THE NATURAL FLAVONOID LUTEOLIN INDUCES APOPTOSIS IN COLON CANCER CELLS BY DYSREGULATING THE SPHINGOLIPID RHEOSTAT

Colorectal cancer is one of the most common malignancies and a leading cause of cancer death in the world. More powerful and safer therapeutic approaches are urgently needed to reduce mortality and garner better curative effects. In this regard, dietary supplements capable of preventing carcinogenesis and inhibiting the growth of colon carcinoma cells have generated intense interest. Luteolin (LU), a common dietary flavonoids, has emerged as a powerful anticancer agents able to sensitize different cancer cells, including colon cancer ones, to therapeutic-induced cytotoxicity. However, the molecular mechanisms underlying LU effects in colon cancer are largely unknown. Sphingolipids have critical functions as signaling molecules. In particular, Ceramide (Cer) and Sphingosine-1-phosphate (S1P) are involved as key antagonist mediators in regulating crucial cell responses such as proliferation and apoptosis. Cer can act as a second messenger, and, by activating a variety of signaling pathways, is able to promote growth arrest, apoptosis, or cell differentiation. To the opposite, the sphingoid mediator S1P can act as a potent mitogen and survival factor for a variety of cell types. These two lipids together form the “sphingolipid rheostat” regulating the balance between cell growth and cell death. The objective of our study was to investigate the effects and the molecular mechanisms of LU in colon cancer, focusing on the role of the bioactive sphingoid molecules Cer and S1P. To this purpose, we used the Caco-2 cell line, obtained from a human colon adenocarcinoma, as cell model of both colon cancer cells, and differentiated intestinal enterocytes. Indeed, in long-term culture, these cells undergo a spontaneous differentiation into polarized cells, representing so far the best available in vitro model of absorptive enterocytes. These two models might thus allow to distinguish the potential LU effects on colon cancer cells in comparison to their healthy counterparts. At first, we characterized the morphological and biochemical features of both models. We found that Caco-2 cell differentiation was accompanied by the formation of “domes” across the monolayer, known as characteristic structures of differentiated cells, and indicative of their property of absorptive epithelium. In addition, the activity of the alkaline phosphatase, a membrane-associated hydrolase was found significantly increased in differentiated Caco-2 cells compared to the tumoural cancer ones. Furthermore, we found that Caco-2 differentiation was associated with a reduction of the phospholipids/protein ratio, and whereas phosphatidylcholine was the most abundant phospholipid species of tumoural cells, phosphatidylethanolamine prevailed in the differentiated ones. Moreover, phosphatidylethanolamine plasmalogen, a quantitative relevant species in the cancer cells, exhibited a marked decrease with intestinal differentiation. As far as the sphingolipid composition concerns, we first demonstrated that Caco-2 differentiation was associated to an increase in the total amount of sphingolipids, including Sphingomyelin (SM), the major component and precursor of both Cer and S1P, and above all ceramide. Since the SM hydrolysis, triggered by Sphingomyelinases (SMases), has been implicated in colon tumourigenesis, we then evaluated whether changes in the activity of the known different SMases (the neutral (N-SMase) and alkaline (Alk-SMase) enzymes) occur with cell differentiation. We found that the Alk-SMase activity which was barely detectable in tumoural cancer cells, significantly increased in differentiated ones. The high level of this enzyme is consistent with its presence in the apical brush border, characteristic of intestinal cells. Interestingly a significant increase in the N-SMase was observed in the differentiated Caco-2 cells suggesting a role for this enzyme in intestinal cells. Prompted by these data indicating that the SM hydrolytic potential is enhanced in differentiated Caco-2 cells, we then investigated SM metabolism in both tumoural and differentiated cells. We found that the tumoural cells were much more rapid in the production of Cer from SM than the differentiated ones, and that Cer turnover was present and rapid in both cell types. Inhibition of N-SMase activity, the