Outer membrane vesicles as carriers for bacterial extracellular ATP modulate local and remote inflammation in peritoneal sepsis

Abstract Objective Adenosine triphosphate (ATP) may be released to the extracellular space and activate specific purinergic receptors to modulate inflammatory responses. In addition to eukaryotes, ATP is also released by bacteria. This bacterial extracellular ATP (eATP) subsequently may act as an inter-kingdom signaling molecule between bacteria and its host. How ATP is secreted by bacteria and its consequences in systemic infection such as sepsis remains unclear. This study aims to determine how ATP is secreted by bacteria and to elucidate its interaction with the host immune system in a model of peritoneal sepsis. Methods Bacteria from septic patients were isolated from peritoneal washes. Peritoneal sepsis in mice was induced by cecal ligation and puncture. Peritoneal fluid was collected, (an)aerobically plated and colonies were identified by 16S-sequencing. EATP was periodically measured from bacteria culture supernatant using a luciferin-luciferase assay. Bacteria in culture were counted by plating and the amount of live/dead bacteria was assessed using flow cytometry. To specifically assess the role of bacterial eATP in vivo, E. coli K-12 BW25113 parental strain from the keio collection was used, transformed with a plasmid coding for an apyrase and i.p. injected. Furthermore, a hypervesiculation mutant of this keio collection was used to collect outer membrane vesicles (OMV). These OMV were then loaded with ATP using electroporation and injected i.p in mice. Results All bacteria that were isolated from human or mouse peritoneal washes were able to secrete ATP in a growth-phase dependent manner in vitro. In addition to passive release by dying bacteria, we have identified that the amount of eATP is connected with the formation of outer membrane vesicles (OMV) as shown by the E. coliΔtolB hypervesiculation mutant. Whole bacterial eATP led to a significant worse outcome in mice suffering from peritoneal sepsis. In addition, bacterial eATP specifically delivered via OMV worsened septic outcome as well, critically modulating neutrophil recruitment. At the infection site, ATP-loaded OMV led to an impaired recruitment of neutrophils whereas in the lung they led to an elevated infiltration. Conclusion OMV serve as a carrier of bacterial wall LPS on its surface as well as bacterial eATP as its content and specifically alter the outcome during sepsis influencing neutrophils locally as well as at remote sites.