The intestinal epithelium is the most vigorously self-renewing tissue in adult mammals. Perturbations of normal tissue homeostasis attributable to genetic lesions or environmental insults can lead to hyperproliferative diseases of the intestinal tract such as cancer . Intestinal epithelial cell regulation has been studied extensively and has revealed canonical Wnt/β-catenin, KRAS/MAPK and PI3K/Akt signaling pathways as key regulators of cell division and differentiation [16, 18, 20]. Normal cell division is a tightly controlled process that only allows cells to divide in a timely and restricted manner. As such, E2F transcription factors control cell division by activating the transcription of numerous genes involved in G1 and S phases [1, 2]. Many previous analyses of E2F proteins in intact intestinal epithelium or in cultured crypt cells have demonstrated that nuclear E2F4 may be determinant in the control of proliferation. Indeed, in mice, deletion of E2F4 gene resulted in a significant decline in proliferative zones (crypts) and a shortening of intestinal villi . Accordingly, double staining experiments in intact human intestine revealed that crypt epithelial cells expressing high levels of nuclear E2F4, were all positive for Ki67 and cyclin A [9, 10]. In this respect, decreased expression of E2F4 by RNA interference reduced the proliferation rate of normal intestinal epithelial crypt cells and colorectal cancer cells in culture . Interestingly, in contrast to E2F1, which resides constitutively in the nucleus throughout the cell cycle, E2F4 protein is mostly distributed in the cytoplasm of quiescent cells and translocates into the nucleus upon serum stimulation . Taken together, these results suggest that the nuclear translocation of E2F4 may represent a critical step in promoting G1/S transition in intestinal epithelial crypt cells.
In the present study, we show that activation of the MEK/ERK pathway by serum is required for E2F4 nuclear translocation as well as for G1/S phase transition of human intestinal epithelial crypt cells. Accordingly, several studies on cultured intestinal epithelial cells and many other cell types have revealed a close correlation between ERK activation and DNA synthesis while pharmacological or molecular inhibition of ERK activity has been shown to block cell cycle progression [15, 25]. Notably, like E2F4 , phosphorylated and activated forms of ERK1/2 have been mostly detected in the nucleus of undifferentiated proliferative crypt cells in human fetal small intestine , hence supporting the role of these kinases in cell cycle control of intestinal crypt cells. Therefore, our results indicate that one of the mechanisms by which ERK1/2 MAP Kinases induce intestinal epithelial proliferation may be by promoting E2F4 nuclear translocation. Once into the nucleus, E2F4 may control the expression of proteins necessary for entry into S phase including cdc6, dihydrofolate reductase (dhfr), thymidine kinase, cyclin E, cyclin A, mcm3 and DNA polymerase α as we have previously shown [2, 10].
The exact molecular mechanism by which ERK1/2 promotes E2F4 nuclear translocation however remains unclear. Herein, E2F4 was found to be rapidly phosphorylated on serine residue(s) in serum-treated cells and that this phosphorylation was MEK-dependent. Of importance, we observed a strong correlation between the rapid phosphorylation of E2F4 and its subsequent nuclear translocation. Hence, one could speculate that phosphorylation of E2F4 by ERK1/2 is necessary to induce its translocation into the nucleus. However, although E2F4 phosphorylation was observed within 30 minutes after serum addition, nuclear accumulation of E2F4 only began after 4 hours . This suggests that ERK-dependent E2F4 phosphorylation may represent an initiating event for nuclear translocation and that other mechanisms are also likely implicated. A possible mechanism could be the heterodimerization of E2F4 with its transcriptional partner DP-2, reported to promote E2F4 nuclear localization and activation . One might speculate that E2F4 phosphorylation could promote its association with DP-2 and subsequently, its nuclear re-localization. To our knowledge, few studies have demonstrated phosphorylation of E2F4 [27–30], and none have linked phosphorylation to the stimulation of E2F4 function or transcriptional activity. We show herein that ERK kinases can efficiently and directly phosphorylate E2F4 protein in vitro, thus identifying E2F4 as a novel target of ERK kinases, adding to the list of ERK substrates implicated in cell proliferation control . Our analysis of putative ERK1/2 phosphorylation sequences revealed that S244 and S384 were both implicated in the transcriptional activity of E2F4 and in its nuclear localization. Nevertheless, more studies are needed to clearly elucidate the molecular mechanisms of E2F4 activity and localization by phosphorylation.
