner (37,38). Together, these studies emphasize the ability of NK cells to naturally control viral infection, which, in humans, may be enhanced through KLRG1 blockades and promoting NK cell infiltration into B cell follicles in the lymph nodes.

Increasing evidence supports the existence of specific NK cell subsets with “memory-like” or “antigen-specific” properties. Memory NK cells were first identified in mice in response to haptens (39), and have since been identified in humans (40). In the context of HIV, TCF7-dependent NK cells exhibiting memory-like markers are increased in PLWH (41). Strikingly, NK cells from SIV and SHIV-infected RMs show specific lysis of Gag- and Env-pulsed DCs ex vivo, which was dependent on NKG2A/C expression (42). These memory NK cell responses could also be elicited through Ad26 HIV Env or SIV Gag vaccination (42). A potential mechanism of the apparent anitgen specificity is via a conserved Gag epitope, which, when presented on HLA-E, is thought to block the interaction between HLA-E and the inhibitory receptor, NKG2A, resulting in increased NK-mediated killing (43). Recently, an exciting study by Jost and colleagues observed the presence of HIV antigen-specific NK cells in PLWH, driven by the CD94/NKG2C receptor and viral peptides that stabilize HLA-E surface expression (44)**. Mechanisms of recognition and function were confirmed in assays with clonal memory-like NK cells, which showed high expression of KLRG1, NKG2C, and α4β7. Importantly, no HIV peptide-specific NK cell responses were observed in people without HIV, further supporting an antigen-specific memory mechanism.

HARNESSING IMMUNE CELLS FOR VIRAL CONTROL

To achieve undetectable viral loads in the absence of cART, therapeutics would need to induce sustained immune responses, since latently infected cells would continue to reactivate over time. While “shock-and-kill” approaches remain an attractive strategy to make meaningful reductions in the HIV reservoir, clinical trials have shown no delay in time to viral rebound following cART interruption with latency reversing agents (reviewed in (45)). Given the extensive amount of work to address latency reversal, we defer to a review by Rodari et al for a more comprehensive analysis of the current strategies around this approach (46). Below we will focus on methods to enhance the antiviral activity of CTLs and NK cells in the context of productively infected cells with the goal of achieving immune-mediated control of infection.

HLA-E: a common element for CTL and NK cell therapies

Insights from HIV controllers and CD8 depletion studies in NHPs (47,48) highlight the role of CD8⁺ T cells in controlling infection. However, determining which viral antigens to employ as immunogens remains a challenge for universal T cell vaccine design. While targeting mutation-resistant regions may overcome this hurdle (reviewed in (49) and (50)), an added level of complexity arises from HLA diversity across different populations. Targeting conserved viral epitopes presented on HLA-E may overcome this obstacle, as only two alleles are expressed. For more than 12 years, much attention has been paid to the rhesus CMV (strain 68–1) vaccine platform, which is unique in its ability to prime MHC-E-restricted T cell responses. The viral vector expresses SIV antigens, and, remarkably, provides NHPs an overall 59% protection from mucosal challenge with SIVmac239 (51–56). Protection is mediated by vaccine-induced MHC-E-restricted CTLs, which only develop if VL9 peptide in the viral protein, Rh67, is present in the vaccine vector (54), and if the vector’s myeloid tropism is maintained (53). Emphasizing that viral control is multifaceted, the development of MHC-E-restricted CTL responses in vaccinated NHPs is not enough to distinguish protected versus unprotected animals (55). Recent studies suggested that the change-from-baseline of an IL-15 response signature correlates with vaccine protection (55)*, but how this can be exploited to yield 100% vaccine efficacy is unknown. While HIV-specific HLA-E restricted CD8⁺ T cells have been observed in PLWH (57), and proof-of-principle studies show that these antiviral human responses can be primed in vitro (58), it’s unclear whether induction of these responses in humans will provide protection similar to what’s seen in NHPs. In an exciting step forward, VIR-1388 (HCMV-HIV vaccine) is currently being tested for safety and immunogenicity in a Phase 1 clinical trial in adults without HIV (NCT05854381). While initially considered a prophylactic, VIR-1388 may also possess therapeutic potential. If the vaccine were given during cART, could HLA-E-specific CTL responses prevent viral rebound during ATI?

