HIV Vaccine Research and Discovery in the NHP model: a unified theory inacquisition prevention and control of SIV infection

Abstract

Purpose of the review

Here we highlight the latest advances in HIV vaccine concepts that will expand our knowledge on how to elicit effective acquisition-prevention and/or control of SIV replication in the NHP model.

Recent findings

In the context of the promising analyses from the RV144 Thai Trial and the effective control of SIV replication exerted by rhCMV-(SIV) elicited EM CD8 T cells, the HIV field has recently shifted toward vaccine concepts that combine protection from acquisition with effective control of SIV replication. Current studies in the NHP model have demonstrated the efficacy of HIV-neutralizing antibodies via passive transfer, the potential importance of the CD4 Tfh subset, the ability to effectively model the RV144 vaccine trial and the capacity of an Ad26 prime and MVA boost to elicit Env-specific antibody and cellular responses that both limit acquisition and control heterologous SIVmac251 challenge.

Summary

The latest work in the NHP model suggests that the next generation HIV-1 vaccines should aim to provoke a comprehensive adaptive immune response for both prevention of SIV acquisition as well as control of replication in break through infection.

Introduction

Over the last 30 years the quest for an HIV vaccine has taken us on a journey from classical empirical vaccine strategies to approaches employing sophisticated structural biology and virally vectored vaccines. Since the first clinical HIV vaccine candidate was evaluated in 1987 there have been nearly 100 clinical vaccine trials and thousands of willing volunteers. Many, if not all, of these vaccine approaches were initially tested in non-human primates (NHPs) prior to advancement into humans(1). In general, SIV and chimeric HIV/SIV (or SHIV) challenge models are used to test immunogenicity and efficacy of vaccine strategies designed to principally elicit T cell or antibody responses, respectively. The NHP model has also been employed extensively to investigate SIV pathogenesis and questions of immunological importance, which have directly advanced the design of prototype HIV vaccines. A key point remains that the selection of challenge viruses and the challenge methodologies are pivotal in the interpretation of NHP challenge studies. The field has evolved from challenging NHP with large doses of cloned viruses, to lower dose swarm based mucosal challenges with SIVmac251 or SIVsmE660 to more accurately recapitulate human HIV acquisition and infection. Furthermore, this strategy permits simultaneous analysis of SIV specific T cell and antibody responses after vaccination and challenge. For example, it has recently been demonstrated that the SIVsmE660 quasi-species is comprised of more neutralization sensitive viruses as compared to SIVmac251 or SIVmac239, emphasizing the significance of the challenge virus in NHP model analysis (2). As highlighted in Table 1, the evolution of the NHP challenge model has informed HIV vaccine development decisions, and continued development will lead to more effective products entering the clinical pipeline. However, until an absolute correlate of immunity is defined or a vaccine regimen proves effective in clinical trials and iterative studies are conducted in a relevant NHP model, (like those NHP studies highlighted below) there will be continued debate on the interpretation of data, SIV challenge virus and routes of infection in relation to human clinical vaccine trials.

Table 1.

