Structure-Function Studies of the HIV-1 Envelope Glycoproteins and Development of a VLP-forming mRNA vaccine for HIV-1

点击翻译按钮获取中文摘要

The development of an HIV vaccine has been hampered by unprecedented challenges, primarily due to the unique properties of the HIV-1 envelope (Env) trimer, a cleverly engineered entry machinery that features an extraordinary assortment of immune-evasion tactics, including antigenic variation, heavy glycosylation of exposed surfaces and conformational masking of key neutralization sites. Further insights into the complex structure-function relationships in the HIV-1 Env trimer and its protective shield may be critical to guide the rational design of a protective vaccine and other immunoprophylaxis measures. 1) Enhancement of broadly HIV-neutralizing antibodies. Passive transfer of bNAbs is a promising alternative to ART for HIV-1 prevention and treatment. We previously reported the discovery of a second CD4-binding site in the HIV-1 Env trimer, that we defined as CD4-binding site 2 (CD4-BS2), which opened new perspectives for HIV vaccine and therapy (Liu et al., NSMB 2018). We also identified selected anti-CD4-BS antibodies, such as VRC03 and VRC06, which mimic the quaternary-binding mode of CD4 and establish contact with two adjacent gp120 protomers via an extended loop in their heavy chain framework region 3 (FR3). Thus, we rationally engrafted the extended FR3 loop of VRC03 onto different CD4-supersite bNAbs and developed chimeric antibodies with enhance activity against a wide panel of global HIV-1 strains (Liu et al. 2019). Previously, we devised a structure-guided strategy for improving bNAbs against the CD4-binding site (CD4bs) by engraftment of the elongated framework region 3 (FR3) loop from antibody VRC03, which confers the ability to establish quaternary interactions with a neighboring gp120 protomer (Liu et al., Nat. Comm. 2019). We applied this strategy to a new series of anti-CD4bs bNAbs (i.e., the N49 lineage) that already possess high neutralization potency and breadth. The resultant chimeric antibodies bound the Env trimer with a higher affinity than their parental forms and displayed increased neutralizing activity against a global panel of HIV-1 Envs, with a remarkable mean half-maximal inhibitory concentration below 0.01 g/mL against a wide panel of 86 global HIV-1 strains. In contrast, engraftment of the elongated CDR1 loop of bNAb 1-18, another potent quaternary-binding antibody, failed to improve neutralization, illustrating the selective requirements for the establishment of quaternary contact with the HIV-1 Env trimer. Due to their increased potency associated with reduced autoreactivity and prolonged in vivo half-life, FR3-loop-chimeric bNAbs are under consideration for clinical use in HIV-1 prevention and treatment. 2) Development and pre-clinical evaluation of a VLP-forming mRNA vaccine against HIV-1. a. Co-expression of HIV-1 env with gag mRNA leads to the efficient production of VLPs VLP-forming vaccines may offer advantages in terms of both antigen structure and size. We evaluated the ability of different HIV-1 Env immunogens to generate extracellular VLPs upon co-expression with SIV Gag using mRNA. The assembly and extracellular release of Env-coated VLPs were readily documented with all the Envs tested, with an inferred mean number of spikes of 32 per particle, which is more than twice the number of spikes present on wild-type HIV-1 virions. b. Design and pre-clinical evaluation of a VLP-forming mRNA vaccine against HIV-1 in macaques We designed an mRNA vaccine with the following features: i) use of mRNA as a vehicle to instruct host cells to endogenously express membrane-bound Env glycoproteins decorated with native N-linked glycosylation; ii) use of full-length or minimally truncated Envs that do not expose distractive immunodominant epitopes; iii) co-expression of Env with Gag to induce the in vivo production of virus-like particles (VLP), which closely mimic native viral particles produced by HIV-1 infection; iv) initial priming with an Env capable of engaging germline bNAb precursors; v) intensive heterologous boosting with tier-2 Envs from different clades (A,B,C) to selectively expand antibody responses against shared bNAb epitopes. After a preliminary study in mice, which validated the superior efficacy of our VLP-forming vaccine platform, we performed a complex pre-clinical immunization-challenge trial in Rhesus macaques. The animals were sequentially immunized with co-formulated HIV-1 env (from clades A, B and C) and SIV gag mRNA for a total of 10 immunizations over a 52 week period. After 3 autologous immunizations, the animals received multiple boosters with mixed heterologous Envs. A subgroup of animals received final protein boosters (instead of mRNA) with homologous Envs in a soluble trimer (SOSIP) form. Despite the high dose of mRNA the vaccine was well tolerated. The vaccine was highly immunogenic as documented by the rapid induction of trimer-binding antibodies and robust polyfunctional Env-specific CD4+ T-cell responses, including T-follicular-helper (TFH) cells. Strikingly, after the third heterologous boost, we started to detect the appearance of bNAbs, albeit at relatively low titers. Most important, vaccinated animals were protected from repeated low-dose mucosal challenges with a difficult-to-neutralize heterologous tier-2/3 SHIV (SHIV-AD8) with a calculated per-exposure risk reduction of 79%. Protection was correlated with the induction of antibodies against the CD4-binding site. These results provided evidence that a multiclade VLP-forming mRNA vaccine can be effective in protecting against heterologous tier-2 infection. c. Optimization of the platform to produce mature VLPs by inclusion of the viral protease Although co-expression of Env with Gag leads to the assembly and release of VLPs, in the absence of the viral protease Gag remains uncleaved and only immature VLPs are formed. Since mature VLPs are believed to be better immunogens than immature VLPs, we aimed at introducing the viral protease (pro) into our env-gag mRNA vaccine platform. Initially, we attempted to add pro mRNA to the formulation, but the results were unsatisfactory due to toxicity associated with excessive enzyme activity and scarce VLP generation due to untimely enzyme activation. Thus, we decided to employ the physiological mechanism utilized by HIV-1, which transcribes the pro gene as part of the long gag-pol mRNA transcript which undergoes a -1 ribosomal frameshift to translate Pro. After optimization of the system, we were able to produce high titers of well-shaped mature VLPs using env:gag:gag-pol at a ratio of 1.5:1:0.1. In vitro, mature VLPs induced a more efficient activation of B cells expressing anti-Env antibodies on their surface compared to immature VLPs, while no difference was seen in their relative ability to activate specific T cells. d. In vivo validation of the env-gag-pro mRNA vaccine platform in mice and macaques To validate the beneficial effects of the protease inclusion, we tested the triple mRNA formulation in mice. After the second immunization, mice immunized with the gag-pol-supplemented vaccine developed higher titers of neutralizing antibodies than those immunized with env alone or env+gag mRNA. Moreover, they were able to neutralize a more protected form of HIV-1. Based on these encouraging results, we performed a second macaque study which compared env-gag with env-gag-gag-pol mRNA. The results of this study confirmed both the induction of bNAbs and the protective effect of the vaccine from heterologous SHIV-AD8 challenge, with an increased protection conferred by the addition of gag-pol mRNA. In addition, we found a significant correlation between protection and titers of bNAbs. Together, the results of these studies validated our VLP-forming HIV-1 env-gag-gag-pol mRNA vaccine, paving the way for the clinical evaluation of our vaccine.