The human immunodeficiency virus (HIV)-specific cytotoxic T-lymphocyte (CTL) response is important in the control of HIV-1 infection. Due to the virus having a high rate of mutation, immune pressure can select for variants that are no longer recognised by CTLs to dominate the viral quasispecies. This is similar to how antiretroviral resistance emerges. HIV-1 is therefore adapting to both human leukocyte antigen (HLA)-restricted immune responses and antiretroviral therapy. This thesis initially focused on the natural CTL response to an HLA-B*51-restricted epitope in integrase. This HLA class I allele is associated with slow progression to AIDS; however, as no CTL-driven escape mutation has been fully defined within this integrase epitope, we cannot determine what contributes to the association of HLA-B*51 and natural control of infection. By longitudinally studying a cohort of early HIV-infected individuals, we observed the emergence of polymorphisms that abrogate a CTL response to this epitope. CTL escape may also prove to be the downfall of current immunotherapy strategies attempting to combat HIV infection. T cell receptors (TCRs) have been genetically modified to enhance their binding affinity to an HLA-A*02-peptide complex and transduced into CD8+ cells to create an HIV adoptive therapy. We demonstrate through in vitro selection pressure assays that escape from these cells may be a difficult task for the virus given that the TCR is able to recognise the majority of variants of this epitope. Antigen processing mutations may represent the only option for escape. How this may translate clinically will only be determined through in vivo studies, which must be meticulously monitored. Finally, when this high affinity TCR was fused to an anti-CD3 single chain variable fragment to create proteins capable of redirecting non-HIV-specific CTLs to HIV-infected cells, we found that the result was specific lysis. These proteins may supersede the use of TCR-transduced cells when used in synergy with antiretroviral therapy.

Consistent efficacy in human clinical trials has been achieved by two leading malaria vaccine candidates encoding either the P. falciparum circumsporozoite (CSP) or thrombospondin-related adhesion (TRAP) proteins. However, protection in humans relies on high antibody (Ab) titers or potent T cells, respectively, when these antigens are used individually. Therefore, a concurrent induction of both anti-CSP antibodies and TRAP-specific T cells may further improve the protective efficacy associated with these immunogens. This thesis investigates the protective potential associated with CSP and TRAP combination vaccines by employing a pre-clinical Plasmodium berghei (P. berghei) mouse malaria challenge model. Depletion of CD4⁺ and CD8⁺ T cells indicated that protection by the CSP antigen relied solely on antibodies while TRAP protected through CD8⁺ T cell responses. Moreover, the administration of relatively small amount of anti-CSP monoclonal antibodies (mAb) substantially enhanced the sub-optimal protection elicited by TRAP. A way to maximize both, anti-CSP antibodies by adenovirus/protein (Ad-P) prime-boost vaccinations and anti-TRAP T cell responses using adenovirus/modified vaccinia strain Ankara (Ad-M), was explored. Combination of the two approaches stimulated optimal humoral and cellular responses that afforded sterile protection in the C57BL/6 animal model. Additional analyses of the effector functions of Abs indicated that the efficacy of this vaccine relied on a two-stage attack against the parasite: anti-CS antibodies reduced the number of sporozoites reaching the liver, thus decreasing the number of infected hepatocytes required to be eliminated by the dual action of natural killer (NK) cells and CD8⁺ T cells. These proof-of-concept studies using a wild-type (wt) P. berghei challenge model allowed progression towards the development of human malaria Plasmodium vivax (P. vivax) vaccines using PvCSP and PvTRAP antigens. The immunogenicity of PvCSP and PvTRAP-based vaccine candidates was assessed in multiple mouse strains. Additionally, transgenic (tg) P. berghei parasites expressing P. vivax genes were generated and used to demonstrate the protective efficacy of Ad-P PvCSP and Ad-M PvTRAP vaccines.

