FoxO3a is regulated by posttranslational modifications such as phosphorylation, acetylation and ubiquitination, each of which affects the transcriptional activity of FoxO proteins[11-16] (Figure 2). The potency of FoxO3a is carefully regulated by phosphorylation. The phosphorylation of FoxO3a by several kinases is well established. Among them, protein kinase B (Akt) is an important kinase that directly phosphorylates FoxOs. In the case of FoxO3a, T32, S253 and S315 residues are phosphorylated by Akt and, in particular, the phosphorylation of S253 is a crucial residue regulating the nuclear/cytoplasmic shuttling of FoxO3a. For example, when cells are cultured in the presence of growth factors or insulin, FoxO3a is phosphorylated by Akt and mainly localized to the cytoplasm, which prevents its transcriptional activity. The phosphorylation event of FoxO3a by Akt facilitates FoxO3a interaction with the 14-3-3 nuclear export protein, further preventing nuclear re-import by concealing nuclear localization signals[13]. Furthermore, the phosphorylation of FoxO3a by activated Akt promotes an association with an ubiquitin E3 ligase, subsequently polyubiquitinating FoxO3a, which facilitates FoxO3a degradation by proteasomes[13-17]. Thus, the activation of Akt is thought to be critical in FoxO3a regulation. However, in some tumors, FoxO3a remains in the cytoplasm even in the absence of active Akt[14]. It has been found that IkB kinase (IKK) phosphorylates FoxO3a at serine 644, thereby inhibiting its transcriptional activity in an Akt-independent manner[15]. The phosphorylation of FoxO3a by IKK also leads to its cytoplasmic localization, although the underlying export mechanism is not understood. The insulin/IGF-1 and integrin-dependent signaling pathways activate Akt via PTEN suppression which phosphorylates FoxO3a, thereby rendering it functionally inactive. In contrast, FoxO3a is localized to the nucleus to activate its target genes when growth factors or serum are deprived. Additionally, serum and glucocorticoid regulated kinase (SGK), casein kinase 1 (CK1), dual specificity tyrosine-phosphorylated and regulated kinase 1A (DYRK1A), janus N-terminal kinase (JNK), mitogen-activated protein kinases (MAPKs), IkappaB kinase (IKKβ), mammalian sterile 20-like kinase 1 (MST1) and AMP activated protein kinase (AMPK) are also known to regulate FoxO3a and other family members[18-23] by phosphorylating multiple residues. Interestingly, SGK1 is transcriptionally up-regulated in response to a variety of external stimuli, including growth factors. SGK1 is also known to phosphorylate the pivotal ser 253 residue, which triggers its location to the cytoplasm, thereby inhibiting its function[19]. In contrast, AMPK activates FoxO3a function. 6 threonine/serine residues (T179, S399, S413, S555, S588 and S626) in mammalian FoxO3a are found to be phosphorylated by AMPK[24,25]. Mutation of these phosphorylation residues to alanine severely impairs its function, yet it does not alter its ability to bind to cognate sequences or to participate in nucleocytoplasmic shuttling depending on external cues[25]. Likewise, JNK also phosphorylates FoxO3a, activating FoxO3a function by enhancing its location into the nucleus which subsequently increases its transcriptional activity[18,22].

Unlike kinases, very few phosphatases have been found to regulate FoxO3a. One particular phosphatase, protein phosphatase-2A (PP2A), has been shown to regulate FoxO3a function. Nho et al[26] showed that when fibroblasts attach to 2D type collagen coated plates, PP2A activity is suppressed, which facilitates FoxO3a inactivation by enhanced Akt, promoting fibroblast proliferation. But the over-expression of PP2A reverses this inactivation and increases dephosphorylated FoxO3a, thereby suppressing their proliferation. Singh et al[24] also demonstrated that FoxO3a interacts with PP2A C/A subunits in HeLA cells, dephosphorylating its T32/S253 residues, which subsequently inhibits the interaction of the 14-3-3 protein to FoxO3a by Akt. This study showed that PP2A is required for the reactivation of FoxO3a by promoting its translocation to the nucleus (Figure 3). Interestingly, recent studies also showed that the adenovirus E1A stabilizes FoxO3a by inducing the expression of PP2A/C, which suppresses βTrCP-mediated degradation of FoxO3a[25]. Thus, these studies clearly suggest that the imbalance between kinases and phosphatase(s) can greatly affect a cell’s fate by curbing FoxO3a function and the alteration of these kinases and phosphatases are directly linked to certain disease progression.

