We investigated the effects of rolipram, a selective cyclic adenosine 3',5'-monophosphate phosphodiesterase type IV inhibitor, and isobutylmethylxanthine, a nonselective phosphodiesterase inhibitor, on purposeless spontaneous chewing movements and tongue protrusions produced by 24 weeks treatment with haloperidol decanoate (25 mg/kg every 4 weeks i.m.) in rats, to examine our hypothesis that restoration of striatal cyclic adenosine 3',5'-monophosphate levels previously reduced due to dopamine D2 receptor supersensitivity, may suppress these movements. Tests were performed 8 weeks after the final injection. Haloperidol treatment significantly increased dyskinetic movements and striatal dopamine D2 receptor density compared with controls. Rolipram (0.1-1.0 mg/kg i.p.) suppressed these movements in a dose-dependent manner, whereas isobutylmethylxanthine (2 mg/kg i.p.) only slightly suppressed the syndrome and doses higher than 5 mg/kg i.p. produced other intensive movements. These results support our hypothesis and suggest that rolipram may have a therapeutic effect on tardive dyskinesia.