Insights from a different vantage point: new modelling of vaccination and epidemiology of Chlamydia infections

Chlamydia are intracellular pathogens that infect a broad range of host species including humans and koalas. Depending on the bacterial strain, chlamydial infections can lead to severe reproductive or ocular disease, potentially resulting in infertility or blindness. Many infections are subclinical and may persist, complicating control strategies for this pathogen. Vaccination is a potential control strategy that could result in protecting vaccinated individuals against Chlamydia-related disease. The purpose of this thesis was to evaluate efficacy of candidate Chlamydia vaccines using novel biostatistical approaches to determine the effects of vaccination on measures of success, including changes in chlamydial load, host immune cytokine expression or anti-Chlamydia antibody production, and chlamydial disease. Vaccination success can be measured for its protective (against pathogen infection or disease from non-disease hosts) or therapeutic effect (reducing pathogen shedding or disease from diseased hosts). An additional research chapter using similar methodologies and focussing on Chlamydia, but departing from the vaccine focus, evaluated epidemiological factors likely to affect repeat chlamydial infection in women in Australia. A chlamydial vaccine for humans does not yet exist but would be an ideal management strategy for controlling chlamydial infections. The number of vaccine-development studies published in recent years has made it difficult to determine trends and an objective review of the literature is important to identify the most promising vaccine candidate against chlamydial infection. In Chapter Two I performed a meta-analysis on systematically selected studies that aimed to develop a chlamydial vaccine either against C. trachomatis or another chlamydial species. Over 4,400 standardized effect sizes were calculated between control and chlamydial vaccination groups. Mice have most often been used in chlamydial vaccine research (78%) and most vaccines against Chlamydia reduced chlamydial load and increased host immune parameter markers, including the antibodies IgA and IgG1, and the cytokine, IFN˜í‚â•. Mice are frequently used in vaccine research due to the known inbred pedigree of each mouse, availability of the murine immunological toolkit, and a smaller demand of veterinary resources compared to non-mouse models. There are, however, limitations to mouse models of chlamydial infection that are used in experiments to develop a chlamydial vaccine for humans, most notably the unrealistically controlled conditions in laboratory settings. Koalas are infected with C. pecorum in the wild and these infections have parallels with human C. trachomatis infections. In Chapter Three I investigated vaccine-immune-chlamydial load-disease relationships from previously collected data from MOMP (major outer membrane protein) vaccine trials in free-ranging koalas. Using structural equation modelling I created a priori hypotheses about perceived direct and indirect interactions from koalas vaccinated six months prior. I found MOMP vaccination had a strong effect on increasing interleukin 17 (IL17) mRNA expression, and that urogenital chlamydial load was positively associated with disease and negatively associated with IL17. Despite multiple potential sources of variation, owing to the koalas being free-ranging, these analyses helped illuminate a link between MOMP vaccination, urogenital chlamydial load and the cytokine IL17, enhancing previous investigations. In Chapter Four I investigated individual variation in the immune response of koalas to MOMP vaccination, with a focus on immunoglobulin G (IgG) antibodies. I undertook this investigation in recognition that an ideal property of vaccines in development (from a veterinary and medical practitioner perspectives) is for them to elicit predictable immune responses with minimal variation among individuals. However, many studies instead focus on average group effects (cohort effects). Using mixed effects models and methods adapted from the behaviour literature I examined previously collected IgG abundance data from koalas spanning three vaccine studies. I found significant heterogeneity in the individual variation of koala IgG levels in response to vaccination. Individual variation was minimised in vaccine trials undertaken on captive koalas measured over more timepoints after vaccination. This particular investigation presents strong evidence that chlamydial vaccine studies should consider examining both the average cohort effects and the individual variability in vaccine development trials. Finally (Chapter Five), I undertook an epidemiological investigation of chlamydial reinfection risk in humans using similar structural equation modelling approaches to those used in Chapter three. Most genital chlamydial infections in humans can be treated with antibiotics, yet repeat infections of treated individuals in some populations remains significant (~20%). Multiple direct and indirect factors are associated with repeat infections and these associations are often complex and not well understood. I utilised data from the Australian Chlamydia Treatment Study, from which 239 women were recruited and 33 (13.8%) repeat infections were documented. My models confirmed that repeat chlamydial infections were predicted directly and positively by inconsistent condom usage. Importantly, I found repeat chlamydial infections were indirectly associated with participant age, use of anal sex, sexual network size, and vaginal sex frequency. These indirect factors highlight important factors for healthcare providers to consider for controlling repeat chlamydial infections. My PhD research has advanced our understanding of the efficacy of candidate chlamydial vaccines and aspects of chlamydial epidemiology. The findings have contributed to: 1) the identification of promising directions toward the development of a chlamydial vaccine, 2) the direct and indirect factors associated with chlamydial disease, 3) the individual variability among systemic antibody responses to vaccination, and 4) the direct and indirect factors associated with repeat genital Chlamydia infections. These findings have been achieved through the use of novel biostatistical approaches in the chlamydial research field, and through the application of these techniques to existing laboratory, wildlife and human studies. More broadly, the modelling approaches used in this thesis are also applicable to other fields of vaccinology and epidemiology.