Therapeutic Vaccines for the Treatment of HIV

Abstract

Despite the success of anti-retroviral therapy (ART) in transforming HIV into manageable disease, it has become evident that long-term ART will not eliminate the HIV reservoir and cure the infection. Alternative strategies to eradicate HIV infection, or at least induce a state of viral control and drug-free remission are therefore needed. Therapeutic vaccination aims to induce or enhance immunity to alter the course of a disease. In this review we provide an overview of the current state of therapeutic HIV vaccine research and summarize the obstacles that the field faces while highlighting potential ways forward for a strategy to cure HIV infection.

Introduction

In 2018, approximately 38 million people were living with HIV globally. Despite the success of anti-retroviral therapy (ART) in limiting HIV replication, decreasing transmission rates, lowering AIDS-related morbidities and improving the quality of life for HIV-infected individuals, it has become evident that long-term ART will not eliminate the HIV reservoir to cure the infection. As a consequence, people living with HIV (PLWH) are forced to stay on life-long therapy with long-term side effects and high cost. Moreover, only an estimated 23.3 million PLWH have access to ART (1), and specifically in resource-constrained settings, access to ART is often limited. Alternative strategies to eradicate HIV infection, or at least induce a state of viral control and drug-free remission are therefore needed.

HIV-1 infection remains incurable as it integrates into the host genome of long-lived memory CD4+ T cell populations where replication-competent virus persists as integrated proviral DNA. From this latent state, viral reemergence and disease progression can rapidly develop if ART is interrupted and virus can disseminate (2–4). These latently infected cells also persist as they are invisible to the immune system due to the lack of active viral replication. The latent viral reservoir therefore presents one of the obstacles for cure approaches. During the natural course of infection, however, a small proportion of HIV-infected individuals are able to spontaneously control HIV replication to undetectable levels in the absence of ART, so called elite controllers (5). Some individuals even demonstrate none or very low-level plasma virus and no clinical disease progression for more than 25 years (6). These elite controllers have therefore been suggested as proof that a functional cure of HIV-1 infection is possible. The underlying mechanisms contributing to this control are likely heterogenous, and a variety of host and viral factors have been associated with the controller phenotype ((5, 7–11). A common feature within elite controllers is the robust T-cell response that can often been found in these individuals. Specifically, enrichment of polyfunctional T-cell populations with the ability to secrete multiple cytokines and/or suppress viral replication in-vitro have been described in the HIV controllers (12–16). It has therefore been suggested that induction of similar immunity in non-controllers, using the elite controllers as ‘blue-print’ for possible immune modulatory approaches, might be a possibility (reviewed in (17)).

One promising strategy for an immunological treatment against HIV infection, either to induce permanent viral control or to even eliminate the viral reservoir, is the development of therapeutic vaccines. Therapeutic vaccines aim to induce or enhance immunity to alter the course of a disease in patients. These vaccines could theoretically provide durable, non-invasive and cost-effective treatment solutions to a vast population of HIV-infected individuals. Indeed in other medical fields, specifically in cancer, therapeutic vaccines have been developed and are being utilized. Currently, there are three therapeutic vaccines approved by the FDA. Bacillus Calmette-Guélin (BCG), initially developed as a vaccine to prevent tuberculosis, proved suprisingly effective in reducing recurrence and progression of non-muscle-invasive bladder cancer (NMIBC) by increasing the antigen presentation in urothelial tumor cells and boosting local immune responses (18–20). Sipuleucel-T (trade name Provenge), was developed to treat men with prostate cancer by collecting and training their T cells ex-vivo to recognize and eradicate tumor cells after re-infusion (21). It was followed by talimogene laherparepvec (T-VEC, trade name Imlygic), a genetically modified oncolytic viral therapy that enhances T-cell function to treat advanced melanoma (22). While the mechanisms of action for these cancer vaccines remain to be completed elucidated, all three vaccines improve specific cytotoxic T lymphocyte (CTL) responses that contribute to the control of disease progression (23, 24).

Similar to therapeutic cancer vaccines, therapeutic HIV vaccines aim to boost the magnitude, breadth of antigen-specificities and functionality of anti-HIV T cell responses to eliminate infected cells and facilitate long-term viral control in the absence of ART. In this review we provide an overview of the current state of T-cell-based therapeutic HIV vaccine research and summarize the obstacles that the field faces while highlighting potential ways forward.

