The wide therapeutic spectrum of fluoxetine (e.g., antidepressant, antipanic, antiphobic, antiobsessive, analgesic, antimigraine) requires long-term administration and adaptive changes. To test whether adaptation involves the serotonin (5-HT) transporters, we measured the effects of fluoxetine on the regional cerebral metabolic rate for glucose (rCMRglc) in control rats or in rats pretreated for 2 weeks with fluoxetine (8 mg/kg, i.p., daily, 2 days wash out); rCMRglc was measured in 56 brain regions, using the quantitative [14C]deoxyglucose technique, at 30 min after i.p. administration of fluoxetine 0.4, 4 or 40 mg/kg, i.p., to non-pretreated rats or fluoxetine 4 mg/kg to pretreated rats. In non-pretreated rats, fluoxetine reduced rCMRglc in a dose-dependent fashion in 4 (7%, mean decrease 11%), 28 (50%, mean decrease 23%) and 37 (66%, mean decrease 32%) brain regions. In chronic fluoxetine-pretreated rats, fluoxetine decreased rCMRglc to a substantially lesser degree (eight regions, 14%; mean decrease, 10%). Subcortical brain regions (i.e., hypothalamic paraventricular, locus coeruleus and basal ganglia nuclei) that mediate the physiological responses to stress were very sensitive to fluoxetine acutely and subsensitive after chronic treatment. As kinetic tolerance to fluoxetine does not occur during chronic administration, the diminished rCMRglc responsivity to fluoxetine reflects dynamic, adaptive tolerance of 5-HT transporters and, consequently, increased synaptic 5-HT concentrations; the findings suggest that fluoxetine may be therapeutic by increasing the 5-HT-negative modulation upon areas that drive the abnormally hyperactive responses to stress found in several neuropsychiatric conditions.