|Home | About | Journals | Submit | Contact Us | Français|
FOXO3a, a member of the Forkhead box O (FoxO) transcription factor family, is believed to be a tumor suppressor because it was found that FOXO3a inactivation promoted cell transformation and tumor progression. There are also a few studies showing that FOXO3a protected cells under stress conditions, including oxidative stress, serum deprivation, and hypoxia. It was reported that FOXO3a promoted invasion of cancer cells. Thus, the role of FOXO3a in cancer is complicated. Here, we report that FOXO3a is a positive regulator of nuclear factor κB (NF-κB) signaling. We found that overexpression of FOXO3a increased and knockdown of FOXO3a repressed NF-κB activities. Mechanistic studies indicate that FOXO3a activated NF-κB via inducing expression of B-cell lymphoma/leukemia 10 (BCL10), an upstream regulator of IκB kinase (IKK)/NF-κB signaling. We found that the serum deprivation activated NF-κB, which was blocked by inhibition of FOXO3a. Knockdown of FOXO3a enhanced cell apoptosis under serum-free conditions, which was inhibited by overexpression of BCL10. These results suggest that FOXO3a promotes cell survival via BCL10/NF-κB in serum starvation. Our findings may add another layer to the complexity of the role of FOXO3a in cancer. Therefore, caution should be taken when FOXO3a is employed as a target for cancer therapy.
The Forkhead box O (FoxO)3 subclass of transcription factors, including FOXO1 (FKHR) FOXO3a, FOXO4 (AFX) and FOXO6, are homologous to the Caenorhabditis elegans transcription factor DAF-16, which regulates life span downstream of the C. elegans insulin receptor (1, 2). They activate or repress multiple target genes and play important roles in multiple biological processes such as cellular metabolism, cell cycle regulation, apoptosis, and regulation of stress response (3, 4). Simultaneous deletion of FOXO1, FOXO3, and FOXO4 has been shown to develop lymphomas and hemangiomas, whereas loss of individual genes resulted in a less widespread and more modest neoplasia phenotype (5, 6). FOXO3a was believed to be a tumor suppressor because inhibition of FOXO3a expression promotes cell transformation, tumor progression, and angiogenesis (5, 7, 8). FOXO3a is regulated by AKT-mediated phosphorylation, which results in FOXO3a nuclear exclusion and inhibition of their transcriptional activity (9). FOXO3a can also be phosphorylated by serum- and glucocorticoid-induced kinase (SGK) (10) and IκB kinase (IKK) (11), leading to suppression of its activity. Phosphorylation of FOXO3a induces its nuclear export. Once exported from the nucleus, the phosphorylated FOXO3a may then be ubiquitylated, undergoing degradation. Deregulation of these kinases is frequently observed in tumors and can contribute significantly to tumorigenesis by promoting the nuclear export and/or proteasomal degradation of FOXO3a (12). These studies suggest that FOXO3a functions as a tumor suppressor and, therefore, may serve as a direct or indirect target for cancer therapy. However, there are also some studies demonstrating that FOXO3a promotes cell survival under stress conditions (13, 14). A recent publication demonstrates that FOXO3a promotes invasion of cancer cells through induction of matrix metalloproteinases (15). Thus, the role of FOXO3a in cancer is complicated.
NF-κB is a family of transcription factors that has been associated with diverse biological and pathological process (16, 17). In most cells, NF-κB proteins are normally inactive because they are sequestered in the cytoplasm by IκB family of inhibitory proteins. Extra-cellular stimuli such as TNFα and IL-1 lead to activation of IKK. The activated IKK phosphorylates IκB proteins, resulting in release of IκB proteins from NF-κB (16–18). This allows the translocation of NF-κB from the cytoplasm to the nucleus, where it regulates gene expression. IKK contains two catalytic subunits, IKKα and IKKβ, and one regulatory scaffold protein, IKKγ (16). Activation of the IKK requires phosphorylation of the T loop serines of IKKα and IKKβ (19).