Challenges in developing a successful T cell based therapeutic HIV vaccine

One major challenge for the immune response against HIV-1 is the enormous viral sequence diversity, even within each infected individual, and the rapid evolution of the virus during active replication. This allows the virus to permanently escape the immune response (25) and as a consequence viral escape mutations accumulate early on and are achieved in the latent reservoir (26). As HIV primarily infects CD4+ T-cells, progressive CD4+ T cell depletion and subsequent immune dysfunction are a hallmark for this infection. Furthermore, in the presence of uncontrolled antigenemia, and with the increasing lack of T-helper cell help, immune exhaustion becomes a dominating characteristic of the anti-HIV immune response (27, 28). While ART can restore some of the immune dysfunction, it has been shown that the immune system never fully recovers, and that CD8+ T-cells often maintain an epigenetically engraved exhaustion program (29). Before this background, therapeutic HIV vaccine development has been ongoing for several decades. Initial studies used envelope-depleted, inactivated virus to boost immunity in HIV-infected individuals (30) or subunit vaccines with recombinant envelope glycoprotein (rgp160) of HIV-1 IIIB to immunize HIV-infected patients (31). Since then, multiple new vaccine concepts have been developed using DNA; viral vectors such as modified vaccinia Ankara (MVA), adenovirus (Ad), vesicular stomatitis virus (VSV) and canary pox virus (ALVAC); RNA; lentiviral vectors and dendritic cells as vaccine vehicles. While all these vaccine candidates have been aimed at improving existing anti-HIV immune responses to improve viral control, the majority of vaccine candidates have specifically focused on the optimization of T cell responses. For example the pTHr.HIVA DNA prime, MVA.HIVA boost vaccine expressing consensus HIV-1 clade A Gag p24/p17 sequences and a string of CD8+ T-cell epitopes (HIVA) to induce and/or boost both HIV specific CD4+ and CD8+ T cell response was extensively trialed in in several hundred healthy or HIV-1-infected volunteers in Europe and Africa (reviewed in (32)).

Although many of these therapeutic HIV vaccines resulted in an improvement of autologous HIV-specific T-cell responses, their effects on viral control, as defined by delayed time to viral rebound or reduced viral load set-point after stopping ART, have often been limited. For example, ALVAC-HIV (vCP1452), a modified recombinant canary pox virus vaccine encoding the HIV-1 genes Env, Gag, Pol and Nef, and recombinant gp160, administered to ART suppressed individuals increased HIV-specific CD8+ T cell responses in 11 of 14 study participants but resulted in no significant difference in HIV virus rebound compared to unvaccinated individuals when ART was interrupted suggesting that the vaccine was ineffective in eliciting immune responses sufficient for viral control (33, 34). Similarly, Kinloch-de Loes et al. evaluated in a randomized controlled trial in 78 individuals during acute HIV infection whether the addition of ALVAC-HIV (vCP1452) or ALVAC-HIV and Remune (inactivated envelope-depleted whole virus) to standard ART versus ART alone would result in enhanced viral control. Again, participants in the vaccine arms had significantly increased HIV-1-specific IFN-γ+ CD8+ and CD4+ T cell responses compared to the ART alone treated group however at the study endpoint, 24 weeks after discontinuation of ART, there was no difference in the frequency of study participants that had </=1000 HIV-1 RNA copies/mL between the vaccine or ART alone groups (35).

Additional studies using other vaccine concepts resulted in similar outcomes: Schooley et al. vaccinated 114 ART-treated individuals with a recombinant adenovirus serotype 5 (Ad5)-based vaccine expressing Gag followed by an analytical treatment interruption (ATI) for 16 weeks. The vaccine elicited modest Gag-specific CD4+ and CD8+ T-cell responses, however, no differences in HIV-1 RNA level setpoint during ATI that met pre-defined levels of significance were observed (36). Moreover, Jacobson et al. used autologous dendritic cells from 54 ART suppressed individuals stimulated with RNA encoding for Gag, Rev, Vpr, and Nef from the study participant’s autologous virus. The vaccine induced multifunctional effector-memory CD8+ T-cell responses, able to express IFN-γ, IL-2, TNF-α, CD107a and granzyme B. Following the ATI, no difference in the end-of-ATI viral load (VL) between the two arms or the VL setpoints post-ATI compared to pre-ART was observed (37). More recently, Sneller et al. reported that a DNA vaccine containing genes encoding the clade B proteins Gag/Pol/Nef/Tat/Vif and Env, followed by a boost with a live attenuated recombinant vesicular stomatitis virus encoding clade B Gag did not prevent viral rebound following ATI in 31 participants in whom ART was initiated during the early stage of HIV infection (38). The vaccine regimen produced a modest increase in HIV specific CD4+ T-cells but did not augment HIV specific CD8+ T-cells.