Despite 30 years of research, a prophylactic vaccine against HIV-1 is still lacking and is urgently needed in order to control the global AIDS pandemic. The discovery of broadly neutralising antibodies (BNAbs) was an important step for HIV-1 research but no vaccine candidate tested so far has been able to reproduce responses containing such antibodies, and it remains unclear how this could be achieved via immunisation. In this thesis, I attempted to explore this gap of knowledge in two ways. First, certain features (‘signatures’) of the Env protein that were associated with a broadly neutralising response were identified through machine learning. Further characterisation of these signatures revealed several ways by which these naturally-occurring mutations might alter the immunogenicity of the Env protein that could result in the elicitation of a broadly neutralising response. The incorporation of such signatures in future vaccine design could be useful as the Env protein might adopt a conformation that encourages the elicitation of a broadly neutralising response. Second, 3 novel vaccination approaches were proposed aiming to induce a BNAb antibody response. The development of 2 approaches proved to be difficult and was not continued. For the third approach, non-neutralising immunogen-derived antibodies were used to mask immunodominant epitopes on the Env protein (i.e. ‘antibody-shielding’), thus allowing the antibody response to be focused to the highly conserved CD4 binding site (CD4bs). Subsequent immunisation of the antibody-shielded gp120 proteins in mice and rabbits demonstrated that antibody-shielding was able to significantly dampen the V3-specific antibody response while retaining the CD4bs-specific response. However, the antibody response to the V1/V2 loop was enhanced upon V3-dampening which indicates that further optimisation of the antibody-shield is needed in order to eliminate any antibody response towards the immunodominant regions. In conclusion, these results are the first description of a number of novel vaccination ideas and provide valuable insights into how these approaches could be optimised to become effective HIV-1 vaccines that can lead to the elicitation of a broadly neutralising antibody response.

The enormous mortality burden exerted by P. falciparum malaria has evolutionarily selected for red blood cell (RBC) polymorphisms which confer protection against the severe manifestations of this disease. Although the epidemiological protection by these polymorphisms has been well-established for the past half-century, the mechanisms underlying this protection are still being uncovered. Recent studies implicate impaired cytoadherence to microvascular endothelial cells (MVECs) due to reduced surface levels and altered display of Plasmodium falciparum erythrocyte membrane protein 1 (PfEMP1) as a mechanism of protection against severe malaria by sickle hemoglobin (Hb) S and HbC. Consequently, in this thesis, I have described three separate, but related investigations into whether hemoglobins S and C influence a parasite’s cytoadherence binding phenotype (Chapter 3), the PfEMP1 variants that parasites express in vivo (Chapter 4), and the IgG recognition of PfEMP1 domains in Malian children (Chapter 5). We found that parasites from HbAS children show statistically insignificant increased binding to MVECs and that parasites did not express a restricted subset of var genes in HbAS and HbAC children. Compared to HbAA and HbAC children, HbAS children demonstrated a slower rate of acquisition of IgG responses to a repertoire of PfEMP1 domains. These findings suggest that, although hemoglobin type influences the binding phenotype of P. falciparum isolates and the acquisition of PfEMP1-specific IgG responses, other factors more likely determine the expressed var gene repertoire within parasites than hemoglobin type.

