Mechanism of neutrophil recruitment to the lung after pulmonary contusion

Abstract

Blunt chest trauma resulting in pulmonary contusion is a common but poorly understood injury. We previously demonstrated that lung contusion activates localized and systemic innate immune mechanisms and recruits neutrophils to the injured lung. We hypothesized that the innate immune and inflammatory activation of neutrophils may figure prominently in the response to lung injury. To investigate this, we used a model of pulmonary contusion in the mouse that is similar to that observed clinically in humans and evaluated postinjury lung function and pulmonary neutrophil recruitment. Comparisons were made between injured mice with and without neutrophil depletion. We further examined the role of chemokines and adhesion receptors in neutrophil recruitment to the injured lung. We found that lung injury and resultant physiological dysfunction after contusion were dependent on the presence of neutrophils in the alveolar space. We show that CXCL1, CXCL2/3, and CXCR2 are involved in neutrophil recruitment to the lung after injury and that intercellular adhesion molecule 1 is locally expressed and actively participates in this process. Injured gp91-deficient mice showed improved lung function, indicating that oxidant production by neutrophil NADPH oxidase mediates lung dysfunction after contusion. These data suggest that both neutrophil presence and function are required for lung injury after lung contusion.

Fig. 1. Neutrophil depletion improves post-injury pulmonary function

Mice were pre-treated with anti-Ly6G (α-Ly6G) or IgG at 24H prior to injury. Samples from uninjured (Control, open bars) and injured animals (filled bars) were isolated after injury. (A) Serum was assayed by ELISA for CXCL1 as described in methods. There is a significant increase in CXCL1 in injured mice (*p<0.001). CXCL1 levels in α-Ly6G pretreated, neutropenic injured mice were not significantly different from untreated injured mice. N=7,7,7,5 for experimental groups, respectively, as shown. (B) The BAL cell pellet was counted and differentiated as described in methods. There is a significant increase in PMN in the BAL in injured mice (*p<0.001). α-Ly6G pretreated, neutropenic injured mice showed a significant decrease in PMN (^#p<0.001). There is no significant difference in PMN between injured and IgG treated mice. n=6,6,6,3 for experimental groups, respectively, as shown. (C) Blood gas analysis showed significantly decreased PaO₂:FiO₂ in injured animals (*p≤0.001). α-Ly6G pretreated, neutropenic injured mice showed significantly better PaO₂:FiO₂ (^#p<0.001). There is no significant difference in PaO₂:FiO₂ between injured and IgG treated mice. n=5 for all experimental groups. (D) Serum was assayed by ELISA for IL-6 as described in methods. There is a significant increase in IL-6 in injured mice (*p<0.001). IL-6 levels in α-Ly6G pretreated, neutropenic injured mice were significantly decreased compared with untreated injured mice (^#p<0.001). IgG pretreatment did not show a significant difference in IL-6 levels compared with untreated injured mice. N=7,7,7,5 for experimental groups, respectively, as shown.

Fig. 2. CXCR-2 dependent neutrophil recruitment to the injured lung after lung contusion

Mice were pre-treated with anti-leukinate (α-leukinate) or vehicle (PBS) at 30min. prior to injury. BAL and arterial blood from uninjured (Control, open bars) and injured animals (filled bars) were isolated at 24H after injury. (A) The cell pellet was counted and differentiated as described in methods. There is a significant increase in PMN in the BAL in injured mice (*p<0.001). α-Leukinate treated injured mice showed a significant decrease in PMN (^#p<0.001). There is no significant difference in PMN between injured and PBS treated mice. n=6,6,6,3 for experimental groups, respectively, as shown. (B) Blood gas analysis showed decreased PaO₂:FiO₂ in injured animals (*p≤0.001). α-Leukinate treated injured mice showed significantly better PaO₂:FiO₂ (^#p<0.05). There is no significant difference in PaO₂:FiO₂ between injured and IgG treated mice. n=5,5,5,3 for all experimental groups, respectively, as shown.

Fig. 3. CXC chemokine dependent neutrophil recruitment to the injured lung after lung contusion

Mice were pre-treated with α-CXCL1, α-CXCL2/3 or both (α-CXCL1±2/3) at 30 min. prior to injury. BAL from injured animals was isolated at 24H after injury. The cell pellet was counted and differentiated as described in methods. There was no significant difference in PMN the BAL in injured mice or injured mice treated with either antibody alone. In contrast, injured mice treated with both antibodies showed a significant decrease in PMN (^#p<0.05) when compared with untreated injured mice. n=6 for all experimental groups.

Fig. 4. ICAM-1 dependent neutrophil recruitment to the injured lung after lung contusion

(A) Immunostaining for ICAM-1 in lungs from uninjured (left panel), and injured (right panel) animals at 3H after injury (40X magnification). Specimens are representative of immunohistochemistry for least 3 animals. (B) Mice were treated with α-ICAM-1 antibody either intraperitoneally (IP), intratracheally (IT), or both routes (IP and IT) at 30min. prior to injury. BAL was counted and differentiated as described in methods. There is a significant decrease in PMN in the BAL between injured mice with and without treatment (*p<0.001). There is no significant difference in decreased PMN level with different routes of administration. n=6 for experimental groups.

Fig. 5. gp91phox dependent lung dysfunction after injury

(A) BALs from injured WT (closed) and Injured gp91phox deficient (gp91^phox-, crosshatched) mice were analyzed for neutrophil infiltration (PMN) at 24H. Injured gp91^phox- mice show a significant increase (*p≤.01) in PMN compared with injured WT mice. (B) CXCL1 was measured in the serum from Control and Injured mice at 3H after injury. There is a significant increase (*p<0.001) in CXCL1 in the serum of Injured gp91phox deficient (gp91^phox-, crosshatched) compared with Injured WT mice. Control uninjured serum CXCL1 levels are shown for comparison. (C) Blood gas analysis showed decreased PaO₂:FiO₂ in Injured WT mice (*p≤0.01). Injured gp91phox deficient (gp91^phox-, crosshatched) showed no significant difference in PaO₂:FiO₂ from uninjured Control mice. n=5-6 mice for all experiments.