These examples suggest that therapeutic immunization and ART, compared with ART alone, generated HIV-1-specific cellular immunity but did not lead to better virological control of HIV-1 after discontinuation of ART. These data furthermore indicated that in addition to solely expanding the magnitude of HIV specific T-cell responses, other factors, including more defined T-cell characteristics might be critical for post-vaccine viral control. Several potential key requirements for vaccine-induced T-cell immunity have therefore been identified:

Specificity:

The HIV-specific T-cell response is often narrow and directed against HIV epitopes that are already escaped.

Broad T-cell responses targeting both dominant and subdominant epitopes are essential in the context of high viral diversity and fast-emerging viral escape (39, 40), specifically for archived viral resistance in the HIV reservoir (26). Thus, a therapeutic vaccine strategy will need to induce broad and ideally new epitope-specific responses that have not already experienced immune selection pressure.

Quantity:

The HIV specific T-cell compartment contracts in the absence of viral antigen during ART.

Maintaining sufficient frequencies of HIV specific CTLs in peripheral blood but also within tissues is a goal of any therapeutic vaccine concept.

Function:

Immunodominant T-cell responses elicited are often of poor quality.

Studies in HIV controllers suggest effective viral control may be rendered by polyfunctional cytotoxic T cells that have superior cytotoxic activity, are able to release multiple cytokines simultaneously and inhibit viral replication (12, 13, 41).

Location:

The HIV reservoir resides in tissues lacking CTL penetration.

It has been shown that HIV-infected cells are frequently found in B-cell follicles, where follicular CD4+ T-cells are more permissive to HIV infection. Due to the lack of follicular homing receptors, i.e. CXCR5, CTLs fail to accumulate in these “sanctuary” sites (42, 43), and therefore cannot execute their antiviral function. In general, as the HIV reservoir is likely ubiquitous in the host, the vaccine induced immune response needs to likely be present within all tissues and compartments at sufficient numbers to be able to rapidly respond to reactivating reservoir cells.

Durability:

T-cell response longevity is necessary for maintenance of viral control.

Due to HIV’s prolonged period of latency and the difficulty of eliminating all HIV reservoirs in vivo, it is important to maintain long-lived CTL responses that can suppress HIV revitalization over the remaining decades of a PLWH’s life. Otherwise, repeated revaccination will be necessary throughout a PLWH’s lifetime.

A therapeutic vaccine concept that succeeds in achieving a functional, if not sterilizing, cure must likely overcome each of these obstacles, and a number of approaches are being developed to conquer these challenges.

Inducing a broad repertoire of T-cell specificities

During primary infection, HIV-specific CTLs have a significant role in suppressing HIV-1 replication and control viremia (44, 45). Due to this strong selective pressure, HIV often quickly acquire mutations to escape from CTL recognition. These escaped viral strains not only circulate in plasma but also exist in latent reservoirs (16, 46, 47). Unless ART is initiated very early in infection, the latent reservoir becomes almost completely dominated by variants resistant to dominant CTL responses (26). To overcome viral escape to CTL responses, two strategies are generally considered: 1. Target a broad range of HIV epitopes simultaneously so that broad CTL responses can overcome reservoir escape mutations and eliminate virus reservoirs before new mutations arise; 2. Target only the most conserved regions of HIV genome, where CTL escape mutations may significantly lower viral fitness. Both strategies have been extensively explored and important progress has been made so far.

One way to elicit broad CTL response is to improve the presentation of antigens by antigen-presenting cells (APCs), particularly dendritic cells (DC). In HIV-infected individuals, DCs are greatly reduced in both quantity and function due to viral infection (48, 49). DC-based therapeutic vaccines aim to improve presentation of HIV antigens through in vitro DC priming with HIV antigens in order to elicit broad T cells responses, preferentially against previously untargeted epitopes. Indeed, treatment of ART suppressed individuals with ex vivo-generated IFNα dendritic cells loaded with LIPO-5 (HIV-1 Nef, Gag and Pol lipopeptides) induced and/or expanded HIV-specific CD8+ and CD4+ T cells directed against dominant and subdominant epitopes across all vaccine regions (50). García et al. reported a clinical trial of a therapeutic DC-based vaccine, in which chronically infected, untreated individuals were given autologous monocyte-derived dendritic cells (MD-DCs) pulsed with heat-inactivated autologous virus. A decrease in plasma viral load setpoint of >/=1 log was observed in conjunction with increased HIV-1-specific CD4+ and CD8+ T-cell responses against Gag, Nef and Env. The increase in HIV-specific T cell responses observed after vaccination tended to be correlated with the decrease in plasma viral load in vaccinated patients (51). Coelho et al. systematically reviewed twelve DC-based immunotherapies, 38% of which showed some immunogenicity that was associated with transient plasma viral load control (52). However, due to the natural advantage of immunodominant HIV-1 epitopes in endogenous processing and MHC presentation (53), it is difficult to specifically direct CTL responses towards e.g. subdominant epitopes. To this point, an ongoing clinical trial (NCT03758625) aims to test if autologous DCs loaded with a conserved HIV Gag and Pol peptide pool or inactivated autologous HIV will yield broader T-cell responses. Another vaccine aimed at eliciting broad CTL responses was HIVAX, a replication-defective HIV-1 strain pseudotyped with vesicular stomatitis virus G protein (VSV-G). VSV-G pseudotyped HIV have increased infectivity and can infect other immune cells including DC and Langerhans cells (54). Again, the idea was to improve the presentation of HIV epitopes from APC to T cells to elicit a broad range of T cell responses. Tung et al. conducted a clinical trial using HIVAX in HIV-1 infected individuals under active ART and found that HIVAX induced a higher magnitude of CD8+ T cell response than CD4+ T cell response when re-stimulated with Gag peptides and demonstrated that five of the seven participants had a significant reduction of viral load in comparison to their pre-HAART levels after ATI (55).