In the search for a vaccine providing comprehensive protection against meningococcal disease, one vaccine currently under development contains the immunogenic proteins PorA and FetA in meningococcal outer membrane vesicles (OMVs). To achieve high levels of coverage against disease-causing isolates, the antigenic variability of these proteins could be overcome using knowledge of meningococcal epidemiology and population structure. In this study, the possible implications of variable expression levels of PorA and FetA on vaccine efficacy were investigated. Producing OMVs containing consistent amounts of FetA is difficult due to iron-repressed expression; therefore, meningococcal strains were constructed which constitutively expressed FetA at increased levels for OMV vaccine production and analysis. In mice, OMVs from modified strains induced antibodies against both PorA and FetA. These antibodies acted synergistically in a serum bactericidal assay; however, antibodies against FetA were weakly bactericidal alone. The potential to increase levels of PorA- and FetA-specific bactericidal antibodies with a prime-boost strategy, using OMV and protein inoculums, was also tested. While successful for a weakly-immunogenic PorA variant, a similar strategy did not increase bactericidal activity against FetA. Although antibodies against FetA can be induced following OMV immunisation, sufficient antigen expression in target bacteria is also required for bactericidal killing; therefore, the variability and regulation of porA and fetA transcription was investigated in a range of isolates. Despite differences in regulation among clonal complexes, variable expression is unlikely to be an issue for vaccine coverage. In particular, regulation of fetA by iron is reduced in many isolates due to a deletion in the sequence bound by the regulatory protein, Fur. Therefore, a vaccine targeting PorA and FetA may provide high levels of protection against meningococcal disease; however, an alternative formulation or immunisation strategy is required to improve coverage against FetA.

Sialic acid (SA) is a known major virulence factor of Haemophilus influenza (Hi). This study aims to analyse the functions of some candidate sialometabolism genes, and to further our current understanding on the Hi sialometabolism gene regulation. Two candidate sialometabolism genes (HI0227 and HI0148) and their adjacent ORFs (HI0228, HI0148.1 and HI0149) were studied. HI0148.1 and HI0149 are transcribed as a single gene in screened NTHi strains, and we refer to the combined ORF as NTHI0236 (the designation in strain 86-028NP). Across Hi strains screened, the sequences of HI0227, HI0148 and NTHI0236 are conserved. However, the sequence of HI0228 is heterogeneous. Mutants that lack the functions of HI0227, HI0228, HI0148 and NTHI0236 were compared to their respective wild type parent strains for ability to grow on SA (in aerobic and microaerophilic conditions), their ability to sialylate LPS and their ability to resist complement mediated killing. The mutants did not exhibit major differences in the tested aspects of sialometabolism compared to their respective wild type strains. Changes observed in some of the mutants in serum bactericidal assays and LPS profiles were due to the effect of phase variable genes. The sialometabolism functions of HI0227, HI0148, and NTHI0236 remain obscure, and we postulate that HI0228 is a pseudogene. Investigation of Hi sialometabolism gene regulation was conducted using mutants that lack different steps of the Neu5Ac catabolism pathway and the Neu5Ac activation pathway. The expression of nanE and siaP, respectively representing the Neu5Ac catabolism and transport operons, were assessed using RT-PCR and qPCR. We investigated a temporal/concentration effect of Neu5Ac on the expression of sialometabolism operons, which highlights the importance of studying the Hi sialometabolism gene regulation as a dynamic process. We further demonstrated that GlcN-6P, a Neu5Ac intermediate from the catabolism pathway, is likely the SIS sugar that interacts with SiaR, the repressor protein of the Hi sialometabolism operons. We postulate that upon binding of GlcN-6P to SiaR, the SiaR-mediated repression on the Hi sialometabolism operons is relieved, resulting in the induction of the expression of Neu5Ac catabolism and transport genes.

