Optimization and Pre-Clinical Evaluation of New Strategies for HIV-1 Prevention and Treatment

National Institute Of Allergy And Infectious Diseases

Investigators

Abstract

We have previously developed and tested in preclinical models an HIV-1 vaccine platform that induces the production of VLPs in vivo. VLP formation was achieved by co-expression of HIV-1 envelope (Env) and SIV Gag. Endogenous expression by host cells guarantees the production of membrane-anchored Env proteins and their decoration by native glycosylation, which is critical to ensure a native antigenic conformation. In rhesus macaques, priming with a transmitted-founder clade-B Env lacking the critical N276 glycan, followed by multiple booster immunizations with glycan-repaired autologous and subsequently bivalent heterologous Envs (clades A and C), was highly immunogenic and eventually led to the induction of heterologous tier-2 neutralization accompanied by robust anti-Env CD4+ T-cell responses. Most notably, immunized animals were protected from repeated low-dose mucosal challenges with a difficult-to-neutralize heterologous tier-2 SHIV (SHIV AD8). These results indicate that immunization with membrane-anchored Env that does not expose off-target immunodominant epitopes such as the trimer base, the endogenous production of Env with its native glycosylation pattern, the formation of VLPs that are more effective stimulators of germinal center reactions, and finally an intensive multi-clade heterologous boosting schedule, which focuses antibody responses on shared conserved epitopes, may all have contributed to the success of this innovative vaccine platform. Our platform was further improved by the addition of a retroviral protease to produce mature VLPs, which were shown to be better immunogens than immature VLPs and by utilizing as a priming immunogen the clade-C Env 426c, which is a potent inducer of the VRC01 germline lineage. These two innovations, combined, resulted in significantly higher levels of protection from heterologous tier-2 SHIV infection in a second pre-clinical study in macaques, where protection was correlated with the titers of heterologous tier-2 neutralizing antibodies, as well as with the mucosal levels of Env-specific IgA and IgG. Despite these promising results, however, our vaccine platform can be further optimized with the objective of: i) enhancing the breadth and titer of neutralizing antibodies; ii) prolonging the durability of protective immune responses; iii) augmenting the levels of specific antiviral immunity at the mucosal level; iv) augmenting the effectiveness of cellular immune responses, in particular CD8+ T cells. Thus, the major goals of the project include to further optimize both the priming phase (via ad hoc-designed mRNA dose-escalation schemas) and the “first-boost” phase of immunization (via selection of most effective Envs that may selectively stimulate B-cell clones with the potential of developing into bNAbs), as well as by developing and testing viral vectors (such as MVA) co-expressing the VLP-forming combination of Env and Gag antigens, which may be utilized for mucosal immunization. Besides antibody responses, cellular immune responses also play a key role in protection from HIV-1 infection. Our laboratories have recently adapted the techniques for measuring CD8+ T-cell avidity, a critical functional marker associated with protection conferred by certain vaccines, to the Rhesus macaque system. While serial immunizations with various vaccine platforms increase CD8 avidity, CD8+ T cells do not reach levels that have previously been associated with effective elimination of HIV-infected cells or immunologic control in chronically infected elite controllers. Thus, a major goal of the project is to further enhance the avidity of CD8+ T cells using heterologous vectors in order to achieve avidity levels analogous to those detected in LTNP/EC macaques or those associated with HIV-1 control in humans. During this year, we have obtained the following results: 1. Selection of most effective HIV-1 Envs for the critical “first-boost” immunization phase. To identify the most effective “first boost” Env immunogens, we have started to screen a wide panel of polyclonal sera and monoclonal antibodies obtained from the early post-priming phase of our previous mouse and macaque studies to test their ability to recognize “intermediate” Env forms that elevate the bar for neutralization from the original germline-engaging priming Env immunogen, but still feature facilitated access to specific neutralization target epitopes. These include Envs derived from the original priming immunogen in which we re-introduced critical glycans (e.g., N276) that represent the major hurdles to overcome along the bNAb maturation path. In addition, we analyzed heterologous tier-2 Envs from which we removed critical glycans such as N276 in order to identify B-cell clones with the potential to develop neutralization breadth. This extensive screening lead to the identification of a restricted number of Envs that are recognized by the largest number of B-cell clones from the largest number of animals early after priming. These Envs will be used, either alone or in combination, as optimized “first-boost” Env immunogens in macaques study. 