Treatment of prostate cancer (PCa), a leading cause of cancer among males, lacks successful strategies especially in advanced, hormone-refractory stages. Some clinical studies have shown an increase in neuroendocrine-like cells parallel to the tumor progression but their exact role is a matter of debate. The prostate is a well-known target for melatonin, which reduces PCa cells proliferation and induces neuroendocrine differentiation. To evaluate the mechanisms underlying the indole effects on neuroendocrine differentiation and its impact on PCa progression, we used a cell culture model (LNCaP) and a murine model (TRAMP). Persistent ERK1/2 activation was found in both, melatonin and androgen-deprived cells. Melatonin blocked nuclear translocation of androgen receptor (AR), thus confirming anti-androgenic actions of the indole. However, using a comparative genome microarray to check the differentially expressed genes in control, melatonin, or androgen-deprived cells, some differences were found, suggesting a more complex role of the indole. By comparing control cells with those treated with melatonin or depleted of androgen, a cluster of 26 differentially expressed genes (±2.5-fold) was found. Kallikreins (KLK)2 and KLK3 (PSA) were dramatically downregulated by both treatments whereas IGFBP3 and IGF1R were up- and downregulated, respectively, in both experimental groups, thus showing a role for IGF in both scenarios. Finally, melatonin prolonged the survival of TRAMP mice by 33% when given at the beginning or at advances stages of the tumor. Serum IGFBP3 was significantly elevated by the indole in early stages of the tumor, confirming in vivo the role of the IGF signaling in the oncostatic action of the indole.