Our research group has developed an ex vivo
perfused human lung model to provide more clinically relevant and mechanistic information regarding potential new therapies for human ALI/ ARDS. We found that the perfused human lung was stable for up to 6 hrs; furthermore, basal alveolar fluid clearance proceeded at a level of 18% to 20% per hour, which was much faster than clearance in the nonperfused human lung (27
). In addition, we found that it was possible to stimulate alveolar fluid clearance with β-agonist therapy to a level of approximately 35% to 40% (27
). Subsequently, we adapted this preparation to study the potential therapeutic mechanistic effects of allogeneic MSC therapy for E. coli
endotoxin-induced lung injury.
For all of these studies, human lungs were donated by the Northern California Transplant Donor Network (Oakland, CA). We used lungs that were not considered suitable or appropriate for lung transplantation. The mean age of the donor lungs was 48 ± 13 yrs, and the ischemia time was 21 ± 13 hrs (mean ± sd). Perfusion was established with a physiologic solution with 5% albumin. Then, approximately 100 mL whole blood was removed from healthy donors and added to the perfusate (with a final hematocrit of 4%). The whole blood was not cross-matched or type-specific with the donor’s human blood type because the human lungs contained almost no residual blood and the lung was flushed initially during reperfusion with a perfusate solution. There were no adverse hemodynamic or pulmonary effects associated with the addition of the whole blood to the perfusate. The preparation was stable and suitable for experimental studies for a period up to 7 hrs with an appropriate baseline (28
Lung injury was induced by the instillation of 0.1 mL/kg of E. coli endotoxin into either the right middle lobe or the left lower lobe. The preparation was performed with perfusion of one lung. We compared the effects of instilling MSCs or the cultured medium of the MSCs or control normal human lung fibroblasts in this preparation. The results showed that endotoxin caused an acute severe neutrophilic inflammation in the right middle lobe with a three-fold increase in lung vascular permeability and a marked increase in lung water. Furthermore, alveolar fluid clearance was reduced to zero with endotoxin compared to normal levels of nearly 18% to 20% per hour. In addition, endotoxin exposure produced a marked increase in the alveolar compartment of proinflammatory cytokines, including IL-1β, tumor necrosis factor-α, and IL-8.
Treatment with MSCs was administered 1 hr after the endotoxin instillation, with instillation of the MSCs into the same region of the right middle lobe where the endotoxin had been instilled. The MSCs completely restored both lung endothelial permeability and lung water to a normal level (). Histologic analysis showed that the inflammatory infiltrate was markedly decreased. In addition, alveolar fluid clearance was restored to a normal level by the MSCs.
Figure 1 Effect of human mesenchymal stem cells or their conditioned medium on lung endothelial permeability to protein and wet-to-dry (W/D) ratio in the ex vivo perfused human lung injured with Escherichia coli endotoxin. Instillation of mesenchymal stem cells (more ...)
Interestingly, the cultured medium of the MSCs was equally effective in reversing the effects of ALI/ARDS from endotoxin. The cultured medium normalized lung endothelial permeability, lung water, and alveolar fluid clearance, and it decreased the histologic injury. There was no effect on cytokine levels with the exception of a small decrease in IL-1β levels with the cultured milieu.
In searching for the mechanisms that might account for these interesting effects, we speculated that the release of a growth factor could be important because we and other investigators (29
) have found that keratinocyte growth factor (KGF) was capable of preventing a wide variety of acute lung injury. We found that the production of KGF by the allogeneic human MSCs was substantial. Therefore, we performed a small interfering RNA (siRNA) inhibition strategy to eliminate the secretion of KGF in the MSCs. We then instilled the cultured medium from the MSCs and found that approximately 80% of the benefit was lost when KGF release was inhibited. To be certain that KGF was the critical factor, we also added recombinant KGF (100 ng) as rescue therapy to the cultured medium that had no KGF present (because of the KGF siRNA) and found that the recombinant KGF restored alveolar fluid clearance to a normal level under these conditions (29
). Additional in vitro
studies suggested that KGF might work, in part, by upregulating the total quantity of epithelial sodium channel in alveolar type 2 cells and potentially increasing the transport of epithelial sodium channel to the apical cell membrane. It is well-known that KGF has cytoprotective effects, which may help to explain the beneficial effects on lung endothelial and epithelial permeability.
Although the results of these studies are interesting, there are important limitations. These experiments were performed in the perfused human lung, so the preparation is restricted to the lungs themselves. The duration of the experiments was short-term (4 hrs of injury). The absence of immunomodulatory organs such as the liver, spleen, and lymph nodes makes it difficult to know what the net effect might be in the intact human with ALI/ARDS. Based on animal studies, one could speculate that the anti-inflammatory properties might have been more substantial. However, although the perfusate contained white blood cells and platelets, the quantity of cells in the perfusate was below normal. Nevertheless, the results indicated that MSCs themselves and their paracrine products may be effective in reducing the severity of lung injury.