Two kinds of immunity underlie pathogen competition for hosts as a function of history of exposure: specific immunity depending on memory of specific variants that have been seen before, and generalized immunity resulting only from the number of previous infections regardless of their antigenic identity. Depending on its strength, selection imposed by specific immunity can lead to different kinds of coexistence regimes in antigenically diverse pathogens, including single strain dominance, cyclical or chaotic replacements, and stationary coexistence of non-overlapping strains. The role of immune selection remains unclear in rotavirus, the most common cause of diarrheal disease worldwide which contributes 40% of total hospitalizations in young children. With a process-based model that allows for both demographic and measurement noise, we analyzed over 10 years of monthly incidence data at the serotype level from Dhaka, Bangladesh. We specifically investigated the role of specific and generalized immunity in temporal patterns of antigenic diversity by fitting and comparing models that represent these different hypotheses, including consideration of variation in the two antigenic determinants on the surface of the virus.
Results/Conclusions
Our findings show that strong generalized immunity is needed to recover the serotype dynamics of rotavirus in Dhaka, a process dominated by exposure to VP4 (P-type), the outer layer protein that mediates cell attachment. A strong generalized response produced against VP4 is consistent with the highly conserved features and functional constraints of this protein. Our results further indicate a role of specific immunity via the VP7 (outer capsid protein, G-type), whose effect is weak by comparison to that of VP4 but is nevertheless required to capture the epidemiological dynamics and antigenic diversity patterns of the virus. A higher number of G than P types observed in common G-P combinations of human rotavirus strains support these findings. Specific immunity could thus explain why G-types not targeted in the vaccine are highly prevalent in areas of the developing world with low vaccine efficacy. Our findings have the potential to contribute to the improvement of intervention strategies, especially in developing countries such as Bangladesh where the efficacy of the rotavirus vaccine is only 41%.