Human models of acute lung injury

Abstract

Acute lung injury (ALI) is a syndrome that is characterised by acute inflammation and tissue injury that affects normal gas exchange in the lungs. Hallmarks of ALI include dysfunction of the alveolar-capillary membrane resulting in increased vascular permeability, an influx of inflammatory cells into the lung and a local pro-coagulant state. Patients with ALI present with severe hypoxaemia and radiological evidence of bilateral pulmonary oedema. The syndrome has a mortality rate of approximately 35% and usually requires invasive mechanical ventilation. ALI can follow direct pulmonary insults, such as pneumonia, or occur indirectly as a result of blood-borne insults, commonly severe bacterial sepsis. Although animal models of ALI have been developed, none of them fully recapitulate the human disease. The differences between the human syndrome and the phenotype observed in animal models might, in part, explain why interventions that are successful in models have failed to translate into novel therapies. Improved animal models and the development of human in vivo and ex vivo models are therefore required. In this article, we consider the clinical features of ALI, discuss the limitations of current animal models and highlight how emerging human models of ALI might help to answer outstanding questions about this syndrome.

Fig. 1.

Pathogenesis of ALI. (A,B) Direct (e.g. pneumonia) or indirect (e.g. sepsis) injuries (A) result in sloughing of alveolar epithelial cells, causing the formation of protein-rich hyaline membranes (B). Neutrophils adhere to the activated capillary endothelium and migrate to the alveolus. Loss of alveolar-capillary membrane barrier integrity facilitates the accumulation of a neutrophilic inflammatory exudate in the interstitium and air space. Resident and recruited cells secrete inflammatory mediators that disrupt epithelial fluid transport and impair surfactant production by alveolar type II cells. Capillary microthrombi and extravascular fibrin deposition potentiate pulmonary dysfunction and the acute inflammatory response. This interplay between inflammation and capillary thrombosis increases pulmonary dead space and shunt, contributing to the severe hypoxaemia observed clinically.

Fig. 2.

Potential roles of human models of ALI in the development of new therapeutics. Experiments performed in ex vivo models guided by animal data and human in vivo observational data will allow safety and basic efficacy testing of novel therapies before Phase I clinical trials are carried out in healthy volunteers. LPS and cardiopulmonary bypass (CBP) in vivo models can be used to provide proof-of-concept early Phase II data, whereas the oesophagectomy surgical in vivo model has the potential to interrogate clinical outcomes.

Fig. 3.

Schematics illustrating ex vivo human models of ALI. The isolated perfused lung model (Frank et al., 2007) (A) is a simplified version of the ex vivo lung perfusion model (EVLP) (Steen et al., 2007) (B). (A) In the isolated perfused lung model, whole human lungs are suspended by the trachea in a humidified chamber. The lungs are perfused with a buffer solution, at a low flow rate, that drains passively via the pulmonary vein. The lungs are either left unventilated or continuous positive airway pressure (CPAP) is applied. (B) The EVLP model is a more elaborate system involving cardiopulmonary bypass circuits. A closed system is employed with cannulae placed in both the pulmonary vein and artery. The lungs are perfused with a high oncotic pressure perfusate (Steen solution) with or without red blood cells, circulated via a centrifugal pump. A membrane oxygenator is supplied with a gas mixture high in nitrogen and carbon dioxide to create a ‘venous’ type perfusate prior to returning to the pulmonary artery. A polyurethane filter is employed to prevent recirculation of leukocytes into the circuit and a protective strategy (5 ml/kg tidal volume) is used to ventilate the lungs. LAP, left atrial pressure; PAP, pulmonary artery pressure; pO₂, partial pressure of oxygen in the perfusate. Figures reproduced with permission (Steen et al., 2001; Frank et al., 2007).