Akt Modulates STAT3-mediated Gene Expression through a FKHR (FOXO1a)-dependent Mechanism*

The phosphatidylinositol 3-kinase/Akt pathway plays an important role in the signaling of insulin and other growth factors, which reportedly attenuate the interleukin-6 (IL-6)-mediated stimulation of acute phase plasma protein genes. We investigated the effect of the protein kinase Akt on IL-6-mediated transcriptional activation. The transient expression of constitutively active Akt inhibited the IL-6-dependent activity of the α2-macroglobulin promoter in HepG2 cells, whereas expression of an inactive mutant of phosphatidylinositol-dependent kinase 1 had the opposite effect. Since Akt is known to regulate gene expression through inactivation of the transcription factor FKHR (forkhead in rhabdomyosarcoma), we examined the effect of FKHR on STAT3-mediated transcriptional regulation. Indeed, the overexpression of FKHR specifically enhanced the activity of STAT3-dependent promoters but not that of a STAT5-responsive promoter. The effect of FKHR required the presence of functional STAT3 and was abrogated by the expression of dominant negative STAT3 mutants. Furthermore, FKHR and STAT3 were shown to coimmunoprecipitate and to colocalize in the nuclear regions of IL-6-treated HepG2 cells. Our results indicate that FKHR can modulate the IL-6-induced transcriptional activity by acting as a coactivator of STAT3.

IL-6 1 is the major regulator of acute phase protein (APP) synthesis by the liver during the inflammatory response (1). It exerts its actions through binding to the receptor complex consisting of a ligand-specific IL-6R ␣-chain (gp80) and two signal-transducing ␤-subunits (gp130). Activation of the gp130associated Janus kinases Jak1, Jak2, and Tyk2 results in the tyrosine phosphorylation of several cellular substrates, including signal transducer and activator of transcription 3 (STAT3), the major mediator of IL-6-induced signaling (2,3). Phosphorylated STAT3 dimerizes and translocates to the nucleus, where it regulates the transcription of multiple target genes.
The STAT3-dependent action of IL-6 appears to be modulated by a variety of stimuli, including insulin and epidermal growth factor that have been reported to inhibit the APP production by cultured hepatic cells (4,5). The potential inhibitory role of growth promoting signals on the IL-6-inducible Jak/ STAT3 pathway corresponds well with the suppressed acute phase response in regenerating liver (6). However, the mechanisms whereby growth factors mediate this effect remained unclear. One of the major effects of signaling via the insulin receptor and other growth factor receptors is the activation of PI 3-kinase (7). Generation of PI 3-phosphorylated lipids in the plasma membrane leads to phosphorylation and activation of the serine/threonine kinase Akt (also called protein kinase B) by phosphatidylinositol-dependent kinase 1 (PDK1). Activated Akt has been described to translocate to the nucleus (8) and to directly phosphorylate members of the forkhead family of transcription factors (9 -11). Phosphorylation of FKHR or closely related FKHRL1 and AFX by Akt results in their transcriptional inactivation and retention in the cytoplasm (9,12).
In the present study, we investigated a potential cross-talk between the PI 3-kinase/Akt signaling and the IL-6-inducible Jak/STAT3 pathway. We have identified FKHR as a specific transcriptional coactivator of STAT3. This functional interaction reflects the association of both proteins and their colocalization in nuclear regions of HepG2 cells.
Transient Transfections and Luciferase Reporter Gene Assays-HepG2 cells were transiently transfected using FuGene6 reagent (Roche Molecular Biochemicals) or by the calcium phosphate method as described previously (16). 24 h after transfection, cells were stimulated with cytokines for another 18 h. Luciferase activities were determined with the Luciferase Assay System (Promega), and the data were normalized according to co-expressed ␤-galactosidase (pCH110) or Renilla luciferase (pRLTK) activities. All of the experiments were repeated at least three times with similar results. Shown are the means Ϯ S.D. of one representative experiment performed in triplicate.
