Transcriptome profiles discriminate between Gram-positive and Gram-negative sepsis in preterm neonates.

BACKGROUND Genome-wide expression profiles have been previously employed as clinical research diagnostic tools for newborn sepsis. We aimed to determine if transcriptomic profiles could discriminate between Gram-positive and Gram-negative bacterial sepsis in preterm infants. METHODS Prospective, observational, double-cohort study was conducted in very low birth weight infants with clinical signs and culture-positive sepsis. Blood samples were collected when clinical signs became apparent. Total RNA was processed for transcriptomic analysis. Results were validated by both reverse-transcription polymerase chain reaction and a mathematical model. RESULTS We included 25 septic preterm infants, 17 with Gram-positive and 8 with Gram-negative bacteria. The principal component analysis identified these two clusters of patients. We performed a predictive model based on 21 genes that showed an area under the receiver-operating characteristic curve of 1. Eight genes were overexpressed in Gram-positive septic infants: CD37, CSK, MAN2B2, MGAT1, MOB3A, MYO9B, SH2D3C, and TEP1. The most significantly overexpressed pathways were related to metabolic and immunomodulating responses that translated into an equilibrium between pro- and anti-inflammatory responses. CONCLUSIONS The transcriptomic profile allowed identification of whether the causative agent was Gram-positive or Gram-negative bacteria. The overexpression of genes such as CD37 and CSK, which control cytokine production and cell survival, could explain the better clinical outcome in sepsis caused by Gram-positive bacteria. IMPACT Transcriptomic profiles not only enable an early diagnosis of sepsis in very low birth weight infants but also discriminate between Gram-positive and Gram-negative bacteria as causative agents. The overexpression of some genes related to cytokine production and cell survival could explain the better clinical outcome in sepsis caused by Gram-positive bacteria, and could lead us to a future, targeted therapy.