Glycogen synthase kinase-3β (GSK-3β)-modulated IFN-γ-induced inflammation has been reported; however, the mechanism that activates GSK-3β and the effects of activation remain unclear. Inhibiting GSK-3β decreased IFN-γ-induced inflammation. IFN-γ treatment rapidly activated GSK-3β via neutral sphingomyelinase- and okadaic acid-sensitive phosphatase-regulated dephosphorylation at Ser9, and proline-rich tyrosine kinase 2 (Pyk2)-regulated phosphorylation at Tyr216. Pyk2 was activated through phosphatidylcholine-specific phospholipase C (PC-PLC)-, protein kinase C (PKC)-, and Src-regulated pathways. The activation of PC-PLC, Pyk2, and GSK-3β was potentially regulated by IFN-γ receptor 2-associated Jak2, but it was independent of IFN-γ receptor 1. Furthermore, Jak2/PC-PLC/PKC/cytosolic phospholipase A2 positively regulated neutral sphingomyelinase. Inhibiting GSK-3β activated Src homology-2 domain-containing phosphatase 2 (SHP2), thereby preventing STAT1 activation in the late stage of IFN-γ stimulation. All these results showed that activated GSK-3β synergistically affected IFN-γ-induced STAT1 activation by inhibiting SHP2.
Glycogen synthase kinase-3β (GSK-3β),3 a serine/threonine kinase, regulates cellular inflammation (1, 2, 3, 4). Inhibiting GSK-3β protects cells from inflammatory stimuli, including TNF-α (5), endotoxemia (6), experimental colitis, type II collagen-induced arthritis (7), OVA-induced asthma (8), experimental autoimmune encephalomyelitis (9), and bacterial infection (10). Notably, recent studies have reported that GSK-3β is a potent regulator of TLR- (11, 12) and IFN-γ-mediated inflammation (13, 14). However, the molecular targets of GSK-3β and the mechanism of GSK-3β activation after IFN-γ stimulation remain unclear.
IFN-γ, an immune IFN produced by T cells and NK cells, is a potent macrophage-activating factor that promotes Ag processing and presentation, microbial killing, and proinflammatory cytokine production (15, 16). IFN-γ activates Jak2-STAT1, which then regulates IFN-γ-inducible gene expression (15, 17, 18). The IFN-γ receptor (IFNGR) consists of IFNGR1 and IFNGR2, which interact with Jak1 and Jak2, respectively (15, 17, 18). IFN-γ binding induces Jak2 autophosphorylation and activation, which leads to Jak1 transphosphorylation. Activated Jak1 then phosphorylates IFNGR1, creating a docking site for STAT1, which, mediated by Jak2, then phosphorylates tyrosine residue 701 (Tyr701) (15, 17, 18). Subsequently, IFN-γ-activated MAPKs, which include ERK1/2 and p38 MAPK, cause serine phosphorylation of STAT1 (Ser727). STAT1 phosphorylation (Tyr701 and Ser727) is essential for its dimer formation, nuclear translocation, and DNA binding stability (19, 20). It is also essential for STAT1 to attain maximal capacity and initiate or suppress the transcription of IFN-γ-inducible genes (15, 19, 20). For feedback regulation, it is now known that suppressor of cytokine signaling (SOCS) proteins SOCS1 and SOCS3 interact with Jak2 and inhibit its catalytic activity to suppress IFN-γ signaling (21, 22). In addition to SOCS, dual-phosphatase Src homology-2 domain-containing phosphatase 2 (SHP2) deactivates STAT1 (23). However, the mechanisms for feedback regulation are not well understood.
IFN-γ causes the production of TNF-α, IFN-inducible protein 10, MCP-1, and RANTES and the expression of adhesion molecule ICAM-1, but it decreases the production of IL-10 (15, 24, 25). IFN-γ also induces inducible NO synthase (iNOS) expression and then NO generation (26, 27). These proinflammatory responses by IFN-γ are activated through a mechanism involving the activation of transcription factors: STAT1 (28), IFN regulatory factor 1 (IRF-1) (29), and NF-κB (30). The recent finding that GSK-3β is involved in the IFN-γ-induced production of TNF-α and inhibition of IL-10 (13, 31) suggests a novel role for GSK-3β in IFN-γ signaling. GSK-3β is also involved in IFN-γ-activated STAT3 and STAT5 (14). In general, serine phosphorylation (Ser9) negatively regulates GSK-3β primarily through PI3K-Akt pathways (1, 2, 3, 4, 32, 33). Okadaic acid (OA)-sensitive serine/threonine protein phosphatases (PPases) such as PP1 and PP2A may concomitantly dephosphorylate and activate GSK-3β directly or indirectly by dephosphorylating Akt (34, 35, 36, 37, 38, 39, 40). IFN-γ-induced iNOS expression requires the activity of PP1 and PP2A (41). Ceramide activates PP1 and PP2A (42, 43, 44, 45, 46) and is also involved in activating GSK-3β (37, 39, 40). Ceramide is a bioactive lipid generated in response to various stresses, such as apoptotic and inflammatory stimuli, by de novo synthesis or sphingomyelinase (SMase)-mediated hydrolysis of sphingomyelin (39, 46). Intracellular levels of ceramide and iNOS expression correlatively increase after LPS and IFN-γ treatment (47). After IFN-γ treatment, neutral SMase- but not acid SMase-mediated ceramide generation is critical for iNOS/NO biosynthesis (48, 49). Cytosolic phospholipase A2 (cPLA2), a phospholipase that liberates arachidonic acid (AA) from phospholipid diacylglycerol (DAG), is essential for activating neutral SMase in cytokine signaling (50, 51). Thus, the role of cPLA2/neutral SMase/ceramide/PPase/GSK-3β cascade activation in IFN-γ signaling requires further investigation.
