Dissecting breathing control dysregulation in neonatal sepsis

Abstract

BackgroundSepsis is a common and frequently fatal condition affecting neonates. The global burden of neonatal sepsis is underappreciated especially regarding breathing and maintenance of adequate ventilation which is essential to life. Systemic inflammation induced by sepsis triggers activation of immune cells peripherally, which can be sensed and further propagated by multiple cell types in the CNS, including those in the brainstem involved in the neural control of breathing. Understanding how neonatal sepsis alters breathing plasticity is crucial for promoting earlier diagnosis and better therapeutic interventions. Here, we investigated how systemic inflammation caused by distinct TLRs ligands in neonates affects breathing plasticity.MethodsCD1 mice were I.P. treated at postnatal day 5 (PD5) with saline (control), LPS, or Pam3CSK4 (PAM), and after 3 hours, they were either euthanized or submitted to respiratory tests using Buxco unrestrained whole‐body plethysmography (WBP). After euthanasia, brainstem was dissected out and slices containing PreBötC and hypoglossal pool were prepared and electrophysiological recordings were performed simultaneously under baseline and hypoxic (95%N2/5%CO2) conditions. Microarray analysis of whole brainstem and isolated neurons, astrocytes, and microglia were performed with Affymetrix Mouse Gene ST2.0 arrays.ResultsRespiratory response to hypoxia, hypercapnia and, anoxia gas challenges were preserved and functional in neonatal PD5 pups. The ventilatory response in TLR4 and TLR1/2 ligands treated pups was reduced during baseline, and an inhibited response to hypoxic and hypercapnic conditions was observed. Recovery time after each anoxic episode was calculated as gasping latency and PAM treated pups demonstrated a disrupted ability to recover after anoxic episodes. The central respiratory activity measured in isolated preBötC slices from PAM treated pups showed that networks failed to transmit rhythmic activity to the hypoglossal nucleus. In PAM treated networks hypoxic rhythmogenesis from the preBötC was weaker, yet in network from LPS treated pups, hypoxic rhythmogenesis continued. Next, we assessed gene modulation in the brainstem PD5 pups. The majority of differentially expressed genes (DEGs) were related to innate immunity in both LPS or PAM groups. The enrichment analysis showed activation of T helper cells, whereas CIBERSORT analysis indicated enrichment of Treg and B cells memory. The molecular pathways activated in neurons, astrocytes, and microglia were founded distinct and specific to the TLR ligand. Neurons were strongly modulated in both of the groups. Surprisingly, astrocytes coming from LPS treated pups exhibited only a few DEGs whereas PAM did not induce any modulation.ConclusionsTaken together, these results shows that the ventilatory response is detrimentally affected in neonatal sepsis within only a few hours of an inflammatory insult. Not only the ability to sense O2/CO2 levels and to respond accordingly is disrupted, but also the central respiratory rhythmic activity in preBötC slices. Both TLRs ligands induced a strong innate immune response in the brainstem which is amplified in neurons especially in pups treated with TLR1/2 ligand.