Open in New Tab   Full Size Figure   Download Figure

Figure 3 Forkhead box O3a localization by phosphorylation and dephosphorylation. Forkhead box O (FoxO)3a becomes translocated to the cytoplasm when phosphorylated on ser 253 residue by Akt or SGK. FoxO3a is then bound to 14-3-3 and this interaction promotes its degradation by the proteasome. In contrast, FoxO3a is dephosphorylated by protein phosphatase-2A and this opposite event facilitates its re-location into the nucleus, thereby activating its target genes. SGK: Serum-and glucocorticoid-induced protein kinase; Akt: Protein kinase B.

As we briefly described above, FoxO3a degradation is also an important step to regulate its function. The single molecule RING-finger E3 ligase murine double minute 2 (MDM2) promotes ubiquitination of FoxO3a as well as FoxO1 and FoxO4, facilitating their degradation[27]. Intriguingly, knockout or knockdown of MDM2 alone increases FoxO3a protein levels. This effect was shown to be mediated by MDM2-induced polyubiquitination of FoxO proteins[27,28], whereas another study showed that MDM2 catalyzes multiple monoubiquitination of FoxO4 rather than polyubiquitination[28]. When FoxO3a is located to the cytoplasm by Akt, FoxO3a becomes ubiquitinated and this event triggers a proteasome-dependent degradation process. Like MDM2, FoxO3a phosphorylation by IKK also leads to its ubiquitination and degradation[15]. Thus, these studies document that FoxO3a localization in the cytoplasm not only deactivates FoxO3a function but also becomes a crucial step leading to FoxO3a degradation.

Acetylation also plays an important role in regulating FoxO3a. Oxidative stress triggers FoxO3a acetylation/deacetylation and affects the localization of FoxO3a. For example, protein acetylase CREB binding protein (CBP)[29-31], p300[32,33] and deacetylase Sirt are known to modulate FoxO3a function[34-38], although a precise mechanism describing the effects of acetylation and deacetylation is not known. A recent piece of evidence suggests that the FoxO family is also regulated by microRNA. mir155, mir96 and mir21 are thought to directly regulate FoxO3a, while mir205 regulates FoxO3a via its upstream target PTEN[39-43]. FoxO3a is also known to be regulated by a transcription factor. E2F-1 can bind to the promoter region of FoxO1 and FoxO3a, thereby regulating FoxO3a at the mRNA level[44]. FoxO3a mRNAs are modulated as a function of age in rat muscle, peaking at 6 and 23 mo, suggesting that FoxO3a may also affect longevity in mammals[45].

Perhaps the two most significant cellular processes that are regulated by FoxO transcription factor are the suppression of cell cycle progression and the promotion of apoptosis[46-50]. FoxO3a activation increases cell cycle inhibitor proteins p21 and p27, both of which subsequently suppress G1 to S cell cycle transition[51-54]. Although p27 is transcriptionally regulated by FoxO3a via the PI3K/Akt-dependent axis, it has been shown that p27 is also regulated via the FoxO3a/NF-κB/c-Myc-dependent pathway. Chandramohan et al[55] showed that in WEHI 231 cells, the suppression of PI3K activity promotes a decrease in c-Myc dependent p27 expression via NF-κB inhibition. Since NF-κB is frequently altered in many types of cancers and NF-κB transcriptionally activates c-Myc gene expression, this finding suggests that p27 is reciprocally regulated by FoxO3a and c-Myc. A recent study further suggests that FoxO3a inhibits NF-κB function and that the alteration of FoxO3a is associated with hyper-proliferative helper T cells, cigarette smoke-induced inflammation, airspace enlargement and chronic obstructive pulmonary disease[56,57]. Likewise, FoxO3a also increases several target genes, such as Bim, TRAIL, PUMA and Fas ligand, which all promote cell apoptosis. For example, FoxO3a directly binds to the promoter region of Bim, causing sympathetic neuron cell death[44]. The activation of the transcription factor FoxO3a led to increased TRAIL transcription and induction of G1 arrest in the absence of v-Abl inhibition; this effect could be inhibited by the expression of a constitutively active Akt mutant in BCR-Abl-transformed human cells. Ghaffari et al[49] also demonstrated that cytokine and BCR-Abl suppression of TRAIL transcription is mediated through phosphorylation and inhibition of the FoxO3a transcription factor. This study showed that BCR-Abl-induced inhibition of TRAIL transcription is linked to the tumorigenicity in chronic myeloid leukemia[50]. FoxO3a is also associated with the regulation of PUMA and Noxa proteins in lymphoid and neuroblastoma cells, respectively[58,59]. Thus, these findings clearly demonstrate that FoxO3a-dependent cell cycle arrest and apoptosis induction are important for tumor suppression (Table 1) and further indicate that the pathological alteration of FoxO3a can potentially contribute to the acquisition of uncontrolled cell proliferation and an apoptosis-resistant cell phenotype.