Along these lines, achieving simultaneous coverage of diverse viral strains and clades by a “global” vaccine has been pursued. Polyvalent ‘mosaic’ antigens have been designed to optimize cellular immunologic coverage of global HIV-1 sequence diversity (56). Compared with consensus or natural sequence HIV-1 antigens, mosaic HIV-1 Gag, Pol and Env antigens expressed by recombinant, replication-incompetent adenovirus serotype 26 (Ad26) vectors markedly improved both the breadth and depth of antigen-specific T cell responses in rhesus monkeys without compromising the response magnitude (57, 58). Phase I/IIa clinical trials of an Ad26-based mosaic vaccine conducted in 393 HIV-uninfected participants demonstrated its ability in eliciting broad T-cell responses. In this study, the mosaic Ad26 plus a high-dose gp140 boost vaccine induced cellular immune breadth covering a median of nine to ten epitopes which was substantially greater than that reported for other Ad5-based and Ad26-based vaccines expressing natural sequence antigens (median of one epitope) (59–61). A phase I/IIa clinical trial completed in 2018 to test the safety and efficacy of Ad26 prime and modified vaccinia Ankara (MVA) boost combination with mosaic inserts in a HIV-infected cohort in Thailand, that initiated ART during the acute infection (NCT02919306). Preliminary data from this trial demonstrated that 100% of the vaccinees responded with significant increase in CD8+ and CD4+ T-cell responses compared to the placebo group. Our group is currently conducting a similar clinical trial of Ad26 prime/MVA boost mosaic vaccine with and without gp140 protein boost (NCT03307915) on individuals who started ART during the chronic phase of the infection but with preserved immunity.

Other groups have focused their immunogen design on more conserved epitopes with the idea that these epitopes are constrained and cannot mutate without affecting viral fitness, thereby increasing antiviral effectiveness of T-cell responses that target such epitopes (62, 63). The first-generation conserved region T-cell immunogen HIVconsv was designed by assembling the 14 most conserved regions of HIV-1 proteome into one chimeric protein (63). In the HIV-CORE002 trial in HIV-negative adults, HIVconsv vaccine vectored in simian adenovirus and MVA induced high levels of effector CD8+ T cells that recognized virus-infected autologous CD4+ T cells and inhibited HIV-1 replication of several HIV-1 isolates by up to 5.79 log₁₀ in vitro. The virus inhibition was mediated by both Gag- and Pol- specific effector CD8+ T cells targeting epitopes that are typically subdominant in natural infection (64). Another clinical trial HIV-CORE004 tested this pan-clade HIVconsv vaccine in HIV-1-negative adults in Nairobi and demonstrated efficacy in inducing high frequencies of broadly HIVconsv-specific plurifunctional T cells, which inhibited viruses from clades A, B, C, D in vitro (65). These results demonstrated the potency of HIVconsv in eliciting broad CTL responses, at least in healthy HIV negative individuals, and thus led to their clinical exploration in HIV-1-infected individuals for evaluation of therapeutic efficacy. The BCN01 trial was a phase I, non-randomized study in individuals of early HIV-1 infection to evaluate the safety and efficacy of HIVconsv vectored on ChAdV63 (chimpanzee adenovirus serotype 63) prime and MVA boost. In the vaccinated group, there was a marked shift in immunodominance profiles of HIV-1-specific CD8+ T cell responses towards conserved T-cell epitopes, along with high in vitro viral inhibition capacity without signs of immune exhaustion (66). However, the efficacy of HIVconsv vaccine in reduction of virus reservoirs or control of viral load after ATI remains to be evaluated in further studies.