Although serine phosphorylation (Ser9) negatively regulates GSK-3β, tyrosine phosphorylation (Tyr216) positively regulates the catalytic activity of GSK-3β. It is generally regulated either by tyrosine kinases, such as proline-rich tyrosine kinase 2 (Pyk2), MEK, and Src-like kinase (1, 2, 3, 4, 52), or by autophosphorylation (53). We have previously (31) shown that IFN-γ induces phosphorylation of GSK-3β (Tyr216) via a Pyk2-mediated pathway. Furthermore, Pyk2 is also essential for IFN-γ-induced MAPK activation and inflammation (54, 55). Pyk2 is activated by protein kinase C (PKC) and Src (56, 57, 58) as well as by calcium/calmodulin-dependent protein kinase II (59). While Jak2-STAT1 is essential for IFN-γ signaling, bioactive lipids and their enzymatic generators are also involved (50, 60). Phosphatidylcholine-specific phospholipase C (PC-PLC)-mediated DAG generation is required for IFN-γ-induced iNOS/NO biosynthesis (61). Moreover, DAG directly activates PKC (60, 61) and then activates Src (62). PKC regulates neutral SMase by increasing the generation of AA after it has activated cPLA2 (63). We show that activated GSK-3β synergistically facilitates IFN-γ-induced STAT1 activation, which then regulates the biosynthesis of iNOS/NO and the production of TNF-α and RANTES.
Materials and Methods
Six-week-old progeny of wild-type (WT) C57BL/6 mice and mice deficient in IFN-γ receptor 1 (B6.129S7-Ifngrtm1Agt/J or IFNGR1−/−) (The Jackson Laboratory) were used for study. The mice were fed standard laboratory chow and water ad libitum in the Laboratory Animal Center of National Cheng Kung University. They were raised and cared for according to the guidelines set by the National Science Council, Taiwan. The experimental protocol adhered to the rules of the Animal Protection Act of Taiwan and was approved by the Laboratory Animal Care and Use Committee of National Cheng Kung University.
Abs and reagents
Mouse mAb specific for β-actin was purchased from Chemicon International. Anti-mouse iNOS mAb was from BD Biosciences. Alexa Fluor 488- and HRP-conjugated goat anti-mouse, goat anti-rabbit, and donkey anti-goat IgG were from Invitrogen. Abs against phospho-GSK-3β (Ser9), phospho-GSK-3β (Tyr216), GSK-3α/β, phospho-GS (Ser641), GS, phospho-Akt (Ser473), Akt, phospho-Pyk2 (Tyr402), Pyk2, phospho-STAT1 (Tyr701), phospho-STAT1 (Ser727), STAT1, phospho-Jak2 (Tyr1007/1008), Jak2, phospho-Src (Tyr416), Src, phospho-ERK1/2 (Thr202/Tyr204), ERK1/2, phospho-p38 MAPK (Thr180/Tyr182), p38 MAPK, phospho-SHP2 (Tyr542), SHP2, SOCS1, and SOCS3 were purchased from Cell Signaling Technology. Polyclonal goat anti-mouse proliferating cell nuclear Ag was from Santa Cruz Biotechnology. Recombinant mouse and human cytokine IFN-γ were from PeproTech. (E)-2-cyano-3-(3,4-dihydrophenyl)-N-(phenylmethyl)-2-propenamide (AG490), PP1, and 8-hydroxy-7-(6-sulfonaphthalen-2-yl)diazenyl-quinoline-5-sulfonic acid (NSC-87877) were purchased from Tocris Bioscience. Sphingolactone-24 (Sph-24) was from Alexis Biochemicals. U73122, Gö6976, and calphostin C (Cal C) were purchased from Calbiochem. 3-(2,4-Dichlorophenyl)-4-(1-methyl-1H-indol-3-yl)-1H-pyrrole-2,5-dione (SB216763), 6-bromo-indirubin-3′-oxime (BIO), OA, 3,5-di-tert-butyl-4-hydroxybenzylidenemalononitrile (tyrphostin A9), and tricyclodecan-9-yl-xanthogenate (D609) were from Sigma-Aldrich. All drug treatments in cells were assessed for their cytotoxic effects using cytotoxicity and viability assays. Doses determined to be harmless were used.
RAW264.7 and MHS murine macrophages and A549 human lung epithelial cells were obtained from Dr. C. C. Huang (Department of Pediatrics, National Cheng Kung University Hospital, Tainan, Taiwan) and Dr. W. C. Su (Department of Oncology, National Cheng Kung University Hospital, Tainan, Taiwan), respectively. The cells were grown in DMEM (Invitrogen) supplemented with 10% heat-inactivated FBS (Invitrogen), 50 U/ml penicillin, and 50 μg/ml streptomycin in a humidified atmosphere with 5% CO2 and 95% air. For primary splenocyte culture, the mice were i.p. injected with a lethal dose of pentobarbital (200 mg/kg) and their spleens were harvested to prepare a single-cell suspension. Isolated splenocytes were grown in RPMI 1640 (Invitrogen) supplemented with 10% heat-inactivated FBS, 50 U/ml penicillin, and 50 μg/ml streptomycin before the experiment.
