Introduction: The etiology of ventilator-associated lung injury (VALI) is unknown. Mechanical stress, such as over-distension or airways opening and closing, and its interaction with inflammation have been thought to cause progression of injury in already damaged lung tissues. Prior studies with CT imaging in injured lungs have demonstrated regional changes following lavage injury. We hypothesized that VALI does in fact have its origin in inflammatory or other cellular responses in lung tissue exposed to mechanical stress, and that these regional cellular responses would correlate with the specific regional mechanical stresses.

Methods: In four anesthetized, supine dogs, lung isolation was achieved with a standard double lumen endobronchial tube via tracheostomy. After position was confirmed by bronchoscopy, independent lung ventilation with FiO2 1.0 was performed via a double-piston ventilator (Harvard Apparatus, Holliston, MA). Individual lung Vt was adjusted to produce a baseline ETCO2 of 30-35 mmHg for each lung. Next, the left lung was mildly injured by saline lavage (20 ml/kg repeated 4 times). Ventilation remained at baseline, with the exception of 5 cm H2O PEEP added to further protect the control lung. After 5 hours, the animal was sacrificed by exsanguination, the lungs removed, and tissue samples were taken from 5 corresponding regions in both lungs (apex and base). These regions were chosen to correspond to areas of diverse mechanical stress based on previous CT studies.

Conclusions: Microarray analysis can be used to explore ventilation related cellular responses in a large animal model. In addition, regional differences in gene expression in our model of ALI are reproducible and are in a pattern consistent with known variation in mechanical stresses in the injured lung. Specific cellular responses to mechanical ventilation such as up-regulation of PBEF may correlate with progressive lung injury, and may be a genomic biomarker of VALI. The powerful combination of regional genomics with functional CT imaging in a clinically relevant model has the potential to fuel major advances in the investigation of mechanisms behind VALI.