Northern Blots-Total RNA was extracted from HepG2 cells with the RNeasy kit from Qiagen. 10 g of RNA were separated by 1% agarose gel electrophoresis and transferred onto nylon membranes (Nytran; Schleicher and Schü ll). Detection with the 32 P-labeled ␣ 2 -macroglobulin (␣ 2 -M) probe was performed as described previously (18), and the signals were detected using a Personal Molecular Imager FX (Bio-Rad).
Cell Fractionation, Immunoprecipitation, and Western Blot Analysis-Nuclear extracts were prepared as described (13) with modifications. Cells were lysed in hypotonic buffer A (10 mM Hepes/KOH, pH 7.9, 1.5 mM MgCl 2 , 10 mM KCl, 1 mM Na 3 VO 4 , 0.2 mM phenylmethylsulfonyl fluoride) for 10 min at 4°C and centrifuged at 300 ϫ g for 2 min at 4°C. The crude nuclei were incubated in lysis buffer B (1% (w/v) BRIJ-97, 20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 10 mM sodium fluoride, 1 mM EDTA, 1 mM Na 3 VO 4 , 1 mM phenylmethylsulfonyl fluoride, 5 g/ml aprotinin, and 5 g/ml leupeptin) for 30 min at 4°C. Short sonication and centrifugation at 14,000 ϫ g for 2 min at 4°C yielded the nuclear extracts. Total cellular lysates were prepared by direct lysis of the harvested cells in buffer B and processed as described above. Protein extracts were incubated overnight at 4°C with polyclonal antiserum against STAT3 (C-20), SMAD2/3 (N-19), or extracellular signalregulated kinase 2 (C-14; Santa Cruz Biotechnology, Inc., Santa Cruz, CA) or monoclonal antibodies against STAT5 (MGF; Transduction Laboratories). The precipitates were later collected with Protein A-Sepharose (Amersham Biosciences), washed three times with lysis buffer, and resolved on 7.5% SDS-PAGE gels. After transfer to polyvinylidene difluoride membranes (GelmanSciences), the blots were probed with the respective antibodies and detected for signals using the ECL system (Amersham Biosciences).
Immunocytochemistry and Confocal Fluorescence Microscopy-Cells grown on coverslips to subconfluence were serum-deprived for the last 12 h, fixed, and permeabilized as described before (20). For intracellular staining, polyclonal anti-STAT3 antibodies (C-20; Santa Cruz Biotechnology) or rabbit antiserum against FKHR and then secondary donkey polyclonal fluorescein isothiocyanate-or rhodamine-coupled antibodies were used (Santa Cruz Biotechnology). After mounting, fluorescence images were visualized by confocal laser-scanning microscopy (LSM 510; Apochromat ϫ63 objective lens; Zeiss). Argon and helium-neon lasers were switched between the excitation wavelengths for fluorescein ( ex ϭ 488 nm) and rhodamine ( ex ϭ 543 nm) fluorescences, which were detected using a 505-530-nm band pass or 560-nm long pass emission filter, respectively.

Activation of PDK1/Akt Signaling Inhibits the IL-6-mediated
Gene Activation through a FKHR-dependent Mechanism-In agreement with previous studies (4, 5), we found that insulin inhibits the IL-6-induced expression of several acute phase plasma proteins, including ␣ 2 -M in hepatoma cells. The levels of ␣ 2 -M mRNA in IL-6-treated HepG2 hepatoma cells were significantly decreased in the presence of insulin (Fig. 1A). One of the well known mediators of insulin signaling is PI 3-kinase, which further downstream activates the serine/threonine ki-nase Akt (7). Interestingly, the inhibitory effect of insulin on ␣ 2 -M expression could be partially relieved by an inhibitor of PI 3-kinase, wortmannin (Fig. 1A). As shown in Fig. 1B, similar effects were also observed on the level of ␣ 2 -M promoter activity using a fragment of the 5Ј-regulatory sequence of the rat ␣ 2 -M gene, which is known for its high sensitivity to STAT3-mediated signaling (16). Therefore, we decided to investigate a potential cross-talk between the PI 3-kinase/Akt pathway and the IL-6-inducible Jak/STAT3 signal transduction cascade. Transfection of HepG2 cells with a constitutively active variant of Akt resulted in a significant reduction of the IL-6-mediated activation of a transfected ␣ 2 -M promoter-luciferase reporter gene construct ( Fig. 2A), comparable with the degree of inhibition by insulin (Fig. 1B). Consistently, overexpression of a kinase-dead variant of PDK1 exerted the opposite effect and augmented the reporter gene activity. The effects observed using wild type forms of both kinases were less pronounced (data not shown). These results suggested that Akt kinase activity can negatively influence IL-6-dependent gene expression.