most abundant enzyme, resulted in no variation of Cer formation in both types of cells, suggesting that the lysosomal acid SMase (A-SMase) is involved in SM degradation. After this characterization, we then evaluated the effect of LU on the tumoural and intestinal cells. Interestingly, we found that the flavonoid exhibited a potent cytotoxic effect on tumoural cells, by inducing apoptosis, without affecting the viability of differentiated cells. Instead, we found that LU induces an increase in intracellular Cer level with a more pronounced trend in tumoural cells. Based on these results we examined whether Cer is involved in the mechanism of LU-induced apoptosis. Notably we obtained that conditions leading to enhance the Cer content in colon cancer cells, as treatment with short-chain Cer analogues and inhibitor of sphingomyelin synthase (SM-synthase) were succeeded by inducing apoptosis in tumoural cells, thus mimicking the LU effect. Furthermore, we evaluated the possible effect of LU on the formation of Cer and S1P in tumoural Caco-2 cells. Pulse experiments showed that treatment with LU induced a dose-dependent increase in Cer, paralleled by a concomitant decrease of both SM and glycosphingolipids synthesis. This result suggested that LU most probably acts by reducing the availability of Cer as a substrate for both SM-synthase and glucosylceramide synthase enzymes localized in the Golgi complex, possibly by inducing a defect in the common vesicular route responsible for complex SLs biosynthetic process. Fluorescent studies using a BODIPY-C5-CER, a ceramide analogue known to mimic the ER-Golgi traffic of natural Cer in living cells, supported the presence of an aberrant traffic of Cer during LU treatment. Furthermore, pulse experiments using treatment with Brefeldin A, known to induce a retrograde merging of Golgi membranes with the ER, demonstrated that LU was not more able to exert alterations of Cer metabolism, thus indicating that the ER-Golgi traffic of Cer is the site of LU action. In parallel, different protein kinases, including Akt have been shown to regulate ER to Golgi traffic. In addition, a study in our laboratory reported for the first time a putative role of PI3K /Akt pathway in the regulation of the vesicle-mediated movements of Cer in the ER-Golgi district. Prompted by these findings, pulse experiments using LY294002, a representative PI3K inhibitor, showed that, LY294002 and subsequently the inhibition of PI3K/Akt had the same effect on the Cer metabolism observed during LU treatment. Furthermore, we obtained that LU strongly inhibited the Akt phosphorylation as a downstream response to PI3K inhibition. Taken together, it emerges that blockade of PI3K by LU and subsequently the downregulation of PI3K/Akt pathway is considered at least one of the mechanisms responsible for the effect of this flavonoid on Cer trafficking observed in our tumoural cell model Further analyses revealed that LU was able to alter not only Cer content but also to decrease the endogenous S1P level inducing thereby a shift of the “sphingolipid balance” to the side of death. Based on this result, we demonstrated for the first time that LU was able to act as inhibitor of Sphingosine kinase (SPHK) activity in tumoural cells, especially SPHKII, the up-regulated isoform in our cancer cell model, exhibiting only a modest effect on SPHKI. Furthermore, in order to gain deeper insights into the role of the “Sphingolipid rheostat” in mediating the effect of LU, we attempted to push the balance towards S1P with addition of exogenous S1P. The results demonstrated that S1P conferred a significant resistance of colon cancer cells to the cytotoxic effect of LU. The mechanism by which S1P acts as LU-antagonist was proved to be mainly by the up-regulation of the PI3K/Akt pathway, which is able to rescue colon cancer cells from the apoptotic effect of the LU-induced increase of Cer. Taken together, our results demonstrate, for the first time, that the dietary natural flavonoid LU induces apoptosis in colon cancer cells by targeting the “sphingolipid rheostat”, and directing the balance in favor of Cer. As the balance between Cer and S1P appears to be an important target for development of new and effective therapeutic strategies against tumour progression, LU could be a novel antitumoural agent not only in colon cancer but possibly in the treatment of other solid tumours.