In this work, we demonstrate that FOXO3a functions as a positive regulator of NF-κB signaling. We found that FOXO3a activated NF-κB through inducing expression of BCL10, an upstream regulator of IKK/NF-κB signaling. Although FOXO3a may serve as a target for cancer therapy, recent studies point out that the targeting of FOXO3a proteins may be complicated by potential feedback mechanisms (12). The previous reports and our results suggest that caution should be taken when FOXO3a is employed as a target for cancer therapy.
Human embryonic kidney 293T cells, colon cancer HCT116 cells, and cervical cancer HeLa cells were maintained in DMEM containing 10% serum. Human prostate cancer PC3 cells and lung cancer A549 cells were cultured in RPMI 1640 medium with 10% serum. The β-actin antibody was from Sigma. The antibodies against FOXO3a, BCL10, c-Myc, PARP-1, IKKγ, and α-tubulin were from Santa Cruz Biotechnology (Santa Cruz, CA). The antibodies against pFOXO3a(T32), pIKKα(S176)/IKKβ(S177), IκBα, p65, IKKα, and IKKβ were purchased from Cell Signaling Technology (Danvers, MA).
To generate the Myc-FOXO3a plasmid, the full-length human FOXO3a cDNA was amplified by RT-PCR and inserted into the pcDNA3.1 vector. HA-FOXO3aTM, a non-phosphorylatable form of FOXO3a, was purchased from Addgene (Addgene plasmid 1788). The Myc-FOXO3aTM plasmid was constructed using HA-FOXO3aTM as a template and subcloned into pcDNA3.1. The dominant-negative FOXO3a (FOXO3a-DN) was generated by deleting the transactivation domain of FOXO3a (amino acids 305–673) from the C terminus. Myc-FOXO3aΔH3 was constructed by deleting the fragment of the 206–216 amino acids of FOXO3a. This fragment is critical for FOXO3a DNA binding (20). The vector encoding IκBαSR was a gift from Dr. Wang Chen (The Institute for Biological Sciences, Shanghai, China). IκBαSR is a dominant-negative variant of IκBα containing Ser-to-Ala mutations at positions 32 and 36 that prevents IκBα phosphorylation and degradation. The domain-negative IKKβ (IKKβ-DN) (Ser-to-Ala mutations at positions 177 and 181) was constructed by site mutagenesis. The NF-κB-driven reporter plasmid was as described (21). The BCL10 reporter was constructed by inserting the BCL10 promoter (-432 ~ +1) into the pGL3-Basic Vector.
The cells were lysed in radioimmune precipitation assay buffer (150 mm NaCl, 100 mm Tris (pH 8.0), 1% Triton X-100, 1% deoxycholic acid, 0.1% SDS, 5 mm EDTA, and 10 mm NaF) supplemented with 1 mm sodium vanadate, 2 mm leupeptin, 2 mm aprotinin, 1 mm phenylmethylsulfonyl fluoride, 1 mm dithiothreitol, and 2 mm pepstatin A. After centrifugation at 4 °C for 10 min (10,000 × g), the supernatant was collected and subjected to Western blot analysis as described previously (22).
Transient transfection of the cells was done using Lipofectamine 2000 from Invitrogen according to the instructions of the manufacturer. The relative luciferase activity was determined by measuring firefly luciferase activity and normalizing it to β-galactosidase or Renilla luciferase activity.
siRNA oligos were purchased from Gene Pharma (Shanghai, China). The cells were transfected with siRNA oligos as described above. The sense sequences of these oligos were as follows: control, 5′UUCUCCGAACGUGUCACGUTT3; FOXO3a(siRNA-1), 5′GGAACGUGAUGCUUCGCAATT3′; FOXO3a(siRNA-2), 5′AGGGAAGUUUGGUCAAUCATT3′; and siBCL10, 5′GAAGGACGCCUUAGAAAAUTT3′.
Total cellular RNA was prepared and used for cDNA synthesis. RT-PCR was performed in a regular way. The primers used were as follows: BCL10, 5′AAGGAGAATCCAGCACGACG3′ (forward) and 5′ATCTGGTGGCAAAGGAGGAG3′ (reverse) and GAPDH, 5′CCACCCATGGCAAATTCCATGGCA3′ (forward) and 5′TCTAGACGGCAGGTCAGGTCCACC3′ (reverse).
