27.01 Differential Effects of Bacteria on Endothelium and Epithelium in the Ex Vivo Perfused Human Lung

J. T. Ross1, N. Nesseler2, E. Minus3, J. E. Gotts4, M. A. Matthay4  1University Of California – San Francisco,Department Of Surgery,San Francisco, CA, USA 2University Hospital of Rennes,Intensive Care Unit, Department Of Anesthesia And Critical Care,Rennes, BRITTANY, France 3Amherst College,Amherst, MA, USA 4University Of California – San Francisco,Departments Of Medicine And Anesthesia, Cardiovascular Research Institute,San Francisco, CA, USA

Introduction: ARDS is an important complication of sepsis, but develops less frequently in non-pulmonary than in pulmonary sepsis. Patients with non-pulmonary sepsis often develop pulmonary edema, but rarely progress to ARDS. We hypothesized that there may be a differential susceptibility of the lung endothelium and alveolar epithelium to bacterial-induced injury. The goal of these studies was to measure endothelial and epithelial injury in the ex vivo lung when exposed to S. pneumonia instilled into the air spaces or injected into the perfusate.

Methods: Human lungs not suitable for transplantation were received, and the right or left lung was selected based on gross appearance. The pulmonary artery and main bronchus were cannulated. A 3 Fr catheter was passed via the bronchus into the distal airspaces. The lung was perfused with DME-H21 and 5% bovine serum albumin at 37°C with a mean pulmonary arterial pressure of 10 mmHg (Figure 1). Perfusate drained passively from the pulmonary veins and collected in a reservoir for recirculation. Lungs were inflated with room air at a continuous positive airway pressure of 8 cmH2O. 100ml of fresh human blood was added to the perfusate. In a subset of lungs, 1010 cfu S. pneumonia were added to either the perfusate or distal airspaces. Alveolar fluid clearance (AFC) was measured at 0 and 5 hours by instilling 100ml of 5% albumin into the distal airspaces and comparing protein concentrations at 5 and 35 minutes. Endothelial and epithelial permeability were evaluated by measuring weight gain of the lung over 6 hours and the accumulation of IgM from the perfusate into the airspaces.

Results: Of the 48 lungs perfused, 77% had a normal starting AFC. Lungs in the control group had minimal weight gain (mean 321 ± 212g, 68 ± 28% of start weight), preserved AFC (normal AFC at 5h in 91%) and minimal accumulation of IgM in the airspaces (mean 0.004g, IQR 0.0002-0.004g). Addition of S. pneumoniae into the distal airspaces was associated with significant weight gain compared to controls (mean 127 vs 68%, p = 0.022), significant decrease in AFC (p<0.0001) and a trend towards accumulation of IgM in the airspaces. S. pneumoniae in the perfusate was associated with significant weight gain (mean 155 vs 68%, p = 0.012) but preserved AFC and minimal IgM accumulation.

Conclusion: The results indicate that the human lung endothelium is susceptible to injury from S. pneumonia in either the vascular compartment or the airspaces. In contrast, the alveolar epithelium is susceptible to injury from S. pneumoniae in the airspaces, but resistant to high doses of S. pneumoniae in the vascular compartment. This pattern suggests one explanation for the lower incidence of ARDS after non-pulmonary sepsis.