Environmental drivers of low vaccine responsiveness in a lab-to-wild rodent model

Vaccination is the most effective way to prevent infectious diseases and safeguard public health. Yet, most new vaccines fail in late clinical trials, and even established ones often underperform in populations apart from those in which they were initially tested. This can lead to reduced vaccine responsiveness, breakthrough infections, and prevent or delay herd immunity. While the causes of vaccine hyporesponsiveness remain difficult to identify, quantify, and therefore address, numerous reports indicate a predominant role of environmental factors. This has notably been demonstrated by a reduction in the immunogenicity and efficacy of various vaccines when transitioning from urban to rural human populations. Here, we tested whether and, if so, how the environment can cause vaccine hyporesponsiveness. We hypothesised that if the leading causes of vaccine hyporesponsiveness were environmental, then environmentally driven hyporesponsiveness would be exacerbated when individuals are under nutritional stress; specifically predicting that high-quality diet supplementation would increase vaccine responsiveness. Finally, we predicted that parasitic helminth infections, which are more common in rural populations, would degrade vaccine responsiveness, e.g., due to their ability to modulate host immunity, and that anthelmintic treatment could rescue vaccine responsiveness in infected individuals. To test these hypotheses, we coupled lab and field experiments with structural causal modelling, and quantified diphtheria toxoid-specific IgG1 optical density (OD) in paired conspecific cohorts of laboratory-reared and wild wood mice (Apodemus sylvaticus) given a single or two doses of diphtheria toxoid vaccine formulated with alum, with and without diet supplementation. We found that anti-toxoid IgG1 OD was ∼ 47 % lower in thewildwoodmice compared to the laboratory-reared population. We also demonstrated that, across both habitats (wild and lab), substantial variation in vaccine responsiveness was caused by diet. However, contrary to our predictions, this high-quality dietary supplementation resulted in lower vaccine responsiveness. Further, once the effects of habitat, diet, and sex were adjusted for, increasing helminth infection burdens negatively affected anti-toxoid IgG1 OD. Counterfactual predictions from our structural causal model suggested that targeting anthelmintic treatment at heavily infected individuals could have improved their anti-toxoid IgG1 OD responses by approximately 2 to 4-fold. Our results indicated that the wild environment and access to a high-quality diet played a dramatic role in shaping the immune system’s response to immunisation. Further, we showed that laboratory settings, even when using a genetically diverse, non-traditional model, systematically yielded higher IgG1 OD than was observed in free-living conspecifics on the same protocol. We provide a causally explicit modelling approach to quantify how habitat, diet, and parasites jointly shape anti-diphtheria toxoid IgG1 levels in a focal population, and to prioritise adjunct interventions such as deworming where model assumptions hold.Author summary: Vaccines often work less well outside tightly controlled trials, especially where infection and nutrition vary. Standard laboratory animals typically miss that heterogeneity and thus fail to capture important characteristics of intended recipient populations. We vaccinated wood mice (Apodemus sylvaticus) with diphtheria toxoid in the laboratory and in Scottish woodlands, with or without high-quality diet supplementation. We read out humoral responses as diphtheria toxoid-specific IgG1 optical density in serum (ELISA), a standard correlate of immunisation rather than a direct measure of clinical protection. Wild mice showed markedly lower vaccine responsiveness than laboratory conspecifics on the same protocol, while high-quality supplementation unexpectedly lowered IgG1 OD in both habitats. Among naturally infected animals, higher gastrointestinal nematode burdens tracked with weaker vaccine responses. Structural causal models integrated these data and simulated parasite removal before vaccination; under stated assumptions, the models predicted large gains in IgG1 OD for heavily infected individuals. This work highlights how environment and parasites shaped IgG1 levels in our experiment and illustrates causal modelling for immunisation in diverse hosts.