Phase	Trial ID	Status	Strategy Prime/Boost	NHP challenge virus
I	Ad26.ENVA.01	Ongoing	Viral Vector – Adeno	SIVmac251 [3■■]
I	Ad5HVR48.ENVA.01	Ongoing	Viral Vector – Adeno	SHIV-SF162P3 [4]
I	HVTN 087	Ongoing	DNA/Viral Vector – VSV	SIVsmE660 [5]
I	HVTN 090	Ongoing	Viral Vector – VSV	SIVsmE660 [5]
I	HVTN 094	Ongoing	DNA/Viral Vector – Pox	SIVsmE660 [6]
I	IAVI B003	Ongoing	Viral Vector – Adeno/Viral Vector – Adeno	SIVmac251 [3■■]
I	IAVI B004	Ongoing	DNA/Viral Vector – Adeno	SIV/DeltaB670 [6]
I	IAVI S001	Ongoing	Viral Vector – Replicating/Viral Vector – Adeno	SIVmac239 [7]
I	RV 306	Scheduled	Viral Vector – Pox/Protein	SIVmac251 [8■]; SHIV KU2 [9]
I	RV 328	Scheduled	Protein	SIVmac251 [8■]; SHIV KU2 [9]
I	RV262	Ongoing	DNA/Viral Vector – Pox	SHIV-E [10]
Ib	HVTN 078	Ongoing	Viral Vector – Pox/Viral Vector – Adeno	SIVmac251 [11]
II	RV 305	Ongoing	Viral Vector – Pox/Protein	SIVmac251 [8■]; SHIV KU2 [9]
IIb	HVTN 505	Ongoing	DNA/Viral Vector – Adeno	SIVmac251 [12,13]; SIVsmE660 [12]
II	AVEG 202/HIVNET 014	Completed	Viral Vector – Pox/Protein	SIVmac251 [11]
II	HIVNET 026	Completed	Viral Vector – Pox/Protein	SIVmac251 [11]
II	HVTN 203	Completed	Virol Vector – Pox/Protein	HIV SF2 [1]; SIVmac251 [11]
II	V520-023	Terminated	Viral Vector – Adeno/Viral Vector – Adeno	SHIV89.6P [14]; SIVsmE660 [15]; S1Vmac239 [16]; SIVmac251 [17]
III	RV 144	Completed	Viral Vector – Pox/Protein	SIVmac251 [8■]; SHIV KU2 [9]
III	VAX 003	Completed	Protein	SIVmac251 [8■]; SHIV KU2 [9]
III	VAX 004	Completed	Protein	HIV SF2[1]

NHP, nonhumon primates; VSV, vesicular stomatitis Indiana virus.

Nevertheless, two significant vaccine studies have established that a HIV vaccine is possible, one performed in the NHP model and the other a large-scale clinical trial. These studies have strongly influenced the field in demonstrating efficacy either by eliciting HIV Env-specific antibodies or high frequency effector memory (EM) CD8 T cells(18, 19).

Antibodies are a crucial component for protection from HIV/SIV infection

The much deliberated RV144 Thai HIV vaccine trial, involved the establishment of a functional HIV-specific antibody response that blocked the ability of the virus to infect target cells at the site of exposure, thereby resulting in ‘sterilizing protection’. Results from this trial demonstrated modest protection from acquisition (31.2%, p=0.04) in humans, and this effect strongly correlated with antibody responses directed towards the Env-V1/V2 loop of HIV(20, 21). These intriguing data challenged the common conception that antibody-mediated neutralization was required to confer protection (21, 22), an interesting conclusion for the vaccine field considering the challenges ahead for inducing broadly neutralizing antibodies (bNab). Exactly how non-neutralizing antibodies mediated effects on acquisition in the RV144 vaccine trial is the subject of discussion and reviewed in detail (21, 23-25), but the NHP model provides an opportunity to address hypotheses regarding the mechanisms behind protection. However, modeling the modest protective vaccine effect of RV144 in NHP is problematic, requiring a large number of NHPs coupled with a suitable virus challenge regimen designed to limit the number of transmitted virus variants and an optimized repeated intra-rectal challenge (26). Recently, Pegu and colleagues performed a pilot study that mimicked the RV144 regimen in a small number of NHP with a SIVmac251 optimized repeated challenge regimen. Three of eleven NHP immunized with ALVAC-SIV/gp120 remained uninfected after 5 SIVmac251 challenge doses(8). Consistent with RV144 recipients, NHPs in this pilot study had no appreciable SIV specific CD4 or CD8 T cell activity yet uninfected NHP demonstrated high avidity binding for the Env-V1/V2 loop with no associated antibody dependent cellular cytotoxicity (ADCC) activity. Consequently, an adequately powered NHP study is underway to confirm this interesting result and address the role of non-neutralizing antibody in protection from acquisition.