A novel and intriguing finding of this study is that stimulation with EGF was not sufficient to induce G1/S phase transition of human non immortalized intestinal epithelial cells. These results contrast with those observed in rodent immortalized intestinal epithelial cell lines in which EGF induced DNA replication and proliferation . Furthermore, many studies have reported proliferative properties of EGF in organotypic cultures of human fetal intestinal epithelium, although these studies did not exclude the contribution of other mesenchymal and epithelial factors in this effect [32–34]. Many studies from our laboratory and others have clearly demonstrated that HIEC are useful and relevant in analyzing the regulation of proliferation of intestinal epithelial crypt cells in humans [9, 10, 35, 36]. Indeed, the expression of intestinal epithelial-specific keratins , of components specific to cell junctions , cell cycle-related proteins [9, 10, 39], as well as typical intestinal cell markers of undifferentiated lower crypt cells [37, 39, 40] indicate that these cells behave as cells representative of the bottom of the human crypt . Their epithelial cryptal origin was also confirmed by their ability to express the 350-kD crypt cell-specific marker MIM-1/39 . Interestingly, we show herein that although EGF induced a rapid activation of ERK1/2 similarly to serum and LPA, this action was not sufficient for S phase entry as visualized by the absence of pRb hyperphosphorylation, cyclin A protein expression and E2F4 nuclear translocation. In addition, EGF did not trigger the degradation of the cell cycle inhibitor p27, an event necessary to exit the quiescent state and pass the restriction point . The failure of EGF to induce G1/S transition in HIEC could be explained by its inability to promote a sustained phosphorylation of Akt. Indeed, stimulation of Akt by growth factors is known to be required for G1 progression and S phase entry of many cell types  including intestinal epithelial cells . Several hypotheses could be suggested to explain why EGF alone is not sufficient to promote a strong and sustained Akt activation in HIEC. First, the nature of EGFR-associated Ras proteins, i.e. H-Ras or K-Ras, can define the selective activation of ERK or Akt pathway by EGF . In addition, in colon cancer, activation of the Rho/Rho-kinase pathway inhibits the capacity of EGF to promote Akt activation, but not ERK1/2 . Finally, Erk5 was recently shown to be necessary for sustained PDGF-induced Akt phosphorylation in endothelial cells . Further studies are thus needed to verify whether these different hypotheses could explain why EGF alone did not promote a sustained Akt activation in human intestinal epithelial cells.
Our results also suggest that phosphorylation and inhibition of GSK3β play an important role in the nuclear translocation of E2F4 and the proliferative response of HIEC. Indeed, serum and LPA, but not EGF, inactivated GSK3β as visualized by the sustained phosphorylation on serine 9, probably triggered by Akt or protein kinase C . Furthermore, when GSK3β was pharmacologically inhibited, EGF induced pRb hyperphosphorylation, cyclin D1 expression, p27 degradation and E2F4 nuclear translocation, four events associated with G1/S phase transition. It is noteworthy however, that inhibition of GSK3 only partially rescued the inability of EGF to induce E2F4 nuclear translocation and cell proliferation in comparison to serum. This could be explained by the fact that serum contains a number of different additional factors (including LPA) and hormones in high concentrations that may activate various signaling pathways including calcium mobilization and PKCs, other signaling events known to promote proliferation . Nevertheless, our results suggest that GSK3β, which is tonically active in quiescent cells, must be phosphorylated and inactivated to enable cell cycle progression of HIEC. GSK3β im-plication in E2F4 nuclear localization control adds to the previously described role of GSK3β on E2F1 regulation by ubiquitination and degradation resulting in a reduced transcriptional activity . This is also reminiscent of the observed decreased expression of several other cell cycle regulated proteins following GSK3 activation, including c-myc, cyclin D1 and β-catenin . In this regard, accumulation of these GSK3 substrates has been linked to increased intestinal proliferation and notably frequently observed in colorectal cancers [20, 23, 49, 50].