We used a colorimetric assay (cytotoxicity detection kit; Roche Diagnostics), according to the manufacturer’s instructions, to measure lactate dehydrogenase activity, an indicator of cell damage caused by IFN-γ and all inhibitors. Aliquots of the incubation medium were transferred to 96-well microplates and the absorbance was measured using a microplate reader (SpectraMax 340PC; Molecular Devices).
We assessed mRNA expression using RT-PCR. Using a reagent (TRIzol; Invitrogen) according to the manufacturer’s instructions, we extracted total cellular RNA from cells. We quantified RNA concentrations using a spectrophotometer (U-2000; Hitachi) at 260 nm. cDNA was prepared using reverse transcription, and PCR was done using a thermal cycler (GeneAmp PCR system 2400; PerkinElmer). Based on published sequences (64) and sequences that we designed using Primer3 online software (65), we used the following oligonucleotide primers: mouse iNOS, sense, 5′-CCCTTCCGAAGTTTCTGGCAGCAGCG-3′ and antisense, 5′-GGCTGTCAGAGCCTCGTGGCTTTGG-3′; TNF-α, sense, 5′-AGCCCACGTCGTAGCAAACCACCAA-3′ and antisense, 5′-ACACCCATTCCCTTCACAGAGCAAT-3′; RANTES, sense, 5′-ATATGGCTCGGACACCACTC-3′ and antisense, 5′-CCCACTTCTTCTCTGGGTTG-3′; and β-actin, sense, 5′-TGGAATCCTGTGGCATCCATGAAAC-3′ and antisense, 5′-TAAAACGCAGCTCAGTAACAGTCCG-3′.
The PCR products were analyzed with 1.5% agarose gel electrophoresis, stained with ethidium bromide, and viewed with UV light using a gel camera (UVP).
Western blot analysis
We harvested the cells and lysed them with a buffer containing 1% Triton X-100, 50 mM Tris (pH 7.5), 10 mM EDTA, 0.02% NaN3, and a protease inhibitor cocktail (Roche Boehringer Mannheim Diagnostics). Additionally, the nuclear lysates were prepared using a compartment ProteoExtract subcellular proteome extraction Kit (Calbiochem) according to the manufacturer’s instructions. After they had been freeze-thawed once, the cell lysates were centrifuged at 9000 × g at 4°C for 20 min. The supernatants were then collected and boiled in sample buffer for 5 min. Following SDS-PAGE, proteins were transferred to polyvinylidene difluoride membrane (Millipore), blocked at 4°C overnight in PBS-T (PBS plus 0.05% Tween 20) containing 5% skim milk, and probed with primary Abs at 4°C overnight. After they had been washed with PBS-T, the blots were incubated with a 1/5000 dilution of HRP-conjugated secondary Abs at 4°C for 1 h. The protein bands were visualized using ECL (Pierce Biotechnology). Graphical analysis of band density was performed using ImageJ software (version 1.41o) from W. Rasband (National Institutes of Health, Bethesda, MD) (http://rsb.info.nih.gov/ij/).
We assessed NO production by measuring the accumulated levels of nitrite in the supernatant with the Griess reagent as previously described (31). Briefly, 100 μl of the culture supernatant was reacted with 100 μl of Griess reagent (1% sulfanilamide, 0.1% naphthylethylenediamine dihydrochloride, and 2.5% H3PO4) for 10 min at room temperature. The concentration of nitrite was measured using a microplate reader (SpectraMax 340PC; Molecular Devices) at 540 nm, and the nitrite concentration was calculated using a standard curve of sodium nitrite with ELISA software (SoftMax Pro; Molecular Devices).
Small interfering RNA (siRNA) and lentiviral-based RNA interference (RNAi) transfection
GSK-3β expression was silenced using commercial GSK-3β siRNA (Santa Cruz Biotechnology; catalog no. sc-35527) in A549 and lentivirus-based RNAi in RAW264.7 cells. siRNA transfection was performed by electroporation using a microporator (Digital Bio Technology). Before transfection, the A549 cells were washed with serum-free DMEM and mixed with siRNA in Opti-MEM medium (Invitrogen) in a volume of 100 μl. After they had been transfected, and before they were stimulated, the cells were incubated for 48 h in DMEM supplemented with 10% FBS at 37°C. A nonspecific scramble siRNA was the negative control. GSK-3β knockdown in RAW264.7 cells was performed using lentiviral transduction to stably express short hairpin RNAs (shRNA) that targeted GSK-3β. shRNA clones were obtained from the National RNAi Core Facility located at the Institute of Molecular Biology/Genomic Research Center, Academia Sinica, Taiwan. The mouse library should be referred to as TRC-Mm 1.0. The construct that was most effective in RAW264.7 cells (TRCN0000012615 containing the shRNA target sequence 5′-CATGAAAGTTAGCAGAGATAA-3′ for mouse GSK-3β) was used to generate recombinant lentiviral particles. Human TE671 cells were cotransfected with two helper plasmids, pCMVdeltaR8.91 and pMD.G (gift from Dr. H. K. Sytwu, Graduate Institute of Life Sciences, National Defense Medical Center, Taiwan), plus pLKO.1-puro-shRNA, using GeneJammer transfection reagent (Stratagene). The transfected cells were incubated at 37°C in an atmosphere of 5% CO2 for 24 h, and then the medium was replaced with fresh medium. Cell supernatants containing the viral particles were harvested at 36, 48, 60, and 72 h after transfection. The supernatants were filtered using a 0.45-μm low-protein-binding filter and concentrated by centrifugation at 20,000 × g at 4°C for 3 h using a JA25.50 (Beckman Coulter) rotor. The virus pellets were resuspended with fresh medium and stored at −80°C. RAW264.7 cells were transduced by lentivirus with appropriate multiplicity of infection in complete growth medium supplemented with 8 μg/ml polybrene. After transduction for 24 h, protein expression was monitored using Western blot analysis.