Based on this first-generation conserved-region vaccine, Ondondo et al. developed a second-generation conserved vaccine expressing novel immunogens tHIVconsvX which combines regions of HIV-1 proteins functionally conserved across all M group viruses, bivalent complementary mosaic immunogens and epitopes associated with low viral load in HIV controllers. When vectored by a combination of DNA, simian adenovirus and MVA, tHIVconsvX elicited broad CTL responses, whose magnitude and breadth correlate with low viral load and high CD4+ T cell count in HIV-infected treatment-naïve individuals (67). As a result, an MVA-vectored vaccine expressing tHIVconsv3 and tHIVconsv4 immunogens (derived from tHIVconsvX) was initiated last year to evaluate its safety and efficacy in HIV-infected adults (NCT03844386). Other approaches that target HIV conserved regions are showing promises as well. For example, the PENNVAX-B vaccine, synthetic plasmids expressing multiclade HIV Gag and Pol as well as a consensus Clade B Env, elicited expanded CTL responses towards the multiclade Gag/Pol/Env antigens and improved CTL lytic granule loading activity in a small cohort of HIV-infected individuals (68). A larger-scale phase II trial is currently enrolling to further assess if PENNVAX could lead to a significant reduction of HIV reservoir size (NCT03606213). Another DNA vaccine that carried HIV-derived conserved element (CE) p24 Gag DNA was shown to redirect the cellular responses to subdominant Gag epitopes in rhesus macaques (69). A clinical trial to test its potency in eliciting novel CTL responses in PLWH is fully enrolled and results are pending (NCT03560258).

Overall, all strategies are still under active development and novel concepts are being proposed. As an example, Gaiha et al. suggested recently an alternative structure-based network analysis that combined protein structure data and network theory to quantify the topological importance of each amino acid residue and evaluate and rank HIV CTL epitopes. CTL epitope targets of high topological importance successfully distinguished HIV controllers from progressors irrespective of HLA alleles (70) suggesting that these key conserved epitopes could be targeted by a vaccine. The future will show if one or another or potentially even a combination of consensus, conserved, mosaic or networked epitope approaches will be most successful.

In addition, no vaccine strategy even if it elicited broad CTL responses, has proven efficacious in substantially reducing the HIV reservoirs or inducing sustained virus control in human clinical trials. It therefore seems that broad CTL responses are necessary but that vaccine regimens have to also focus on additional T-cell characteristic, e.g. functionality.

Quality of CTL responses including polyfunctionality, proliferation and cytotoxicity is crucial

As the size of the CTL response is important in cellular immunity, the quality of CTL responses has been shown to be crucial in controlling viral load. Studies in HIV controllers demonstrated that although the overall breadth and magnitude of CD8+ T-cell responses induced in elite controllers is comparable to HIV progressors (HIV-infected individuals with high-level viremia and progressive immune destruction), several qualitative differences in T-cell characteristics have been described (11): CD8+ T cells from elite controllers can effectively inhibit HIV-1 replication in ex vivo-infected autologous CD4+ T cells (12, 71), mediated by increased expression of cytotoxic granule components (13, 72); controller CD8+ T cells are more likely to be polyfunctional, that is, to simultaneously execute multiple effector functions (degranulation, secretion of IFN-γ, MIP-1β, TNF-α and IL-2, cytotoxicity etc.) (41); in some cases, within an HLA restricted epitope response, differences in T-cell receptor clonotypes between controllers and progressors lead to differences in effectiveness in inhibiting virus infection (73). These cumulative data suggest that high-quality cytotoxic T cells might be a correlate of protective immunity; however, inducing these responses, particularly in an immune environment that has often been damaged by persistent HIV infection, is challenging.

Several studies have shown that the quality of HIV-specific CTL responses can be enhanced by DC-based vaccines. Lu et al. conducted a clinical trial utilizing a vaccine consistent of autologous monocyte-derived DCs pulsed with autologous inactivated HIV-1 in patients with chronic HIV-1 infection. This vaccine induced HIV-1 Gag-specific CD8+ T cells by more than threefold at day 112, whereas the percentage of HIV-1 Gag-specific CD8+ T cells expressing perforin increased around twofold. While after 1 year the percentages of total Gag-specific CD8+ T cells did not correlate with plasma viral load decrease, HIV-1 Gag-specific CD8+ T cells expressing perforin were correlated with lower plasma viral loads (74). This study was consistent with previous reports about the importance of perforin expression in effector CD8+ T cells, and their enrichment in HIV non-progressors compared to individuals with progressive disease (72, 75). Another study using a DC-based vaccine also led to increased functionality in HIV-specific T-cell immunity: Lévy et al. administered autologous DC pulsed with HIV lipopeptides from Gag, Pol and Nef to patients on ART, and the vaccine regimen increased both the breadth of epitope specific immune responses and the frequency of functional CD8+ T cells producing at least two cytokines across IFN 2. In this study a correlation was found specifically between lower viral loads following ATI and the frequency of polyfunctional CD4+ T cells (76).