Perhaps an optimal vaccine would stimulate both HLA-E-restricted CTL responses and the newly identified HIV antigen-specific NKG2C⁺ NK cells. Indeed, earlier work by Reeves et al showed that memory NK cell responses were observed in Ad26/SIV-vaccinated NHPs (42). With the recent study by Jost and colleagues confirming that HIV antigen-specific NK cells are found in PLWH, it is plausible that both CTL and NK cell responses could be targeted against more conserved (or potentially “networked”), viral peptides on HLA-E (44). Methods to enhance IL-15 signaling during the time of vaccination may be required to ensure proper development of effective HLA-E-restricted CTL responses and boost NK cell activity. The IL-15 superagonist, N-803, developed by Immunity Bio, has not only been shown to act as a latency reversal agent (47,59,60) but can also enhance activation of both T and NK cells in PLWH on cART (60)*. This could be supplemented with a STAT5 agonist, HODHBt, which synergizes with IL-15 to enhance activation of HIV-specific T cells (61)*. Separately, as the antigen-specific NK cells are KLRG1⁺ (44), blocking this pathway during chronic infection may be a method to further enhnace the expansion and antiviral activity of this NK cell population (21), but how this would affect the antigen-specific CTL populations is unknown. Finally, the evaluation of these strategies should consider whether HIV nef would adapt to elicit stronger HLA-E downregulation (as is seen with primary strains of HIV (62)) in response to CTL/NK selective pressure. This would make the case for also engaging NKG2A⁺ NK cell populations.

HIV-specific antibodies - neutralizing virus and engaging cellular immunity

One promising area for HIV therapeutics is in broadly neutralizing antibodies (bNAbs). In addition to their ability to neutralize the virus, preventing entry into new target cells, the Fc portion of the antibodies can engage in effector functions. While NK cells primarily function as innate immune cells, their expression of the Fc receptor FcRyIII (CD16) allows them to engage and kill antibody opsonized infected cells through antibody-dependent cellular cytotoxicity (ADCC) (reviewed in (63)). Of note, while ADCC enhances NK cell killing of HIV-infected CD4⁺ T cells (64–66), NK cell ADCC is less efficient towards infected macrophages in vitro, indicating that particular cell types may have mechanisms of limiting ADDC responses (65,67). In addition, Fc receptor expression on professional antigen-presenting cells allows them to capture and process antigen to stimulate T cells. Together, this emphasizes the potential for antibodies to engage mutliple immune compartments to help control infection.

In NHPs, treatment with bNAbs resulted in significant delays to viral rebound following ATI (68). In humans, the bNAb 3BNC117 is effective at delaying time to viral rebound during ATI as a monotherapy (69,70), with efficacy further enhanced when used in combination with another bNAb, 10–1074 (71). Ultimately, rebound was correlated with a loss of circulating antibody (71). The observed efficacy may be due to enhanced HIV-specific CTL activity (72,73)(* for 73). While vaccinal effects, where bNAbs form immune complexes and stimulate antigen processing and presentation by antigen-presenting cells (74), may play a role in control, to date the precise mechanism by which cellular immunity is enhanced is not known.

Combining bNAb treatment with other therapies to achieve more durable responses is an active area of research. One such approach is stimulation of the innate immune system via TLR agonism, which could enhance NK cell activity (75,76). This approach has shown promise. Lefitolimod, a TLR9 agonist, caused increased NK cell activation and HIV transcription in PLWH on cART (77). However the recent TITAN trial showed no additional benefit from treatment with lefitolimod in addition to 3BNC117 and 10–1074 (78)*. Additional studies are in process using the TLR7 agonist Vesatolimod (GS9620), which has shown slightly delayed time to viral rebound following ATI as a monotherapy (79), and is under investigation as a combination with a T cell vaccine (80). Finally, additional clinical trials are underway to assess whether N-803, to enhance CTL and NK cell activity, could contribute to delayed viral rebound in combination with bNAbs during ATI (NCT05245292 and NCT04340596).

