A complex and intricate network of signaling pathways determines whether a cell will either proliferate, differentiate, survive or die. Retinoids, due to their strong differentiative potential, have been widely used for both cancer therapy and cancer prevention . There are many examples in the literature of distinct cell types whose differentiation is under the control of retinoids: embryonal carcinoma cells, promyelocytic leukemia cells, neuroblastoma cells, normal erythroid progenitors, etc. [3, 44–48]. In addition to differentiation induction, retinoids are able to initiate several other programs that may contribute to its therapeutical potential. Indeed, it has been shown that retinoids induce apoptosis of APL cells and blasts of APL patients through selective paracrine action of the death ligand TRAIL . In breast cancer cells, we provide evidence that retinoic acid induces cell growth inhibition and depending on cell-context, promotes a sort of differentiation without affecting viability or makes the cells enter a fully apoptotic program. The finding that 9-cis-RA causes differentiation of T47D cells is in agreement with the previously reported accumulation of lipid droplets in cytoplasmic vesicles  and milk protein casein  in normal mammary epithelial cells, and in the breast cancer cell lines MCF7 and AU565  treated with retinoids. However, further studies are needed to determine whether the differentiation characterized by accumulation of cellular lipid depots contributes to the antiproliferative effects of retinoic acid in breast cancer cells.
A circuitry of several apoptotic programs is induced in breast cancer cells by retinoic acid. We have previously provided evidence that retinoids promote the induction of TRAIL not only in hematopoietic but also in breast cancer cells . In the current study, we have shown that induction of TRAIL and FAS by retinoic acid in the breast cancer cell line H3396 correlates with an increase in the number of apoptotic cells. In accordance with studies that report that TRAIL and FAS signal through caspase-8 activation, the activity of this enzyme is induced in H3396 cells treated with 9-cis-RA or with exogenous TRAIL. Although additional studies will be required to clarify the possible involvement of the extrinsic death pathway in retinoic-induced apoptosis in H3396 cells, activation of downstream caspases like caspase-9, as well as the release of cytochrome c and SMAC/DIABLO from the mitochondria to the cytosol and the loss of the mitochondrial membrane potential prove that the intrinsic pathway is dominantly involved in retinoic acid-induced apoptosis.
Paradoxically in certain breast cancer cells, retinoic acid induces concomitantly to TRAIL upregulation, the activation of a gene program of apparently opposite functionality, characterized by the induction of the antiapoptotic IAP family member, cIAP2, a NF-κB target gene. cIAP2 expression was significantly modulated at the mRNA and protein levels by retinoic acid in a cell context dependent manner. Using promoter mapping, promoter site-directed mutagenesis, EMSAs and chromatin immunoprecipitation analysis we show that retinoic acid induces the recruitment of NF-κB proteins to NF-κB binding sites in the proximal region of the cIAP2 promoter, thereby causing induction of cIAP2 expression. In agreement with our data, the induction of NF-κB proteins binding and activity by retinoic acid has been reported in several cell systems such as neuroblastoma or leukemia cells [49, 54, 55]. Importantly, in addition to NF-κB proteins, the retinoid receptors, RAR and RXR, are also recruited in vivo to the cIAP2 promoter upon retinoic acid treatment, despite the absence of bona fide RARE sites in this promoter by in silico analysis. Protein-protein interaction between p50/p65 and RXR that could contribute to stabilize the transcriptional activation complex have been described . Despite the finding that mutation of an AP-1 motif decreases 9-cis-RA inducibility, we could not detect in vivo recruitment of cJUN to the cIAP2 promoter in response to the retinoid. Although we cannot totally dismiss the possibility that cJUN takes part of the transcriptional complex induced by retinoic acid, other AP-1 binding factors could be recruited to the promoter. Importantly, although our data suggest that ligand-bound RAR/RXR may be recruited directly to the transcriptional activation complex we cannot discard that, in addition, retinoic acid induction of cIAP2 expression proceeds via regulatory circuits, which are likely to involve retinoic acid-target genes as well as cross-talk with other signaling pathways. Thus, as reported for neuroblastoma cells , retinoic acid could induce the activation in breast cancer cells of the phosphatidylinositol 3-Kinase/Akt signaling pathway that finally results in NF-κB activation.
