Statistical Modeling of Genetic Susceptibility to Immune-Driven Disease in Collaborative Cross-Derived Mouse Populations

Immune-driven diseases, including severe viral disease, pose a large and growing public health burden, and their progression is shaped by a combination of environmental and genetic factors. Improved mechanistic understanding of genetic susceptibility to immune-mediated disease can inform the development of more effective therapeutic and preventive strategies. Experimental mouse models enable precise control of environmental variables that are especially important in the development of immune-driven disease, including prior immune exposure, diet, and pathogen strain and dose, while also allowing tissue-specific sampling of immune responses that are difficult or impossible to obtain in human cohorts. Recombinant inbred genetic reference populations such as the Collaborative Cross (CC) combine extensive genetic diversity with the replicability of inbred strains to enable the effective study of complex traits. Crosses between phenotypically divergent CC strains are particularly powerful for mapping genetic loci underlying disease variation and for linking these loci to intermediate immune phenotypes that can inform causal mechanisms of susceptibility. Here, we use CC-derived experimental crosses to study mechanisms of genetic susceptibility to immune-driven disease. First, we use a CC027xC3H cross to study genetic regulation of susceptibility to orally driven anaphylaxis exhibited by CC027, a novel mouse model for peanut allergy. We identify several loci associated with this more human physiology-relevant mode of susceptibility, and we identify a convincing candidate gene in the T-cell developmental gene, Themis. Next, we use a cross between coronavirus-susceptible CC006 and coronavirus-resistant CC044 to study genetic susceptibility to severe coronavirus disease. We identify host susceptibility loci associated with disease severity that are conserved across viruses and between mouse and human. We then perform an integrative analysis of immune profiles from these F2 mice, including QTL mapping and Bayesian variable selection and mediation analysis, to identify immune mediators of genetically driven disease severity. We observe an immune response that is consistent with that observed in humans, and we detect genetic regulation of this immune response that affects disease severity through both virus-specific and shared immune mechanisms. These studies provide an experimental and statistical framework for the study of complex, multi-trait, and treatment-dependent disease outcomes.