Nevertheless, empirically developed vaccines for common viral diseases, such as influenza, afford protection by eliciting neutralizing antibody(27). In the past few years, highly potent and broadly neutralizing antibodies (bNabs) capable of neutralizing the majority of circulating strains of HIV-1 have been identified from HIV infected individuals. The bNabs target four major components of the viral spike (CD4bs, V1/V2, glycan-V3 and MPER)(28). Design of immunogens capable of eliciting bNabs is complex and challenging, especially in light of the extraordinarily high levels of somatic hypermutation observed in these antibodies(29). It remains unclear if HIV Env-antibodies elicited during vaccination can reproduce the same degree of affinity maturation, especially in comparison with a currently licensed influenza vaccination regimen (30). Therefore, alternative strategies for delivery of bNabs are now being considered and evaluated in NHP. Direct vectored-mediated gene transfer of bNabs into muscle tissue allows the continued secretion of the desired antibody into the bloodstream(31, 32). Whether this strategy will deliver a long lasting optimal concentration of bNabs remains to be determined. In the interim, passive immunization of bNabs by intravenous transfer into NHP followed by subsequent mucosal SHIV challenge evaluates the serum concentration and potential efficacy of bNabs. Broadly neutralizing antibodies such as b12, 2G12, 4E10 and 2F5 can all confer protection against mucosal SHIV challenge in NHP models(33-36). A recent study by Moldt and colleagues showed that passive immunization with a highly potent bNab-PGT121 (glycan-V3 specific) completely protected all 5 NHP against SHIV_SF162P3 challenge atan average serum concentration of 15μg/ml, significantly lower than previously reported for other bNabs (37, 38) (Fig. 1). If potent bNabs can indeed be efficacious at low concentrations, they could be delivered singly or as combination therapy(28, 39), perhaps limiting the pathways of viral escape. Moreover, it has been shown that passive immunization with low levels of 2G12 reduces viral load in newborn macaques(40), thereby highlighting the prospects for an additional, cost effective therapeutic strategy to combat mother to child transmission.

Fig 1. Broadly neutralizing antibodies used in passive immunization studies.

A glycosylated gp120 model of the HIV viral spike derived from (28)highlights five antibodies that have been used in passive transfer studies in NHP. The epitopes of these antibodies are indicated by color: CD4bs in red, V3 in blue and membrane-proximal region (MPER) in cyan. Average serum concentration of each mAb in protected animals before is indicated as well as the dose-model and challenge strain.

For an effective HIV-Env immunogen to mediate protection by the induction of neutralizing antibodies, it will most likely require the generation of Env-specific antibodies with exceptional levels of somatic mutation and long complimentary determining regions in the heavy chain (CDRH3s), which are hallmarks of currently identified HIV-bNabs(29). Although there may be no way to induce long CDR3H3 loops solely by vaccination since they are formed by recombination events (41), the problem of extensive affinity maturation is another hurdle unique in current vaccine designs. For example, multiple studies have demonstrated that Influenza vaccination elicits vaccine-induced antibodies with mutation rates in the V-gene that are about the same, or slightly higher as non-specific memory B cells (5%) (30, 42). However, CD4 binding site-specific bNab, VRC01 demonstrates a mutation rate around 30%(43).