The concentrations of TNF-α and RANTES in cell-conditioned culture medium were determined using ELISA kits (R&D Systems) according to the manufacturer’s instructions.
Sphingomyelinase activity was determined from cellular extracts (Amplex Red sphingomyelinase assay kit; catalog no. A12220, Invitrogen) according to the manufacturer’s instructions. Briefly, each reaction contained 50 μM Amplex Red reagent, 1 U/ml HRP, 0.1 U/ml choline oxidase, 4 U/ml alkaline phosphatase, and 0.25 mM sphingomyelin in 1× reaction buffer. Reactions were incubated at 37°C for 1 h. Fluorescence was measured using a microplate reader (Fluoroskan Ascent; Thermo Electron) with excitation at 530 nm and emission at 590 nm.
PC-PLC activity was determined from cellular extracts (Amplex Red PC-PLC assay kit; catalog no. A12218, Invitrogen) according to the manufacturer’s instructions. Briefly, each reaction mixture contained 200 μM reagent (10-acetyl-3,7-dihydroxyphenoxazine; Amplex Red), 1 U/ml HRP, 4 U/ml alkaline phosphatase, 0.1 U/ml choline oxidase, 0.5 mM lecithin, and 20–100 mU/ml PC-PLC in 50 mM Tris-HCl (pH 7.4)/140 mM NaCl/10 mM dimethylglutarate/2 mM CaCl2. Reactions were incubated at 37°C for 1 h. Fluorescence was measured using the microplate reader with excitation at 530 nm and emission at 590 nm.
Immunostaining of STAT1 nuclear translocation
Cells were fixed with 1% formaldehyde in PBS at room temperature for 10 min. After they had been washed twice with PBS, they were stained with rabbit anti-mouse STAT1 at a final concentration of 1 μg/ml at room temperature for 1 h and then incubated with a mixture of Alexa Fluor 488-conjugated goat anti-rabbit IgG plus 4,6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich) at a concentration of 5 μg/ml at room temperature for 1 h. After another washing with PBS, the cells were covered with mounting fluid and visualized under a confocal laser microscope (Leica TCS SPII).
Student’s t test was used to analyze the data (SigmaPlot 8.0 for Windows; Systat Software). Statistical significance was set at p < 0.05.
Inhibiting GSK-3β reduces IFN-γ-induced iNOS/NO biosynthesis as well as TNF-α and RANTES production
RAW264.7 cells treated with 10 ng/ml IFN-γ showed significantly higher NO production as well as iNOS mRNA and protein expression via a time-dependent process (data not shown). We also found that the GSK-3 inhibitors SB216763 and BIO decreased IFN-γ-induced iNOS mRNA and protein expression (Fig. 1,A) as well as NO production in RAW264.7 (Fig. 1,B) and MHS macrophages (data not shown). Co-treatment with SB216763 significantly reduced IFN-γ-induced TNF-α (Fig. 1,C) and RANTES (Fig. 1,D) secretion as well as mRNA expression (data not shown) in RAW264.7 cells. To exclude the indeterminate effects of GSK-3 inhibitors, we silenced GSK-3β expression in RAW264.7 and A549 epithelial cells (supplemental Fig. S1)4 and then treated the cells with IFN-γ. Western blot analysis showed down-regulated GSK-3β expression in GSK-3β RNAi (Fig. 1,E). ELISA analysis showed that silencing GSK-3β significantly reduced IFN-γ-induced TNF-α production (Fig. 1 F). These results provide evidence that IFN-γ induces iNOS/NO biosynthesis and TNF-α and RANTES production through GSK-3β-regulated mechanisms.
IFN-γ induces neutral SMase-, PPase-, and Pyk2-regulated GSK-3β activation and inflammation
GSK-3β is negatively and positively regulated by phosphorylation at Ser9 and Tyr216, respectively (1, 2, 3, 4, 32, 33). Western blot analysis showed that IFN-γ rapidly induced GSK-3β dephosphorylation (Ser9) and then caused time-dependent phosphorylation of GSK-3β (Tyr216) (Fig. 2,A). To further examine the mechanism of activation of GSK-3β by IFN-γ, we analyzed the phosphorylation of glycogen synthase (GS), a GSK-3β substrate. Western blotting showed that IFN-γ time-dependently induced GS phosphorylation (Ser641) (data not shown). Additionally, we confirmed the dependence of GSK-3β on IFN-γ-induced GS phosphorylation using GSK-3β inhibitor BIO (Fig. 2 B). The accumulation of GS expression in BIO-treated cells was also showed as consistent with previous studies (66). These results provide evidence that IFN-γ causes GSK-3β activation.
To further investigate the mechanism of GSK-3β activation after IFN-γ stimulation, the involvement of Akt (1, 2, 3, 4, 32, 33) and OA-sensitive PPases (34, 35, 36, 37, 38, 39, 40) were tested. Western blot analysis showed that IFN-γ induced Akt dephosphorylation (Ser473) in the early stage of stimulation (Fig. 2,C). Furthermore, treating cells with OA had the opposite effect (Fig. 2 D). These results demonstrate that IFN-γ-activated OA-sensitive PPases are essential to inactivate Akt and then activate GSK-3β.
