|Home | About | Journals | Submit | Contact Us | Français|
Serine/threonine kinase Akt regulates key cellular processes such as cell growth, proliferation, and survival. Activation of Akt by mitogenic factor depends on phosphatidylinositol 3-kinase (PI3K). Here, we report that IKBKE (also known as IKKϵ and IKKi) activates Akt through a PI3K-independent pathway. IKBKE directly phosphorylates Akt-Thr308 and Ser473 independent of the pleckstrin homology (PH) domain. IKBKE activation of Akt was not affected by inhibition of PI3K, knockdown of PDK1 or mTORC2 complex. Further, this activation could be inhibited by Akt inhibitors MK-2206 and GSK690693 but not the compounds (perifosine and triciribine) targeting the PH domain of Akt. Expression of IKBKE largely correlates with activation of Akt in breast cancer. Moreover, inhibition of Akt suppresses IKBKE oncogenic transformation. These findings indicate that IKBKE is an Akt-Thr308 and -Ser473 kinase and directly activates Akt independent of PI3K, PDK1, and mTORC2 as well as PH domain. Our data also suggest that Akt inhibitors targeting the PH domain have no effect on the tumors in which hyperactive Akt resulted from elevated IKBKE.
Akt has been shown to be a major cell survival and growth pathway by regulation of a number of proteins (1–5). Growth factor-induced Akt activation has been well documented and shown to depend on the integrity of the pleckstrin homology (PH)2 domain, which mediates its membrane translocation, and on the phosphorylation of Thr-308 and Ser-473 (6–9). Phosphoinositides, PtdIns-3,4-P2 and PtdIns-3,4,5-P3, produced by phosphatidylinositol 3-kinase (PI3K) bind directly to the PH domain of Akt (7), driving a conformational change in the molecule, which enables the activation loop of Akt to be phosphorylated by PDK1 at Thr-308 (10). However, full activation of AKT is associated with phosphorylation of Ser-473 (11) within a C-terminal hydrophobic motif characteristic of kinases in the AGC kinase family. Although the role of PDK1 in Thr-308 phosphorylation is well established, the mechanism of Ser-473 phosphorylation is controversial. A number of candidate enzymes responsible for this modification have been put forward, including integrin-linked kinase (12), DNA-dependent kinase (13), and the rictor-mTOR (mTORC2) complex (14). Recent studies indicated that mTORC2 phosphorylation of Akt-Ser473 also requires Akt translocation to plasma membranes through mTOR binding to small GTPase Rac1 (15). The activation of Akt by growth factor is negatively regulated by tumor suppressor PTEN, encoding a dual-specificity protein and lipid phosphatase that reduces intracellular levels of PtdIns-3,4,5-P3 by converting them to PtdIns-4,5-P2 (16). Accumulating studies demonstrated that Akt is activated by inflammatory agents lipopolysaccharide (LPS) and phorbol myristate acetate (PMA). However, the underlying mechanism remains elusive. For instance, Salh et al. (17) showed that both PI3K and Akt were activated by LPS, but LY294002 and Wortmannin inhibited the activation of PI3K but not Akt, suggesting that other signal molecule(s) mediates the Akt activation induced by inflammatory agent.
IκB kinase (IKK) regulates a wide variety of cell functions through activating NF-κB pathway, which links inflammation to tumor formation and progression (18–20). The IKK family contains the serine/threonine kinase IKKα, IKKβ, IKBKE, TBK1, and an adaptor protein IKKγ. In the canonical pathway, inflammatory stimulation induces IKKα, IKKβ, and adaptor protein IKKγ to form a complex which phosphorylates IκB-Ser32/Ser36 leading to degradation of IκB and subsequent activation of NF-κB (21). IKBKE is a non-canonical IKK family member and plays an important role in the regulation of inflammatory signaling pathway. In response to inflammatory factors such as LPS and PMA, IKBKE is activated and phosphorylates p65/RelA, interferon response factors 3 and 7 (IRF3 and IRF7) and STAT1 (21–23). A recent study also showed that activated IKBKE regulates CYLD to activate the canonical NF-κB pathway (24). Moreover, other signaling molecules beyond NF-κB cascade regulated by IKK family members have recently been emerged for their critical role in oncogenesis. IKKα and IKKβ could directly phosphorylate FOXO3a and mTOR to induce cell proliferation and survival (25–27). Activated IKBKE induces tamoxifen resistance by phosphorylation and activation of ERα (28). It is noted that among IKK family members, only IKBKE has been shown to be frequently altered in human cancers (29–31) and functions as an oncogene to substitute Akt for malignant transformation (29).