How are these affinity matured antibodies generated? Specialized CD4 T follicular helper cells (T_FH) are intimately associated with B cell differentiation and survival signals in the germinal centers(44, 45). Also, CD4 T_FH cells promote somatic mutation (46), isotype switching and generation of high affinity memory B and plasma cells (47, 48). The possibility of eliciting antigen-specific T_FH in the context of vaccination has yet to be fully explored or realized. Interestingly, most current vaccine designs tested in NHP and in clinical trials are down-selected for the ability to elicit IFNγ and/or IL-2 responses, which potently inhibit CD4 T_FH cells(45, 49). So what is currently known about the role of CD4 T_FH in NHP and clinical models? During chronic SIV and HIV infection, T_FH cells accumulate in lymphoid tissues (50, 51). This accumulation may promote the generation of high affinity and broadly reactive antibodies, and studies have observed a correlation between the frequency of T_FH cells and the production of IgG antibodies during chronic infection (50, 51). It is, however, very important to understand the potential risk of increasing the frequency of T_FH cells. This study by Petrovas and colleagues observed that T_FH cells were the major targets of SIV infection during the acute phase of infection, although there was not a demonstrable relationship between SIV infection and the frequency of T_FH cells in the chronic phase of infection. Although the importance of T_FH cells for B cell help is clear, it is still unclear whether the accumulation of T_FH cells during the chronic phase of HIV/ SIV infection drives B cell dysfunction, or if accumulation of T_FH cells, especially HIV-specific T_FH cells, contribute to antibody responses. Further studies in NHP are ongoing to determine the contribution of antigen-specific CD4 T_FH cells toward eliciting bNabs. For future vaccine studies, it will be important to investigate whether it will be scientifically pragmatic to track T_FH cells in the peripheral blood of humans and NHPs. There is currently much discussion regarding which cell markers define T_FH or T_FH-like cells in peripheral blood and whether those T_FH-like cells are representative of germinal center TFH cell responses; however, studies have indicated that antigen-specific T_FH-like cells can be detected in human blood(52, 53). Quite recently Salah-Eddine Bentebibel et al. showed that antigen-specific T_FH-like detected in the periphery associated with protective antibody responses after seasonal flu vaccination, suggesting this population may be useful for the evaluation of vaccine efficacy although further research is still required in this area(54).

CD8 T cells are a crucial component for control of HIV/SIV infection

Analyses of immune correlates in the RV144 vaccine trial and the failure of the Merck STEP trial have shifted the field of HIV vaccine design from CD8 T cell based vaccines to vaccines that induce Env-specific antibodies and prevent HIV acquisition (21, 43, 55). However, epitope specific CD8 T cells remain essential in the initial reduction of virus load and subsequent control following HIV and SIV infection, especially in HIV and SIV long-term controllers (56). Few NHP studies have demonstrated significant control of virus load following SIV challenge (3, 57). In general, older studies were not powered to observe differences in acquisition, despite having SIV-Env included in the regimen. But an evaluation of DNA prime followed by replication incompetent recombinant Ad5 (rAd5) boost vaccine by Letvin and colleagues was powered to address both acquisition-prevention and SIV viral control(12). These animals were challenged with either SIVsmE660 or SIVmac251, and a reduction in acquisition of SIVsmE660 infection correlated with robust Env-specific CD4 T cells responses, neutralizing antibody activity and innate genetic factors. Among SIVsmE660-infected animals, peak viral loads were reduced by approximately 1 log and associated with CD8 T cell responses. SIV-specific effector memory (EM) CD8 T cells, which display strong virus inhibitory activity (VIA), were associated with this SIVsmE660 control in vitro (58). In-vivo, EM CD8 T cells reside in peripheral lymphoid and mucosal tissue with the capacity to rapidly proliferate, and have been strongly associated with early virus control in NHP receiving rhCMV vector (19, 59). In this rhCMV vector study, EM CD8 T cells also exerted persistent control of viral recrudescence in the absence of SIV Env-specific antibody. Extrapolation of the specific CD8 T cell epitopes involved in control of SIVmac239 replication, for other viral vector insert design, is difficult because rhCMV generates persistent high-frequency SIV-specific EM CD8 T cells that target different epitopes as compared to natural SIV infection or vaccination with rAd5 encoding the same genes (60).