In general, ceramide activates OA-sensitive PPases (42, 43, 44, 45, 46) and GSK-3β (37, 39, 40). Because ceramide and PPases are involved in IFN-γ-induced inflammation (39, 46, 47, 48, 49), we next examined their effects on GSK-3β activation. First, we found that neutral SMase- (supplemental Fig. S2, A and B) but not acid SMase- (supplemental Fig. S2, C–E) or ceramide synthase-mediated (supplemental Fig. S2, F–H) ceramide generation was crucial for IFN-γ-induced iNOS/NO biosynthesis. Second, we found that IFN-γ induced ceramide generation (supplemental Fig. S3A). Both the exogenous administration of ceramide (supplemental Fig. S3, B and C) and the endogenous accumulation of ceramide (supplemental Fig. S3, D and E) increased IFN-γ-induced iNOS/NO biosynthesis. Third, we found that IFN-γ up-regulated SMase activity (Fig. 2,E). Neutral SMase inhibitor Sph-24 reduced IFN-γ-induced Akt and GSK-3β dephosphorylation (Fig. 2 F). Finally, we found that inhibiting PPases (supplemental Fig. S4) and neutral SMase (supplemental Fig. S2, A and B) reduced IFN-γ-induced iNOS expression as well as NO production. These results indicate that neutral SMase and OA-sensitive PPases are important for IFN-γ-induced GSK-3β activation as well as for iNOS/NO biosynthesis.
Tyrosine phosphorylation of GSK-3β is also critical for its kinase activity (1, 2, 3, 4, 52). We previous showed that IFN-γ induced GSK-3β phosphorylation (Tyr216) through a Pyk2-mediated pathway in LPS-activated macrophages (31). Western blot analysis showed that IFN-γ time-dependently induced Pyk2 phosphorylation (Tyr402) (Fig. 2,G). Treating cells with Pyk2 inhibitor A9 decreased IFN-γ-induced GSK-3β phosphorylation (Tyr216) (Fig. 2,H) as well as NO generation (Fig. 2,I) and TNF-α production (Fig. 2,J). Inhibiting Pyk2 did not affect IFN-γ-induced GSK-3β dephosphorylation (Ser9) (Fig. 2 H). These results show that IFN-γ-activated Pyk2 is also essential for GSK-3β activation as well as for NO and TNF-α production.
IFN-γ induces PC-PLC-, PKC-, and Src-regulated Pyk2 and GSK-3β activation and inflammation
PC-PLC-, PKC-, and Src-regulated Pyk2 activation is involved in IFN-γ-induced inflammation (67, 68). We first found that PC-PLC inhibitor D609 and PKC inhibitors Cal C and Gö6976, but not PI-PLC inhibitor U73122, significantly reduced IFN-γ-induced NO production as well as iNOS expression (supplemental Fig. S5). We also found that PC-PLC and PKC activated Src (supplemental Fig. S6). Western blot analysis showed that treating cells with D609, Gö6976, or Src inhibitor PP1 decreased IFN-γ-induced phosphorylation of Pyk2 (Tyr402) and GSK-3β (Tyr216), dephosphorylation of GSK-3β (Ser9), and iNOS expression (Fig. 3, A–C). Furthermore, inhibiting PC-PLC, PKC, and Src blocked IFN-γ-induced NO (Fig. 3,D) and TNF-α production (Fig. 3 E). These results indicate that IFN-γ-activated PC-PLC/PKC/Src is also essential for Pyk2 and GSK-3β activation and inflammation.
IFN-γ induces IFNGR2-associated Jak2-regulated PC-PLC, Pyk2, and GSK-3β
We next examined the role of IFNGR1 and IFNGR2-associated Jak2 in IFN-γ-activated PC-PLC, Pyk2, and GSK-3β. In splenocytes obtained from IFNGR1 WT or deficient (IFNGR1−/−) C57BL/6 mice, a PC-PLC activity assay showed that IFN-γ increased PC-PLC activity even in IFNGR1−/− cells (Fig. 4,A). Western blot analysis showed that IFN-γ induced phosphorylation of STAT1 (Tyr701), but not Pyk2 (Tyr402) or GSK-3β (Tyr216), via an IFNGR1-regulated process (Fig. 4,B). Notably, inhibiting IFNGR2-associated Jak2 using specific inhibitor AG490 significantly blocked IFN-γ-induced PC-PLC activity in splenocytes (data not shown) as well as in RAW264.7 cells (Fig. 4,C). We found that AG490 effectively blocked IFN-γ-induced phosphorylation of STAT1, Pyk2, and GSK-3β (Fig. 4 D). These results show that Jak2 acts an upstream regulator role for activating PC-PLC, Pyk2, and GSK-3β.
Signaling of Jak2, PC-PLC, PKC, and cPLA2 are critical for IFN-γ-induced activation of neutral SMase and GSK-3β as well as inflammation
PKC and cPLA2 are neutral SMase-regulating kinases that increase AA generation from DAG (63). An activity assay showed that inhibiting Jak2, PC-PLC, or PKC significantly decreased neutral SMase activity in IFN-γ-stimulated RAW264.7 cells (Fig. 5,A). Cells treated with neutral SMase inhibitor were the positive control. Further results showed that cPLA2 inhibitor BEL blocked IFN-γ-induced SMase activation (Fig. 5,B), GSK-3β dephosphorylation (Ser9) and phosphorylation (Tyr216) (Fig. 5,C), and iNOS/NO biosynthesis (Fig. 5, D and E). All these results show that Jak2/PC-PLC/PKC cross-talk signaling regulates cPLA2-activated neutral SMase and then activates GSK-3β.
