Objective: It has been reported that both normal pituitary and pituitary tumours express PPAR-γ, a nuclear hormone receptor, the expression being more abundant in pituitary tumours, and that this is the basis for the reported antiproliferative effects of the thiazolidinedione, rosiglitazone, in animal models. However, the mechanisms for the responsivity to rosiglitazone have remained unclear. Design and measurements: To investigate this further, 'real-time' PCR was used to assess PPAR-γ mRNA expression, and Western blotting and immunohistochemistry to study its protein expression, in 46 human pituitary tumours and normal pituitary tissue. Cell proliferation of the GH3 pituitary cell line was assessed by [3H]-thymidine-incorporation after 48 h rosiglitazone and pioglitazone (10-4 m - 10-10 m) treatment alone, or rosiglitazone in combination with the PPAR-γ antagonist GW9662. Results: PPAR-γ mRNA and protein was found to be expressed in normal pituitary and was variably expressed in pituitary tumours, but were increased specifically in nonfunctioning pituitary adenomas. However, very little staining was observed with immunohistochemistry, with only occasional cell nuclei stained, and no difference was detectable between controls and tumours. Rosiglitazone at 10-4 m and 10-5 m concentrations inhibited cell proliferation (10-4 m 14.0% ± 1.5% and 10-5 m 67% ± 4% [mean ± SEM] vs Control 100% ± 3%, P < 0.0001) while lower concentrations showed no significant effect. Following withdrawal of rosiglitazone 10-5 m, the cells fully recovered at a further 48 h, while lower doses showed a 'rebound' of stimulation. Pioglitazone was of similar potency to rosiglitazone in inhibiting proliferation. The PPAR-γ antagonist did not show a significant reversal of the antiproliferative effect of rosiglitazone, and indeed suppressed proliferation on its own. Conclusions: Our data suggest that the antiproliferative action of rosiglitazone is probably not via PPAR-γ.