HIV-1 chronically infected individuals represent a large disease burden, making the development of a therapeutic vaccine for use in chronic infection a priority. However, therapeutic vaccination has not been successful to date. Most approaches have employed viral vectored vaccines encoding full length viral proteins, which aimed to boost pre-existing CD8+ T cell responses by mimicking natural HIV-1 infection. Simply boosting these pre-existing CD8+ T cell responses which have previously failed to control the virus may be insufficient. Although HIV-1 has a huge capacity to diversify, certain regions are less tolerant of mutations due to structural and functional constraints. We therefore hypothesised that it would be necessary to redirect the immune response to more vulnerable regions of the proteome, such as conserved regions. HIVconsv is a rationally designed conserved region vaccine. Vaccination of 19 chronically HIV-1 infected HAART treated patients with MVA.HIVconsv was safe and well tolerated. There was a significant increase in the magnitude of HIVconsv-specific T cell response following vaccination (p = 0.001), as measured by IFN-γ Elispot assay, but changes observed in vaccinees did not reach statistical significance when compared with placebo recipients (p = 0.48). The capacity of CD8+ T cells to inhibit HIV-1 replication in vitro is highly predictive of virus control in vivo and is thus a possible surrogate of vaccine efficacy. There was a trend towards increased CD8+ T cell mediated inhibition following vaccination with 2x10⁸pfu MVA.HIVconsv (17% inhibition pre- vs 54% inhibition post-vaccination, p = 0.06). However, measurement of the latent HIV-1 reservoir by quantification of total HIV-1 DNA in circulating CD4+ T cells by droplet digital PCR showed no reduction in size upon vaccination, indicating CD8+ T cells induced by vaccination with MVA.HIVconsv were not of sufficient potency to impact the reservoir. In a second cohort of HIV-1 infected individuals, antiviral inhibitory activity was measured in 36 HIV-1-infected STEP and Phambili trial participants. Sustained potent CD8+ T cell antiviral inhibitory responses were rare but were strongly correlated with IFN-γ responses to so-called ‘beneficial’ low entropy regions in HIV-1 Gag and Pol, that had been reported previously to be associated with HIV-1 control, (r = 0.69, p = 0.0001). This correlation was still significant after controlling for protective HLA alleles, whereas responses to conserved elements were only weakly correlated with viral inhibition (r = 0.41, p = 0.04). These data indicate that immunogens that are based on the selection of regions within the viral proteome by conservation score alone may not induce the most effective HIV-1-specific T cell responses and they highlight the importance of systematically selecting specific regions associated with HIV-1 control, together with exclusion of immunodominant decoy epitopes.

Human immunodeficiency virus-1 (HIV-1) and tuberculosis (TB) are significant global health issues today and the immunological correlates of protection to both diseases remain poorly defined. HIV-1 is the greatest risk factor for the development of TB and HIV-associated TB contributes significantly to the global TB burden, particularly in sub-Saharan Africa. The host T cell response to Mycobacterium tuberculosis (Mtb) is critical for control and is weakened in HIV-1 disease. However, the extent and precise mechanisms remain incompletely elucidated. Antiretroviral therapy (ART) for HIV-1 reduces TB risk in treated individuals significantly, although not to levels seen in HIV-uninfected individuals. This thesis studies HIV- and Mtb-specific T cell responses in 2 cohorts of individuals from a community with high HIV-1 and TB prevalence in Bloemfontein, South Africa. An analysis of HIV-specific CD8 T cell responses in Chapter 4, confirms the superior role of HIV-1 Gag-specific CD8 responses in controlling HIV-1 viraemia and demonstrates that the loss of such responses contributes to HIV-1 disease progression and likely susceptibility to opportunistic infection. I investigated the impact of HIV-1 infection on the Mtb-specific T cell response through a cross-sectional comparison of T cell responses in HIV-infected and HIV-uninfected individuals (Chapters 5 and 6). HIV-infected individuals had a significant depletion of both Th1 and Th17 CD4 responses to Mtb-specific antigens as well inhibition of Mtb-specific CD8 T cell responses, in comparison to those uninfected with HIV-1. PPD- and Rv2031c-specific responses were particularly reduced in HIV-infected individuals. Over a 12 month period of therapy, ART partially restored the Mtb-specific CD4 T cell response (Chapters 5 and 7). This effect was greater for Th1 than Th17 responses and had no detectable effect on the Mtb-specific CD8 response. However, despite some evident restoration, there remains a significant quantitative deficit in individuals on ART that is likely to contribute to persistent elevated TB risk. Overall, these data contribute to a better understanding of the mechanisms of susceptibility to TB during HIV-1 disease and ART, as well as of the correlates of protective immune responses to both pathogens.