GSK-3β facilitates IFN-γ-induced persistent STAT1 activation by inhibiting SHP2
We next examined the effects of bioactive lipids and their enzymatic generators—PC-PLC, PKC, Src, Pyk2, cPLA2, neutral SMase, and OA-sensitive PPases—on IFN-γ signaling. Immunocytochemical staining and Western blot analysis showed that inhibiting these proteins considerably reduced IFN-γ-induced STAT1 nuclear translocation (supplemental Fig. S7A) and phosphorylation (supplemental Fig. S7, B–H). These results suggest that bioactive lipids and their enzymatic generators are indispensable for sustaining IFN-γ signaling such as STAT1 activation.
We next examined the potential effects of GSK-3β. Using immunocytochemical staining (Fig. 6,A) and Western blot analysis of nuclear proteins (Fig. 6,B), we found that BIO inhibited STAT1 nuclear translocation. Co-treatment with BIO inhibited IFN-γ-induced STAT1 phosphorylation (Tyr701 and Ser727) in the late stage, but not in the early stage (Fig. 6,C). In IFN-γ-stimulated RAW264.7 and A549 cells (supplemental Fig. S8), silencing GSK-3β (Fig. 6,D) reduced STAT1 phosphorylation (Tyr701 and Ser727). Actually, tyrosine kinases Jak2 and Src and MAPKs ERK1/2 and p38 MAPK were critical for IFN-γ-induced NO production (data not shown) as well as for STAT1 phosphorylation (supplemental Fig. S9) (15, 16, 17, 18, 19, 20, 69). However, Western blot analysis showed that inhibiting GSK-3β did not decrease the activation of Src, ERK1/2, and p38 MAPK but blocked Jak2 phosphorylation (Tyr1007/1008) in the late stage (3 h posttreatment) of IFN-γ stimulation (Fig. 6 C). These results strongly suggest that GSK-3β is critical for IFN-γ-activated Jak2 and STAT1 but has no effect on Src, ERK1/2, or p38 MAPK.
Feedback from both SHP2 and SOCS protein down-regulate IFN-γ-induced STAT1 activation (21, 22, 23). Interestingly, Western blot analysis showed that using BIO in IFN-γ-stimulated RAW264.7 cells (Fig. 7,A) or RNAi in IFN-γ-stimulated RAW264.7 (Fig. 7,B) and A549 cells (supplemental Fig. S10) to inhibit GSK-3β caused higher phosphorylation of SHP2 (Tyr542). However, inhibiting GSK-3β had no effect on SOCS1 or SOCS3. Furthermore, a protein phosphatase activity assay showed that inhibiting GSK-3β increased IFN-γ-induced protein phosphatase activation (Fig. 7 C). These results suggest that GSK-3β negatively regulates SHP2; however, the mechanism remains unclear.
We next examined the potential role of up-regulated SHP2. Immunocytochemistry and Western blot analysis showed that using specific inhibitor NSC-87877 to inhibit SHP2 reversed BIO-induced inhibition of STAT1 nuclear translocation (Fig. 7,D) as well as phosphorylation (Tyr701 and Ser727) (Fig. 7,E). Notably, co-treatment with NSC blocked the BIO-induced inhibition of iNOS expression (Fig. 7,F) and NO production (Fig. 7 G) in IFN-γ-stimulated RAW264.7 cells. All of these results indicate that inhibiting GSK-3β reduces IFN-γ-induced iNOS/NO biosynthesis and persistent STAT1 activation by up-regulating SHP2.
GSK-3β regulates IFN-γ signaling and is involved in IFN-γ-induced inflammation (13, 14, 31, 67). IFN-γ activates the proinflammatory expression of TNF-α, RANTES, and iNOS (15, 24, 25, 26, 27). We provide the first evidence that RANTES and iNOS, as well as TNF-α (31), expression is GSK-3β-dependent. Additionally, IFN-γ-inducible protein such as caspase-1 is also induced via a GSK-3β-regulated manner (data not shown). The specific targets for GSK-3β in IFN-γ signaling need further investigation. However, the mechanism of GSK-3β activation and its effects on IFN-γ signaling remain unknown. Our findings lead us to hypothesize, however, that, after IFN-γ stimulation, neutral SMase- and PPase-mediated dephosphorylation at Ser9 and Pyk2-mediated phosphorylation at Tyr216 regulate the activation of GSK-3β. These processes are universal with various stimuli (1, 2, 3, 4, 31, 34, 35, 36, 37, 40). Thus, neutral SMase and PPases act synergistically with Pyk2 to regulate GSK-3β activation. We also showed that a PC-PLC/PKC/Src regulated process-activated Pyk2, that an IFNGR2-associated Jak2-mediated process potentially regulated PC-PLC/Pyk2/GSK-3β activation, and that Jak2/PC-PLC/PKC/cPLA2 signaling activated neutral SMase. An unknown mechanism may inhibit SHP2 when GSK-3β is activated. We hypothesize that bioactive lipids and their enzymatic generators regulate GSK-3β activation, and that GSK-3β in turn inhibits SHP2, which facilitates IFN-γ-induced persistent Jak2-STAT1 activation and inflammation. Based on these findings, we have created a schematic summary for GSK-3β activation and its effects on IFN-γ signaling and inflammatory activation (Fig. 8).
