Both thalidomide and dexamethasone have been shown to inhibit the production of tumour necrosis factor α (TNF-α), but little is known of their cellular selectivity. Inhibition of monocyte TNF-α expression has been implicated in the clinical efficacy of thalidomide, and it has been suggested that the drug modulates only monocyte-derived cytokines. Given the importance of T lymphocyte responses in immunological disorders in which treatment with thalidomide has been successful, it is pertinent to study the effects of this drug on T cell-derived TNF-α. In the present investigations we have examined the influence of both thalidomide and dexamethasone on mitogen-induced elaboration of TNF-α by CD3 + peripheral blood mononuclear cells (PBMC) and the T cell line MOLT-4. PBMC from healthy human volunteers were stimulated optimally with phytohaemagglutinin (PHA) in the presence of varying concentrations of thalidomide or dexamethasone, and supernatants assayed for TNF-α and interleukin 2 (IL-2). Concurrently, PHA-stimulated PBMC were treated with 1×10 −1 mM thalidomide or dexamethasone and the cells fixed, permeabilised, stained with anti-CD3 and anti-TNF-α fluorescently labelled antibodies and analysed by flow cytometry. MOLT-4 cells were cultured in the presence or absence of the drugs following activation with phorbol myristate acetate (PMA)/ionophore, and supernatants analysed by enzyme-linked immunosorbent assay (ELISA) for cytokine expression. Thalidomide was found to inhibit PBMC-derived TNF-α, but not IL-2. In contrast, dexamethasone down-regulated both TNF-α and IL-2 in a dose-dependent manner. Thalidomide and dexamethasone both suppressed intracellular levels of TNF-α in CD3 + PBMC, reducing percentages of double positive staining cells by 28 and 52%,respectively, compared with controls. In addition, TNF-α production by CD3 − PBMC was inhibited by 31% by thalidomide and by 47% by dexamethasone. In order to determine whether thalidomide was acting directly on T cells, or indirectly through effects on accessory cells, TNF-α production in the T cell line MOLT-4 was investigated. TNF-α secretion by PMA/ionophore activated MOLT-4 cells was reduced by 80% following thalidomide treatment and close to background levels following dexamethasone treatment. To verify that thalidomide was acting selectively to down-regulate TNF-α, IL-2 production by MOLT-4 cells was also measured and found to be unaffected by the drug. In contrast, dexamethasone reduced MOLT-4-derived IL-2 levels by 20%. These observations suggest that thalidomide, in addition to its known inhibitory effect on monocyte-derived TNF-α, is capable also of down-regulating T cell-derived TNF-α in a direct and selective manner. In addition, the inhibition of intracellular levels of TNF-α strengthens the evidence that the inhibitory effect of thalidomide is at the level of transcription and/or translation and does not reduce cellular TNF-α secretion. Such effects could explain the efficacy of thalidomide treatment in various immunological disorders where T cell activation plays an important role in the pathogenesis of the disease.