#64: An odyssey Beyond the Capsule: Genetic Determinants of Pediatric Invasive Streptococcus pneumoniae Empyema

Abstract Background Streptococcus pneumoniae is among the most common causes of invasive bacterial infections in children, including pneumonia, bacteremia, and meningitis. Over 90 different serotypes (ST) of pneumococcus exist, with enrichment of some ST within specific invasive phenotypes. Other than capsular genes, molecular determinants of particular invasive phenotypes remain largely unknown. Although vaccination targeting especially invasive ST capsular antigens has successfully decreased the incidence of invasive pneumococcal disease (IPD), new ST have emerged, suggesting methods to target other aspects of pneumococcal invasiveness are needed. Methods Pneumococcal isolates from IPD were collected from children presenting to Primary Children’s Hospital from 1996–2018. All viable isolates underwent next-generation sequencing (Illumina), quality control filtering for contamination and low coverage, de novo genome assembly with SPADES, and annotation with PROKKA. Clinical phenotypes were manually validated with physician chart review. Isolates were serotyped via Quelling and in silico using SeroBA. ROARY was used for pan-genome assembly, and SCOARY for microbial genome-wide association studies. RAxML was used for phylogenetic analysis. Results A total of 354 viable pneumococcal isolates were available for genomic analysis including a spectrum of invasive phenotypes: pneumonia (n = 138, of which 54 were complicated by empyema), CNS infection (n = 50), SSTI/bone infections (n = 42), and isolated bacteremia (n = 68). Thirteen samples were censored for poor coverage or genetic contamination. Invasive isolates spanned 37 capsular ST. The pneumococcal pan-genome comprised 6462 genes, of which only 23% were shared by at least 99% of samples. Phylogenetic relatedness resulted in clustering of some ST (e.g., ST1, ST3), whereas others (eg ST19A) were more broadly distributed. Empyema and meningitis phenotypes were distributed across the phylogenetic tree, but enriched in distinct clusters that crossed ST clusters. Genes involved in empyemagenic pneumococcal capsule production, and those implicated in sensing of preferred sugars or non-preferred sugar metabolism were statistically correlated with the empyema phenotype. Conclusion There is marked genetic diversity among invasive pneumococcal isolates, potentially contributing to the variability of disease phenotypes observed. Clustering of invasive phenotypes across ST suggests a genetic signature for invasive phenotypes other than capsule genes alone, further supported by enrichment of specific genes associated with alternative sugar metabolism in empyema isolates. Critical determinants of invasive phenotypes will inform future efforts at disease prevention and treatment.