Bioactive lipids and their enzymatic generators (ceramide, cPLA2, and PC-PLC) are generally involved in cytokine (TNF-α, IL-1, and IFN-γ)-mediated signaling (45, 46, 47), and their downstream targets and effects are abundant. The generation of ceramide through neutral SMase-mediated SM hydrolysis is essential for IFN-γ synergized with LPS to induce iNOS/NO biosynthesis and cytokine production (47, 48, 49). We have provided evidence that IFN-γ alone causes neutral SMase-regulated signaling and inflammation. Furthermore, ceramide increases IFN-γ-induced inflammation. Ceramide also activates PPases such as PP1 and PP2A (42, 43, 44, 45, 46), which are required for IFN-γ-synergized with LPS to induce iNOS/NO biosynthesis (41). We showed the potency of OA-sensitive PPases in IFN-γ-induced STAT1 activation and inflammation.
The targets of OA-sensitive PPases are diverse. Notably, ceramide induces GSK-3β activation (37, 39, 40). For GSK-3β activation, PP1- or PP2A-mediated serine residue dephosphorylation is essential (34, 35, 36, 37, 38, 39, 40). PPases may cause GSK-3β dephosphorylation directly or indirectly by dephosphorylating Akt or other GSK-3β-phosphorylating kinases (1, 2, 3, 4, 32, 33, 34, 35, 36, 37, 38, 39, 40). Consistent with previous studies (37, 39, 40) on IFN-γ signaling, we showed the involvement of neutral SMase- and PPase-mediated Akt inactivation followed by GSK-3β activation.
cPLA2, which is activated by PKC, is required for DAG to generate AA (63). Furthermore, cPLA2 is important for ceramide generation (50, 51). Others (60, 61, 62) have reported that both PC-PLC- and PI-PLC-mediated DAG generation are involved in IFN-γ-induced PKC activation. Our results, however, show an independent role for PI-PLC in IFN-γ-induced iNOS/NO biosynthesis. Consistent with previous studies (50, 51, 61, 63), we showed the essential roles of PC-PLC, PKC, and cPLA2 for neutral SMase activation and GSK-3β dephosphorylation (Ser9) and inflammation. Based on our findings, we hypothesize that PC-PLC and ceramide synergistically activate GSK-3β through IFN-γ signaling.
Pyk2 acts upstream of IFN-γ-induced STAT1 activation (54). However, the mechanism of Pyk2 activation and its effects remain unclear. We provide evidence that PKC is also essential for IFN-γ-activated Pyk2 via a Src-mediated process. This is not consistent with current studies suggesting that type I IFN-α activates Pyk2 via calcium/calmodulin-dependent protein kinase II (56). After IFN-α stimulation, Pyk2 activation positively regulates Jak-STAT signaling. Pyk2 activates GSK-3β by phosphorylating its tyrosine residue (1, 2, 3, 4, 52). Consistent with our previous study (31), we further show that IFN-γ activates GSK-3β via PC-PLC-, PKC-, Src-, and Pyk2-regulated signaling. Thus, we hypothesize that IFN-γ regulates GSK-3β activation either by PPase-mediated dephosphorylation (Ser9) or by Pyk2-mediated phosphorylation (Tyr216). Furthermore, we showed that IFNGR2-associated Jak2, but not IFNGR1, is important for activating PC-PLC, Pyk2, and GSK-3β. In contrast, Chang et al. (62) showed that IFN-γ-mediated activation of Jak1/Jak2 was essential for PI-PLC-mediated PKC and Src activation and ICAM-1 expression. Because our results and previous studies (61) showed that PC-PLC, but not PI-PLC, is required for IFN-γ-induced iNOS/NO biosynthesis, the regulation of PC-PLC activation by Jak2 requires further investigation.
Proinflammatory IFN-γ, a Th1 cytokine, down-regulates the signaling of antiinflammatory Th2 IL-10 (15, 24, 25). IFN-γ facilitates TLR-mediated inflammation through a mechanism involving GSK-3-mediated CREB inactivation and then IL-10 down-regulation (13, 31, 67). In contrast, IL-10 also inhibits IFN-γ-induced Jak-STAT signaling by inducing SOCS proteins (21, 22). In general, IFN-γ automatically induces feedback regulation that diminishes Jak2-STAT1 activation through a mechanism involving the activation of SOCS1, SOCS3, and SHP2 (21, 22, 23, 68, 69, 70). However, the mechanism that activates feedback regulation remains unclear. We showed that, in the early stage of IFN-γ stimulation, STAT1 activation is fully regulated by IFNGR1 as well as IFNGR2-associated Jak2, but that it is independent of GSK-3β. This is consistent with current studies (14) suggesting that GSK-3 dependence is selective for activation of STAT3 and STAT5, whereas STAT1 and STAT6 activation are GSK-3-independent. Although our results and those of Beurel and Jope (14) showed that inhibiting GSK-3 does not inhibit Jak2 phosphorylation (Tyr1007/1008) and STAT1 phosphorylation (Tyr701) in the early stage of IFN-γ stimulation (within 0.5–1 h posttreatment), we provide further evidence showing that, after the early stage of IFN-γ stimulation, activated GSK-3β is critical for extending Jak2-STAT1 activation. Indeed, we showed that inhibiting GSK-3β did not reduce IFN-γ-activated IRF-1 (data not shown), an early transcriptional regulator of IFN-γ signaling, but blocked iNOS, TNF-α, and RANTES expression as well as caspase-1 (data not shown). Thus, we hypothesize a mechanism by which GSK-3β facilitates persistent Jak2-STAT1 activation after the early stage of IFN-γ stimulation. Moreover, our results exclude the possible effects of inhibiting GSK-3β on IFN-γ-activated Src (Tyr416), ERK1/2 (Thr202/Tyr204), and p38 MAPK (Thr180/Tyr182), even though they are required for IFN-γ-induced STAT1 activation. Since NF-κB is activated by IFN-γ-activated Jaks, PKR, and IKKα/β (71, 72), we currently demonstrate that inhibiting GSK-3β decreases NF-κB activation through SHP2-regulated pathway (data not shown). In addition to Jak2-STAT1, the role for GSK-3β-inhibited SHP2 in regulating NF-κB signaling and the targets of activated GSK-3β in IFN-γ signaling require further investigation.
