Anti-inflammatory activity of salmeterol: down-regulation of cytokine production.

Abstract

Elevation of intracellular cAMP levels has been shown previously to inhibit cytokine secretion by various cell types in vitro. Since salmeterol is a beta 2-agonist which activates adenylate cyclase, its ability to inhibit cytokine production was evaluated. Though salmeterol, and the related drug albuterol, did not inhibit IL-1 beta production in vitro, both drugs did inhibit tumour necrosis factor-alpha (TNF-alpha) secretion by lipopolysaccharide (LPS)-activated THP-1 cells with similar IC50s of approximately 0.1 microM. This inhibition was effectively reversed by the beta 2-antagonist oxprenolol, indicating that the inhibition was mediated through the beta 2-adrenergic receptor. A strikingly different reactivity profile was seen with T cells. Salmeterol was able to inhibit the activation of both mouse and human T cells, as measured by proliferation and IL-2 secretion in response to anti-CD3 antibody, whereas albuterol was completely inactive in these assays. This T cell inhibition by salmeterol was about 10-fold less potent than that for TNF-alpha production, and was not reversed by a beta 2-antagonist, indicating that a different mechanism was involved in the effect of salmeterol on T cells. Paralleling the TNF-alpha inhibitory activity in vitro, oral dosing of salmeterol and albuterol inhibited LPS-induced increase in murine serum TNF level in vivo, with ED50s of approximately 0.1 mg/kg. This inhibition could be abrogated by dosing orally with the beta-blocker propranolol. The long-acting pharmacological profile of salmeterol was apparent in that it maintained its efficacy for 3 h, while albuterol had a much shorter duration of action. Salmeterol also had some protective effects in the galactosamine/LPS model of endotoxic shock, which is dependent upon TNF-alpha production. Though salmeterol inhibited serum TNF-alpha levels by up to 94% in this assay, it protected less than 50% of the animals from the lethal effects of the LPS/galactosamine mixture. This observation suggests that functional levels of TNF-alpha localized in tissues may not be accurately reflected by serum levels.