While Env-specific CD8 T cell responses have found to be associated with higher HIV viremia in people (61), HIV Gag-specific CD8 T cell responses are consistently associated with effective control of viremia, especially when restricted by protective MHC class I molecules (61, 62). In a recent study, it has been shown that CD8 T cell cytotoxic capacity is strongly associated with disease outcome in SIV-infected NHPs and potentially in vaccines(58, 63). Although, further defining the characteristics of effective T cell epitopes restricted by class I molecules associated with slow or non-progression of HIV or SIV could possibly enlighten vaccine mediated CD8 control for future vaccine design(64).

Unified adaptive immune response approach to HIV Vaccine design

Recently, Barouch and colleagues extensively analyzed both humoral and cellular responses in the NHP immunized with a rAd26/MVA prime boost vector-based vaccine expressing SIVsme543 Gag, Pol and Env antigens (3). They observed an 80% or greater reduction in the per-exposure probability of low dose intra-rectal heterologous SIVmac251 challenge that correlated with Env-specific antibody. Interestingly, the central memory (CM) or EM CD8 T cell responses did not correlate with this protection from acquisition. Yet, in the same study, NHP with break-through SIV replication after rAd26/MVA vaccination developed robust SIV specific CD4 and CD8 responses and reduced virus replication by >2 logs. Based on these studies, arAd26 vector expressing HIV-Env has advanced to clinical evaluation eliciting promising immunogenicity results (65, 66). Similar findings were recently generated by Patel et al using a different vaccine regimen(67). In this study rhesus macaques received DNA vectors expressing SIVmac239 followed by inactivated SIVmac239 viral particles (AT-2), effectively targeting both arms of the immune system. After challenge with neutralization-sensitive SIVsmE660, 2 of the 8 co-immunized monkeys were protected from acquisition that correlated with mucosal Env-specific antibody. Furthermore, 73% of total vaccinated NHP that became infected controlled virus replication to low levels, and this control was associated with SIV-specific CD8 T cell responses.

Given this finding and other data now available, vector based HIV vaccine designs will include an HIV-Env component to reduce acquisition (8, 68) although additional components that potentially mediate vaccine control of HIV replication remain to be defined. As the current pipeline of vaccine candidates near the clinic, there will be a requirement for induction of Env-specific antibodies to reduce acquisition, and the activation of CD8 T cells to reduce virus replication, in a similar manner to rAd26/MVA approach in NHP challenge studies, and these compound effects will be an essential requirement for new vaccines.

Conclusion

HIV vaccines that are currently undergoing large-scale clinical evaluation were originally designed to contain viral components best suited to elicit T cell responses, although the majority included HIV-Env antigens. The unexpected outcome and correlates analyses of the RV144 Thai vaccine trial has prompted intense HIV-acquisition correlate evaluations and scrutiny of responses elicited by the HIV-Env components of the vaccines. It is clear that prototype HIV vaccines under evaluation in the NHP will ostensibly utilize viral vector delivery platforms delivered as homologous prime and boost or in heterologous combinations with DNA or protein constructs. Current road blocks in vaccine development are the delivery-method or elicitation of bNabs that demonstrate high rates of somatic hypermutation, as well as the safety considerations of novel viral vectors such as human CMV. Ideally a successful HIV vaccine will develop robust antigen specific CD4 T_FH T cells that will help induce and maintain a functional antibody response while also activating only the most effective antigen-specific CD8 response. Most importantly, these new vaccines will need to be tested in a well-designed and sufficiently powered NHP studies in order to best test candidates for large-scale clinical trials in humans.

Key points.

• Broadly neutralizing antibodies have yet to be induced by a vaccine, so current studies have focused on passive immunization as well as vectored-delivery methods.

• One group has successfully modeled the RV144 vaccine trial for the use in future vaccine studies.

• The use of a rAd26/MVA prime-boost has shown promise in eliciting Env-specific antibody and cellular responses that limit acquisition and control heterologous SIVmac251 challenge.