The protein kinase Akt is known to directly phosphorylate and inhibit transcription factors of the forkhead family (9, 21). Therefore, we asked whether FKHR (FOXO1a), a forkhead transcription factor that has been reported to be regulated in an insulin-dependent manner in hepatic cells (22), is involved in ␣ 2 -M promoter regulation. Transfection of HepG2 cells with increasing amounts of FKHR cDNA led to a marked and dosedependent increase of the IL-6 responsiveness of the ␣ 2 -M promoter (Fig. 2B) without a significant effect on the promoter activity in untreated cells. In order to confirm that FKHR links Akt to IL-6 signaling, we examined whether its overexpression affects the inhibitory action of PDK1 and Akt. As shown in Fig.  2C (left panel), wild type PDK1 decreased the IL-6-induced reporter gene activity by about 50% in FKHR-overexpressing cells, whereas the kinase-dead PDK1 mutant produced the opposite effect. In addition, overexpression of FKHR remarkably enhanced the effect of Akt. Whereas wild type Akt, similarly to wild type PDK1, decreased IL-6-mediated gene expression, expression of constitutively active Akt further reduced the reporter gene activity to about 25% of the control level (Fig. 2C, right panel).
The inactivation of FKHR by Akt results from a direct phosphorylation of three regulatory sites Thr 24 , Ser 256 , and Ser 319 (10,23,24), which in turn leads to cytoplasmic retention of FKHR (9,12). In order to investigate the influence of FKHR phosphorylation on ␣ 2 -M promoter activity, we compared the effects of wild type FKHR and mutated FKHR variants lacking the first, the first two, or all three phosphorylation sites (Fig.  2D). All three examined FKHR mutants were more efficient than wild type FKHR in augmenting the IL-6-induced ␣ 2 -M promoter activity. Interestingly, mutation of threonine 24 to alanine was already sufficient for a nearly maximal effect; introduction of additional point mutations (S256A and S319A, respectively) did not lead to a significant further increase in promoter activity. These results indicate that PDK1/Akt signaling can modulate the transcriptional activation of IL-6responsive genes by targeting FKHR.
IL-6 Does Not Induce Akt Activity and FKHR Phosphorylation in HepG2 Cells-IL-6 has been reported to activate Akt in Hep3B and different cells from myeloma patients (25,26). In this case, IL-6 stimulation should result in nuclear translocation of activated STAT3 and concomitant repression and nu-clear exclusion of FKHR. However, whereas insulin induced a robust and sustained activation of Akt, IL-6 did not significantly stimulate Akt kinase activity over its basal level in HepG2 cells (Fig. 3A). The low Akt activity after IL-6 stimulation was reflected by the lack of its phosphorylation on two critical regulatory residues, Thr 308 and Ser 473 (Fig. 3B). This also correlated with the low phosphorylation status of Thr 24 and Ser 256 , two residues of FKHR, which become directly targeted by Akt (Fig. 3C). In contrast, Akt and FKHR were both strongly phosphorylated after treatment of the cells with insulin, and their activation was not significantly affected by the presence of IL-6. However, as expected, IL-6 strongly induced tyrosine phosphorylation of STAT3, the major transcription factor mediating IL-6-dependent signal transduction (2, 3) (Fig.  3D). The pattern of Akt, FKHR, and STAT3 phosphorylation in HepG2 cells after 16 h of IL-6 and/or insulin treatment was qualitatively similar, although the phosphorylation of STAT3 was less prominent (data not shown). We conclude that in HepG2 cells, IL-6 cannot induce Akt activity and consequently does not repress FKHR.