Here, we present evidence that IKBKE is a critical regulator of Akt. Depletion of IKBKE results in a decrease of Akt activity. IKBKE activates Akt by direct phosphorylation of Thr-308 and Ser-473. Notably, PI3K, PDK1 and mTORC2 as well as PH domain of Akt are dispensable for IKBKE-induced Akt activation. Inhibition of Akt largely reduces oncogenic transformation of IKBKE. Thus, Akt is a bona fide substrate of IKBKE and mediates IKBKE oncogenic activity.
The breast cancer cell lines (MCF7, MDA-MB-231) were purchased from ATCC and lung cancer cell lines (H1299, H157) were obtained from Lung Cancer Cell Core at Moffitt Cancer Center. The cells were maintained in RPMI 1640 medium supplemented with 5% fetal bovine serum and 100 units/ml penicillin/streptomycin. Ikbke-knock-out mouse fibroblast (MEF-Ikbke−/−) and wild-type MEF (MEF-W) were kindly provided by Tom Maniatis (Harvard University). PDK1-knock-out HCT116 and the parental cells are gifts from Bert Vogelstein (Johns Hopkins University). These cells were maintained in DMEM medium supplemented with 10% fetal bovine serum. Cell transfection was carried out with either ExGen (Fenmentas Life Science) or Lipofectamine 2000 (Invitrogen). Fresh-frozen primary human breast cancer specimens were obtained from patients who underwent surgery at Moffitt Cancer Center, and were approved by Institutional Review Board. Each sample contained at least 80% tumor cells, confirmed by microscopic examination.
The IKBKE, IKKα, IKKβ, and Akt plasmids were previously described (31, 32). pRK5-myc-TBK1 was kindly provided by Dr. Ivan Dikic (Goethe University Medical School, Germany). ΔPH-Akt was created by deletion of PH domain of HA-Akt. Lentiviral short hairpin (sh)RNAs targeting IKBKE or Akt1/2/3 were purchased from Open Biosystems and the siRNAs of mTOR, Rictor and IKBKE were from Siegen. The antibodies anti-IKBKE, -my, -Flag, and -HA were from Sigma. Anti-pAkt-S473, -pAkt-Th308, -pGSK3β, -pFOXO3a, -pTSC2, and -p-p70S6K were purchased from Cell Signaling. Recombinant proteins IKBKE and Akt were from Invitrogen and New England Peptide, respectively. Akt inhibitors MK-2206, perifosine, and GSK690693 were from Sellect Chemical and Sigma.
Western blot, immunoprecipitation, and IHC were performed as previously described (33). Briefly, cell lysates were prepared in lysis buffer and subject to immunoprecipitation and immunoblots. Each experiment was repeated three times. Phospho-Akt in autoradiograph was quantified relative to total Akt using ImageQuant software. For IHC, breast cancer TMAs were immunostained with anti-IKBKE (1:250) and -pAkt-Ser473 (1:200) antibodies. The expression of IKBKE and pAkt was evaluated and scored as previously described (31).
