Ed with melatonin. Although the precise mechanism by which GSK-126 melatonin induces ROS in cancer cells remains unknown, our results together with data from other authors suggest that ROS produced by the mitochondrial electron transport chain constitutes a key factor in melatonin-induced cell death and differentiation. Excessive ROS contributes to mitochondrial outer membrane permeabilization, which is mainly controlled by proteins from the BCL-2 family and is an important factor in mediating the intrinsic apoptosis. Previous studies have noted that melatonin 
alters the balance between BAX and BCL-2 in some cancer cells by up-regulating BAX expression, resulting in MOMP and cytochrome c release. However, in other cancer cells, melatonin induces a decrease in BCL-2. Here, BAX remained unaltered whereas BCL-2 was down-regulated in the cells grown in galactose media and treated with melatonin, altering the BAX/BCL-2 balance but without causing cytochrome c release, probably due to increased mitochondrial membrane PF-562271 site potential. Accordingly, we have not detected an increase in caspase-3-like activity. Nonetheless, the Live/Dead assay suggests that the observed decrease in cell proliferation is due to a cytotoxic rather than a cytostatic action of melatonin. Melatonin increased cytosolic AIF levels in GalCSCs and Gal-dCCs, as well as in Glu-CSCs. However, the observed band corresponds to a 67-kDa form of AIF, which is the precursor form containing a mitochondriallocalizing sequence, and which is unable to cause cell death. On the contrary, melatonin seems to exert another undescribed mitochondrial effect by inducing de novo synthesis of the AIF precursor protein. After being imported into mitochondria, the mitochondrial localizing sequence contained in 67-kDa AIF is cleaved, resulting in the accumulation of the mature 57-kDa form of the AIF protein. This is described to translocate to the nucleus where it triggers a caspase-3-independent type of cell death. We found a higher cytosolic localization of this ~57 kDa form of AIF in cells cultured in galactose 
media and treated with melatonin, alone or in combination with dichloroacetate, and in Glu-dCCs treated with melatonin and dichloroacetate. Nonetheless, the increased percentage of dead cells after 72 hours of treatment with melatonin, PubMed ID:http://www.ncbi.nlm.nih.gov/pubmed/19858355 although moderated, was statistically significant in cells with a more oxidative metabolism. In these cells, cultured in the galactose medium, more than 50% www.impactjournals.com/oncotarget 17090 of the cellular mass was lost after 72 hours of treatment with melatonin. Due to this, the measurement of the percentage of live/dead cells in the remaining population probably represents early events of cell death as well as processes of selection of the more resistant subpopulations. In fact, we cannot exclude divergent effects of melatonin in the different cell subpopulations. In accordance to this hypothesis, we observed a different pattern of AIF expression in some cells which may indicate the presence of different cell subpopulations with dissimilar susceptibility to activate this pathway. Taking all our results into account, we can infer that melatonin induces a toxic effect in P19 embryonal carcinoma cells via the inhibition of mitochondrial metabolism as described in other types of tumor cells. Thus, P19 Glu-CSCs present a strong resistant phenotype, which seems to be linked to their glycolytic metabolism. In fact, we previously observed that only the P19 cells with a.Ed with melatonin. Although the precise mechanism by which melatonin induces ROS in cancer cells remains unknown, our results together with data from other authors suggest that ROS produced by the mitochondrial electron transport chain constitutes a key factor in melatonin-induced cell death and differentiation. Excessive ROS contributes to mitochondrial outer membrane permeabilization, which is mainly controlled by proteins from the BCL-2 family and is an important factor in mediating the intrinsic apoptosis. Previous studies have noted that melatonin alters the balance between BAX and BCL-2 in some cancer cells by up-regulating BAX expression, resulting in MOMP and cytochrome c release. However, in other cancer cells, melatonin induces a decrease in BCL-2. Here, BAX remained unaltered whereas BCL-2 was down-regulated in the cells grown in galactose media and treated with melatonin, altering the BAX/BCL-2 balance but without causing cytochrome c release, probably due to increased mitochondrial membrane potential. Accordingly, we have not detected an increase in caspase-3-like activity. Nonetheless, the Live/Dead assay suggests that the observed decrease in cell proliferation is due to a cytotoxic rather than a cytostatic action of melatonin. Melatonin increased cytosolic AIF levels in GalCSCs and Gal-dCCs, as well as in Glu-CSCs. However, the observed band corresponds to a 67-kDa form of AIF, which is the precursor form containing a mitochondriallocalizing sequence, and which is unable to cause cell death. On the contrary, melatonin seems to exert another undescribed mitochondrial effect by inducing de novo synthesis of the AIF precursor protein. After being imported into mitochondria, the mitochondrial localizing sequence contained in 67-kDa AIF is cleaved, resulting in the accumulation of the mature 57-kDa form of the AIF protein. This is described to translocate to the nucleus where it triggers a caspase-3-independent type of cell death. We found a higher cytosolic localization of this ~57 kDa form of AIF in cells cultured in galactose media and treated with melatonin, alone or in combination with dichloroacetate, and in Glu-dCCs treated with melatonin and dichloroacetate. Nonetheless, the increased percentage of dead cells after 72 hours of treatment with melatonin, PubMed ID:http://www.ncbi.nlm.nih.gov/pubmed/19858355 although moderated, was statistically significant in cells with a more oxidative metabolism. In these cells, cultured in the galactose medium, more than 50% www.impactjournals.com/oncotarget 17090 of the cellular mass was lost after 72 hours of treatment with melatonin. Due to this, the measurement of the percentage of live/dead cells in the remaining population probably represents early events of cell death as well as processes of selection of the more resistant subpopulations. In fact, we cannot exclude divergent effects of melatonin in the different cell subpopulations. In accordance to this hypothesis, we observed a different pattern of AIF expression in some cells which may indicate the presence of different cell subpopulations with dissimilar susceptibility to activate this pathway. Taking all our results into account, we can infer that melatonin induces a toxic effect in P19 embryonal carcinoma cells via the inhibition of mitochondrial metabolism as described in other types of tumor cells. Thus, P19 Glu-CSCs present a strong resistant phenotype, which seems to be linked to their glycolytic metabolism. In fact, we previously observed that only the P19 cells with a.