Beurel and Jope (14) also reported that inhibiting STAT3 by blocking GSK-3 is not mediated by phosphatases. However, we provide further strong evidence that inhibiting tyrosine phosphatase SHP2 is essential for GSK-3β-regulated STAT1 activation. After GSK-3β was inactivated in the end-stage of IFN-γ stimulation or treatment with a GSK-3β inhibitor, SHP2 might be activated and cause STAT1 inactivation. Thus, in IFN-γ signaling, we hypothesize the novel mechanism of SHP2 inactivation by bioactive lipids and GSK-3β regulated pathways.
In this investigation, inhibiting Jak2, PC-PLC, PKC, Src, Pyk2, cPLA2, neutral SMase, and PPase blockaded IFN-γ-induced STAT1 nuclear translocation and phosphorylation. This indicates that other signal molecules cooperate with or act in parallel to Jak-STAT signaling to regulate IFN-γ signaling (73). Nevertheless, in addition to Jak2-STAT1 signaling, their effects on STAT1 activation are controversial until GSK-3β is involved. Although our results show that GSK-3β regulates STAT1 by negatively regulating SHP2, the mechanism by which GSK-3β negatively regulates SHP2 is not fully understood. SHP2 inactivates STAT1 by dephosphorylating STAT1 both at Tyr701 and Ser727 (23, 68, 69, 70, 74, 75). Modulation of SHP2 by phosphorylation at Tyr542 and Tyr580 has been previously demonstrated in a variety of activation of ligand-receptor protein tyrosine kinases (70, 74, 75) and cytokine-mediated signaling through an unclear mechanism (70). We found that inhibiting GSK-3β increased SHP2 phosphorylation (Tyr542) and phosphatase activity. It is still unclear whether GSK-3β directly regulates SHP2 through phosphorylation or indirectly by affecting the SHP2-phosphorylating tyrosine kinases or phosphatases or both. Notably, inhibiting GSK-3β increases the intracellular levels of cAMP and the activities of cAMP-dependent protein kinase (PKA) (76, 77). The cAMP-dependent PKA positively regulated SHP2 has been demonstrated (78, 79). Furthermore, the interregulation of PKA and GSK-3β is still controversial (1, 2, 3, 4, 80). We hypothesize that GSK-3β regulates SHP2 posttranscriptionally, but this needs further investigation.
We conclude that GSK-3β and bioactive lipids and their enzymatic generators synergistically facilitate IFN-γ-activated STAT1 by inhibiting SHP2. To summarize the regulation of GSK-3β, we propose a model for the activation of GSK-3β based both on our previous studies (31, 39, 40) and the work of many others (1, 2, 3, 4, 13, 14, 67). In our model, GSK-3β is effectively regulated by multiple pathways, including IFNGR2/Jak2-mediated subsequent activation of PC-PLC/PKC/Src/Pyk2 and PKC/cPLA2/neutral SMase/PPase activation, and GSK-3β has prolonged effects on IFN-γ signal transduction, including SHP2 inactivation, Jak2-STAT1 activation, and inflammation. For the feedback regulation of IFN-γ signaling, inhibiting GSK-3β is a potent mechanism after IFN-γ stimulation. Reducing IFN-γ-induced inflammatory activation by inhibiting GSK-3β suggests that it may be possible to use GSK-3β inhibition as a novel antiinflammatory strategy. We therefore hypothesize that GSK-3β is important in IFN-γ-induced anticancer, antimicrobe, and immunomodulation. This hypothesis needs further investigation.
The authors have no financial conflicts of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by Grant NSC 96-2320-B-006-018-MY3 from the National Science Council, Taiwan, and the Landmark Project C020 of National Cheng Kung University, Taiwan.
Abbreviations used in this paper: GSK-3β, glycogen synthase kinase-3β; IFNGR, IFN-γ receptor; SOCS, suppressor of cytokine signaling; SHP2, Src homology-2 domain-containing phosphatase 2; iNOS, inducible NO synthase; IRF-1, IFN regulatory factor 1; OA, okadaic acid; PPase, protein phosphatase; SMase, sphingomyelinase; cPLA2, cytosolic phospholipase A2; AA, arachidonic acid; DAG, diacylglycerol; Pyk2, proline-rich tyrosine kinase 2; PKC, protein kinase C; PC-PLC, phosphatidylcholine-specific phospholipase C; WT, wild type; BIO, 6-bromo-indirubin-3′-oxime; siRNA, short interference RNA; RNAi, RNA interference.
The online version of this article contains supplemental material.