In vitro IKBKE kinase assay was performed as described previously (28). Briefly, the reaction was carried out in a kinase buffer in the presence of 10 μCi of [γ-32P]ATP using recombinant Akt as substrate. After incubation at 30 °C for 30 min, the reaction was stopped by adding protein-loading buffer and separated in SDS-PAGE gels. The relative amounts of incorporated radioactivity were determined by autoradiography and quantitated with a Phosphoimager (Molecular Dynamics). In vitro Akt kinase assay was carried out as described (34). For in vivo labeling, HEK293 cells were transfected with HA-Akt-K179M or HA-Akt-K179M-Thr308/Ser473–2A together with or without myr-IKBKE. After serum starvation overnight, cells were labeled with [32P]Pi (0.5 mCi/ml) in phenol red-free MEM without phosphate for 4 h. Akt was immunoprecipitated with anti-HA antibody. The immunoprecipitates were separated on SDS-PAGE and transferred to membrane; the phosphorylated Akt was detected by autoradiography and quantified. Each in vitro kinase and in vivo labeling experiment was repeated three times.
Mass spectrometry was performed as previously described (28). Briefly, in vitro IKBKE/DN-Akt kinase reactions were separated in SDS-PAGE and DN-Akt bands were excised. Following wash and digestion with trypsin, peptides were extracted and subjected into a nanoflow liquid chromatograph that coupled to an electrospray hybrid ion trap mass spectrometer (LTQ Orbitrap, Thermo, San Jose, CA). Oxidized methionine, deamidation, carbamidomethyl cysteine, and phosphorylated serine, threonine, and tyrosine were selected as variable modifications, and as many as 2 missed cleavages were allowed. Assignments were manually verified by inspection of the tandem mass spectra and coalesced into Scaffold.
After stable transfection with myr-IKBKE together with and without DN-Akt and shRNA/Akt, NIH3T3 cells were plated in triplicate in 0.4% Noble agar and DMEM containing 10% bovine calf serum. Colonies were counted following culture for 3 weeks. For tumorigenecity, IKBKE transformed and pcDNA3 vector-transfected (control) NIH 3T3 cells were subcutaneously injected into nude mice (1 × 106/mouse, 8 mice/group) and tumor growth was monitored daily for 6 weeks. At the end of experiment, tumor weight was calculated.
Differences between control and testing cells were evaluated by Student's t test. The correlation of IKBKE expression with Akt activation in breast cancer specimens was analysis with Chi-square test. All analyses were completed with SPSS software, version 10.0. p < 0.05 was considered statistically significant.
A recent study has shown that constitutively active IKBKE could substitute Myr-Akt to transform hTERT/SV40/MEKDD immortalized human mammary epithelial cells (29), suggesting a direct link between IKBKE and Akt. To test this, we performed in vitro kinase and immunoblotting assays and found that the kinase activity and phospho-Thr308 and -Ser473 of Akt were induced by expression of constitutively active (Myr) or wild-type IKBKE but not dominant-negative IKBKE (Fig. 1, A–C and supplemental Fig. S1). Phosphorylation levels of Akt downstream targets p70S6K and GSK3β were also elevated upon expression of IKBKE (Fig. 1, A–C). Further, IKBKE activated Akt in a dose-dependent manner (Fig. 1, B and C). Knockdown of IKBKE considerably reduced the kinase activity and phospho-Thr308/Ser473 levels of Akt (Fig. 1, D and E and supplemental Fig. S2). Moreover, we observed increase of phospho-Akt-Thr308/Ser473 levels by restoration of IKBKE in Ikbke-knock-out MEFs (Fig. 1F). These data indicate that IKBKE is a key upstream activator of Akt.