Little is known about the anti-apoptotic potential of retinoic acid [58–61]. We provide evidence that in a cellular context, present in T47D, ZR-75-1 and SK-BR-3 cells, retinoic acid markedly upregulates cIAP2 expression. Retinoic acid significantly mitigates the apoptosis induced by chemotherapeutic agents in T47D and ZR-75-1 cells, while it is able to increase apoptosis by these compounds in H3396 cells where retinoic acid does not induce cIAP2 expression. Many antiapoptotic proteins, such as Bcl-2, Mcl-1 and Bcl-XL, have been shown to inhibit chemotherapeutic agent-induced apoptosis in diverse cell system models including hematopoietic and neuroblastoma cells. Additionally, it has been shown that the activation of genes encoding TRAF and IAP proteins by NF-κB serves to block apoptosis promoted by different insults including chemotherapy-induced apoptosis in different cell types [30, 32, 39, 62]. In particular, overexpression of cIAP2 inhibits etoposide-induced apoptosis, processing of caspase-3 and generation of caspase-like protease activity in 293T cells . Accordingly, it has also been shown that cIAP2 overexpression blocked etoposide-induced processing of caspase-3 and apoptosis in HT1080 cells under NF-κB-null conditions . Thus, cIAP2 emerged as a likely candidate to mediate the antiapoptotic effect of retinoic acid in our cell system. To test the involvement of cIAP2 in retinoic acid action, we performed siRNA studies to selectively suppress cIAP2 expression. Notably however, these studies did not show sensitization of T47D cells to etoposide-induced apoptosis in conditions of retinoic acid pretreatment, despite effective cIAP2 downregulation. These findings clearly demonstrated that cIAP2 is not necessary for retinoic acid protection of chemotherapy-induced apoptosis. However, we cannot rule out the possibility that compensatory expression of other members of the IAP family protein could supersede the absence of cIAP2 in our system, explaining the lack of effect of cIAP2 knockdown. Recent data also suggest that neither cIAP1 nor cIAP2 are able to inhibit caspases directly . Thus, these results and ours suggest a more complex role for cIAP2 in antiapoptosis than previously expected. Further studies are required to reveal the precise involvement of cIAP2 on retinoic acid effects in breast cancer cells.
It has been reported that the NF-κB signaling pathway plays a major role in cell survival  and in sensitivity of cancers to chemotherapy . In accordance with these observations, we have found that retinoic acid can activate the NF-κB signaling pathway in certain breast cancer cells, which correlates with the induction of resistance against apoptosis induced by cancer therapy agents, such as etoposide, doxorubicin or camptothecin. Furthermore, we have demonstrated that impairment of NF-κB activation results in a moderate increment of retinoic acid-induced apoptosis and in a similar sensitivity to etoposide in the presence and absence of 9-cis-RA. The multiplicity of mechanisms whereby NF-κB serves the antiapoptotic function is becoming increasingly complex. It has been reported that NF-κB increases the expression of several antioxidant effectors, such as glutathione cysteine synthetase, glutathione, manganese superoxide dismutase, hemeoxygenase, ferritin heavy chain and thioredoxin [65–69]. On the other hand, retinoic acid has been shown to reduce susceptibility to oxidative stress in chick embryonic neurons , in PC12 cells, and in mesangial cells , although the mechanism of the antioxidant effect of retinoic acid remains unclear. Furthermore, it has been reported that retinoic acid treatment represses ROS accumulation by a mechanism involving NF-κB in NB4 cells; in these studies, the impairment of NF-κB activation resulted in increased ROS levels and JNK activation in retinoic treated NB4 cells . Since etoposide induces marked biochemical alterations characteristic of oxidative stress, including enhanced lipid peroxidation and decreased levels of reduced glutathione, it will be of interest to determine the role of different antioxidant effectors in retinoic acid protection of etoposide-induced apoptosis. It is tempting to speculate that retinoic acid is able to regulate the sensitivity to chemotherapeutic agents-induced apoptosis by increasing antioxidant defense components through NF-κB proteins in certain cellular contexts such as T47D breast cancer cells.