Gene microarrays may enable the elucidation of neurobiological changes underlying the pathophysiology and treatment of major depression. However, previous studies of antidepressant treatments were performed in healthy normal rather than 'depressed' animals. Since antidepressants are devoid of mood-changing effects in normal individuals, the clinically relevant rodent transcriptional changes could remain undetected. We investigated antidepressant-related transcriptome changes in a corticolimbic network of mood regulation in the context of the unpredictable chronic mild stress (UCMS), a naturalistic model of depression based on socio-environmental stressors. Mice subjected to a 7-week UCMS displayed a progressive coat state deterioration, reduced weight gain, and increased agonistic and emotion-related behaviors. Chronic administration of an effective (fluoxetine) or putative antidepressant (corticotropin-releasing factor-1 (CRF1) antagonist, SSR125543) reversed all physical and behavioral effects. Changes in gene expression differed among cingulate cortex (CC), amygdala (AMY) and dentate gyrus (DG) and were extensively reversed by both drugs in CC and AMY, and to a lesser extent in DG. Fluoxetine and SSR125543 also induced additional and very similar molecular profiles in UCMS-treated mice, but the effects of the same drug differed considerably between control and UCMS states. These studies established on a large-scale that the molecular impacts of antidepressants are region-specific and state-dependent, revealed common transcriptional changes downstream from different antidepressant treatments and supported CRF1 targeting as an effective therapeutic strategy. Correlations between UCMS, drug treatments, and gene expression suggest distinct AMY neuronal and oligodendrocyte molecular phenotypes as candidate systems for mood regulation and therapeutic interventions.