Functional Interaction between FKHR and STAT3 in Transcriptional Regulation-In order to clarify the role of FKHR for the regulation of the ␣ 2 -M promoter, we investigated whether FKHR modifies the function of STAT3, the crucial mediator of IL-6 signaling (2, 3). Coexpression of STAT3 and FKHR in HepG2 cells indeed synergistically increased the responsiveness of the ␣ 2 -M promoter to IL-6 stimulation (Fig. 4A). In contrast, FKHR expression had no effect on the transcriptional activation of the ␤-casein promoter by STAT5, another member of the STAT family of transcription factors (Fig. 4B). Likewise, FKHR did not enhance the activity of transcription factors unrelated to STATs such as SMADS, the mediators of tumor growth factor-␤ signaling (data not shown). Hence, FKHR appears to be a specific transcriptional coactivator of STAT3responsive promoters.
The stimulatory effect of FKHR on the ␣ 2 -M promoter activation essentially depends on the presence of functional STAT3, since overexpression of dominant negative STAT3 factors (STAT3F or STAT3D, respectively) (28) almost completely abrogated the FKHR-mediated induction of ␣ 2 -M promoter activity (Fig. 4C). This is also consistent with the observation that expression of FKHR efficiently up-regulates IL-6 responsiveness of an artificial promoter, comprising a tandem of isolated STAT3 consensus binding sites (Fig. 4D). These results suggest that the transcriptional effects of FKHR are indirect and result from an enhanced activity of STAT3.
The Complete C Terminus of FKHR Is Crucial for the Coactivation of IL-6-dependent Gene Expression-The PAX3-FKHR fusion protein from alveolar rhabdomyosarcoma was shown to possess a strong transactivation domain localized in its Cterminal FKHR-derived part (29,30). We found that the complete C-terminal part of FKHR was also required for its effect on the ␣ 2 -M promoter activation (Fig. 4, E and F). FKHR was recently shown to interact in HepG2 cells with the coactivator p300/CREB-binding protein (CBP), which is also essential for its transcriptional activity (31). STAT3 is also capable of recruiting p300/CBP, but this interaction and the level of transactivation are relatively weak in comparison with other STATs like STAT2 (32). It is widely accepted that activating proteins like the STATs most often do not act alone but rather in combination with other site-specific or more general DNAbinding proteins as well as with coactivators of transcription (33,34). Therefore, the transcriptional function of STAT3 might be further enhanced by FKHR acting as an accessory factor that directs coactivator complexes to STAT3 sites in the promoter.

FKHR and STAT3 Colocalize in the Nuclei of HepG2
Cells-In order to estimate if both proteins may interact in vivo, we assessed the localization of endogenous STAT3 and FKHR in HepG2 cells. We found by indirect immunofluorescence and confocal microscopy that in unstimulated cells the localization of FKHR (Fig. 5, right panels) was predominantly nuclear, whereas STAT3 (Fig. 5, left panels) was distributed equally in the nuclear and cytoplasmic regions. It is noteworthy that, although STAT3 was present in the nuclei of untreated HepG2 cells, it did not bind DNA as measured in gel shift experiments (data not shown). Consistent with our previous observations, IL-6 stimulation did not affect FKHR localization but induced STAT3 migration to the nucleus, so that both factors accumulated in the nuclear regions of IL-6-treated HepG2 cells. In contrast, insulin treatment with or without IL-6 costimulation led to partial nuclear exclusion of FKHR. However, this effect was significantly weaker in HepG2 cells than in HeLa cells used as positive controls (27) (data not shown). Taken together, these results suggest that FKHR might associate with STAT3 in the nuclei of IL-6-stimulated HepG2 cells.
