Detecting Neonatal Sepsis by Immune-Metabolic Network Analysis

Several studies have shown that changes in host gene expression may occur pre- symptomatically in response to infection in any part of the body, with the continuous interaction between blood and tissue allowing blood cells to act as biosensors for the changes (Manger and Relman, 2000, Liew et al., 2006). Genome wide analysis reveals coordinate expression that develop networks causally linked through pathways.

Earlier studies from our group investigating optimal methods for the sampling and extraction of neonatal whole transcription products (RNA) demonstrated the first feasibility studies for using genome wide RNA analysis as a methodological approach for identifying host biomarkers of infection and vaccination in early life (Smith et al., 2007, Flanagan et al., 2013). The sampling methods were further refined in 2015 with the development of a single drop methodology that has been extensively tested in a wide range of settings including the collection of neonatal samples in the home, at the point of Guthrie testing by mid-wives. Details of these methods have been recently submitted to the Nature Protocols journal. Further, we conducted early on virtual clinical trials using a super computing framework that simulated several 100,000 neonatal whole blood samples for predicting infection (Khondoker et al., 2010). Those investigations showed the requirement for multiple markers and ideally in discrete biological pathways underpinning causality (Khondoker et al., 2010, Watterson and Ghazal, 2010). Those investigations provided a strong foundation for initiating a case-control of neonatal sepsis (Dickinson et al., 2015). Accordingly, we were the first group to publish studies investigating the systemic immune response in neonates to sepsis by measuring the activity of all known human genes (Smith et al Nature comm. 2014). These computationally intensive investigations led to uncovering, for the first time, the pathway biology underlying neonatal sepsis with blood samples taken at the first clinical signs of suspecting an infection. A combination of machine learning, statistical and deep pathway biology analyses led to the identification of a 52-gene panel of biologically connected network modules. The modules comprise three central pathways, innate-immune or inflammatory, adaptive-immune and unexpectedly metabolic. The expression levels of particular combinations of biomarkers, and specifically those of a pathway previously unconnected to immune responses, gives an unusually high diagnostic quality. Despite patient heterogeneity, the 52-node dual biomarker network had greater than 99% accuracy for detecting bacterial infection with 100% sensitivity showing superior performance to previously characterised markers. Furthermore, these specific combinations of biomarkers allowed the detection of neonatal sepsis in samples which had displayed blood-culture negative results, illustrating the specific diagnostic benefits of the particular combinations of biomarkers. The unexpectedly high accuracy and sensitivity values could not have resulted from the investigation of any of the individual biomarkers alone, nor could they have been predicted. A critical part of these findings is the requirement of metabolic pathways for increasing both sensitivity and specificity. A subset of the metabolic markers encompass ligands (specifically small and medium chain fatty acids) that are derived from microbial metabolism, in particular from commensals and which are reflected in the faecal microbiome. To date these studies provide a proof of concept but need independent confirmatory studies as well as investigating specificity against non-bacterial (viral and fungal) infections and sterile inflammation. The urgent unmet medical question is whether predictive host pathways of infection can be used to first identify whether a patient is infected at or before clinical presentation and, to further discriminate between the type of infection (in particular bacterial or viral) and predictability of sepsis.