To improve understanding of the mechanisms of action of oral corticosteroids in asthma, we have conducted a double-blind, placebo-controlled study with prednisolone (20 mg for 2 wk followed by 10 mg for 4 wk) or placebo in 14 and 13 atopic corticosteroid-naive asthmatic subjects, respectively. Before and after treatment subjects underwent bronchoscopy with bronchoalveolar lavage (BAL) and bronchial biopsy. Treatment with prednisolone, but not placebo, significantly reduced asthma symptoms (from mean +/- SEM total weekly score of 34 +/- 6.2 to 15.7 +/- 3.2, p = 0.02) and albuterol usage (from mean +/- SEM number of puffs/wk of 29.7 +/- 6.2 to 18.2 +/- 3.7, p = 0.01) and significantly increased FEV1 (from 89.8 +/- 4.4% to 99.3 +/- 4.1% of predicted, p = 0.03). There were no significant changes in inflammatory or epithelial cell counts, levels of T-cell activation or albumin concentration in BAL. However, immunohistochemistry of bronchial biopsies showed that in the submucosa prednisolone significantly decreased numbers of mast cells by 62% (from median 45 to 17/mm2, p = 0.01), eosinophils by 81% (from median 30.1 to 5.7/mm2, p = 0.004), and CD4+ T-cells by 68% (from median 64.6 to 18.5/mm2, p = 0.02). In the epithelium only the reduction in the numbers of eosinophils was significant (from median 1.1 to 0/mm of epithelium, p = 0.02). There were no significant changes in any cell counts in the subjects receiving placebo, and comparison of the changes between the treatment groups identified a significant prednisolone-related reduction in submucosal eosinophil and mast cell counts (p = 0.003 and 0.03, respectively). The temporal association between the clinical and physiologic improvement, and the correlation between the magnitude of change in CD4+ T-cell counts in the submucosa and increase in PC20 methacholine (rs = 0.60, p = 0.049) suggests that the reduction in airways inflammatory cell numbers underlies the clinical efficacy of oral corticosteroids.