The antiemetic activity of sendide, a new peptide tachykinin NK1 receptor antagonist, against cisplatin-induced emesis was investigated using ferrets. The frequency of cisplatin (10 mg/kg, i.p.)-induced retching (104.6 +/- 14.3/6 h) and vomiting (19.0 +/- 3.0/6 h) was significantly reduced by pretreatment with sendide (3.0 mg/kg, s.c.) (14.0 +/- 8.1/6 h and 1.8 +/- 1.2/6 h, respectively). Intravenous bolus injection of substance P (1-10 microg/kg) or 5-hydroxytryptamine (5-HT) (10-50 microg/kg) produced a dose-dependent increase in the abdominal afferent vagus nerve activity. The change from pre-injection level in the afferent nerve activity induced by substance P (1 microg/kg, i.v.) (453.7 +/- 51.5%) was significantly reduced by pretreatment with either sendide (100 microg/kg, i.v.) (276.1 +/- 50.1%, P < 0.05) or granisetron, a 5-HT3 receptor antagonist (1 mg/kg, i.v.) (146.3 +/- 14.0%, P < 0.01). The amount of 5-HT released into the solution during a 1-h exposure to 2-methyl-5-HT (10(-6) M), a 5-HT3 receptor agonist, was significantly increased (317.9 +/- 46.7%, P < 0.05) compared with that of the control tissues (160.4 +/- 8.1%). The 2-methyl-5-HT-induced 5-HT release was significantly inhibited by administration of sendide (10(-6) M) (174.0 +/- 21.6%, P < 0.05) or granisetron (10(-6) M) (186.6 +/- 27.3%, P < 0.05). Since sendide does not penetrate the central nervous system, these results suggest that the antiemetic effects of sendide are due to the inhibition of NK1 and 5-HT3 receptors on the emetic peripheral detector sites.