A number of other therapeutic vaccine candidates also demonstrated improved T cell qualities. Casazza et al. evaluated a vaccine regimen that combined an ad5 vector encoding clade B Gag and Pol, and clade A, B, and C Env with a DNA vaccine encoding the same antigens plus clade B Nef in 17 ART suppressed individuals. The vaccine improved the polyfunctionality of HIV-specific CD8+ T cells, the proportion of CTLs able to co-produce IFN-γ, Mip1b and TNF α increased, but these responses did not impact viral loads, at least at the single copy level (77). Harari et al. showed that a poxvirus-based vector, NYVAC, expressing Gag, Pol, Nef and Env, administered to 10 ART suppressed individuals, enhanced both polyfunctionality, as determined by co-production of IL-2, IFN-γ, or TNF-α and proliferative capacity of HIV-specific CD8+ T cells, but its efficacy in viral control was not measured in this study (78). Lind et al. demonstrated that Vacc-4x, a peptide vaccine derived from conserved domains of HIV-1 p24 Gag, given the 25 ART suppressed individuals, increased Vacc-4x-specific CD8+ T cell degranulation and IFN-γ production and proliferative capacity was even increased in 80% of vaccinated study participants (79).

Although these results suggest, that CTL qualities can be optimized by vaccines, it remains unclear how durable these CTL qualities are. Moreover, the underlying mechanism of CTL polyfunctionality, proliferation capacity and cytotoxicity remain to be fully understood and further research on CTL function may provide new directions for future HIV therapeutic vaccine design. For example, a mechanistic study led by Chiu et al. showed that HIV-specific T-cell polyfunctionality can be enhanced by inhibition of sprout-2, a negative regulator of the MAPK/ERK pathway, offering a new potential target for boosting polyfunctionality (80). Another study pointed out that therapeutic vaccination of HIV can increase the frequency of regulatory T cells that suppress the polyfunctionality of CD8+ T cells, suggesting that the role of Tregs should also be considered in HIV therapeutic vaccine trials (81).

In addition, functional avidity is also an important CTL quality proven to be associated with superior control of HIV-1 replication (82, 83). Functional avidity measures the sensitivity of T cells in recognizing cognate antigens and is determined by the ratio of recruited clones with high versus low avidity among a heterogenous oligoclonal T-cell population. It has been suggested that the lack of an immunodominant response to a specific epitope could be caused by competition between multiple CD8+ T cell clonotypes (84). How to selectively stimulate high-avidity HIV-specific T cells by vaccination remains unclear. Recently, Billeskov et al. showed that vaccination with low-dose antigen combined with a cationic liposomal adjuvant selectively primed CD4+ T cells of higher functional avidity, consequently improving CD8+ T-cell antiviral effects (85). Hu et al. reported that vaccination with vaccinia virus in mice significantly enhanced the functional avidity of antigen-specific CD8+ T cells through the intrinsic MyD88 pathway independent of TCR selection (86). Alternatively, gene therapy approaches that manually select and engineer TCR for high functional avidity are also being explored (87, 88).

Effective CTL penetration to HIV reservoirs in all locations

One of the most challenging barriers for HIV eradication is the prevalence of HIV reservoir cells in a wide variety of tissues and with distinct viral and cellular signatures (89). Even when effective CTL responses are stimulated and maintained, HIV can still evade being cleared given the poor accessibility of some tissue reservoirs and reduced CTL penetration into these compartments.

