| Popis: |
Despite huge strides being made towards decreasing the number of individuals getting newly infected with HIV-1, and in reducing AIDS-related deaths, unfortunately current predictions are that the 2020 UNAIDS goals (90-90-90 targets, where 90% of people living which HIV-1 are diagnosed as such, from which 90% will will receive sustained antiretroviral therapy, resulting in viral suppression in 90% of these individuals by 2020) are out of reach. This of course means that the numbers of newly infected indivuals and AIDS-related deaths will be above the target derived from the 2020 UNAIDS goals. The development of an effective HIV vaccine could therefore be an important step towards realising these objectives. In work done for this thesis, a heterologous HIV-1 vaccine platform regimen was developed using antigen sequences from the predominant circulating HIV-1 subtype (subtype C) in South Africa. Specifically, this involved use of the envelope glycoprotein sequence of the CAP256 superinfecting virus (CAP256_SU) from the CAPRISA 002 cohort, and a mosaic Gag sequence which resulted in robust autologous Tier 2 neutralisation of CAP256_SU pseudovirions. The envelope glycoprotein sequence was modified so as to replace the native leader sequence with the tissue plasminogen activator leader, the furin cleavage site with a glycine rich flexible linker, and to introduce an I559P mutation. DNA and modified vaccinia virus Ankara (MVA) vaccines were generated where Env was truncated to gp150, thereby retaining the transmembrane domain and a partial cytoplasmic tail (Env). The Env sequence for the protein vaccine was further trimmed by removal of the transmembrane domain to give gp140, leading to a soluble, secreted protein (soluble Env). This allowed for the latter vaccine to be affinity purified using lectin (soluble Env (GNL)), and after generating stable cell lines, soluble Env yields were high enough to enable size exclusion chromatography which allowed isolation of the trimeric fraction of Env as determined by molecular weight (soluble trimeric Env). A Cterminal His-tagged version of soluble Env was generated as well. Surprisingly, the folding of Env-His was inferior to soluble Env, with a switch in profile from mainly trimeric Env to mainly monomeric Env. Nevertheless, soluble Env-His (GNL) and soluble trimeric Env-His were assessed for the presence of Env broadly neutralising antibody (bnAb) epitopes in an ELISA assay. The V3-glycan supersite (binding of bnAbs PGT128 and PGT135), the CD4-binding site (VRC01) and the V2-glycan site (PG9) were detected for both Env-His (GNL) and soluble trimeric proteins, whereas low signals for PG16, PGT145 and CAP256-VRC26.08, bnAbs which specifically recognise Env trimers in a native-like conformation, were only detectable for soluble trimeric Env-His. Soluble Env (GNL) was subsequently used as a protein vaccine in rabbits to test the immunomodulatory effects of the two adjuvants AlhydroGel (similar to alum) or the MF59-like squalene-based oil-in-water nano-emulsion AddaVax. Soluble Env (GNL) adjuvanted in AlhydroGel resulted in improved immune response in rabbits, with significantly higher serum binding antibodies to soluble Env (GNL) and scaffolded CAP256 V1V2-loop in comparison to AddaVax and unadjuvanted protein. Furthermore, significantly higher neutralisation titres to Tier 1A subtype C virus (MW965.26), in combination with an improved breadth to subtype C Tier 1A and 1B viruses, were observed in the AlhydroGel group. However, no neutralisation of Tier 2 viruses was detected. Nonetheless, AlhydroGel was selected as the best protein adjuvant for all further rabbit immunogenicity studies. Furthermore, in all subsequent experiments, soluble trimeric Env was used as a protein vaccine. DNA and recombinant MVA vaccines were generated using a membrane anchored gp150 (Env) with the aim that co-expression with mosaic Gag (GagM) would lead to the incorporation of Env into Gag virus-like particles (VLPs). Electron microscopy of cells expressing Env+GagM from DNA and recombinant MVA vaccines verified VLP formation from these constructs, and the presence of Env was observed in VLPs purified using a two-step OptiPrep gradient centrifugation protocol. The presence of Env bnAb epitopes in cellular membrane-bound Env was verified by qualitative immunofluorescent microscopy of live-cell stainings and a quantitative FACS assay. The same bnAb epitopes as for the Env protein vaccine were detectable, including bnAbs recognising only native-like Env trimers (PG16, PGT145 and CAP256-VRC26_08). However, expression levels of native-like Env trimers were lower, at approximately 20% when normalised to VRC01. These HIV-1 DNA, rMVA and soluble trimeric Env protein vaccines were tested in different heterologous vaccine platform immunogenicity studies in rabbits. These consisted of either priming with two recombinant MVA vaccines and boosting with three protein vaccines (MMPPP), or priming with DNA vaccines followed by two MVA vaccines, followed by two protein vaccines (DDMMPP). Furthermore, the inclusion of GagM into the DNA and MVA vaccines was compared to use of Env alone. Both vaccine regimens resulted in binding antibodies to soluble trimeric Env and a scaffolded CAP256 V1V2-loop; however, these were induced by MVA and protein vaccines, but not by DNA vaccines. Despite the lack of Env binding antibodies after DNA vaccination, better neutralisation was observed for the DDMMPP regimen compared to MMPPP, resulting in higher sera neutralisation titres towards vaccinematched, autologous Tier 2 CAP256_SU virus. Most encouragingly, when compared to Env alone, the inclusion of Gag (Env+GagM) into DNA and MVA vaccines improved the immunogenicity of the DDMMPP regimen even further. For Env+GagM DDMMPP, more animals developed Tier 2 neutralising antibodies, and improved titres, whereas Tier 2 neutralisation in general started to develop after fewer vaccinations, as for most rabbits this was observed after the second MVA inoculation. In an attempt to improve the spike density of Env on VLPs and the plasma membrane, two Env chimaeras were made replacing parts of gp41 with the corresponding elements of influenza A H5 haemagglutinin (HA2) (Env:HA2 chimaeras). Increased Env spike density was observed in a previous study using this strategy for the gp41 transmembrane domain and cytoplasmic tail (gp140HA2tr). A similar construct was generated here for CAP256_SU and a second chimaera was included replacing the whole of gp41 with HA2 (gp120HA2). Surprisingly, in experiments where VLPs were purified from OptiPrep gradients or the whole-cell bnAb FACS assay conducted with these Env:HA2 chimaeras, there was no evidence of increased spike density on VLPs or the plasma membrane as compared to Env. Furthermore, the folding of Env was severely impacted, especially regarding gp120HA2 where no binding of PG16, PGT145 and CAP256 VRC26.08 - bnAbs recognising native-like Env trimers - was observed. Although results for gp140HA2tr was improved over gp120HA2, in general the data for gp150 (Env) was superior in both the bnAb live-cell staining and FACS assay. Consequently, when both Env:HA2 chimaeras in combination with GagM were tested in the DDMMPP regimen, no improvement was observed with regard to autologous Tier 2 neutralisation. For rabbits receiving gp120HA2, no animals developed Tier 2 nAbs, whereas for gp140HA2tr, Tier 2 neutralisation in general developed later and to lower titres compared to Env+GagM. In conclusion, different HIV-1 DNA, recombinant MVA and protein vaccines were generated and characterised both in vitro and in vivo, leading to a vaccination regimen that induced both high titre Env binding and vaccine-matched Tier 2 neutralising antibodies in rabbits. Furthermore, a new Env sequence, the first from the South African CAPRISA cohort, has been added to the small list of Env sequences that can induce Tier 2 neutralisation. |
