To determine whether prostaglandins (PGs) mediate the ACTH response to tumor necrosis factor-alpha (TNFalpha), indomethacin (Indo; 0.1 1.0 mg/kg, iv) was administered before TNFalpha (1 mug, iv) in freely moving, alert rats. While Indo alone did not affect plasma ACTH levels, it dose-dependently blocked the ACTH response to TNFalpha. The highest dose of Indo abolished the ACTH response to TNFalpha [peak plasma ACTH values (mean +/- SEM): buffer/buffer, 137 +/- 34 pg/ml; Indo/buffer, 115 +/- 31; buffer/TNFalpha, 469 +/- 77; Indo/TNFalpha, 120 +/- 27] without modifying the ACTH response to CRF 1 mug/kg, iv, demonstrating that pituitary responsiveness was unaffected. Since it has been reported that Indo elevates plasma corticosterone (B) levels, the effect of Indo could reflect rapid negative feedback by B, rather than the involvement of PGs. However, inhibition of ACTH secretion was shown to be dependent on the dose of Indo, whereas plasma B levels were elevated to the same degree, independent of the Indo dose. In addition, Indo failed to block the ACTH response to an unrelated ACTH stimulus, insulin-induced hypoglycemia (area under response curve: insulin alone, 31,131 +/- 2,794 pg/min.ml; Indo/insulin, 32,919 +/- 3,582 pg/min.ml). Finally, in adrenalectomized B-replaced rats, TNFalpha elevated ACTH to levels similar to those seen in sham animals, and Indo inhibited these ACTH responses to the same extent in both groups. Thus, Indo inhibited the ACTH response to TNFalpha by a mechanism independent of B feedback. These results indicate that acute systemic administration of TNFalpha stimulates ACTH secretion through a PG-dependent mechanism.