Identifying potential microbial aetiologies of Kawasaki Disease with metagenomics and metaproteomics

Abstract

Kawasaki disease (KD) is an acute inflammatory disorder of early childhood. It causes coronary artery aneurysms, due to the vasculitis which underlies much pathology. Despite epidemiological evidence supporting an infectious trigger in predisposed individuals, decades of research have found no strong evidence to implicate an individual organism. A diverse range of organisms have been suggested as the cause, frequently with limited or contradictory evidence. Studies are typically small and consider a narrow range of organisms, mostly pathogens. Two promising sites to search for microbial material are the oropharynx and immune complexes (IC). The former is the main microbial entry portal, and the latter form in the subacute phase of illness and may include microbial antigen. I present bioinformatic analyses aiming to identify microbes associated with KD from these two sources. Data comprises oropharyngeal metagenomics (116 cases; 101 controls) and metaproteomics of ICs (112 cases; 128 controls). Extensive method development was required, with a focus on mitigating against laboratory and reference database contamination for metagenomics and finding database search methods which could cope with a vast metaproteomic search space. Metaproteomic analyses yielded no strong evidence of microbial associations, though the data allowed novel, exploratory analysis of immunoglobulin peptides and germline immunoglobulin locus usage. Since immunoglobulin is strongly implicated both genetically and pathologically in KD, the methods developed here show potential for further development and application. Metagenomic analyses found organisms associated with KD, including Abiotrophia defectiva and Lautropia mirabilis (six-fold higher abundance, Q values 0.06 and 0.05 respectively). The strengthening of relationships when age-profiles are modelled, and previous implication of A. defectiva in infective endocarditis add to the potential significance of the findings. I interpret the results in the light of the epidemiology of KD and consider the limitations of this study, and challenges inherent in the search for causes of KD.