Real-time PCR was performed as described (21) on a Step Two real-time PCR system (Applied Biosystems) using the comparative Ct quantization method. Real-time Master Mix (Toyobo) was used to detect and quantify the expression levels of the target gene. β-actin was amplified as an internal control. The primers used were as follows: IL-8, 5′AGGTGCAGTTTTGCCAAGGA3′ (forward) and 5′TTTCTGTGTTGGCGCAGTGT3′ (reverse); TNFα, 5′CCTGCCCCAATCCCTTTATT3′ (forward) and 5′CCCCAATTCTCTTTTTGAGCC3′ (reverse); and β-actin, 5′TGTGATGGTGGGAATGGGTCAG3′ (forward) and 5′TTTGATGTCACGCACGATTTCC3′ (reverse).
A ChIP assay was performed using the ChIP-ITTM Express enzymatic kit (Active Motif, catalog no. 53009). The FOXO3a-specific antibody for the ChIP assay was from Santa Cruz Biotechnology (catalog no. sc-11351X). 293T cells were used for the ChIP assay. The primers used for real-time PCR to quantitate the ChIP-enriched DNA were as follows: BCL10 promoter, 5′AAGCTGCAGTGAGCCGAGAA3′ (forward) and 5′TAGGGCTGCCGGGATCATA3′ (reverse), 5′CCGGCAGCCCTAAAAGGTA3′ (forward) and 5′TCCCACCACTTTAACACCTCATT3′ (reverse), and 5′AATGAGGTGTTAAAGTGGTGGGAT T3′ (forward) and 5′CCCAGACCCTGCCCTACAAT3′ (reverse).
The Biotin 3′ end DNA labeling kit and the LightShift chemiluminescent EMSA kit from Pierce were used in the test. The cells were transfected with FOXO3a-TM for 24 h. The nuclear proteins were isolated, and EMSA was performed according to the instructions of the manufacturer. The probes for the FOXO3a DNA binding region (BCL10 promoter) were as follows: 5′-GGTTTCTTTTTTTCTTTCAAAAATTTGTTTAGTATTGTAGTTAGG-3′ (forward) and 5′-CCTAACTACAATACTAAACAAATTTTTGAAAGAAAAAAAGAAACC-3′ (reverse). The mutant probes were 5′-GGTTTCTTTTTTTCTTTCCCGCCGCCGTTTAGTATTGTAGTTAGG-3′ (forward) and 5′-CCTAACTACAATACTAAACGGCGGCGGGAAAGAAAAAAAGAAACC-3′ (reverse). The mutant bases are shown in italics.
The data represent mean ± S.D. from three independent experiments except where indicated. Statistical analysis was performed by Student's t test at a significance level of p < 0.05.
To know the effect of FOXO3a on NF-κB activity, we overexpressed FOXO3aTM in 293T, HeLa, and PC3 cells and determined the activity of NF-κB. These cells are all epithelial. FOXO3aTM is a non-phosphorylatable form of FOXO3a (with mutations of T32A, S253A, and S315A) and, therefore, is constitutively active (9). We found that overexpression of FOXO3aTM increased the activities of NF-κB in these cells (Fig. 1A). Overexpression of FOXO3aTM activated NF-κB in a dose-dependent manner (Fig. 1B). We determined the effect of FOXO3a on the expression of IL-8 and TNFα, the downstream targets of NF-κB, and found that overexpression of FOXO3aTM increased the transcription of IL-8 and TNFα (Fig. 1C). We also determined the effects of inhibition of FOXO3a on activities of NF-κB. The results demonstrate that knockdown of FOXO3a suppressed the activities of NF-κB in 293T, HCT116, and A549 cells (Fig. 1D). Taken together, these results suggest that FOXO3a functions to activate NF-κB.