Because of the high susceptibility of follicular CD4+ T cells to HIV infection, poor follicular CTL penetration, and significant extracellular HIV virion accumulation on the surface of follicular dendritic cells (FDCs), B cell follicles are considered a critical sanctuary for HIV replication and persistence (90). HIV-specific CTL are abundant within lymphoid tissues, but fail to accumulate within lymphoid follicles where HIV-1 replication is concentrated (43). Recent work by Louis Picker validated that productive SIV infection in rhesus monkey elite controller is restricted to follicular CD4+ T helper cells (T_FH)) (42) due to the physical exclusion of CD8+ T cells from B cell follicles (91). A study of peripheral blood from chronically HIV-infected individuals on ART also showed that within central memory CD4+ T cells, peripheral T follicular helper cells are the most significant HIV reservoirs (92). Follicular cytotoxic T cells, a group of cytotoxic T cells expressing CXCR5, were found to localize to B cell follicles and control viral infection of T_FH cells (93). Thus, boosting cytotoxic CTLs with specific homing signatures like CXCR5 could be valuable in eliminating the viral reservoir in B cell follicles. An alternative strategy could be to chemically induce or genetically engineer CTL to express CXCR5 (94). Indeed, transduction with CXCR5 effectively localized CTL to B cell follicles in rhesus macaques, but if these cells were enough to clear the B-follicular HIV reservoir remains unknown (95, 96). Other strategies include treatment with IL-15, use of bispecific antibodies and disruption of the follicle by medication such as rituximab (97–100).

The central nervous system (CNS) is another compartment that might not be easily targetable for vaccine mediated HIV reservoir elimination approaches. While it is populated with macrophages and astrocytes that might be susceptible to HIV infection these cell types are potentially less responsive to CTL elimination (reviewed in (101)). Furthermore, the blood-brain barrier (BBB) by regulating the trafficking of cells might prevent the free influx of vaccine induced T-cell responses and If prior therapeutic vaccines have induced relevant HIV specific CD8+ T cell response in the CNS has not be examined. Finally it remains questionable if highly cytolytic CD8+ T cell response should be induced in the CNS via therapeutic vaccination, as massive infiltration by polyfunctional CD8+ T-cell could result in uncontrolled astrocytic and microglial activation with detrimental consequences for the vaccinees (102).

Other cell types than CD4+ T-cells, in particular macrophages, serve as harbor for latent HIV. Traditionally, macrophages are believed to have a moderate life span and lack self-renewing potential, but recent data suggested the role of these cells as long-lived HIV reservoirs even during active ART (103, 104). Moreover, Clayton et al. showed that macrophages were intrinsically more resistant to CTL mediated killing relative to CD4+ T cells (105). New strategies to enhance CTL activities towards HIV-infected macrophages remain to be investigated.

Combinatorial strategies to enhance the efficacy of CTL-based therapeutic vaccines

In light of HIV latency, durable viral remission will likely require the combined reactivation of the latent virus coupled with potent HIV reservoir elimination strategies. Latency reversal agents (LRA) can activate latently infected T cells and improve recognition by both CTLs and antibodies. The key to successful implementation of this so-called “shock and kill” method is to find an effective yet safe LRA. Many LRA candidates have been proposed and some have been tested in clinical trials (106, 107). Among them, histone deacetylase inhibitors have initially demonstrated some of the most promise. For example, Elliott et al. revealed that short-course vorinostat use in HIV-infected ART-suppressed patients could activate HIV transcription, demonstrated by increased level of HIV RNA in total CD4+ T-cells from blood (108). A combinatorial study called the RIVER (Research in Viral Eradication of HIV Reservoirs) was conducted with newly HIV-infected adults to assess the combinatorial effect of a therapeutic vaccine ChAdV63.HIVconsv prime with MVA.HIVconsv boost followed by vorinostat administration. As the first randomized human trial using the “shock and kill” cure strategy, the RIVER study regimen induced significantly higher HIV-specific CD4+ and CD8+ T-cell responses in the intervention group, but there was no difference between intervention group and control group for the primary endpoint, that of change in the viral reservoir as measured by total HIV DNA/million CD4+ T cells. Another histone deacetylase inhibitor romidepsin was also under clinical investigation together with HIV therapeutic vaccine candidates. BCN02-Romi enrolled 15 individuals from the aforementioned BCN01 trial to assess the clinical effects of MVA.HIVconsv vaccine combined with romidepsin. Preliminary results revealed that 4 out of 13 subjects (31%) controlled viral rebound beyond four weeks, the first report of viremic control in a large proportion of participants (109, 110).

Toll-like receptor agonists were also discovered to be effective latency reversing agents in vitro (111–113), and a number of them are in clinical trials to evaluate their safety and efficacy. The TLR7 agonist vesatolimod (GS-9620) was shown to reverse latency state and reduce viral reservoir size in simian-immunodeficiency virus (SIV)-infected rhesus macaques on ART (114), and has undergone multiple clinical trials to evaluate its therapeutic effects in HIV-infected individuals on ART and were overall well tolerated, with the expected immune activation, however, no direct effect on virological markers were observed (115). When combined with the Ad26/MVA mosaic vaccine (as described before), the TLR7 agonist GS-9620 improved virologic control and delayed viral rebound following ART discontinuation in SIV-infected rhesus monkey (116). Human clinical trials to test T-cell based therapeutic vaccines combined with TLR7 agonists have started. For example, the ongoing human clinical trial AELIX-003 (EudraCT 2018–002125-30) is aimed to evaluate the safety and efficacy of vesatolimod with AELIX therapeutic’ HTI vaccine that contains conserved HIV antigenic regions against which T-cell responses are enriched in HIV controllers (117, 118). Overall, the combinatorial strategy of a potent CTL-based vaccine with an effective LRA is currently under extensive exploration.

Cytokines can support the expansion of stimulated virus-specific T-cells, when administered as a vaccine adjuvant. A number of cytokines have been combined with T-cell vaccines in clinical trials to boost T-cell expansion and function, including IL-2, IL-7, IL-12, IL-15 and IL-21. IL-2 has been considered as T-cell survival signal, but it also promotes Treg expansion. The ANRS 093 study combined the ALVAC-HIV and Lipo-6T vaccine with IL-2 administration to induce sustained and broad HIV-specific CD4+ and CD8+ T cell responses in chronically HIV infected individuals on ART (119) and the T-cell responses correlated with the reduction of the time off antiviral therapy. Direct administration of IL-7 to ART-treated HIV-infected individuals induced a sustained expansion of naïve and central memory CD4+ and CD8+ T cells (120). This effect could be advantageous in the context of therapeutic immunization by potentially increasing the number of vaccine-elicited HIV-specific T cells, but it could also pose a risk due to an expansion in the absolute number of circulating CD4+ T cells harboring integrated HIV DNA (121). IL-15 might also be a promising tool for HIV immunotherapy given its role in enhancing the survival of HIV specific CD8+ T cells (122): co-immunization with a chimeric Simian-Human Immunodeficiency Virus (SHIV) antigen and IL-15 plasmid elicited protection against SHIV in macaques (123); however, in a different study, IL-15 abrogated a vaccine-induced decrease in virus level in SIV-infected macaques (124), again indicating both pathogenic and therapeutic potentials of immune cytokines (reviewed in (125)). More interestingly, heterodimeric interleukin-15 (hetIL-15) or an IL-15 superagonist were shown to activate and redirect SIV-specific CD8+ T cells from peripheral blood to B-cell follicles by upregulating their CXCR5 level (126, 127). IL-21 produced by CD4+ T cells is considered necessary for the maintenance of CD8+ T cell function, especially in HIV controllers (128); a clinical trial combining IL-21 with ART helped to restore intestinal immune cells and reduced viral reservoirs in SIV-infected macaques (129).

It has been demonstrated that immune exhaustion is under the active control of inhibitory immune checkpoints such as programmed cell death-1 (PD-1) and cytotoxic T-lymphocyte associated protein (CTLA-4), whose selective blockade can reinvigorate antigen-specific CD8+ T cell function (130–132). Therefore, when combined with HIV-specific T-cell vaccines, immune checkpoint blockers (ICB) could help reverse exhaustion and unchain cytotoxic T cells to target and kill infected cells to decrease the virus reservoir. To test this strategy, studies with SIV-infected rhesus macaques have revealed that anti-PD-1 antibody can expand functional virus-specific CD8+ T cells and lower SIV RNA levels in plasma (133). The first clinical trial of an anti-PD-L1 antibody in HIV-infected individuals was conducted in 2017 by Bristol-Myers Squibb, but resulted only in a modest increase in Gag-specific CD8+ T cells expressing interferon-γ in a subset of study participants (134). More studies are needed to validate this strategy.

Conclusion

The past 30 years of HIV research have witnessed huge progress in our understanding of HIV immunology. As the field continues to explore a wide range of approaches for curing HIV/AIDS, therapeutic vaccination remains one of the most cost-effective, non-invasive and promising cure strategies. From studies in HIV controllers, we have observed that potent, polyfunctional T-cell responses can effectively suppress viral load in the absence of ART. However, eliciting polyfunctional T-cell responses in all HIV-infected individuals independent of HLA background, disease state etc. remains a challenge.

Here we have laid out the potential cornerstones for a successful therapeutic vaccine: a vaccine that induces broad, high-quality, follicle-penetrating HIV-specific T-cell responses, that are durable and rapidly responsive to clear reactivating reservoir cells. Together with the help of ICB, LRA and/or other immunomodulatory support, to recover immune exhaustion and reverse viral latency, future studies will show if the concept of therapeutic vaccination will ultimately be successful.