The influence of two selective serotonin reuptake inhibitors (SSRIs), litoxetine and fluoxetine, has been studied on 5-HT4 receptor-mediated relaxation in the rat isolated oesophageal muscularis mucosae. In carbachol-precontracted oesophageal tissues, 5-hydroxytryptamine (5-HT) (0.1 nM-1 microM) induced concentration-dependent relaxations. Concentration-response curves were monophasic and reproducible. Litoxetine at concentrations without antimuscarinic properties (10 nM-1 microM) caused concentration-dependent relaxations, which reduced carbachol tone up to 37%. Higher litoxetine concentrations (3 microM-300 microM) were associated with marked relaxation up to the abolition of carbachol tone. The overall curve profile of litoxetine was biphasic in nature with a high (10 nM-1 microM) and a low (3 microM-300 microM) potency phase. Unlike 5-HT, the second curve of litoxetine was not reproducible, with a reduction involving mainly the low potency phase. Compared to litoxetine, fluoxetine caused minimal relaxation (less than 10% at 1 microM). Treatment of rats with parachlorophenylalanine (pCPA: 375 mg kg-1 per day, for two days), to deplete endogenous 5-HT stores, did not modify the relaxant effect of 5-HT, while it significantly reduced the high potency phase of litoxetine. In tissues from untreated rats, this phase was reduced by the 5-HT4 receptor antagonist GR 125487 (10 nM) to an extent similar (P = 0.20: ANOVA for continuous-by-class effects) to that induced by pCPA treatment. However, in tissues from pCPA treated animals GR 125487 (10 nM) exerted a slight but significant antagonism of litoxetine response (P = 0.037: ANOVA for continuous-by-class effects) mainly involving the high potency phase. In tissues from untreated rats, litoxetine (1 microM) increased the relaxant effects of 5-HT, while in tissues from pCPA treated animals it exerted a small but significant depression of the maximal response to 5-HT, without changing its potency value. Fluoxetine (1 microM) slightly, but significantly, antagonized the relaxant effect of 5-HT in an unsurmountable manner. In conclusion, litoxetine up to 1 microM relaxed the rat isolated oesophageal muscularis mucosae through a mechanism involving release of endogenous 5-HT, which in turn activates 5-HT4 receptors. However, based on results with GR 125487 in tissues from pCPA treated rats, a small component of litoxetine-induced relaxation may involve a direct activation of 5-HT4 receptors. It is unlikely that blockade of 5-HT reuptake can participate in the action of litoxetine, since fluoxetine, a 5-HT reuptake inhibitor equipotent to litoxetine, was ineffective in the same range of concentrations. The antimuscarinic activity of litoxetine, previously demonstrated in the isolated guinea-pig intestine, played a role at concentrations greater than 1 microM. The 5-HT-releasing action of litoxetine could account for the potentation by litoxetine of 5-HT-induced relaxation in tissues from untreated rats, which was reversed by pCPA treatment. Under these conditions, litoxetine depressed relaxations to high 5-HT concentrations only. In tissues from untreated rats, fluoxetine slightly but unsurmountably antagonized 5-HT-induced relaxations, thus confirming previous observations in the guinea-pig small intestine.