Next, we determined the possible mechanisms of NF-κB activation by FOXO3a. Nuclear localization of p65 plays an important role in activation of NF-κB. We found that overexpression of FOXO3aTM increased protein levels of p65 in the nucleus (Fig. 2A). Overexpression of FOXO3aTM had little effect on the total level of p65 (Fig. 2B). Thus, the elevated p65 in the nucleus is not due to increased expression of p65. Moreover, we found that knockdown of FOXO3a repressed the nucleus translocation of p65 (Fig. 2C).
It is known that phosphorylation and degradation of IκBα promotes nuclear transport of p65. We therefore determined the effect of FOXO3a on IκBα. We found that overexpression of FOXO3aTM decreased the protein level of IκBα and that knockdown of FOXO3a increased the protein level of IκBα (Fig. 2D). These results suggest that FOXO3a activates NF-κB through IκBα. To further confirm this, we employed the vector encoding IκBαSR (a dominant-negative form of IκBα) in our work. We found that expression of IκBαSR repressed the FOXO3TM-induced activation of NF-κB (Fig. 2E). Thus, FOXO3a may activate NF-κB through repressing IκBα.
IκBα is a downstream target of IKK. IKK phosphorylates IκBα, leading to its proteasomal degradation. To know whether FOXO3a inhibited IκBα through IKK, we overexpressed FOXO3aTM in 293T cells and determined the phosphorylation of IKKα and IKKβ, the subunits of the IKK complex. We found that overexpression of FOXO3aTM induced phosphorylation of IKKα/β (Fig. 2F). Overexpression of FOXO3aTM had little effect on the protein levels of IKKα, IKKβ, and another IKK subunit, IKKγ (Fig. 2F). We also found that expression of IKKβ-DN, the dominant-negative form of IKKβ, repressed the FOXO3aTM-induced activation of NF-κB (Fig. 2G). Taken together, these results suggest that FOXO3a activates NF-κB activity through IKK.
FOXO3a is a transcription factor and functions to activate gene expression. We asked whether the transcriptional activity of FOXO3a was required to activate NF-κB. We employed the vector that encoded FOXO3aΔH3 in our experiment. FOXO3aΔH3 cannot bind to DNA and lose the capacity to activate gene transcription. We found that expression of FOXO3a induced activation of NF-κB but that FOXO3aΔH3 could not (Fig. 3A). In addition, expression of dominant-negative FOXO3a (FOXO3a-DN) inhibited the basal NF-κB activity and attenuated FOXO3aTM-induced NF-κB activity (Fig. 3B). These data suggest that the transcriptional activity is required for FOXO3a to activate NF-κB.
To know the target gene of FOXO3a that was involved in activation of IKK/NF-κB signaling, we checked the promoters of the genes that are upstream regulators of IKK. We found that the BCL10 promoter had three putative FOXO3a-binding sites (Fig. 3C). Previous studies showed that BCL10 protein was crucial for IKK activation (19, 23), and it is known for its role in mediating T-cell receptor-induced NF-κB activation (24, 25). There are also a lot of studies of BCL10 that were performed in non-immune cells, including epithelial cells (26–28), suggesting that BCL10 plays an important role in NF-κB activation in epithelial cells. We then determined the effect of FOXO3aTM on expression of BCL10. We found that expression of FOXO3aTM increased both mRNA and protein levels of BCL10 (Fig. 3D), implying that FOXO3a is a positive regulator of BCL10 expression. Moreover, we found that knockdown of FOXO3a attenuated expression of BCL10 (Fig. 3E). These data provide more evidence that FOXO3a positively regulates expression of BCL10. AKT functions to phosphorylate FOXO3a, leading to FOXO3a inactivation (9). We treated the cells with insulin to induce activation of AKT. Insulin treatment resulted in induction of FOXO3a phosphorylation and repression of BCL10 expression (Fig. 3F). Finally, we determined the effect of BCL10 on NF-κB. Consistent with a previous report, overexpression of BCL10 induced NF-κB activity (Fig. 3G), and knockdown of BCL10 repressed NF-κB (Fig. 3H). Moreover, we found that knockdown of BCL10 attenuated FOXO3a-induced NF-κB activity (Fig. 3I).
