Granulocyte colony-stimulating factor (G-CSF) expression is often induced during infection, resulting in high concentrations of G-CSF in inflammatory exudates and in the blood. This increased expression suggests that G-CSF may regulate both local and systemic neutrophil responses. Previous studies of G-CSF or G-CSF receptor deficient (G-CSFR−/−) mice challenged with different infectious agents have provided conflicting results concerning the importance of G-CSF in the regulation of stress granulopoiesis. In the present study, we use a physiologically-relevant infectious model in which Pseudomonas aeruginosa coupled to agarose beads is injected intratracheally into wild type or G-CSFR−/− mice inbred onto a C57BL/6 background. This model produces a smoldering P. aeruginosa infection in the lungs, simulating the infection observed in many patients with cystic fibrosis. Importantly, the concentration of G-CSF in the blood and bronchoalveolar lavage fluid (BALF) is significantly elevated in this model. We show that G-CSFR−/− mice display decreased survival in response to P. aeruginosa infection. Extensive pulmonary necrosis was present in G-CSFR−/− mice, and quantitative bacteriology showed decreased clearance of P. aerguinosa from the lungs. Despite this evidence for severe infection, histological studies showed that neutrophil infiltration into the lungs of G-CSFR−/− mice was markedly decreased. To characterize this defect further, we examined the systemic neutrophil response. In wild type mice, neutrophil number in the blood increased from 0.8 ± 0.1 x 10−6/ml at baseline to 2.5 ± 0.3 x 10−6/ml at 48 hours post-infection (all data represent the mean ± SEM). In contrast, G-CSFR−/− mice were neutropenic at baseline (0.1 ± 0.02 x 10−6/ml) and little increase in blood neutrophils was noted at 48 hours post-infection (0.5 ± 0.1 x 10−6/ml). In both groups of mice, a modest decrease in bone marrow neutrophils was observed during infection. These data suggest that the systemic neutrophil response in this model is dependent upon G-CSF signals and is primarily mediated by increased neutrophil release from the bone marrow rather than increased neutrophil production. We next examined the local neutrophil response. Despite normal (mip-2) or increased (KC) expression of the major chemokines regulating neutrophil migration in mice, the number of neutrophils present in the BALF of G-CSFR−/− mice following infection was markedly reduced compared to wild type [number of neutrophils per ml of BALF at 48 hours post-infection ± SEM: 3.5 ± 0.7 x 106 (wild type); 0.7 ± 0.4 x 106 (G-CSFR−/−); p<.05]. Since neutrophil number in a tissue is dependent upon both their emigration and subsequent clearance, we measured the percentage of apoptotic neutrophils in the BALF. Interestingly, the percentage of apoptotic (Annexin V-positive) neutrophils was significantly increased in G-CSFR−/− mice, suggesting that G-CSF signals may play an important role in regulating neutrophil survival at the inflammatory site. Collectively these data provide new evidence that G-CSF signals play important but specific roles in the regulation of the systemic and local neutrophil response following infection.

Author notes

Corresponding author