Giant cell arteritis (GCA) is a vasculitic syndrome that preferentially affects medium and large-sized arteries. Glucocorticoid therapy resolves clinical symptoms within hours to days, but therapy has to be continued over several years to prevent disease relapses. It is not known whether and how glucocorticoids affect the function of the inflammatory infiltrate or why the disease persists subclinically despite chronic treatment. GCA is self-sustained in temporal arteries engrafted into SCID mice, providing a model in which the mechanisms of action and limitations of glucocorticoid therapy can be examined in vivo. Administration of dexamethasone to temporal artery-SCID chimeras for 1 wk induced a partial suppression of T cell and macrophage function as indicated by the reduced tissue concentrations of IL-2, IL-1beta, and IL-6 mRNA, and by the diminished expression of inducible NO synthase. In contrast, synthesis of IFN-gamma mRNA was only slightly decreased, and expression of TGF-beta1 was unaffected. These findings correlated with activation of the IkappaBalpha gene and blockade of the nuclear translocation of NFkappaB in the xenotransplanted tissue. Dose-response experiments suggested that steroid doses currently used in clinical medicine are suboptimal in repressing NFkappaB-mediated cytokine production in the inflammatory lesions. Chronic steroid therapy was able to deplete the T cell products IL-2 and IFN-gamma, whereas the activation of tissue-infiltrating macrophages was only partially affected. IL-1beta transcription was abrogated; in contrast, TGF-beta1 mRNA synthesis was steroid resistant. The persistence of TGF-beta1-transcribing macrophages, despite paralysis of T cell function, may provide an explanation for the chronicity of the disease, and may identify a novel therapeutic target in this inflammatory vasculopathy.