We then investigated the regulation of BCL10 transcription by FOXO3a. We found that expression of FOXO3a but not FOXO3aΔH3 increased BCL10 reporter activity (Fig. 4A). The results suggest that DNA-binding capacity is required for FOXO3a to regulate BCL10 expression. The BCL10 promoter has three putative FOXO3a binding sites (Fig. 3C). We constructed a few BCL10 reporter plasmids that cover all or part of the putative FOXO3a-binding consensus sequences (Fig. 4B). We found that overexpression of FOXO3aTM induced a dramatic increase of activities of the BCL10 promoter reporters except BCL10(-141/+1) (Fig. 4C). It suggests that FOXO3a binds to the region of −255 ~ −141 of the BCL10 promoter, probably at the putative sequence of AAATTTGTTT (Fig. 3C). To verify this, we constructed a mutated BCL10(-432/+1) reporter at the possible binding site (AAATTTGTTT to AAACCGGTTT). Mutation at the site significantly reduced the FOXO3a-indcued BCL10 reporter activities (Fig. 4D).
We performed a ChIP assay to determine whether FOXO3a bound to the BCL10 promoter. Three pairs of primers for the BCL10 promoter were designed, and the first pair of primers was better than the others (data not shown). We used the first primer pair to perform the CHIP assay three times, and the results indicated that FOXO3a bound to the BCL10 promoter (Fig. 4E). Addition of insulin blocked the binding of FOXO3a to the BCL10 promoter (Fig. 4E). This might be due to the insulin-induced phosphorylation and nuclear export of FOXO3a, leading to the decrease of BCL10 promoter binding of FOXO3a.
Finally, we performed EMSA to directly show that FOXO3a binds to the region. The probe corresponding to the BCL10 promoter region was designed as described under “Materials and Methods.” The mutant probe (the mutated site was located between −255 ~ −141) was also employed. Our results showed that there was a shift band when the probe was employed. When the mutated probe was used in the test, the shift band was reduced significantly. The results suggest that FOXO3a might bind to the BCL10 promoter region.
We determined whether FOXO3a was involved in TNFα- and IL-1β-induced activation of NF-κB. Our results showed that TNFα and IL-1β treatment induced phosphorylation of FOXO3a, indicating that these factors inactivated FOXO3a (Fig. 5A). Thus, FOXO3a may not be involved in TNFα- and IL-1β-induced NF-κB activation. We found that NF-κB was activated when cells were incubated under serum-free conditions (Fig. 5B). Serum deprivation induced dephosphorylation of FOXO3a. To know whether FOXO3a was involved in serum deprivation-induced activation of NF-κB, we knocked down FOXO3a and determined NF-κB activities under serum-free conditions. We found that knockdown of FOXO3a blocked the induction of NF-κB activities (Fig. 5C). These results suggest that FOXO3a plays a role in serum deprivation-induced activation of NF-κB.
The function of NF-κB regulation by FOXO3a was determined. It is known that during tumor progression an increase in tumor mass is concomitant with serum deprivation (29) and that FOXO3a is activated under starvation conditions. Therefore, we determined the possible function of NF-κB regulation by FOXO3a under serum starvation conditions. HCT116 cells were transfected with siFOXO3a or siFOXO3a+BCL10. The transfected cells were incubated in serum-free medium. We found that knockdown of FOXO3a enhanced cell apoptosis (Fig. 5D). When the cells were transfected with siFOXO3a plus BCL10, cell apoptosis was partially inhibited (Fig. 5D). The results suggest that FOXO3a prevents cells from apoptosis under serum starvation conditions and that it may function, at least partially, via BCL10.