We next investigated whether IKBKE phosphorylates Akt. In vitro kinase assay was performed by incubation of recombinant wild-type Akt1 and Akt2 proteins, as substrates, with and without recombinant IKBKE. Fig. 2A shows that Akt1 and Akt2 were highly phosphorylated in the presence of IKBKE but the autophosphorylation of Akt is barely detectable, suggesting that Akt is a substrate of IKBKE. To determine IKBKE phosphorylation site(s) of Akt, we carried out in vitro IKBKE kinase assay using truncated Akt (GST-PH, -KD, and -RD) GST fusion proteins as substrates (supplemental Fig. S3). Mass spectrometry analysis of these in vitro kinase reactions revealed three putative phosphorylation sites of Akt induced by IKBKE, including Ser-137 (supplemental Fig. S4A), Thr-308 and Ser-473. To ascertain if IKBKE directly phosphorylates Thr-308 and Ser-473 of Akt, we immunoblotted recombinant Akt1 and Akt2 following incubation with/without recombinant IKBKE and found that pAkt-Thr308 and -Ser473 were robustly induced by IKBKE (Fig. 2B). Furthermore, in vitro kinase showed that IKBKE phosphorylated immunoprecipitated DN-Akt (K179M) but not DN-Akt-2A (T308/S473A; Fig. 2C). In addition, in vivo labeling revealed that DN-Akt, in consistence with in vitro data, was phosphorylated by IKBKE whereas mutation of Thr-308 and Ser-473 to alanine (DN-Akt-2A) abrogated the phosphorylation (Fig. 2D), suggesting Thr-308 and Ser-473 are the primary sites phosphorylated by IKBKE. Immunoblotting analysis further revealed that IKBKE phosphorylated Thr-308 and Ser-473 in vivo and all three members of Akt (Fig. 2, E and F).
We also investigated if IKBKE phosphorylates Ser-137, which is conserved in Akt3 but not Akt2. We created Akt1aa1–200 and Akt1aa1–200-S137A mutants. In vitro kinase assay revealed IKBKE phosphorylation of Akt1aa1–200 but not Akt1aa1–200-S137A (supplemental Fig. S4B). However, full-length Akt1-S137A remains to be fully activated by IKBKE as compared with wild-type Akt1 (supplemental Fig. S4C), suggesting that the phosphorylation of Ser-137 is not required for IKBKE activation of Akt. Based on these results, we concluded that Akt is a bona fide substrate of IKBKE and that IKBKE activates Akt through primarily direct phosphorylation of Thr-308 and Ser-473.
It has been well documented that PI3K transduces extracellular signals to activate Akt (35, 36). Thus, we investigated if IKBKE-activated Akt depends on PI3K. Following introduction of wild-type or constitutively active IKBKE into Ikbke-knock-out MEFs, the cells were serum-starved and treated with and without PI3K kinase inhibitor. Vector-transfected Ikbke−/− MEFs treated with insulin were used as controls. As expected, insulin-induced Akt activation was inhibited by LY294002. However, the PI3K inhibitor had no effect on IKBKE-activated Akt (Fig. 3A). To further confirm this observation, H1299 cells expressing low levels of endogenous IKBKE were transfected with IKBKE and were treated with two different PI3K inhibitors or API-2/triciribine, an Akt inhibitor binding to PH domain and preventing Akt membrane translocation (34, 37). Fig. 3B shows that neither LY294002 nor Wortmannin inhibited IKBKE-induced Akt activation. Interestingly, Akt inhibitor API-2 also had no effect on IKBKE activated Akt, while it inhibited insulin-stimulated Akt (Fig. 3B). In addition, expression of dominant-negative p85α, a regulatory subunit of PI3K, reduces insulin but not IKBKE-induced Akt activation (Fig. 3C). These data suggest that IKBKE activation of Akt is independent of PI3K, and that small molecule inhibitor targeting PH domain of Akt might not inhibit IKBKE-activated Akt.