HIV-specific T cell products and biologics

Taking lessons from the cancer field, several groups have focused on development of HIV-specific chimeric antigen receptor (CAR) T cells. Maldini et al developed HIV-resistant CAR T cells that have both the proliferative benefits of the 4–1BB CAR T and the enhanced effector function of the CD28 CAR. Adoptive transfer of these CAR T cells resulted in a slower decline in CD4⁺ T cells during acute infection in humanized mice, and when given in combination with cART, accelerated the decline in HIV RNA and decreased the frequencies of Gag⁺ cells in the tissues, including Gag⁺ CD14⁺ macrophages (81). This supports an earlier report that HIV-infected macrophages are susceptible to HIV-specific CAR T cells (65). Recently, duoCAR T cells, which are created using autologous T cells transduced with two CAR constructs (mD1.22-CAR and m36.4-CAR), eliminated HIV-infected splenic PBMCs of humanized mice, along with infected CD4⁺ T cells and monocytes in vitro (82)**. These are currently being tested in a Phase I/IIa clinical trial in PLWH (NCT04648046). Another interesting CAR T cell concept is CMV and HIV dual-specific CAR T cells, which have been tested in conjunction with a CMV vaccine in humanized mice; these endogenous CMV-specific T cells, made to express an anti-gp120 CAR, have been shown to expand in vivo and reduce viral load during cART and after cART interruption (83). Finally, promoting CAR T cell infiltration into the B cell follicle via CXCR5 overexpression may enhance delay to viral rebound following cART cessation (84). Overall, CAR T cells with multiple specificities that can persist in vivo, expand in response to a controlled stimulus, and infiltrate viral sancturaies in the B cell follicle may prove useful in reducing the HIV reservoir.

Also an inspiration from the cancer field (85–88), various bispecifics are under investigation to enhance clearance of HIV-infected cells. Bispecific killer cell engagers (BiKEs) (89) and bispecific T cell engagers (BiTEs) (90) show pre-clinical promise in enhancing directed cytotoxicity of NK cells and T cells, respectively, towards HIV-infected targets. Trispecific killer cell engagers (TriKEs), which engage CD3 and CD28 of the effector cell, along with the CD4 binding site N6, will potentially prove to be an improved platform (91)*.

CONCLUSION

Achieving sustained viral control following cART cessation will likely require engagement of multiple immune cell types. Decades of research characterizing PLWH who naturally control infection, mechanistic studies elucidating mechanisms of antiviral cellular immunity, and lessons from the cancer immunotherapy field have yielded inspiration for current strategies to control viral infection in the absence of cART. It’s clear that emerging themes in the cure field include HLA-E coordination of CTL and NK cell responses, IL-15 enhancement of CTL and NK cell function, and therapies that take advantage of bNAb breadth and cellular immunity. Understanding the nature of the HIV reservoir will be essential for optimizing these strategies for maximal antiviral activity. Indeed, over 10 years of research has shown that infected myeloid cells react differently than T cells to cytolytic immune cells, thus, the strategies discussed in this review would greatly benefit from assessments of antiviral activity towards these different cellular reservoirs.

KEY POINTS.

• Rhesus CMV SIV vaccine-induced CD8⁺ T cells and memory NKG2C⁺ NK cells that recognize viral anitgens on MHC-E could be exploited to control infection.

• IL-15 superagonist, N-803, is a potent booster of cellular immunity and latency reversal, the efficacy of which could be enhanced by small molesules that target STAT5 activtaion.

• Inspired by the success of cancer immunotherapies, HIV broadly neutralizing antibodies and antibody fragments to engage cellular immunity show promise in pre-clinical models.

• HIV-infected macrophages and dendritic cells have been observed in several tissues and should be considered as part of the resevroir that needs to be targeted for immune-based control of infection.