High concentrations of IL-8 and IL-1β were detected in BALF of DPB patients in the present study. Furthermore, the correlation between serum sL-selectin and BALF IL-8, and that between serum sE-selectin and the BALF IL-1β were significant in DPB patients. Cellular expression and shedding of E-selectin on endothelial cells are upregulated in vitro by various cytokines, particularly IL-113. L-selectin is shed from the neutrophils surface as it is activated. The present results showed that stimulation of neutrophils with FMLP or IL-8, but not IL-1 β, induced a dose-dependent decrease in L-selectin expression on neutrophils of healthy subjects (Fig 5). In this context, Schleiffenbaum and colleagues demonstrated that the levels of L-selectin were increased in the supernatant fluid after stimulation of neutrophils by FMLP and IL-8, but not IL-1β. These findings indicate that L-selectin shedding from neutrophils by IL-8 occurs in a dose-dependent manner. Thus, these in vitro results are in agreement with those obtained in vivo in DPB. Our findings suggest that the release of cytokines in the lung and microvascular neutrophil-endothelial events are probably involved in the pathogenesis of DPB.

Since Kudoh and colleagues reported the beneficial effects of long-term treatment of DPB patients with erythromycin, this therapeutic regimen has become the treatment of choice for DPB in Japan. Roxithromycin and clarithromycin, new 14-mem-bered macrolide antibiotics, also have similar beneficial therapeutic effects in DPB. Although the mechanisms of macrolide-induced improvement are not fully understood, accumulating evidence suggests that these drugs act by inhibiting intrapulmo-naiy influx of neutrophils and activated T lymphocytes and suppressing the secretion of mucus from secretory epithelial cells, rather than acting as bactericidal antibiotics. Interestingly, we noticed in the present study a reduction in serum levels of these adhesion molecules, together with reduced BALF levels of IL-1β and IL-8 following treatment. Previous in vitro studies have also shown an inhibitory effect of macrolides on IL-8 release from bronchial epithelial cells and macrophage-like cell line in response to stimuli. Together, these results suggest that the reduction in serum levels of soluble adhesion molecules after therapy may be secondary to the inhibition of release of proinflammatory cytokines upregulating these adhesion molecules. This is supported by our finding that incubation of neutrophils with erythromycin did not affect L-selectin shedding from neutrophils on stimulation with IL-8 (data not shown). However, our results do not rule out a possible direct inhibition of macrolides of the expression and shedding of adhesion molecules, since a recent report indicated that erythromycin blocks the release of sICAM-1 from bronchial epithelial cells induced by the endotoxin of Haemophilus influenzae.
In conclusion, we demonstrated the presence of elevated levels of soluble adhesion molecules in serum of patients with DPB. It is possible that serum levels of these molecules reflect disease activity and that their release is regulated by cytokines produced in the lungs.