Whereas previous studies demonstrated that PH domain of Akt is required for Akt activation (37, 38), our data show that neither PI3K nor the PH domain disruption of Akt inhibitor API-2 had effect on IKBKE activation of Akt, suggesting that PH domain of Akt is not required for IKBKE-induced Akt activation. To test this, PH domain deletion mutant Akt (ΔPH-Akt) was introduced into H1299 cells together with and without IKBKE. Immunoblotting analysis revealed that both wild type and constitutively active IKBKE induced phosphorylation of Ser-473 and Thr-308 of ΔPH-Akt and endogenous Akt. As expected, insulin only stimulates endogenous Akt (Fig. 4A). To further confirm these findings, we took a pharmacological approach by treating the cells with different Akt inhibitors. Fig. 4B shows that IKBKE-activated Akt was inhibited by MK-2206, an allosteric Akt inhibitor, and GSK690693, an ATP-competitive Akt kinase inhibitor but not by perifosine, an alkylphospholipid blocking PH domain binding to the plasma membrane, while all of them inhibited insulin-induced Akt (39, 40). Collectively, these data indicate that PH domain of Akt is dispensable for IKBKE-stimulated Akt and that activation of Akt by different mechanisms in human tumors requires different Akt inhibitors. This is clinically important because Akt inhibitors perifosine and API-2/TCN are currently in clinical trial (41, 42) and have no effects on Akt activation induced by IKBKE.
Because PDK1 and mTORC2 are responsible for phosphorylation of Thr-308 and Ser-473, respectively, we also investigated whether IKBKE-stimulated Akt depends on PDK1 and mTORC2. We transfected PDK1-null and parental HCT116 cells (43) with IKBKE and observed that phosphorylation levels of Thr308 induced by IKBKE were no difference between PDK1−/− and PDK1+/+ cells (Fig. 4C). For mTORC2 complex, Rictor and mTOR were knocked down in H1299 cells. Immunoblotting analysis revealed that depletion of either Rictor or mTOR had no effects on IKBKE-induced phosphorylation of Ser-473 (Fig. 4D). These findings indicate IKBKE activation of Akt independent of PDK1 and mTORC2.
Having demonstrated IKBKE direct activation of Akt in cell culture system, we asked if this regulation is seen in vivo. We examined IKBKE protein and pAkt levels in primary breast tumor samples (Fig. 5, A and B). Of the 98 tumors, 54 had overexpression of IKBKE and 64 had increase of pAkt. Of the 54 tumors with elevated IKBKE, 40 (74%) also had elevated pAkt levels (p = 0.035, Fig. 5C). Immunohistochemistry of these tumor samples showed that the co-expression of IKBKE and pAKT are located specifically to the cancer cells and not to the stroma (Fig. 5B). These data further support the findings of IKBKE phosphorylation/activation of Akt.
To determine the functional link between IKBKE and Akt, we transfected constitutively active IKBKE together with and without dominant-negative Akt or shRNA specific for Akt into NIH3T3 fibroblasts (Fig. 6A). After selection, stable Myr-IKBKE clonal cells grew colony in soft agar and formed tumor in nude mice (Fig. 6, B and C). However, co-expression of dominant-negative Akt or knockdown of Akt considerably reduced colony formation, tumor growth and tumor weight induced by Myr-IKBKE (Fig. 6, B–D). In agreement with a previous report that IKBKE and Akt could substitute each other for transformation (29), our finding places Akt as a downstream target of IKBKE and suggests that Akt mediates IKBKE oncogenic function.
It is well documented that activation of Akt by mitogenic factor depends on PI3K whose product PtdIns-3,4,5-P3 binds to the PH domain of Akt leading to the Thr308 phosphorylation by PDK1. We report here that IKBKE directly phosphorylates Akt-Thr308 and -Ser473 and plays a critical role in Akt activation. Moreover, IKBKE activation of Akt is independent of PI3K. Thus, our findings provide a novel mechanism of Akt activation, i.e. IKBKE direct phosphorylation of Thr-308 and Ser-473 to activate Akt.