Physical Association of FKHR and STAT3-Since FKHR specifically contributes to the activation of STAT3-dependent promoters, we investigated whether this effect reflects a physical interaction between these two proteins. Western blot analysis indicated that FKHR was expressed in HepG2 cells at a low level and could be visualized only in highly concentrated total cellular lysates ( Fig. 3C; compare to Figs. 2D and 6A). In contrast, FKHR was easily detectable in immunoprecipitates obtained with anti-STAT3 antibodies from total cell lysates of HepG2 cells (Fig. 6A), indicating the association of both proteins. The reverse immunoprecipitation experiments performed using antiserum against FKHR yielded similar results and led to coprecipitation of STAT3 from total cellular lysates, confirming the previous results (Fig. 6A). To further consolidate the specificity of the interaction between FKHR and STAT3, we carried out several additional control immunoprecipitations. As shown in Fig. 6B, FKHR did not associate with FIG. 4. Functional interaction between FKHR and STAT3 in transcriptional regulation. A and B, FKHR contributes specifically to the function of STAT3. A, synergistic effect of FKHR and STAT3 on the IL-6-induced ␣ 2 -M promoter activity. HepG2 cells were transiently transfected as described in the legend to Fig. 1 together with FKHR and/or wild type STAT3 expression vectors. B, FKHR has no effect on the STAT5-mediated activation of the ␤-casein promoter. STAT5 or other unrelated signaling molecules like SMAD2/3 or extracellular signal-regulated kinase 2. This allows us to conclude that the functional interaction of FKHR and STAT3 reflects a specific binding of both proteins. Although STAT3 and FKHR were found to be associated also in total cell lysates from untreated cells, stimulation with IL-6 strongly enhanced the binding of both factors in the nuclear fraction of HepG2 cells (Fig. 6C). Furthermore, in agreement with our results on insulin-mediated inhibition of ␣ 2 -M gene expression, costimulation with insulin abrogated the IL-6-induced interaction between STAT3 and FKHR in the nucleus. Therefore, the negative regulation of IL-6-dependent gene expression by the activation of the PI 3-kinase/Akt pathway appears to result from the loss of cooperation between STAT3 and its transcriptional partner FKHR and depends on the subcellular localization of both proteins.

FKHR Interacts with STAT3 and Augments IL-6-dependent
Gene Expression-The major finding of the present study is that FKHR (FOXO1a), a member of the forkhead family of transcription factors, can augment IL-6-dependent transcriptional activity by interacting with STAT3. This conclusion is based on three lines of evidence. First, FKHR expression enhanced the IL-6-dependent activation of the ␣ 2 -macroglobulin promoter in HepG2 cells and showed a synergistic action together with STAT3. Second, we observed the physical association of both factors in nuclear extracts of HepG2 cells, which could be further enhanced by IL-6 stimulation and abrogated in the presence of insulin. Third, STAT3 showed similar to FKHR a nuclear distribution in IL-6-treated HepG2 cells.
Our results suggest that FKHR can act as a coactivator in STAT3-mediated transcriptional activation of acute phase protein genes. Since FKHR did not significantly induce basal activity of the ␣ 2 -M promoter and its positive influence on gene expression was dependent on the presence of activated STAT3, we conclude that FKHR indirectly augments IL-6-induced transcriptional activation. To our knowledge, this is the first report demonstrating the cooperation between both factors. However, it is widely accepted that activating proteins like the STATs most often do not act alone but rather in combination with other site-specific or more general DNA-binding proteins as well as with coactivators of transcription (33,34). Examples of such a cooperativity include STAT1 and Sp1 (35), STAT5 and the glucocorticoid receptor (36), and STAT3 and c-Jun (37). The cooperation between STAT3 and c-Jun has been well documented for a number of different genes (34), but the initial observation concerned their interaction on the ␣ 2 -M promoter (37). Therefore, it would be of interest to assess whether c-Jun could also participate in the interaction between STAT3 and FKHR. A recent report (38) documented that STAT3 can act in concert with hepatocyte nuclear factor 1 to enhance the hepatocyte nuclear factor 1-mediated transactivation of hepatic gene expression in HepG2 cells and murine livers. It seems conceivable that this finding represents a more general mechanism whereby tissue-specific and inducible transcription factors cooperate in response to external signals. The results of our study suggest a different scenario, where FKHR transcription factors, closely related to the liver-specific hepatocyte nuclear factor 3 (39), support the STAT3-dependent gene expression.
