GM-CSF differentially regulates eosinophil and neutrophil adhesive interactions with vascular endothelium in vivo

Abstract

Allergic airway inflammation is characterized by elaboration of cytokines and chemokines leading to recruitment of inflammatory leukocytes, predominantly eosinophils, to the airways. Granulocyte macrophage colony stimulating factor (GM-CSF) is generated in the lungs of human subjects with asthma in response to allergen challenge and is necessary for the development of allergen-induced bronchial eosinophilia in mice. The effect of GM-CSF on human eosinophil and neutrophil interactions with the vascular endothelium under conditions of blood flow was investigated in post-capillary venules of the rabbit mesentery by intravital microscopy.While GM-CSF significantly reduced the rolling fraction of neutrophils in vivo and induced consistent shedding of neutrophil L-selectin in vitro, its effect on eosinophil rolling was variable. Eosinophils from 57% of the donors demonstrated inhibition of rolling, while eosinophils from the remaining 43% of donors demonstrated no inhibition or increased rolling. The variable effect of GM-CSF on inhibition of eosinophil rolling was associated with variable shedding of L-selectin in vitro. In contrast to the differential effect of GM-CSF on neutrophils versus eosinophils, stimulation with phorbol myristate acetate demonstrated a similar degree of inhibition of rolling and L-selectin shedding by neutrophils and eosinophils suggesting that there was no defect in L-selectin shedding in the eosinophil donors who did not respond to GM-CSF. Overall, these studies demonstrate that GM-CSF consistently inhibits interaction of neutrophils with endothelium in vivo, whereas its effect on eosinophil-endothelial interactions is variable. GM-CSF may thus be one factor accounting for the varying percentage of eosinophils and neutrophils recruited to sites of allergic inflammation in different individuals.

Figure 1. GM-CSF inhibits neutrophil rolling to a greater level than eosinophil rolling in inflamed post capillary venules in vivo

The effect of GM-CSF (0-20 ng/ml) on the flux of rolling CFDA-labeled neutrophils and eosinophils was compared to PBS-treated cells (n = neutrophils from 8 donors and eosinophils from 14 donors). The values represent mean (% control) ± SD obtained from 3-5 venules/animal from n = 8-14 rabbits.

Figure 2. GM-CSF treatment results in variable inhibition of eosinophil but not neutrophil rolling on inflamed venular endothelium in vivo

The effect of GM-CSF (20 ng/ml) on rolling of CFDA labeled neutrophils (n=8) and eosinophils (n=14) was compared to PBS treated cells (control). The values represent % inhibition of rolling compared to control cells (broken line). The negative y-axis represents an increase in frequency of rolling eosinophils compared to control cells. Values represent data obtained from 3-5 venules/animal.

Figure 3. Activation with GM-CSF results in altered adhesion receptor expression on neutrophils and eosinophils

Neutrophils and eosinophils were stimulated with increasing concentrations of PBS or GM-CSF (0-100 ng/ml) for 30 min at 37°C and the expression of L-selectin and CD18 was determined by flow cytometry using FITC-labeled mAbs DREG-200 and IB4, respectively. MAb 81C6 was used as an irrelevant, isotype matched non-binding control antibody.

Figure 4. Concentration dependent effects of PMA on surface receptor expression and rolling of eosinophils

CFDA-labeled eosinophils were exposed to increasing concentrations of PMA (0 to 10^-6 M) at 37°C for 30 min before evaluating rolling in vivo in the inflamed mesenteric circulation of rabbits (n=3, 7 venules/rabbit) as described in detail in Materials and Methods (A). Eosinophils from the same donors were exposed to PMA as described above and evaluated for the expression of L-selectin and CD18 by flow cytometry (B). Next, eosinophils were treated with PMA at a concentration of 10^-7 M for different time intervals at 37°C and the expression of L-selectin and CD18 by these cells was determined by flow cytometry (C).

Figure 5. Comparison of the effects of PMA on the rolling flux of eosinophils and neutrophils in mesenteric venules

CFDA-labeled eosinophils and neutrophils were stimulated ex vivo with 10^-7 M PMA for 30 min at 37°C before evaluating rolling in vivo in the inflamed mesenteric circulation of rabbits (n=3, 7 venules/rabbit) as described in detail in Materials and Methods.

Figure 6. Effect of GM-CSF and PMA activation on the velocity distribution profiles of eosinophils

CFDA-labeled eosinophils were activated ex vivo with GM-CSF (10ng/ml), PMA (10^-7M), or PBS (control) for 30 min at 37°C prior to administration into the superior mesentery artery of rabbits. Passage of the flowing eosinophils was recorded and velocity of 80-100 consecutive cells was determined by frame-by-frame analysis in each of 5 representative venules in 2 rabbits. Relative velocity of individual cells was determined as described in detail in Materials and Methods and cells were assigned separate sections of normalized velocity, i.e., 0 - < 0.1, 0.1 - < 0.2, and so on up to 0.9 - < 1.0.