The most recent discovery regarding FoxO3a’s function is that it is also associated with stress response and longevity. In contrast to FoxO3a’s better known functions of inhibiting cell proliferation and promoting apoptosis as described above, FoxO3a also participates in protecting cells when exposed to unfavorable conditions. This seemingly contradictory effect of FoxO3a has been observed in various cell models and it has been found that the reactive oxygen species (ROS) are linked to the activation of FoxO3a to protect cells from a stress inducing environment[60,61]. In C. elegans, DAF-16 is thought to regulate 230 genes on the ablated germ cell line background and most of these genes are related to the resistance of external stress[62,63]. Deregulated ROS induce apoptosis and are associated with various diseases and aging. Sirtuin-1 (Sirt1) decreases ROS levels and promotes cell survival under oxidative stress conditions. Interestingly, FoxO3a and other FoxO family members increase superoxide dismutase (SOD) and protect cells from oxidative stress in a Sirt1-dependent manner[34,38]. A Sirt1/FoxO3a-dependent cell regulatory function that has been linked to stress management was previously studied. Brunet et al showed that Sirt1 and FoxO3a form a complex in cells in response to oxidative stress and Sirt1 increases the ability of FoxO3a to induce cell cycle arrest and resistance to oxidative stress but inhibited FoxO3a’s function to induce cell death[38]. These results showed that FoxO3a deacetylation by Sirt1 in response to ROS can be an important self defense mechanism to detoxifying harmful reactive molecules, further suggesting that Sirt1 is linked to protect cells from a stress inducing environment by tipping FoxO dependent response away from apoptosis and toward stress resistance[38]. Studies also found that Sirt3, which belongs to class III of HDACs, is linked to the resistance of stress inducing environments by detoxifying ROS. The role of Sirt3 and FoxO3a function is particularly well described in myocytes[64]. At the cellular level, when cardiomyocytes are exposed to stressful stimuli, Sirt3 levels are elevated, which subsequently deacetylase FoxO3a and facilitate its location into the nucleus to activate anti-oxidant genes[65]. Among them, catalase (Cat) and manganese superoxide dismutase (MnSOD) are direct targets of detoxifying enzymes by FoxO3a. Thus, the increased level of Cat and MsSoD by FoxO3a activation may efficiently and effectively manage ROS, which can be beneficial for reducing stress induced by ROS. Interestingly, a prior study found a potential FoxO activator as a way to protect cells from oxidative stress. Resveratrol, a polyphenolic flavonoid abundant in red wine with potent antioxidant activity, is known to up-regulate the FoxO family and block caspase 3, 8, and 9 activation, protecting photoreceptor cells from oxidative stress[66]. Thus, it is believed that when cells are exposed to a stress inducing environment, FoxO3a protects cells by utilizing SOD, catalase, etc., and this action is ultimately beneficial to cells. Given the fact that FoxO3a is linked to stress response and cells utilize FoxO3a to respond to ROS, it is a plausible scenario that the activation of FoxO3a under stress inducing conditions triggers the cell’s defense system, which can protect cells from harmful environments.

However, perhaps the most intriguing recent discovery in FoxO3a function is that the FoxO3a gene is associated with aging. Because FoxO3a is regulated by insulin-IGF1 signaling (IIS) which influences metabolism and lifespan in model organisms[67], FoxO3a had been proposed to be an ideal candidate to study longevity as the link between FoxO3a and longevity that has previously been described. Willcox et al[68] described 3 single nucleotide polymorphisms (SNPs) in the FoxO3a gene that were statistically significantly associated with longevity and different aging phenotypes in a sample of long-lived Americans of Japanese ancestry. Furthermore, Flachsbart et al[69] found that not only were certain FoxO3a variants very common in 90 year olds, they were even more common in 100 year olds, emphasizing the importance of genetics for aging well. It becomes clear that increases in cellular ROS levels are known to be associated with aging[70-75]. Increased cellular oxidative stress regulates FoxO post-translational modifications and the activation of the FoxO family has been shown to regulate cellular oxidative-stress resistance[76-81]. Interestingly, to support these findings, recent studies suggest a possibility that Sirt3 and FoxO3a have been linked to an extended life span in humans[75-78,82].