Accumulating evidence indicates that phosphorylation of Thr-308 and Ser-473 is required for activation of Akt. PDK1 and mTORC2 are responsible for phosphorylation of Thr-308 and Ser-473 in response to growth factor stimulation, respectively (14). In current report, we found that IKBKE does not mediate growth factor-induced Akt (supplemental Fig. S5A). However, knockdown of IKBKE significantly reduces pAkt-Thr308 and -Ser473 levels in different cell lines (Fig. 1E and supplemental Fig. S2). IKBKE phosphorylates Thr-308 and Ser-473 in vitro and in vivo (Fig. 2). Knockdown of PDK1 and mTORC2 had no effect on IKBKE phosphorylation of Akt-Thr308 and -Ser473 (Fig. 4, C and D). In addition, it has been demonstrated that DNA-dependent protein kinase (DNA-PK) is an Akt-Ser473 kinase in response to stress (13). However, expression of IKBKE induces pAkt-S473 and -Thr308 in DNA-PK−/− cells (data not shown). These data indicate that IKBKE is a bona fide Thr308/Ser473 kinase and plays a pivotal role in Akt activation.
A recent study has shown that TBK1, which exhibits 49% identity and 65% similarity to IKBKE, also directly activates Akt and mediates glucose-induced Akt activation (44). Further, Xie et al. (45) demonstrated TBK1 and IKBKE activation of Akt in a manner dependent on PI3K signaling. However, we did not find that IKBKE is required for glucose- and nutrient-activated Akt (supplemental Fig. S5B). In agreement with TBK1 activation of Akt reported by Ou et al. (44), we present the evidence showing that IKBKE-activated Akt was not interfered by either pharmacological PI3K inhibitors (LY294002 and Wortmannin) or dominant-negative p85α (Fig. 3). Moreover, we did not observe activation of Akt by other IKK family members e.g. IKKα and IKKβ (supplemental Fig. S6).
The significance of IKBKE direct activation of Akt resides in up-regulation of IKBKE in human malignancy. Hyperactivation of Akt is observed in more than 50% human cancers, which has been shown to be resulted from activated mutation of upstream molecules such as PIK3CA and PTEN (46, 47). We demonstrated in this study a close correlation between overexpression of IKBKE and elevated pAkt in breast cancer, indicating that aberrant expression of IKBKE is a causal factor of Akt activation in human cancer. More importantly, IKBKE-induced Akt activation resists to Akt inhibitors that target PH domain of Akt such as perifosine and API-2/TCN, but not Akt inhibitors targeting Akt kinase and/or regulatory domains which include GSK690693 and MK-2206. These findings are clinically important because these inhibitors are currently in clinical trials (40, 48) and numerous efforts have been focusing on development of Akt inhibitor by targeting PH domain which should not work in the tumors with hyperactive Akt resulted from up-regulation of IKBKE. Furthermore, since the activation of Akt by IKBKE is independent of PI3K, PI3K inhibitors would not have effect on the patients whose tumors carry IKBKE-activated Akt.
In conclusion, recent evidence suggests that IKBKE is an oncogenic kinase participating in malignant transformation and tumor development (29). The NF-κB pathway has been shown to be a major target of IKBKE to mediate IKBKE oncogenic function (29). Our study identified Akt as a bona fide substrate of IKBKE. IKBKE activates Akt by direct phosphorylation of Akt-Thr308 and -Ser473 which is independent of PI3K, PDK1 and mTORC2 as well as PH domain of Akt. Thus, Akt inhibitors targeting the PH domain had no effect on IKBKE-induced Akt activation. Further, we demonstrated a functional link between up-regulation of IKBKE and Akt activation in breast cancer. These findings reveal additional mechanism of Akt activation and suggest IKBKE as a therapeutic target for cancer intervention.
We thank the Tissue Procurement, DNA Sequence, Proteomics, and Image Core Facilities at H. Lee Moffitt Cancer Center for providing cancer specimens, sequencing, and cell apoptosis analysis. We also thank Moffitt Cancer Center Lung Cancer SPORE for providing NSCLC cancer cell lines.
*This work was supported, in whole or in part, by National Institutes of Health Grant CA137041 and Florida James & Esther King Biomedical Research Program 1KG02 (to J. Q. C.) and 10KD-04 (to J.-P. G.).
2The abbreviations used are: