Analysis of the dynamics of protective immune responses in human populations with endemic schistosome infection

Abstract

Urinary schistosomiasis, which is caused by the blood fluke Schistosoma haematobium, is a tropical disease infecting over 100 million people in sub-Saharan Africa. Infection typically involves repeated re-infection with long-lived parasites, and field studies have demonstrated that protective immunity takes many years to develop in humans. In communities with endemic schistosomiasis, distinctive patterns of infection and schistosome-specific antibody responses are seen, including a peaked age-infection curve, a highly aggregated distribution of infection intensities among individuals, and an age-related switch in the subclasses of antibody produced. The antibody switch, which occurs naturally in older children, is also seen in younger children following treatment with the antihelminthic drug praziquantel, which kills adult worms. This study aimed to identify the important mechanisms underlying the slow development of protective immunity, using a mathematical modelling approach. Deterministic population-level and stochastic individual-based models were developed that describe how levels of infection and antibody change with age for individuals living in endemic communities. These models were used to explore different hypotheses for the slow development of protective immunity: (1) that schistosome parasites actively suppress protective immune responses; (2) that dying worms provide the main antigenic stimulus for protective immunity and (3) that a threshold level of antigen must be experienced before a protective immune response is initiated. Models were assessed for their ability to simultaneously reproduce different robust patterns of infection and antibody responses identified in cross-sectional and post-treatment field data from Zimbabwe. It was found that significant immunosuppression by schistosomes was not consistent with population-level patterns of infection intensity, including the peaked age-infection curve. In order to explain both age-related and post-treatment changes in infection intensity and antibody responses, including the antibody switch, protective antibody responses had to be stimulated by antigens from dying worms. Additionally, it was shown that these protective responses reduced worm fecundity rather than reducing rates of re-infection. An antigen threshold was found to be consistent with observed field patterns, but was not necessary to explain them. From a large number of possible models that were considered, a single model structure and a subset of parameter combinations were identified that were consistent with field data. This model was used to predict the longer-term impact of mass-treatment programmes upon the development of protective immunity, and the consequent effects on infection levels.