2. Evaluation of dose-escalation/long-interval regimens to enhance GC activation and prolong vaccine durability. There is currently great interest in the field in designing more efficient priming protocols that use a dose-escalation schema followed by a long pause prior to the first boosting, as this protocol seems to potentiate specific responses to the HIV-1 envelope in the GCs and may prolong vaccine durability. The rationale behind dose-escalation protocols is the induction of low-affinity IgM by the low initial doses, which would efficiently trap the Env immunogen within GCs after administration of the increasing doses at the end of the dose escalation. In collaboration with Xuejun Chen and Hongying Holdsworth at the VRC, we have started to test a variety of dose-escalation/long-interval priming protocols in VRC01gl knock-in mouse models using a VLP-forming mRNA vaccine in order to identify optimized schemas. The initial resulta of these studies suggest that a 2-dose and a 4-dose escalating schemas over a period of 14 days are the most effective for the initial stimulation of specific B-cell precursors. 3. In vitro evaluation of MVA vectors expressing VLP-forming immunogens. Vaccines based on a single delivery system, e.g., mRNA, may not be sufficient to achieve all the desired effects for eliciting strong, effective and durable protective immunity. In contrast, there is emerging evidence that vaccine platforms based on live or attenuated viral vectors, such as MVA, may specifically boost certain components of protective immunity and could be effectively delivered at the mucosal level, where protective immune responses are most needed in the case of HIV-1. MVA is a powerful vaccine delivery platform that has been extensively tested in pre-clinical and clinical studies and that is known to induce potent CD8+ T-cell responses, which are particularly low after mRNA vaccination that preferentially induces CD4+ T-cell responses. Moreover, MVA can effectively be delivered by the mucosal route (e.g., intranasally) and thereby can preferentially induce mucosal immunity. Thus, in collaboration with the group of Mariano Esteban in Madrid, a leader in the development of MVA vectors (2), we have produced MVA vectors expressing VLP-forming HIV-1 components (Env+Gag+Pro; with Envs from strains BG505 and REJO). Preliminary in vitro testing documented the efficient expression of the immunogens as well as the production of extracellular VLPs. The effectiveness of boosting with MVA vectors will be tested in macaques by mucosal inoculations (intranasal) in combination with intradermal injections (which also effectively boosts mucosal immunity). 4. Evaluation of boosting with replication-competent VSV vectors expressing Gag and Env to increase specific CD8+ T-cell avidity. We recently described a mechanism by which prior HIV vaccines designed to stimulate CD8+ T cells likely failed. In three separate clinical trials, these vaccines induced CD8+ T cells with antigen receptors that were insufficiently sensitive to degranulate and kill in response to the low levels of antigen presented on an HIV-infected cell. A particularly intriguing finding was that 3 DNA immunizations followed by replication-competent VSV did drive CD8+ T cells that had avidity equivalent to elite controllers in about 1/3 participants. Our labs have recently adapted the techniques for measuring avidity to Rhesus macaques lacking MHC alleles associated with control, that are typically excluded from vaccine cohorts. We have observed progressive increases in T-cell avidity that did not improve after 4-6 doses of mRNA and were well below those of chronically infected macaques or of elite-controller humans. We propose to follow up on the above observations by boosting with replication-competent VSV-Gag to specifically augment CD8+ T-cell avidity. VSV recombinants have been propagated in Vero cell monolayers, clarified by filtration, concentrated and purified by tangential flow filtration, then further purified by sucrose overlays and ultracentrifugation. The virus has been quantified by plaque assays in Vero cells and Gag expression confirmed by flow cytometry. Env on VSV particles have been similarly confirmed for expression and conformation by flow virometry. The rVSV vectors will be given to macaques as 10e8 PFU intrarectally and 10e8 intravenously.