FKHR was originally identified in human rhabdomyosarcomas as a fusion protein composed of the transactivation domain of FKHR combined with the intact DNA binding domain of the transcription factor PAX3 (40). The C-terminal half of FKHR turned out to be very potent in transcriptional activation, although PAX3-FKHR proteins showed impaired DNA binding (41). Our results indicate that the C-terminal part of FKHR was also required for its effect on the ␣ 2 -M promoter activation, but in this case the lack of the 16 last amino acid residues already significantly inhibited the coactivating function of FKHR. This finding implies that the complete C terminus of FKHR is required for an efficient cooperation with STAT3.
Plausible mechanisms of the costimulatory action of FKHR on STAT3-mediated gene expression can include 1) enhanced STAT3 activation, 2) facilitated nuclear migration of STAT3, and 3) recruitment of additional coactivator proteins to the STAT3 transcriptional complex. Enhanced STAT3 activation does not seem likely, since gel shift experiments did not reveal an effect of FKHR on DNA-binding activity of STAT3 (data not shown). Our immunofluorescence results do not support the second possibility of facilitated translocation of STAT3 to the nucleus in the presence of FKHR. We did not observe any effect of insulin on IL-6-induced nuclear STAT3 staining. Another possibility could be that FKHR augments IL-6-mediated signaling by recruiting additional coactivator molecules to the STAT3-containing transcriptosome. In HepG2 cells, FKHR has recently been shown to interact with the coactivator, p300/ CBP, of the constitutive transcription machinery (31). Interestingly, STAT3 has also been found to associate with p300/CBP via its carboxyl terminus. This interaction as well as the level of transactivation are relatively weak compared with other STATs such as STAT2 (32). Therefore, the transcriptional function of STAT3 might be reinforced by FKHR acting as an accessory factor directing coactivator complexes to STAT3 binding sites in the promoter of respective target genes.
Activation of PDK1/Akt Signaling Modulates IL-6-mediated Gene Expression-The present results suggest a novel mechanism of the modulation of IL-6-dependent gene expression by the PI 3-kinase/Akt signaling pathway. We have demonstrated that down-regulation of Akt activity by a kinase-defective mutant of PDK1 and constitutive activation of Akt induce opposite effects on the IL-6-responsive ␣ 2 -M promoter in human HepG2 hepatoma cells. Moreover, these effects could be associated with the expression level and the phosphorylation status of FKHR transcription factors. FKHR has been reported to be regulated by insulin in several cell lines of hepatic origin including SV40-transformed murine hepatocytes (22), rat hepatomas (42), and human HepG2 hepatoma cells (43). Recently, also IL-6 has been shown to activate Akt in a significant proportion of multiple myeloma cell lines (26) as well as in human Hep3B hepatoma cells (25,44). In both cell types, activation of the PI 3-kinase/Akt pathway was suggested to play a role in IL-6-dependent protection against apoptosis. The activation of Akt signaling should result in the direct phosphorylation of three regulatory sites (Thr 24 , Ser 256 , and Ser 319 ) in FKHR and subsequent transcriptional inactivation and relocation to the cytoplasm (10,23,24). However, it is not known whether the reported transient activation of Akt in multiple myeloma and Hep3B hepatoma cells is sufficient for the permanent exclusion of FKHR from nuclei of treated cells. We found that IL-6 neither induced kinase activity of Akt nor stimulated its phosphorylation on the critical regulatory residues Thr 308 and Ser 473 (7). Correspondingly, phosphorylation of FKHR on Thr 24 and Ser 256 was not increased during IL-6 treatment of HepG2 cells for up to 16 h (Fig. 3C and data not shown). Our initial observation that expression of kinase-dead PDK1 (Fig. 2) and treatment with wortmannin (Fig. 1B) can enhance the IL-6induced ␣ 2 -M promoter activity may result from an inhibition of the