NEW APPROACHES TO DESIGN HIV-1 T CELL VACCINES

Abstract

Purpose of review

Following the evidence that T cell responses are crucial in the control of HIV-1 infection, vaccines targeting T cell responses were tested in recent clinical trials. However these vaccines showed a lack of efficacy. This review attempts to define the qualitative and quantitative features that are desirable for T cell induced responses by vaccines. We also describe strategies that could lead to achievement of this goal.

Recent findings

Using the yellow fever vaccine as a benchmark of an efficient vaccine, recent studies identified factors of immune protection and more importantly innate immune pathways needed for the establishment of long-term protective adaptive immunity.

Summary

To prevent or control HIV-1 infection, a vaccine must induce efficient and persistent Ag-specific T cells endowed with mucosal homing capacity. Such cells should have the capability to counteract HIV-1 diversity and its rapid spread from the initial site of infection. To achieve this goal, the activation of a diversified innate immune response is critical. New systems biology approaches will provide more precise correlates of immune protection that will pave the way for new approaches in T cell based vaccines.

Introduction

The HIV-1 pandemic is one of the leading causes of death worldwide and remains a serious challenge to global public health [1]. Although antiretroviral drugs can control HIV/AIDS progression in many patients, they only succeed in reducing viral loads without completely eliminating the virus [2, 3]. The development of an effective HIV-1 vaccine represents the optimal solution for control of the HIV-1 pandemic. While this is a clearly agreed upon goal, its implementation has been a difficult task [4–6]. HIV-1 vaccine studies have led to disappointing results, likely a consequence of the difficulty in generating broadly cross-neutralizing antibodies [7–9]. Recently the focus has shifted towards vaccines that control viral load after infection, thereby reducing secondary transmission [10]. In individuals exposed to HIV-1 yet remaining uninfected, CD8 T cell-mediated immunity was shown to be critical for the resistance to HIV-1 acquisition [11–15]. In non-human primates (NHP), the magnitude and kinetics in the establishment of effector CD8 T cell responses upon exposure to SIV were correlated with the control of acute infection [16–19]. In humans, the initial peak of T cell responses was shown to be temporally associated with a decrease in viremia [20, 21]. Furthermore, the immune selective pressure exerted by T cell responses induced an accumulation of viral mutations concentrated within T cell epitopes [22–25]. These observations that T cell responses are critical in controlling HIV-1 acquisition and infection have led to the development of vaccine strategies targeting T cells that showed promising results when tested in NHP [26–29].

Several candidate HIV-1 vaccines have been tested; however only four of them have reached phase IIb/III (efficacy) clinical trials [30]. The Step and Phambili trials, using as vaccine a replication-incompetent recombinant adenovirus 5 (Ad5) expressing HIV-1 clade B gag, pol and nef, were stopped before completion [6, 31–33]. Protective immunity was not observed in highly exposed individuals despite the induction of HIV-specific T cell responses in 80% of the vaccinees. Moreover, the incidence of HIV-1 infection was increased in subgroups of Ad5-preimmunized and uncircumcised male vaccinees [6, 34, 35]. The more recent RV144 clinical trial in Thailand, based on priming with a canarypox vaccine ALVAC HIV vCP1521 (env, gag, pol) and a boost with the HIV-1 gp120 AIDSVAX B/E recombinant protein, showed encouraging results with an overall reduction in HIV-1 acquisition of 31.2% compared to placebo [36, 37]. Ag-specific CD4 T cell proliferative responses were measured in 60% of the vaccinees and Ag-specific CD8 T cell responses were detectable in around 20% of the vaccinees by IFN-γ ELISPOT and cytotoxic assays. However, the vaccination had no effect on the CD4 T cell count or viral load in subjects subsequently diagnosed with HIV-1 infection [36]. Despite its low efficacy, this vaccine strategy provides hope that protective immunity against HIV-1 acquisition may be achieved.

A successful vaccine against HIV-1 must overcome several obstacles including the diversity of the virus and the early establishment of latent viral reservoirs [38, 39]. The characteristics of HIV-1 and its immunopathology represent a major challenge for immunologists. Moreover, fundamental correlates of immune protection still need to be defined and validated for the design of novel vaccine strategies. In this review, we will examine the T cell immune responses to HIV-1 infection and those elicited by efficient vaccines with the aim of defining desirable T cell characteristics that should be targeted to prevent or control HIV-1 infection. We will focus on vaccine strategies that engage the innate immune compartment due to its crucial role in shaping an efficient T cell response.

How to counter HIV-1 antigenic diversity?

The diversity of circulating viral strains and the rapid generation of viral variants during infection constitute a major obstacle in the development of an HIV-1 vaccine [25, 38, 40–42]. The antigenic diversity must be represented in the vaccine components to provide broad T cell responses. Indeed, in SIV infection, T cell correlates of protection have been associated with a broader epitope-specific repertoire prior to heterologous SIV challenge [43]. It was shown that HIV-1 infected individuals whose CD8 T cell responses are dominantly and broadly directed against the gag protein exhibit lower plasma viral load [44]. To date, only 1 to 3 HIV-1 strain sequences were used in vaccine design and the lack of representation of these actual sequences in the infecting virus isolates could be one of the reasons behind the inefficacy of such vaccines [6, 30, 33, 45]. New strategies were recently developed to improve “immunological coverage” by consensus and optimized mosaic Ags, which assemble synthetically designed antigenic sequences within several clades to generate full-length mosaic HIV-1 proteins (Table 1 and Figure 1) [46]. Studies in NHP showed that T cell responses induced by these mosaic Ags increased the breadth of the response, as Ag-specific T cells were cross-reactive to multiple HIV epitopes and variants [47–49]. However, the specificity of T cells against natural epitopes and the protective effect induced by such vaccines against HIV-1 infection still need to be demonstrated [50–52].

Table 1.

Comparison between the CD8 T cell immune response during HIV-1 infection and an efficient immune response that should be induced by a protective vaccine.

CD8 T cell response characteristics	HIV-1 infection	Ideal Vaccine
Magnitude	Not associated to protection	Moderate or High
Breadth	Narrow/escape mutations	Broad and cross reactive
Avidity	Moderate/low	High
Functionality	Mono-functional responses	Polyfunctional responses
Proliferation	Impaired	Sustained
Memory	Impaired	Sustained
Timing of homing	Too late	Early

Figure 1.

To elicit an efficient T cell response, a vaccine must target innate immune cells, and in particular the DCs: viral vectors that infect DCs, adjuvants, TLR-ligands or costimulatory molecules should be included in the vaccine composition to improve Ag-presentation and T cell priming. The breadth of the vaccine-induced T cell response can be improved by using optimal vaccine antigens, as mosaic Ags that enhance the recognition of multiple HIV-1 epitopes and variants within theses epitopes by cross-reactivity. Defined as correlates of HIV-1 disease control, the polyfunctionality, proliferative capacity and functional avidity of CD8 T cells should be induced. To achieve a long-term immune protection, HIV-1 T cell vaccine must induce the generation of long-lasting central memory T cells (TCM) and mostly effector memory T cells (TEM) that are more likely to act rapidly in the mucosal site of HIV-1 transmission. The development of novel in vitro models, immuno-monitoring and systems biology approaches will provide a comprehensive analysis of the complex networks between both arms of the innate and the adaptive immunity and allow to define precise correlates of immune protection and key targets for an efficient HIV-1 T cell vaccine.

What characterizes an efficient HIV-specific T cell response ?

Live attenuated viruses such as Yellow Fever 17D (YF-17D) or vaccinia virus (VV) are amongst the most efficient vaccines and studying immune responses to these vaccines should reveal correlates of immune protection. YF-17D mimics an acute viral infection and induces innate and adaptive immune responses, with a balanced Th1/Th2 response, leading to a long-term (10 to 60 years) efficient immune protection [53–55]. In HIV-1 infection, the maintenance of proliferative CD4 T cell responses has been associated to the long-term control of HIV-1 infection [56]. However, the role of CD4 T cells in the disease control is still unclear, as the induction of HIV-specific CD4 T cells has also been suggested to enhance HIV-1 infection by providing an activated pool of target cells for viral replication [39, 57, 58]. Therefore, as the correlates of immune protection against HIV-1 disease progression are mostly demonstrated for CD8 T cell responses, HIV vaccine strategies mainly focused on the induction of strong CD8 T cells responses have been employed and should be pursue (Figure 1) [59–61].

Vaccination by vaccinia virus (Dryvax) or modified vaccinia virus Ankara (MVA) was shown to be particularly efficient in eliciting Ag-specific CD8 T cells with a high degree of polyfunctionality [62]. Following vaccination with YF-17D, Miller et al. showed that more than 10% of total circulating CD8 T cells are activated within 2 weeks [63]. In contrast to persistent viruses as HIV-1 that lead to chronic infection and T cell exhaustion, YF-17D and VV vaccines elicit a rapid expansion of highly specific and polyfunctional Ag-specific CD8 T cell expressing only transient levels of inhibitory receptors including PD-1 and CTLA-4 (Table 1) [62–68]. Importantly, it appears that T cell mediated immune protection is more likely related to the quality than the quantity of Ag-specific T cell responses. Indeed, high levels of HIV-1 responding CD8 T cells could be seen both in progressors and in long term non progressors (LTNPs) [69–73]. During HIV-1 infection, the proportion of polyfunctional CD8 T cells, as evidenced by their capacity to produce several cytokines (IFN-γ, MIP1-β, TNF-α, IL-2), was shown to inversely correlate with viral load [69]. Similarly, the response to HIV-2, which displays a slower disease progression, was characterized by the generation of polyfunctional CD8 and CD4 Ag-specific T cells [74]. Nevertheless, the polyfunctional T cell responses demonstrated in the Step trial was not sufficient to confer any protection [75]. Indeed, if the correlation between T cell polyfunctionality and HIV-1 disease control is now well established, it is still unclear if this feature is sufficient to provide T cell immune protection [51, 76–78]. Other parameters such as proliferative capacity and functional sensitivity have also been associated with the efficiency of Ag-specific CD8 T cells to suppress HIV-1 replication [79] (Table 1 and Figure 1). For inducing an efficient CD8 T cell response and a balanced Th1/Th2 response, several critical parameters should be considered as the Ag dose or the vaccine regimen (e.g. heterologous prime-boost strategies) [80–82]. Importantly, valuable correlates of immune protection still need to be defined and will be elucidated by systems biology approaches.

How to induce long-term immune protection?

The generation of a specific immunological memory that can protect individuals throughout their lifespan, represents one of the major features that determines the success of vaccines and strongly depends on the efficiency of the primary effector response [83–85]. Long-lasting immunological memory is based on heterogeneous CD4 and CD8 T cell sub-populations classically divided into the long-lived central memory T cells (TCM) and the effector memory T cells (TEM) [86, 87]. YF-17D and VV vaccination result in the rapid and massive expansion of effector CD8 T cells that gradually differentiate into a memory pool providing immune protection for more than 10 years [63]. Furthermore, a progressive downregulation in activation and proliferation markers and effector functions (HLA-DR, CD62L, CD38, Ki-67, Granzyme B) is observed as well as the acquisition of memory markers (CD127 and Bcl-2) [63, 67]. While protective immune responses are generally attributed to central memory T cells, the YF-17D specific memory CD8 T cells have a particular phenotype of effector memory T cells (CD45RA+, CD27+, CCR7−), which are able to maintain a proliferative capacity. Moreover, a recent study by Vezys et al. showed that, upon heterologous prime-boost immunization in mice, CD8 TEM cells exhibit an extensive expansion capacity in response to new infection by Lymphocytic Choriomeningitis Virus (LCMV) compared to TCM [88]. Prolonged Ag presentation might be required to increase the size of the effector memory T cell pool and to elicit a protective immune response as was recently demonstrated for the malaria vaccine in mice [89]. Indeed, the most successful vaccine strategies have been obtained using modified viral vectors (MVA and LCMV) that permit the persistence of Ags [27, 29, 62, 90–93]. Vaccine strategies should promote primarily CD8 persisting TEM. However, the induction of a long-term immune protection by a vaccine might also need the establishment of a central memory T cell pool (Table 1 and Figure 1) [94, 95].

How to block HIV-1 dissemination from mucosal tissues?

To counteract HIV-1 infection, a vaccine must induce a robust immune response before the establishment of chronic infection and prior to its sequestration into the latent viral reservoir [2, 23]. Defects in CD8 T cell function and survival have been shown to occur within the first few weeks of SIV infection [96, 97]. As HIV-1 transmission occurs most commonly through sexual transmission and the first immunopathologic events take place at mucosal sites, vaccine-induced HIV-specific T cell responses that can be recruited rapidly to those sites is critical [12, 18, 57]. The magnitude and quality of effector CD8 T cells at mucosal sites has been correlated with viral load [18]. Although TCM are considered as the stem cells for the memory pool this subset undergoes expansion only 3 days after re-exposure to their cognate Ag [98]. Moreover, in contrast to TEM, the TCM subset is poorly represented among intraepithelial lymphocytes and in gut associated lymphoid tissue and is mostly localized in the lymph nodes (LN) as dictated by the expression of the LN homing receptor CCR7 [51, 99, 100]. Therefore, vaccines inducing TEM would promote a rapid response at mucosal sites upon HIV-1 exposure (Table 1 and Figure 1). Masopust et al. recently demonstrated the induction of the transient expression of α4β7 integrin on early specific effector T cells and on memory T cells shortly after their activation by systemic re-stimulation with LCMV in mice, allowing their homing to intestinal and epithelial tissues [99]. Furthermore, a study using SIV mac239(delta)nef immunized animals showed that lung CD8 T cells, unlike peripheral CD8 T cells, suppressed viral replication by up to 80% in vitro [101]. These results suggest that this effective vaccine was eliciting functional immunity at mucosal sites with a distinct behavior from circulating cells. The Rhesus CMV, used as a vector for SIV gag, rev, tat, nef and env when injected subcutaneously, successfully induced multifunctional CD4 and CD8 effector memory T cells in mucosal tissues and provided protection of NHP against repeated low-dose SIV intrarectal challenge [27]. The route of immunization might be of particular importance for early prevention and inhibition of viral dissemination. Few studies have addressed this parameter and its impact is still debated [102].

How to shape HIV-specific T cell responses?

Innate immune cells are needed to generate an efficient adaptive immune response. Dendritic cells (DCs) are known to activate naíve T cells to generate effector and memory T cells. Therefore, the types of signals DCs provide to T cells could result in different fates of the subsequent adaptive immune response [103, 104]. However, the exact molecules and mechanisms involved in the priming of T cells and in the generation of memory are still unknown. The SIV mac239(delta)nef is the most successful vaccine tested in macaques [105, 106]. The deletion of nef avoids the inhibitory effect of this protein on MHC class I presentation and other negative regulatory mechanisms. Likewise, early activation of innate immunity might have a major role in modulating the efficacy of the vaccine. YF-17D was shown to infect and activate multiple DC subsets via TLRs 2, 7, 8, and 9, enhancing the presentation of vaccine Ags and the production of pro-inflammatory cytokines [53, 54, 107]. TLR ligands have been used to increase the frequency of CMV and HIV-specific CD8 T cell responses in vitro and also to modulate T cell responses in vivo [108–110]. Immunization with gag protein conjugated to a TLR7/8 agonist in mice or NHP enhanced the magnitude and quality of gag-specific Th1 and CD8 T cell responses [111]. Likewise, the same authors studied the use of different TLR ligands to influence Th1 and CD8 T cell responses in NHP in a prime boost immunization regimen [112]. Longhi et al. demonstrated that TLR3 agonist strongly induces type I IFN production that promotes maturation of DCs and generates CD4 Th1 immunity [113]. Type I interferons have also been shown to dictate clonal expansion, attrition and memory formation of CD8 T cells [114, 115]. The fate of CD8 T cells is also dictated by other mechanisms. For example, IL-15 trans-presentation by DCs has been shown to promote effector CD8 T cell survival, while on the other hand, CD137 signaling in DCs could lead to T cell activation induced cell death [116, 117]. Deciphering DC signals that dictate the features of T cells is needed for the development of new vaccines. Targeting DCs during vaccination would increase the magnitude but also the quality of the immune response.

The most efficient vectors for HIV-1 vaccine are those that infect and activate DCs such as the Canarypox virus (ALVAC vaccine), the LCMV or MVA [62, 91, 93, 118]. Other viral vectors, virus like particles (VLP) or recombinant viral proteins should be used in combination of costimulatory molecules and TLR ligands to activate DCs and stimulate an efficient immune response (Figure 1). These strategies have already been tested in different models and have shown promising results; for example the incorporation of CD40 ligand into simian-HIV VLP enhanced DC activation and boosted immune responses against HIV-1 [119]. Moreover, the use of an anti-DCs (DEC-205) HIV gag fusion antibody vaccine led to an intensified and protective CD4 T cell immunity in mice [120]. Nevertheless, the mechanisms that are used to activate DCs need to be clearly defined and understood as activation of a DC-T cell axis by Ad5 immune complexes has been shown to create an improved environment for replication of HIV-1 in T cells [121].

How to monitor a good protective immune response?

Few T cell assays have been used to analyze T cell responses induced by vaccination strategies and predict vaccine efficacy. They include the commonly used IFN-γ ELISPOT assay, but also tetramer staining, in vitro proliferation assays, intracellular cytokine staining assays, staining for markers of activation and cytotoxic related molecules. However, experimental procedures as well as analysis and interpretation still remain controversial as most of these assays have yet to be standardized. Moreover, considering that few memory T cells could be sufficient for a robust secondary immune response, the detection level of these assays should be improved to study the frequencies of memory cells as their frequencies rarely exceed 10⁻³ to 10⁻⁴ of total T cells. The field has begun to standardize the immuno-monitoring for clinical trials [122–124] and this will help by defining reliable assays to predict vaccine efficacy. In that context, both the IFN-γ ELISPOT as well as polyfunctional T cells failed to predict the unprotective effect of the Step trial [31, 75]. New parameters of the induced T cell responses have to be monitored, such as the differentiation status, migration patterns, proliferation potential and survival capacity. For example, the expression levels of transcription factors such as T-bet, Eomes, Blimp-1 and Bcl-6 could be investigated to define the effector and memory CD8 T cell status [125, 126]. Moreover T cells should be monitored at mucosal sites, and the capacity of memory T cells to migrate in situ should be assessed.

As the very early events in HIV-1 infection in the mucosal sites are not available for study in vivo, the impact of vaccines as well as HIV-1 infection have to be studied in vitro. Some in vitro models have been already used for mucosal reconstruction to follow virus behavior from its site of entry to the lymph nodes [127]. Randolph et al. have reported a tissue-engineered in vitro model to promote autonomous generation and maturation of DCs from PBMCs, without adding exogenous cytokines [128]. In our previous work [53], we found comparable results in the immune responses stimulated by the YF-17D vaccine in this system versus in vivo responses after vaccination. Such new models would serve as a surrogate human immune system to elucidate the first events of HIV-1 infection and would enable the testing of the response to a vaccine directly in a human system rather than in animal models.

An effective protection in response to a virus or a vaccine involves a complex polyvalent and coordinated immune response. The recent advances in biotechnological methods, associated with computational tools, permit a quantitative, qualitative and integrated analysis at genomic, proteomic and cellular levels [129, 130]. This systems biology approach offers the possibility to examine the immune status of one individual, revealing the complex networks between all of the innate and adaptive immune components over time and to define key molecules and signatures of immune protection (Figure 1) [53, 54].

Conclusion

As discussed in this review, some correlates of HIV-1 disease control have been identified and new promising T cell vaccines strategies are emerging [27, 29, 91, 93]. Results from current clinical trials such as HTVN 205 with a prime DNA vaccine containing gag, pol, env, tat, rev, vpu and a boost MVA vaccine containing gag, pol, env will provide new advances in T cell vaccine development [131]. Key factors of immune protection have also been defined in response to efficient licensed vaccines such as YF-17D [53, 54]. Further studies are still needed to understand the underlying mechanisms of immune protection, especially the crucial role of the innate immune system in activating and shaping an efficient T cell response. The characterization of innate and adaptive immune responses in novel in vitro models and in highly exposed non-infected individuals should also give important clues to the identification of correlates of immune protection. Furthermore, acute immune responses against other viruses that persist in humans but are naturally controlled should also bring insights in the first immune events that lead to viral control. In all these studies, systems biology approaches would allow for the analysis and integration of innate and adaptive immune responses, providing the tools to build models of immune protection. These findings could then be employed to define composition, dose and administration regimens in novel HIV-1 vaccination strategies. Parameters of immune protection and disease control could be used to assess the efficacy of vaccines in phase I clinical trials. Therefore, more valuable and meaningful data could emerge from phase I studies, avoiding the entry into phase II/III trials without sufficient evidences of vaccine efficacy. New correlates of protection and disease control could emerge and would provide key elements to develop a potent HIV-1 cell vaccine. As many arms of the immune system work in concert, combinations of vaccines that induce effective T cell responses along with neutralizing antibodies or other strategies inducing innate immune responses could achieve successful results [132].