Next-generation sequencing technologies are revolutionising our ability to characterise and investigate infectious diseases. Utilising the power of high throughput sequencing, this study reports, the development of a sensitive, non-PCR based, unbiased amplification method, which allows the rapid and accurate sequencing of multiple microbial pathogens directly from clinical samples. The method employs Φ29 DNA polymerase, a highly efficient enzyme able to produce strand displacement during the polymerisation process with high fidelity. Problems with DNA secondary structure were overcome and the method optimised to produce sufficient DNA to sequence from a single bacterial cell in two hours. Evidence was also found that the enzyme requires at least six bases of single stranded DNA to initiate replication, and is not capable of amplification from nicks. Φ29 multiple displacement amplification was shown to be suitable for a range of GC contents and bacterial cell wall types as well as for viral pathogens. The method was shown to be able to provide relative quantification of mixed cells, and a method for quantification of viruses using a known standard was developed. To complement the novel molecular biology workflow, a data analysis pipeline was developed to allow pathogen identification and characterisation without prior knowledge of input. The use of de novo assemblies for annotation was shown to be equivalent to the use of polished reference genomes. Single cell Φ29 MDA samples had better assembly and annotation than non-amplification controls, a novel finding which, when combined with the very long DNA fragments produced, has interesting implications for a variety of analytical procedures. A sampling process was developed to allow isolation and amplification of pathogens directly from clinical samples, with good concordance shown between this method and traditional testing. The process was tested on a variety of modelled and real clinical samples showing good application to sterile site infections, particularly bacteraemia models. Within these samples multiple bacterial, viral and parasitic pathogens were identified, showing good application across multiple infection types. Emerging pathogens were identified including Onchocerca volvulus within a CSF sample, and Sneathia sanguinegens within an STI sample. Use of Φ29 MDA allows rapid and accurate amplification of whole pathogen genomes. When this is coupled with the sample processing developed here it is possible to detect the presence of pathogens in sterile sites with a sensitivity of a single genome copy.

Severe acute respiratory syndrome (SARS) was defined as an invasive respiratory disease in 2002, which originally came from China and rapidly spread all over the globe. Acute pneumonia and lower respiratory tract involvement most affected the middle aged individuals and elderly with a mortality rate of 11%. While SARS Corona virus (SARS-CoV) has maintained its potential capacity to reemerge, clinical study of the immune system of SARS patients, as well as controlled studies may lead to application of new treatment strategies in future. Throughout this work, I have focused on early immune responses to SARS-CoV in humans and in ferrets. CXCL0 has been associated with alterations in the clinical course of several infectious diseases, including SARS and influenza. Here I have cloned ferret CXCL10 gene and have expressed its recombinant protein. I demonstrate that the CXCL10 plasma level in SARS patients is associated with the severity of disease. I also show that endogenous ferret CXCL10 exhibits similar mRNA expression patterns in the lungs of deceased SARS patients and ferrets experimentally infected with SARS-CoV. Type I interferons (IFNs) are indispensable parts of the innate immunity during early stages of infection. A clear distinction between genes upregulated by direct virus-cell interactions and genes upregulated by secondary IFN production has not been made yet. Here, I have investigated differential gene regulation in ferrets upon subcutaneous administration of IFN-a2b and during SARS-CoV infection. In vivo experiments revealed that IFN-a2b causes upregulation of abundant IFN response genes (IRGs), chemokine receptors, and other genes that participate in phagocytosis and leukocyte migration. SARS-CoV infection of ferrets leads to upregulation of varieties of IRGs and a broad range of genes involved in cell migration and inflammation. This work allowed dissection of several molecular signatures present during SARS-CoV infection, which are part of a robust IFN antiviral response. Since localization of CD8+ Tcells may contribute to tissue injury, I have characterized ferret CD8 gene and have generated reagents that can be used in future studies with the aim of evaluating CD8+ T cells localization in the ferret lung during infection with SARS-CoV.