basal, IL-6-independent Akt activity. These results were further corroborated by immunofluorescence data showing no FIG. 6. Physical association of FKHR and STAT3. A, coimmunoprecipitation of FKHR and STAT3. HepG2 cells were serum-starved for 24 h and incubated in the presence of LY294002 (100 M) for the last 1 h. Thereafter, cells were washed with medium and stimulated for 10 min with IL-6 (200 units/ml), lysed, and incubated with polyclonal anti-STAT3 (left panel) or polyclonal anti-FKHR antibodies (right panel). The precipitates together with 30 g of total cell lysates (TCL) were separated by SDS-PAGE, blotted onto polyvinylidene difluoride membranes, and detected with anti-FKHR or anti-STAT3 (upper panels). After stripping, the blots were redeveloped with the antibodies used for the immunoprecipitations (IP) to control loading (lower panels). B, FKHR does not interact with signaling molecules other than STAT3. Starved HepG2 cells were stimulated with IL-6 (200 units/ml), interferon-␥ (1000 units/ml), tumor growth factor-␤ (10 units/ml), or 10% fetal calf serum, lysed, and incubated with anti-STAT3, anti-STAT5, anti-SMAD2/3, or anti-extracellular signal-regulated kinase 2 antibodies, respectively. The precipitates together with a fraction of total cell lysates collected before the immunoprecipitation were separated by SDS-PAGE and analyzed as described above. Upper panel, sections of the same blot developed with anti-FKHR serum. Lower panel, control detection of the stripped membranes with the respective antibodies used for the immunoprecipitations. C, treatment with IL-6 enhances the association of FKHR and STAT3 in nuclear fractions of HepG2 cells. Cells were serum-starved and pretreated with LY294002 as previously and then stimulated with IL-6 (200 units/ml) or IL-6 together with insulin (1 M). After fractionation, nuclear extracts were incubated with polyclonal anti-STAT3 antibodies or control normal rabbit serum (NRS). The precipitates were separated by SDS-PAGE, blotted onto polyvinylidene difluoride membranes, and detected with anti-FKHR (upper panel). After stripping, the blots were redeveloped with the antibodies used for the immunoprecipitations to control loading (lower panel). WB, Western blot. change in nuclear localization of endogenous FKHR as well as of overexpressed green fluorescent protein-tagged FKHR upon stimulation of HepG2 cells with IL-6 ( Fig. 5 and data not shown). Therefore, we conclude that FKHR during IL-6 treatment remains active and may participate in IL-6-induced gene expression.
Phosphorylation of FKHR by Akt on Thr 24 , Ser 256 , and Ser 319 was reported to attenuate its nuclear import possibly by binding of FKHR to 14-3-3 proteins (9,12,27). Several groups demonstrated the resistance of FKHR threonine/serine mutants to both Akt-mediated phosphorylation and PI 3-kinasestimulated nuclear export (23,24). This could well explain the enhanced potency of the mutated FKHR variants T24A, T24A/ S256A, and T24A/S256A/S319A to augment IL-6-induced ␣ 2 -M promoter activity as a result of nuclear retention of these proteins. Our observations indicate that the single mutation of Thr 24 is sufficient for maximal transcriptional stimulation. They are in agreement with previous reports on the crucial role of Thr 24 , a residue that lies within a 14-3-3 consensus motif, involved in the interaction with 14-3-3 proteins (9,45).
Conclusions-In the present study, we have investigated the potential cross-talk between the PI 3-kinase/Akt and Jak/ STAT3 signaling pathways at the level of transcriptional regulation. Our results reveal a novel function of FKHR, suggesting that it acts as a specific coactivator of STAT3. This functional interaction correlates with the physical association of both proteins and their colocalization in the nuclear regions of human HepG2 hepatoma cells. Taken together, our data demonstrate that activation of the PI 3-kinase/Akt pathway can modulate IL-6 signaling by targeting and inactivating FKHR. This can apply to APP genes like the ␣ 2 -macroglobulin gene, which are negatively regulated by insulin and growth factors.