The fact that cyclin D1 levels are suppressed during S phase as a result of phosphorylation on Thr-286  suggests that the kinase responsible might be GSK3 (see [15,17]). If GSK3 were responsible for the cell cycle specific suppression of cyclin D1 levels, the activity of this kinase and those signaling molecules which regulate its activity would be expected to vary in activity through the cell cycle. Thus, inhibition of these signaling molecules would be expected to abolish the cell cycle variations in cyclin D1 levels. We found, however, that inhibitors of PI3K, rapamycin and serum removal all suppressed cyclin D1 levels, but that this suppression was uniform through the cell cycle. The typical cell cycle expression pattern of cyclin D1 was not altered by any of these treatments. Moreover, cyclin D1 expression was not significantly suppressed in any cell cycle phase by expression of a dominant inhibitory AKT mutant protein, which is able to directly regulate GSK3 activity. These results fail to support the notion that these signaling molecules are responsible for specific suppression of cyclin D1 during S phase. To extend this result, the activity of signaling molecules was tested throughout the cell cycle. PI3K activity was assessed in individual cells with a GFP chimeric protein able to bind phosphatidylinositol-3 phosphate. The activating phosphorylation of AKT and the inactivating phosphorylation of GSK3β, together with the enzymatic activity of total GSK3, were assessed biochemically in synchronized populations. There was no evidence that any of these signaling activities vary thorough the normal cell cycle.
These results not only raise questions regarding the role of proliferative signaling in regulating cyclin D1 levels through the cell cycle, they bring into question the overall role of GSK3 in regulating cyclin D1 levels in living cells in general. To directly analyze the involvement of GSK3 in the control of cyclin D1 levels, its activity was inhibited by a variety of small molecule inhibitors, while its cellular levels were suppressed by siRNA. In no case was there any quantitative change in cyclin D1 levels in any cell cycle phase. LiCl failed to alter the phosphorylation of cyclin D1 on Thr-286 even though it efficiently blocked phosphorylation of β-catenin in the same cells. Finally, constitutively activated GSK3 β was introduced into living cells and found to have no influence upon cyclin D1 levels. We conclude that GSK3 does not play a role in the suppression of cyclin D1 during S phase, and that it is unlikely to be involved in the regulating cyclin D1 levels in the nucleus during any cell cycle phase. The one potential uncertainty with these results is the fact that in all the above experiments the cells were in the presence of normal serum levels. These might promote the constant suppression of GSK3 activity in these cells, masking the potential involvement of this enzyme in cyclin D1 regulation in the absence of proliferative signaling. In a final analysis, serum was removed from cycling NIH3T3 cells to allow GSK3 to become active, and then in some cultures it was again inhibited by the addition of LiCl. There was a slight, if minor elevation of cyclin D1 in a small proportion of G2 phase cells in the serum-deprived cultures treated with LiCl. It is not clear, however, if this was the direct result of cyclin D1 phosphorylation by GSK3, or the ability of GSK3 to interfere with passage through G2 phase.
The possibility that GSK3 might have an effect upon cell cycle progression is supported by its ability to influence a variety of cellular processes in addition to insulin and glycogen metabolism. It is directly involved in regulating cell proliferation and survival [33,34]. GSK3 is also able to phosphorylate transcription factors involved in growth regulation, including c-Myc, c-Jun and c-Myb ; and has been implicated in the action of growth factors on neural cell growth . It is also able to regulate development directly , and as a member of the Wnt signaling pathway [38-41]. GSK3 is able to phosphorylate the Tau protein, involved in Alzheimer's disease ; play a role in mitotic spindle function ; repress Hedgehog signaling , and regulate cyclin E stability . Recent studies implicate its action in cell polarity determination . The results presented here suggest that the ability of GSK3 to phosphorylate a protein in vitro might not indicate that it has a role in the normal regulation of that protein within a living cell [47,48]. This conclusion, however, should be considered cautiously in cases where cell based assays indicate a connection between GSK3 and the regulation of a cellular protein .
Even though GSK3 has been implicated in a number of regulatory pathways, our view of GSK3 as a modulator of a wide variety of cellular processes must be reconsidered in light of recent genetic studies. With a reverse genetics approach in Drosophila, the inhibitory phosphorylation of GSK3 was shown to play a critical role in the action of the insulin/PI3K pathway, but was not involved in the stimulation of growth by PI3K . In addition, knock-in studies in mice with homozygous activating mutations of both GSK3α and GSK3β demonstrated that while glycogen metabolism was altered in some tissues, the animals were otherwise normal in growth and development . While developmental alterations can mask the normal role of a mutant protein, these studies emphasize the importance of carefully analyzing the role of GSK3 in signaling processes. In the studies reported here, we find that despite the fact that GSK3 is able to phosphorylate cyclin D1 on Thr-286 , this molecule does not play a role in the cell cycle regulated expression of cyclin D1. Our evidence further suggests that it may not play any major role in the expression of cyclin D1 in human and murine fibroblast cells. The possibility remains that GSK3 plays some role in other cell types, particularly during the growth factor induced increase in nuclear cyclin D1 during G2 phase. It has been shown, however, that ubiquitination of cyclin D1 can efficiently take place following phosphorylation of another site , or without the apparent requirement for phosphorylation [50,51]. Consistent with this conclusion, studies by others demonstrated that Thr-286 does not play a major role in the overall regulation of cyclin D1 levels in a variety of cells, although it is reported to play a role in the intracellular localization of this protein through the cell cycle . Our studies on intracellular localization of cyclin D1 with the use of quantitative image analysis, however, have failed to produce any evidence of relocation from the nucleus to the cytoplasm during passage into S phase .