The importance of NF-κB in tumorigenesis has been well studied (30). Activation of NF-κB was found in multiple types of tumors, and it plays a crucial role in promoting tumorigenesis by inducing expression of many genes involved in cell proliferation and survival, tumor invasion, and angiogenesis. NF-κB can be activated through canonical or non-canonical pathways in response to different signals (31). Stimulation of various receptors activates different signaling pathways, leading to activation of IKK. BCL10 is essential for activation of IKK/NF-κB downstream of either the T cell or B cell antigen receptors (19, 23), and it has been characterized as an important component to activate the IKK complex (25, 33). BCL10 was also found to play an important role in NF-κB activation in non-immune cells, including epithelial cells (26–28). In this work, we report that FOXO3a functions to activate IKK/NF-κB signaling through inducing expression of BCL10. Our results demonstrate for the first time that FOXO3a activates NF-κB and disclose the mechanism for the activation of NF-κB by FOXO3a.
A functional role for FOXO3a in human cancer progression has been established by analysis of the signaling pathways that control its transcriptional activity. In response to growth factors such as IGF-1 or genetic lesions in cancer cells such as phosphatase and tensin homolog (PTEN) mutations that activate AKT, FOXO3a is phosphorylated and maintained in an inactive state by nuclear exclusion and cytoplasmic retention. The inactivation of FOXO3a may provide an advantage in proliferation of cancer cells and accelerates tumor growth. The adenovirus-mediated transfer of constitutively active FOXO3a has been reported to induce apoptosis in melanoma cells (34), indicating that direct activation of FOXO3a might be a clinical strategy. FOXO3a may also serve as an indirect therapeutic target (35–37).
Recent studies demonstrate that the targeting of FoxO proteins may be complicated by potential feedback mechanisms. For example, Hui et al. (13) reported that FOXO3a promoted survival of L562 myelogenous leukemia cells by amplifying PI3K/AKT signaling in response to doxorubicin treatment. DAF16 has been shown to induce the expression of the insulin receptor in C. elegans, resulting in increased PI3K activity and cell growth under low-nutrient conditions (38).
In many cell types, activation of FOXO3a leads to cell quiescence but not apoptosis (39, 40). Kops et al. (14) reported that FOXO3a protected quiescent cells from oxidative stress. Jensen et al. (41) found that FOXO3a was activated and that it promoted cell survival in hypoxia. FOXO3a promoted cell survival through inducing metabolic adaptation to hypoxia (41). They showed that FOXO3a knockdown xenograft tumors exhibited impaired growth and increased apoptosis. Consistently, Storz et al. (15) demonstrated recently that knockdown of FOXO3a in breast cancer cells injected into nude mice led to a significant decrease in tumor volume. Our results showed that, under serum-free conditions, knockdown of FOXO3a promoted apoptosis of HCT116 cells (Fig. 5D). Thus, the role of FOXO3a in tumorigenesis is likely to be context-dependent. The inactivation of FOXO3a in the early stage of tumors by increased signaling through growth factors may offer a proliferative advantage. At later stages, limitations in the access to serum in combination with intra-tumoral hypoxia may reactivate FOXO3a and thus promote tumor cell survival. Our results suggest that under serum deprivation conditions the activation of FOXO3a promotes cell survival, at least partly, via the BCL10/NF-κB pathway.
Here, we demonstrated that FOXO3a activated NF-κB via BCL10. It was reported that FOXO3a inhibited activation of NF-κB in helper T cells (32). The difference might be due to the different systems employed in these two works. Fox proteins are potential targets in therapeutic strategies. However, many of these strategies are in their infancy, and evidence that Fox proteins are clinically viable as direct targets for intervention has yet to be obtained (12). And the rationale for either attenuating or promoting individual Fox proteins may be unclear because of the pleiotropic roles of Fox proteins and the systems that they regulate. During tumor progression, an increase in tumor mass is concomitant with hypoxia and serum deprivation. Results from our laboratory and others demonstrated that FOXO3a promoted cell survival under stress conditions such as serum-deprivation, hypoxia, and oxidative stress. Therefore, caution should be taken when FOXO3a is employed as a target for cancer therapy.
*This work was supported by Natural Science Foundation of China Grant 30970586, by Innovation Program of the Chinese Academy of Sciences Grant KSCX2-EW-R-08, and by Chinese Ministry of Science and Technology Grant 2007BAC27B02.
3The abbreviations used are: