Title: Salicylic Acid Reverses Phorbol 12-Myristate-13-Acetate (PMA)- and Tumor Necrosis Factor α (TNFα)-induced Insulin Receptor Substrate 1 (IRS1) Serine 307 Phosphorylation and Insulin Resistance in Human Embryonic Kidney 293 (HEK293) Cells
Abstract: Salicylates, including aspirin, have been shown to improve insulin sensitivity both in human and animal models. Although it has been suggested that salicylates sensitize insulin action by inhibiting IκB kinase β (IKKβ), the detailed mechanisms remain unclear. Protein kinase C isoforms and tumor necrosis factor α (TNFα) signaling pathways are well described mediators of insulin resistance; they are implicated in the activation of IKKβ and the subsequent inhibition of proximal insulin signaling via insulin receptor substrate 1 (IRS1) and Akt. This study investigated the effect of salicylic acid on phorbol 12-myristate 13-acetate (PMA)- and TNFα-induced insulin resistance in a human embryonic kidney 293 (HEK293) cell line stably expressing recombinant human IRS1. The results showed that both PMA and TNFα inhibited insulin-induced Akt phosphorylation and promoted IRS1 phosphorylation on Ser-307. Salicylic acid pretreatment completely reversed the effects of PMA and TNFα on both Akt and IRS1. Whereas PMA activated protein kinase C isoforms and IKKβ, TNFα activated neither. On the other hand, both PMA and TNFα activated the c-Jun N-terminal kinase (JNK), which has been reported to directly phosphorylate IRS1 Ser-307. SP600125, a JNK inhibitor, prevented PMA and TNFα-induced IRS1 Ser-307 phosphorylation. Finally, salicylic acid inhibited JNK activation induced by both PMA and TNFα. Taken together, these observations suggest that salicylic acid can reverse the inhibitory effects of TNFα on insulin signaling via an IKKβ-independent mechanism(s), potentially involving the inhibition of JNK activation. The role of JNK in salicylic acid-mediated insulin sensitization, however, requires further validation because the JNK inhibitor SP600125 appears to have other nonspecific activity in addition to inhibiting JNK activity. Salicylates, including aspirin, have been shown to improve insulin sensitivity both in human and animal models. Although it has been suggested that salicylates sensitize insulin action by inhibiting IκB kinase β (IKKβ), the detailed mechanisms remain unclear. Protein kinase C isoforms and tumor necrosis factor α (TNFα) signaling pathways are well described mediators of insulin resistance; they are implicated in the activation of IKKβ and the subsequent inhibition of proximal insulin signaling via insulin receptor substrate 1 (IRS1) and Akt. This study investigated the effect of salicylic acid on phorbol 12-myristate 13-acetate (PMA)- and TNFα-induced insulin resistance in a human embryonic kidney 293 (HEK293) cell line stably expressing recombinant human IRS1. The results showed that both PMA and TNFα inhibited insulin-induced Akt phosphorylation and promoted IRS1 phosphorylation on Ser-307. Salicylic acid pretreatment completely reversed the effects of PMA and TNFα on both Akt and IRS1. Whereas PMA activated protein kinase C isoforms and IKKβ, TNFα activated neither. On the other hand, both PMA and TNFα activated the c-Jun N-terminal kinase (JNK), which has been reported to directly phosphorylate IRS1 Ser-307. SP600125, a JNK inhibitor, prevented PMA and TNFα-induced IRS1 Ser-307 phosphorylation. Finally, salicylic acid inhibited JNK activation induced by both PMA and TNFα. Taken together, these observations suggest that salicylic acid can reverse the inhibitory effects of TNFα on insulin signaling via an IKKβ-independent mechanism(s), potentially involving the inhibition of JNK activation. The role of JNK in salicylic acid-mediated insulin sensitization, however, requires further validation because the JNK inhibitor SP600125 appears to have other nonspecific activity in addition to inhibiting JNK activity. IκB kinase insulin receptor substrate c-Jun N-terminal kinase tumor necrosis factor α mitogen-activated protein MAP kinase phorbol 12-myristate 13-acetate human embryonic kidney 293 protein kinase C IκB kinase β (IKKβ),1 a serine/threonine kinase, is a component of the larger IKK signalosome, which also consists of IKKα and IKKγ. Whereas IKKα is also a serine/threonine kinase and is highly homologous to IKKβ, IKKγ is a scaffold protein with no enzymatic activity. Upon exposure to stimuli including stress or cytokines, both IKKα and IKKβ undergo phosphorylation and activation. The activated IKK kinases phosphorylate serine residues in IκB, which subsequently becomes degraded via the proteasome pathway. Prior to its degradation, IκB binds to NFκB and prevents its translocation into the nucleus. Thus, IKKs potentiate NFκB-mediated gene transcription (as occurs in the inflammatory states). Interestingly, IKKβ but not IKKα deficiency appears to result in immunodeficiency in mice, suggesting that IKKβ but not IKKα plays a dominant role in immunoresponses (1Karin M. Ben-Neriah Y. Annu. Rev. Immunol. 2000; 18: 621-663Crossref PubMed Scopus (4106) Google Scholar, 2Karin M. Delhase M. Semin. Immunol. 2000; 12: 85-98Crossref PubMed Scopus (865) Google Scholar). Salicylates, including aspirin, have been shown to be antidiabetic in both animal models and man in both long term (i.e. 3–4 week) and short term (i.e. overnight) treatments (3Williamson R.T. Lond M.D. Br. Med. J. 1901; 1: 760-762Crossref PubMed Scopus (158) Google Scholar, 4Reid J. MacDougall A.I. Andrews M.M. Br. Med. J. 1957; 2: 1071-1074Crossref PubMed Scopus (125) Google Scholar, 5Kim J.K. Kim Y.J. Fillmore J.J. Chen Y. Moore I. Lee J. Yuan M., Li, Z.W. Karin M. Perret P. Shoelson S.E. Shulman G.I. J. Clin. Invest. 2001; 108: 437-446Crossref PubMed Scopus (623) Google Scholar). Several lines of evidence suggest that the antidiabetic effects of salicylates are mediated via direct inhibition of IKKβ. First, salicylates were reported to specifically inhibit the kinase activity of recombinant IKKβ but not IKKα proteins in vitro and to down-regulate NFκB-mediated transcription regulation in cultured cells (6Yin M.J. Yamamoto Y. Gaynor R.B. Nature. 1998; 396: 77-80Crossref PubMed Scopus (1438) Google Scholar, 7Pierce J.W. Read M.A. Ding H. Luscinskas F.W. Collins T. J. Immunol. 1996; 156: 3961-3969PubMed Google Scholar, 8Kopp E. Ghosh S. Science. 1994; 265: 956-959Crossref PubMed Scopus (1621) Google Scholar). Second, IKKβ heterozygous mice have improved insulin sensitivity and reduced plasma glucose levels (9Yuan M. Konstantopoulos N. Lee J. Hansen L., Li, Z.W. Karin M. Shoelson S.E. Science. 2001; 293: 1673-1677Crossref PubMed Scopus (1643) Google Scholar). Third, IKKβ heterozygosity in ob/ob background also improves insulin sensitivity in the mice (9Yuan M. Konstantopoulos N. Lee J. Hansen L., Li, Z.W. Karin M. Shoelson S.E. Science. 2001; 293: 1673-1677Crossref PubMed Scopus (1643) Google Scholar). Finally, IKKβ heterozygosity and salicylate treatment prevents lipid-induced insulin resistance in animal models (5Kim J.K. Kim Y.J. Fillmore J.J. Chen Y. Moore I. Lee J. Yuan M., Li, Z.W. Karin M. Perret P. Shoelson S.E. Shulman G.I. J. Clin. Invest. 2001; 108: 437-446Crossref PubMed Scopus (623) Google Scholar). Serine phosphorylation of insulin receptor substrate proteins (IRSs) such as IRS1 has been strongly implicated as a mechanism of insulin resistance (10Tanti J.F. Gremeaux T. van Obberghen E. Le Marchand-Brustel Y. J. Biol. Chem. 1994; 269: 6051-6057Abstract Full Text PDF PubMed Google Scholar, 11Hotamisligil G.S. Peraldi P. Budavari A. Ellis R. White M.F. Spiegelman B.M. Science. 1996; 271: 665-668Crossref PubMed Scopus (2223) Google Scholar, 12De Fea K. Roth R.A. J. Biol. Chem. 1997; 272: 31400-31406Abstract Full Text Full Text PDF PubMed Scopus (267) Google Scholar, 13Rui L. Aguirre V. Kim J.K. Shulman G.I. Lee A. Corbould A. Dunaif A. White M.F. J. Clin. Invest. 2001; 107: 181-189Crossref PubMed Scopus (492) Google Scholar). Upon serine phosphorylation, IRS1 proteins have a reduced ability to interact with the insulin receptor, to be tyrosine phosphorylated by the insulin receptor, and to bind phosphatidylinositol 3-kinase (PI3K) (10Tanti J.F. Gremeaux T. van Obberghen E. Le Marchand-Brustel Y. J. Biol. Chem. 1994; 269: 6051-6057Abstract Full Text PDF PubMed Google Scholar, 14Kanety H. Feinstein R. Papa M.Z. Hemi R. Karasik A. J. Biol. Chem. 1995; 270: 23780-23784Abstract Full Text Full Text PDF PubMed Scopus (385) Google Scholar, 15Paz K. Hemi R. LeRoith D. Karasik A. Elhanany E. Kanety H. Zick Y. J. Biol. Chem. 1997; 272: 29911-29918Abstract Full Text Full Text PDF PubMed Scopus (452) Google Scholar, 16Delahaye L. Mothe-Satney I. Myers M.G. White M.F. Van Obberghen E. Endocrinology. 1998; 139: 4911-4919Crossref PubMed Scopus (36) Google Scholar, 17Aguirre V. Werner E.D. Giraud J. Lee Y.H. Shoelson S.E. White M.F. J. Biol. Chem. 2001; 17: 17Google Scholar). Such negative phosphorylation in IRS1 has been mapped to several serine residues (13Rui L. Aguirre V. Kim J.K. Shulman G.I. Lee A. Corbould A. Dunaif A. White M.F. J. Clin. Invest. 2001; 107: 181-189Crossref PubMed Scopus (492) Google Scholar,14Kanety H. Feinstein R. Papa M.Z. Hemi R. Karasik A. J. Biol. Chem. 1995; 270: 23780-23784Abstract Full Text Full Text PDF PubMed Scopus (385) Google Scholar, 18Feinstein R. Kanety H. Papa M.Z. Lunenfeld B. Karasik A. J. Biol. Chem. 1993; 268: 26055-26058Abstract Full Text PDF PubMed Google Scholar, 19De Fea K. Roth R.A. Biochemistry. 1997; 36: 12939-12947Crossref PubMed Scopus (223) Google Scholar, 20Ravichandran L.V. Esposito D.L. Chen J. Quon M.J. J. Biol. Chem. 2001; 276: 3543-3549Abstract Full Text Full Text PDF PubMed Scopus (192) Google Scholar, 21Strack V. Hennige A.M. Krutzfeldt J. Bossenmaier B. Klein H.H. Kellerer M. Lammers R. Haring H.U. Diabetologia. 2000; 43: 443-449Crossref PubMed Scopus (52) Google Scholar). It was reported recently that IRS1 Ser-307 (refers to the rat IRS1 protein or Ser-312 in human IRS1) is a key regulatory site (17Aguirre V. Werner E.D. Giraud J. Lee Y.H. Shoelson S.E. White M.F. J. Biol. Chem. 2001; 17: 17Google Scholar, 22Aguirre V. Uchida T. Yenush L. Davis R. White M.F. J. Biol. Chem. 2000; 275: 9047-9054Abstract Full Text Full Text PDF PubMed Scopus (1181) Google Scholar). Moreover, phosphorylation at this site can be mediated by the c-Jun N-terminal kinase (JNK) and may be an important contributor to the inhibitory effects of tumor necrosis factor α (TNFα) on insulin signaling. It was further proposed that salicylate treatment or IKKβ deficiency can improve insulin sensitivity by reducing the serine phosphorylation of IRS proteins (i.e.IRS1) (5Kim J.K. Kim Y.J. Fillmore J.J. Chen Y. Moore I. Lee J. Yuan M., Li, Z.W. Karin M. Perret P. Shoelson S.E. Shulman G.I. J. Clin. Invest. 2001; 108: 437-446Crossref PubMed Scopus (623) Google Scholar, 9Yuan M. Konstantopoulos N. Lee J. Hansen L., Li, Z.W. Karin M. Shoelson S.E. Science. 2001; 293: 1673-1677Crossref PubMed Scopus (1643) Google Scholar). Salicylates inhibit the kinase activity of recombinant IKKβ in vitro and the biological function of endogenous IKKβ in cultured cells with an EC50 of 50–100 μm (6Yin M.J. Yamamoto Y. Gaynor R.B. Nature. 1998; 396: 77-80Crossref PubMed Scopus (1438) Google Scholar). On the other hand, in in vivo experiments in which antidiabetic effects were demonstrated, salicylates reached millimolar concentrations in the plasma (5Kim J.K. Kim Y.J. Fillmore J.J. Chen Y. Moore I. Lee J. Yuan M., Li, Z.W. Karin M. Perret P. Shoelson S.E. Shulman G.I. J. Clin. Invest. 2001; 108: 437-446Crossref PubMed Scopus (623) Google Scholar). Such a discrepancy between in vitro and in vivo potency may be ascribed to, for example, potentially lower salicylate levels in the intracellular environment than in the plasma. On the other hand, it is also possible that at such a high concentration in vivo, salicylates may manifest antidiabetic effects by regulating other targets in addition to or other than IKKβ. In fact, at high concentrations salicylates have been shown to have additional effects such as inhibition of mitochondria function (23Trost L.C. Lemasters J.J. Toxicol. Appl. Pharmacol. 1997; 147: 431-441Crossref PubMed Scopus (91) Google Scholar, 24Al-Nasser I.A. Toxicol. Lett. 1999; 105: 1-8Crossref PubMed Scopus (30) Google Scholar) and inhibition of phosphorylation and nuclear translocation of signal transducers and activators of transcription (STATs) (25Wang Z. Jiang B. Brecher P. Biochem. Pharmacol. 2002; 63: 1197-1207Crossref PubMed Scopus (16) Google Scholar). Furthermore, it was reported that treatment of intact cells with salicylate inhibited TNFα-induced but not interleukin-1-induced IKK activity and that this inhibition was prevented by a p38 MAPK inhibitor, suggesting that salicylate does not directly bind to and inhibit IKKβ in intact cells (26Alpert D. Vilcek J. J. Biol. Chem. 2000; 275: 10925-10929Abstract Full Text Full Text PDF PubMed Scopus (41) Google Scholar). These observations suggest that the antidiabetic effects of salicylates may not be mediated exclusively by the inhibition of IKKβ. The current study was initiated to investigate the effects of salicylic acid on phorbol 12-myristate-13-acetate (PMA)- and TNFα-induced insulin resistance in a HEK293 cell line stably expressing recombinant IRS1. Results from this study suggest the following. 1) PMA and TNFα inhibit insulin-stimulated Akt phosphorylation and promote inhibitory IRS1 serine 307 phosphorylation. 2) Salicylic acid reverses the effects of PMA and TNFα on Akt and IRS1 phosphorylation. 3) Salicylic acid attenuates the inhibitory effects of TNFα on insulin signaling through an IKKβ-independent mechanism. 4) JNK kinase is activated by and may be important for both PMA and TNFα-induced insulin resistance in these cells. 5) Salicylic acid partially inhibits PMA and TNFα-induced JNK activation. Insulin, TNFα, PMA, and salicylic acid were purchased from Sigma. Salicylic acid was prepared as 1 mstock solution in 1 m Tris-HCl, pH7.5, aliquoted, and stored at −20 °C. Rabbit polyclonal antibodies against total insulin receptor β subunit, total IRS1, IRS1 phosphorylated at Ser-307 (IRS1-pS307), and total PKCδ were purchased from Upstate Biotechnology (Lake Placid, NY). Mouse monoclonal antibody (4G10) against tyrosine-phosphorylated proteins was also purchased from Upstate Biotechnology. Rabbit polyclonal antibodies against total Akt, Akt phosphorylated at threonine 308 (Akt-pT308), total IKKα, total IKKβ, both IKKα phosphorylated at serine 180 and IKKβ phosphorylated at serine 181, phosphorylated PKC isoforms (PKC-pan), PKCα/βII phosphorylated at threonine 638 and 641, PKCδ phosphorylated at threonine 505, PKCδ phosphorylated at serine 643, PKCμ phosphorylated at serine 744/748, PKCμ phosphorylated at serine 916, PKCθ phosphorylated at threonine 538, PKCξ/λ phosphorylated at threonine 410 and 403, total PKCμ, total c-Jun, and c-Jun phosphorylated at serine 73 were purchased from Cell Signaling(Beverly, MA). Rabbit polyclonal antibodies against total p38 MAP kinase, p38 MAPK phosphorylated at threonine 180 and tyrosine 182, total p42/44 MAP kinase, p42/44 MAP kinase phosphorylated at threonine 202 and tyrosine 204, total JNK, and JNK phosphorylated at Thr-183, and Tyr-185 were purchased from New England Biolabs (Beverly, MA). Mouse monoclonal antibody against total PKCθ was purchased from Santa Cruz Biotechnology. SP600125, a JNK inhibitor, was purchased from Biomol(Plymouth Meeting, PA). HEK293 cells stably expressing recombinant human IRS1 (HEK293.IRS1 cells) were a gift from Dr. Richard Roth. The cells were maintained in Dulbecco's modified Eagle's medium with 10% fetal bovine serum, 100 units/ml penicillin, and 100 μg/ml streptomycin (Invitrogen) at 37 °C in 10% CO2. For the experiments, the cells were first incubated in serum-free Dulbecco's modified Eagle's medium overnight. The cells were then treated with vehicle, salicylic acid, or SP600125 for 2 h at 37 °C. PMA or TNFα were then added to the media. After incubation for another 30 min, the cells were exposed to insulin for 15 min, rinsed with ice-cold phosphate-buffered saline, and then lysed in ice-cold lysis buffer. The lysis buffer contains 20 mm HEPES, pH7.4, 1% Triton X-100, 20 mm β-glycerophosphate, 150 mm sodium chloride, 1 mm sodium orthovanadate, 10 mmsodium fluoride, and 1× concentration of a protease inhibitor mixture (Roche Diagnostics). Cell lysates were cleared by centrifugation. Protein concentrations were determined using Bradford reagent (Bio-Rad). Cell lysates were resuspended in SDS-loading buffer (Invitrogen) and separated in precast 4–20% gradient NuPAGE SDS-PAGE gels (Invitrogen). The proteins were then transferred to polyvinylidene difluoride membrane and probed with primary antibodies. Detection was performed with ECF Western blotting Kit (Amersham Biosciences) by scanning with a Storm® gel and blot imaging system (Amersham Biosciences) per the manufacturer's recommendation. Both PMA (19De Fea K. Roth R.A. Biochemistry. 1997; 36: 12939-12947Crossref PubMed Scopus (223) Google Scholar, 27Takayama S. White M.F. Kahn C.R. J. Biol. Chem. 1988; 263: 3440-3447Abstract Full Text PDF PubMed Google Scholar, 28Bollag G.E. Roth R.A. Beaudoin J. Mochly-Rosen D. Koshland D.E., Jr. Proc. Natl. Acad. Sci. U. S. A. 1986; 83: 5822-5824Crossref PubMed Scopus (205) Google Scholar) and TNFα (14Kanety H. Feinstein R. Papa M.Z. Hemi R. Karasik A. J. Biol. Chem. 1995; 270: 23780-23784Abstract Full Text Full Text PDF PubMed Scopus (385) Google Scholar, 18Feinstein R. Kanety H. Papa M.Z. Lunenfeld B. Karasik A. J. Biol. Chem. 1993; 268: 26055-26058Abstract Full Text PDF PubMed Google Scholar, 29Lang C.H. Dobrescu C. Bagby G.J. Endocrinology. 1992; 130: 43-52Crossref PubMed Scopus (400) Google Scholar) have been shown to inhibit insulin signaling in cultured cells. Akt phosphorylation (and activation) is a key event in insulin signaling and has been commonly used as a surrogate marker for assessing the level of activation of the proximal insulin signaling pathway (30Shepherd P.R. Withers D.J. Siddle K. Biochem. J. 1998; 333: 471-490Crossref PubMed Scopus (841) Google Scholar). We therefore determined whether PMA and/or TNFα inhibit insulin-stimulated Akt phosphorylation in the HEK293.IRS1 cells. If so, we sought to see if such inhibition could be reversed by salicylic acid. As shown in Fig. 1 A, insulin stimulation resulted in an apparent increase in Akt phosphorylation at threonine 308 (Akt-pT308) in the HEK293.IRS1 cells (Fig. 1 A, lane 2 versus 1). PMA treatment reduced insulin-stimulated Akt phosphorylation (Fig. 1 A, lane 3 versus 2). Although salicylic acid pre-treatment by itself did not affect insulin-stimulated Akt phosphorylation (Fig. 1 A, lane 5 versus 2), it reversed the negative effects by PMA (Fig. 1 A, lane 4 versus 3). Similar observations were made on cells treated with TNFα (Fig. 1 B). Taken together, these results suggest that PMA and TNFα negatively affect insulin signaling in HEK293.IRS1 cells and that such negative regulation can be reversed by salicylic acid. Both PMA and TNFα have been shown to inhibit insulin signaling by promoting IRS1 phosphorylation at multiple serine residues (13Rui L. Aguirre V. Kim J.K. Shulman G.I. Lee A. Corbould A. Dunaif A. White M.F. J. Clin. Invest. 2001; 107: 181-189Crossref PubMed Scopus (492) Google Scholar, 14Kanety H. Feinstein R. Papa M.Z. Hemi R. Karasik A. J. Biol. Chem. 1995; 270: 23780-23784Abstract Full Text Full Text PDF PubMed Scopus (385) Google Scholar, 18Feinstein R. Kanety H. Papa M.Z. Lunenfeld B. Karasik A. J. Biol. Chem. 1993; 268: 26055-26058Abstract Full Text PDF PubMed Google Scholar, 19De Fea K. Roth R.A. Biochemistry. 1997; 36: 12939-12947Crossref PubMed Scopus (223) Google Scholar, 20Ravichandran L.V. Esposito D.L. Chen J. Quon M.J. J. Biol. Chem. 2001; 276: 3543-3549Abstract Full Text Full Text PDF PubMed Scopus (192) Google Scholar, 21Strack V. Hennige A.M. Krutzfeldt J. Bossenmaier B. Klein H.H. Kellerer M. Lammers R. Haring H.U. Diabetologia. 2000; 43: 443-449Crossref PubMed Scopus (52) Google Scholar). Furthermore, it was reported recently that IRS1 Ser-307 phosphorylation mediates the inhibitory effects of TNFα on insulin signaling (17Aguirre V. Werner E.D. Giraud J. Lee Y.H. Shoelson S.E. White M.F. J. Biol. Chem. 2001; 17: 17Google Scholar,22Aguirre V. Uchida T. Yenush L. Davis R. White M.F. J. Biol. Chem. 2000; 275: 9047-9054Abstract Full Text Full Text PDF PubMed Scopus (1181) Google Scholar). We therefore investigated whether PMA and/or TNFα induce IRS Ser-307 phosphorylation in the HEK293.IRS1 cells. We further determined whether salicylic acid could attenuate the effect of PMA or TNFα on IRS Ser-307 phosphorylation. As shown in Fig. 2 A, PMA treatment resulted in ∼4-fold increase in IRS1 Ser-307 phosphorylation. Such phosphorylation was reversed by pre-treatment with salicylic acid. Similar observations were made using cells treated with TNFα (Fig. 2 B). Taken together, these results suggest that PMA and TNFα induce IRS1 Ser-307 phosphorylation and that such phosphorylation can be effectively blocked by salicylic acid treatment in the HEK293.IRS1 cells. PMA is known to activate multiple isoforms of classical and novel PKCs (31Quest A.F. Enzyme Protein. 1996; 49: 231-261Crossref PubMed Scopus (65) Google Scholar, 32Liu W.S. Heckman C.A. Cell. Signal. 1998; 10: 529-542Crossref PubMed Scopus (446) Google Scholar). We performed experiments to identify the PKC isoforms that were activated by PMA in the HEK293.IRS1 cells and determine whether the activation of any of the PKC isoforms was reversed by salicylic acid. To this end, Western blots were performed using antibodies against PKC isoforms that were phosphorylated at their activation loops, autophosphorylation sites, or at hydrophobic sites within the C terminus of the enzyme. Such phosphorylation is known to correlate to the activation and catalytic activity of the PKC isoforms (33Newton A.C. Chem. Rev. 2001; 101: 2353-2364Crossref PubMed Scopus (837) Google Scholar, 34Parekh D.B. Ziegler W. Parker P.J. EMBO J. 2000; 19: 496-503Crossref PubMed Scopus (513) Google Scholar). As expected, PMA treatment did not activate PKCξ/λ, the atypical PKC isoforms, in the HEK293.IRS1 cells (Fig. 3,bottom panel). Although PMA did not appear to activate classical PKCα/βII (Fig. 3, second panel from top), it activated novel PKCμ, PKCδ, and PKCθ (Fig. 3, third through ninth panels from top). Pre-treatment with salicylic acid did not appear to affect PMA-mediated activation of any of these PKC isoforms. Taken together, these results suggest that salicylic acid did not affect PMA-mediated PKC activation. Some of the PMA-responsive PKC isoforms (i.e. PKCθ and PKCζ) have been implicated in the activation of IKKβ (35Lin X. O'Mahony A., Mu, Y. Geleziunas R. Greene W.C. Mol. Cell. Biol. 2000; 20: 2933-2940Crossref PubMed Scopus (229) Google Scholar, 36Trushin S.A. Pennington K.N. Algeciras-Schimnich A. Paya C.V. J. Biol. Chem. 1999; 274: 22923-22931Abstract Full Text Full Text PDF PubMed Scopus (155) Google Scholar, 37Leitges M. Sanz L. Martin P. Duran A. Braun U. Garcia J.F. Camacho F. Diaz-Meco M.T. Rennert P.D. Moscat J. Mol. Cell. 2001; 8: 771-780Abstract Full Text Full Text PDF PubMed Scopus (325) Google Scholar). It has also been reported that TNFα activates IKKβ (2Karin M. Delhase M. Semin. Immunol. 2000; 12: 85-98Crossref PubMed Scopus (865) Google Scholar). We decided to investigate whether PMA and/or TNFα activate IKKβ in the HEK293.IRS1 cells and, if so, whether such activation could be reversed by salicylic acid. To this end, we determined the activation of IKKα and IKKβ using an antibody that specifically recognizes both IKKα phosphorylated at serine 180 and IKKβ phosphorylated at serine 181. Phosphorylation of serine 176 and 180 in IKKα (177 and 181 in IKKβ) are required for activation of the kinases (38Mercurio F. Zhu H. Murray B.W. Shevchenko A. Bennett B.L., Li, J. Young D.B. Barbosa M. Mann M. Manning A. Rao A. Science. 1997; 278: 860-866Crossref PubMed Scopus (1855) Google Scholar). In fact, for IKKβ there is full correspondence between its kinase activity and its Ser-177 and Ser-181 phosphorylation (39Delhase M. Hayakawa M. Chen Y. Karin M. Science. 1999; 284: 309-313Crossref PubMed Scopus (754) Google Scholar). As shown in Fig. 4, PMA treatment led to phosphorylation (and activation) of both IKKα and IKKβ (Fig. 4,top three panels, lane 3 versus 1). In contrast, TNFα stimulation did not result in the phosphorylation of either IKKα or IKKβ (Fig. 4,top three panels, lane 2 versus 1). In addition, TNFα treatment did not activate PKC isoforms based on Western blot using an antibody against phosphorylated pan-PKC (Fig. 4, bottom panel,lane 2 versus 1) or antibodies against phosphorylated individual PKC isoforms (data not shown). The effect of PMA to induce IKKα and IKKβ phosphorylation implies that these two IKKs are functional in these cells. Because the same concentration of TNFα that was sufficient to impair insulin signaling did not activate either IKKα or IKKβ, it appears that IKKβ is not involved in TNFα-mediated intracellular signaling in these cells. Furthermore, these results exclude the possibility that the salicylic acid-elicited blockade of TNFα-induced insulin resistance is mediated via IKKβ in these cells. JNK has been reported to phosphorylate IRS1 at serine 307 and inhibit insulin signaling (17Aguirre V. Werner E.D. Giraud J. Lee Y.H. Shoelson S.E. White M.F. J. Biol. Chem. 2001; 17: 17Google Scholar, 22Aguirre V. Uchida T. Yenush L. Davis R. White M.F. J. Biol. Chem. 2000; 275: 9047-9054Abstract Full Text Full Text PDF PubMed Scopus (1181) Google Scholar). We performed experiments to address whether JNK was activated by PMA and/or TNFα in the HEK293.IRS1 cells. As shown in Fig. 5 A, PMA treatment leads to phosphorylation (and activation) of the 54- and 46-kDa isoforms of JNK. A similar observation was made about TNFα-treated cells (Fig. 5 B). Taken together, these results demonstrated that both PMA and TNFα treatments in the HEK293.IRS1 cells activate JNK. Next we assessed whether JNK is critical or required for PMA and TNFα-mediated effects on insulin signaling. To address this question, we determined the effect of SP600125, a JNK small molecular inhibitor (40Bennett B.L. Sasaki D.T. Murray B.W. O'Leary E.C. Sakata S.T., Xu, W. Leisten J.C. Motiwala A. Pierce S. Satoh Y. Bhagwat S.S. Manning A.M. Anderson D.W. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 13681-13686Crossref PubMed Scopus (2244) Google Scholar), on PMA- and TNFα-induced IRS1 Ser-307 phosphorylation. It was reported that, at 5μm, SP600125 inhibits JNK activity in intact hemopoietic cells (40Bennett B.L. Sasaki D.T. Murray B.W. O'Leary E.C. Sakata S.T., Xu, W. Leisten J.C. Motiwala A. Pierce S. Satoh Y. Bhagwat S.S. Manning A.M. Anderson D.W. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 13681-13686Crossref PubMed Scopus (2244) Google Scholar). We first determined whether SP600125 inhibits JNK activity in the HEK293.IRS1 cells. As shown in Fig. 6, SP600125 did not affect JNK phosphorylation (and activation) upon both TNFα (Fig. 6 B,third panel from top,lanes 5 and 6 versus 3 and4) and PMA treatment (data not shown). On the other hand, SP600125 inhibited phosphorylation of the transcription factor c-Jun, a physiological substrate of JNK, upon PMA treatment (Fig. 6 A,third panel from top, lanes 5 and 6 versus 3 and 4). We had difficulty detecting TNFα-induced c-Jun phosphorylation (data not shown). Taken together, these results confirm that SP600125 indeed inhibits the kinase activity but not the activation process of JNK in the HEK293.IRS1 cells. It was reported that, although SP600125 potently inhibits the kinase activity of recombinant JNK enzymes (IC50 of 40–90 nm), it is rather inactive against the kinase activity of recombinant p42/44, p38, IKKα, IKKβ, and PKC isoforms (including PKCθ) (IC50 of more than 10 μm), indicating a high degree of JNK selectivity for SP600125 (40Bennett B.L. Sasaki D.T. Murray B.W. O'Leary E.C. Sakata S.T., Xu, W. Leisten J.C. Motiwala A. Pierce S. Satoh Y. Bhagwat S.S. Manning A.M. Anderson D.W. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 13681-13686Crossref PubMed Scopus (2244) Google Scholar). Furthermore, it was also reported that at 5 μm or less the effect of SP600125 on JNK activity is specific because SP600125 does not affect the activity or activation of other serine/threonine kinases in the intact hemopoietic cells (40Bennett B.L. Sasaki D.T. Murray B.W. O'Leary E.C. Sakata S.T., Xu, W. Leisten J.C. Motiwala A. Pierce S. Satoh Y. Bhagwat S.S. Manning A.M. Anderson D.W. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 13681-13686Crossref PubMed Scopus (2244) Google Scholar). We decided to confirm such specificity in the HEK293.IRS1 cells. Indeed, SP600125 did not appear to affect PMA-induced activating phosphorylation of IKKα and IKKβ (Fig. 6 A, bottom panel, lanes 5and 6 versus 3 and 4). SP600125 treatment also did not affect PMA- (data not shown) and TNFα-induced activation of both p42/44 MAPK kinases (Fig. 6 B, fourth panel from top, lanes 5 and6 versus 3 and 4) and p38 MAPK (Fig. 6 B, bottom panel,lanes 5 and 6 versus 3 and4). Taken together, these results are consistent with the notion that at 5 μm SP600125 specifically inhibits JNK activity in the HEK293.IRS1 cells (see more in Fig. 6 C and the corresponding discussion for additional nonspecific activity). Finally, we determined whether JNK enzymes are important in mediating PMA- and TNFα-induced IRS1 Ser-307 phosphorylation. The results indicate that SP600125 treatment prevented IRS1 Ser-307 phosphorylation induced by both PMA (Fig. 6 A, top two panels, lanes 5 and 6 versus 3 and 4) and TNFα (Fig. 6 B, top two panels,lanes 5 and 6 versus 3 and4). Taken together, these results suggest that JNK is activated by PMA and TNFα and may be important for PMA and TNFα-induced IRS1 Ser-307 phosphorylation. mediated JNK Activation—The above results suggest that JNK was required for PMA- and TNFα-mediated IRS1 Ser-307 phosphorylation and insulin resistance. A key question was whether the effects of salicylic acid on insulin signaling, including IRS1 Ser-307 phosphorylation, were mediated by the inhibition of JNK. Thus, we determined whether salicylic acid affected JNK activation by PMA and TNFα. As shown in Fig. 5, salicylic acid significantly inhibited the activating phosphorylation of both the 54- and 46-kDa JNK induced by both PMA (Fig. 5 A) and TNFα (Fig. 5 B). Furthermore, the inhibition of JNK phosphorylation by salicylic acid appeared to be specific because salicylic acid did not inhibit the activating phosphorylation of p42/44 MAP kinases induced by both PMA (Fig. 5 A, bottom two panels,lanes 5 and 6 versus 3 and4) and TNFα (data not shown). Salicylate acid also did not affect the activating phosphorylation of the p38 MAPK kinases by PMA and TNFα (data not shown). Taken together, these results suggest that salicylic acid specifically inhibits JNK activation by PMA and TNFα. To further demonstrate that JNK plays a negative role in insulin signaling and that its activation is required for PMA and TNFα-mediated insulin resistance, we decided to determine whether SP600125 can reverse PMA and TNFα-mediated inhibition on insulin signaling. Rather surprisingly, however, we found that SP600125 by itself completely inhibited insulin-stimulated threonine phosphorylation of Akt (Fig. 6 C, top panel) as well as tyrosine phosphorylation of IRS-1 (middle panel) and the insulin receptor β subunit (bottom panel). These results suggest that SP600125 has some nonspecific activity that has not been reported previously (40Bennett B.L. Sasaki D.T. Murray B.W. O'Leary E.C. Sakata S.T., Xu, W. Leisten J.C. Motiwala A. Pierce S. Satoh Y. Bhagwat S.S. Manning A.M. Anderson D.W. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 13681-13686Crossref PubMed Scopus (2244) Google Scholar). Unfortunately, such nonspecific activity excludes us from determining whether SP600126 treatments can reverse PMA- and TNFα-mediated inhibition of insulin signaling. In the current study, several observations have been made using a cellular model system for insulin signaling, i.e. a HEK293 cell line stably expressing recombinant human IRS1. First, both PMA and TNFα inhibited insulin-stimulated activating phosphorylation of Akt and promoted IRS1 Ser-307 phosphorylation. Second, the effects of PMA and TNFα on Akt phosphorylation and IRS1 serine phosphorylation were reversed by salicylic acid pretreatment. Third, PMA treatment led to activation of multiple serine/threonine kinases including PKC isoforms and IKKβ; however, such activation was not affected by salicylic acid. Fourth, TNFα treatment did not active either PKC isoforms or IKKβ. Fifth, JNK was activated by both TNFα and PMA, and its inhibition by a small molecule inhibitor prevented TNFα- and PMA-induced IRS1 Ser-307 phosphorylation. The small molecular inhibitor, however, has some additional nonspecific activity. Finally, salicylic acid also inhibited TNFα- and PMA-induced activating phosphorylation of JNK. The goal of this study was to further elucidate the role of IKKβ and salicylic acid in inhibiting or augmenting insulin signaling, respectively. In our experimental system, PMA and TNFα were chosen to induce insulin resistance for several reasons. First, the negative effects of PMA and TNFα on insulin signaling are well documented, at least in cultured cells in vitro (11Hotamisligil G.S. Peraldi P. Budavari A. Ellis R. White M.F. Spiegelman B.M. Science. 1996; 271: 665-668Crossref PubMed Scopus (2223) Google Scholar, 14Kanety H. Feinstein R. Papa M.Z. Hemi R. Karasik A. J. Biol. Chem. 1995; 270: 23780-23784Abstract Full Text Full Text PDF PubMed Scopus (385) Google Scholar, 18Feinstein R. Kanety H. Papa M.Z. Lunenfeld B. Karasik A. J. Biol. Chem. 1993; 268: 26055-26058Abstract Full Text PDF PubMed Google Scholar, 19De Fea K. Roth R.A. Biochemistry. 1997; 36: 12939-12947Crossref PubMed Scopus (223) Google Scholar, 27Takayama S. White M.F. Kahn C.R. J. Biol. Chem. 1988; 263: 3440-3447Abstract Full Text PDF PubMed Google Scholar, 28Bollag G.E. Roth R.A. Beaudoin J. Mochly-Rosen D. Koshland D.E., Jr. Proc. Natl. Acad. Sci. U. S. A. 1986; 83: 5822-5824Crossref PubMed Scopus (205) Google Scholar, 29Lang C.H. Dobrescu C. Bagby G.J. Endocrinology. 1992; 130: 43-52Crossref PubMed Scopus (400) Google Scholar,41Hotamisligil G.S. Murray D.L. Choy L.N. Spiegelman B.M. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 4854-4858Crossref PubMed Scopus (1052) Google Scholar). Second, IKKβ is a well established downstream kinase for TNFα (2Karin M. Delhase M. Semin. Immunol. 2000; 12: 85-98Crossref PubMed Scopus (865) Google Scholar). It has also been shown that IKKβ can be activated by PKC isoforms (i.e. PKCθ and PKCζ) (35Lin X. O'Mahony A., Mu, Y. Geleziunas R. Greene W.C. Mol. Cell. Biol. 2000; 20: 2933-2940Crossref PubMed Scopus (229) Google Scholar, 36Trushin S.A. Pennington K.N. Algeciras-Schimnich A. Paya C.V. J. Biol. Chem. 1999; 274: 22923-22931Abstract Full Text Full Text PDF PubMed Scopus (155) Google Scholar, 37Leitges M. Sanz L. Martin P. Duran A. Braun U. Garcia J.F. Camacho F. Diaz-Meco M.T. Rennert P.D. Moscat J. Mol. Cell. 2001; 8: 771-780Abstract Full Text Full Text PDF PubMed Scopus (325) Google Scholar). In fact, it was recently reported that PKCζ deficiency in mice impaired IKKβ activation (37Leitges M. Sanz L. Martin P. Duran A. Braun U. Garcia J.F. Camacho F. Diaz-Meco M.T. Rennert P.D. Moscat J. Mol. Cell. 2001; 8: 771-780Abstract Full Text Full Text PDF PubMed Scopus (325) Google Scholar). The results from the current study indicate that PMA activates multiple isoforms of PKC and IKK enzymes in the HEK293.IRS1 cells (Fig. 3). In addition to activating IKKβ (2Karin M. Delhase M. Semin. Immunol. 2000; 12: 85-98Crossref PubMed Scopus (865) Google Scholar), TNFα has also been shown to activate PKC in multiple cell types (42Matsubara N. Fuchimoto S. Orita K. Immunology. 1991; 73: 457-459PubMed Google Scholar, 43Matsubara N. Mizukawa K. Iwagaki H. Fuchimoto S. Kimura Y. Orita K. Cell. Signal. 1993; 5: 811-816Crossref PubMed Scopus (2) Google Scholar, 44Prasanna G. Dibas A. Brown K. Yorio T. J. Ocul. Pharmacol. Ther. 1998; 14: 401-412Crossref PubMed Scopus (13) Google Scholar, 45Ferro T.J. Parker D.M. Commins L.M. Phillips P.G. Johnson A. Am. J. Physiol. 1993; 264: L7-14PubMed Google Scholar, 46Wyatt T.A. Ito H. Veys T.J. Spurzem J.R. Am. J. Physiol. 1997; 273: L1007-L1012PubMed Google Scholar). It was therefore rather surprising that TNFα did not activate IKKα, IKKβ, and the multiple PKC isoforms examined in the HEK293.IRS1 cells (Fig. 4). Based on these observations, we conclude that in the HEK293.IRS1 cells IKKβ is intact but is not involved in TNFα signaling. Therefore, the lack of IKKβ activation by TNFα in the HEK293.IRS1 cells provided us with a unique setting to determine whether IKKβ activation and its inhibition are involved and/or required for salicylate-mediated insulin sensitization. It then follows that the effect of TNFα on insulin signaling in the HEK293.IRS1 cells is independent of IKKβ activation. Because salicylate was readily able to reverse TNFα-induced effects on insulin signaling, it also appears that these effects of the drug must be also IKKβ-independent. In contrast, the current study does not provide sufficient evidence to assess whether the action of salicylic acid to reverse the effects of PMA on insulin action is mediated via IKKβ inhibition. How does salicylic acid reverse insulin resistance in these cells if not via inhibition of IKKβ? We considered that salicylic acid might improve insulin sensitivity by inhibiting kinase(s) other than IKKβ, which were responsive to both PMA and TNFα. To this end, we examined the serine/threonine kinase JNK. JNK has been shown to be activated by both PMA (47Werlen G. Jacinto E. Xia Y. Karin M. EMBO J. 1998; 17: 3101-3111Crossref PubMed Scopus (253) Google Scholar, 48Ghaffari-Tabrizi N. Bauer B. Villunger A. Baier-Bitterlich G. Altman A. Utermann G. Uberall F. Baier G. Eur. J. Immunol. 1999; 29: 132-142Crossref PubMed Scopus (108) Google Scholar, 49Kaneki M. Kharbanda S. Pandey P. Yoshida K. Takekawa M. Liou J.R. Stone R. Kufe D. Mol. Cell. Biol. 1999; 19: 461-470Crossref PubMed Scopus (66) Google Scholar) and TNFα (50Surapisitchat J. Hoefen R.J., Pi, X. Yoshizumi M. Yan C. Berk B.C. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 6476-6481Crossref PubMed Scopus (197) Google Scholar, 51Chan E.D. Winston B.W. Jarpe M.B. Wynes M.W. Riches D.W. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 13169-13174Crossref PubMed Scopus (81) Google Scholar, 52Natoli G. Costanzo A. Ianni A. Templeton D.J. Woodgett J.R. Balsano C. Levrero M. Science. 1997; 275: 200-203Crossref PubMed Scopus (415) Google Scholar). Most interestingly, TNFα was recently shown to inhibit insulin signaling by activating JNK, which subsequently becomes associated with Ser-307 in IRS1 and phosphorylates it (22Aguirre V. Uchida T. Yenush L. Davis R. White M.F. J. Biol. Chem. 2000; 275: 9047-9054Abstract Full Text Full Text PDF PubMed Scopus (1181) Google Scholar). To elucidate the role of JNK in PMA- and TNFα-induced insulin resistance and its reversal by salicylic acid, we performed a separate series of experiments and made several interesting observations. First, by demonstrating that both PMA and TNFα treatment induced activating phosphorylation of JNK (Fig. 5), we confirmed that JNK is involved in signaling in response to both PMA and TNFα in the HEK 293 cells. Second, by showing that SP600125 inhibits PMA- and TNFα-induced IRS1 Ser-307 phosphorylation, we propose that JNK activity may be important for the effects of PMA and TNFα on insulin signaling in these cells (see more discussion later on the nonspecific activity of SP600125). Finally, we showed that salicylic acid inhibits PMA- and TNFα-induced JNK activating phosphorylation (Fig. 6). Thus, these observations suggest that JNK is activated by (and its activity may be important for) PMA- and TNFα-induced insulin resistance. Most interestingly, the results also suggest that salicylic acid improves insulin sensitivity potentially by inhibiting JNK activation. Fig. 7 summarizes the working hypothesis on the effects of salicylic acid on PMA- and TNFα-induced insulin resistance in the HEK293.IRS1 cells. In short, we propose that JNK may be directly responsible for IRS1 Ser-307 phosphorylation. IRS1 phosphorylated at Ser-307 has an impaired ability to undergo tyrosine phosphorylation by insulin receptor, leading to insulin resistance as reflected in reduced Akt phosphorylation (and activation). Salicylic acid reversed the effects of PMA and TNFα in association with reduced JNK activation. Although the detailed mechanism by which salicylic acid inhibits JNK activation in these cells remains unknown, it is most likely to be IKKβ-independent. Although both PMA and TNFα treatment lead to JNK activation, the signaling pathways were apparently different. PMA activates PKC isoforms as well IKK kinases. TNFα did not activate either. Nevertheless, salicylic acid appears to inhibit JNK activation by both PMA and TNFα. In summary, the results from the current study suggest that IKKβ inhibition is not necessarily critical to the mechanism(s) for salicylate-mediated insulin sensitization in the TNFα treatment paradigm. One approach to further validate our results would be to determine whether PMA and TNFα can induce IRS1 Ser-307 phosphorylation in IKKβ-deficient cells and, if so, whether the effects by PMA and TNFα can be reversed by salicylic acid. However, it also remains to be established that the antidiabetic effects of salicylate in vivo are directly tied to inhibition of IKKβ. Nevertheless, the apparent existence of IKKβ-independent alternative pathway(s) underscores the need for additional knowledge regarding the mechanism(s) of salicylate-mediated insulin sensitization Finally, although the current study suggests that JNK activation plays an important role in PMA and TNFα-mediated insulin resistance and that its inhibition may be responsible for salicylic acid-mediated insulin sensitization, such a proposal needs further validation because some of the results were derived from the use of the JNK inhibitor SP600125. We found that although SP600125 does not affect PMA- or TNFα-induced activation of several kinases (Fig. 6,A and B), it inhibits insulin signaling (Fig. 6 C), indicating that the inhibitor may not be as specific as originally reported (40Bennett B.L. Sasaki D.T. Murray B.W. O'Leary E.C. Sakata S.T., Xu, W. Leisten J.C. Motiwala A. Pierce S. Satoh Y. Bhagwat S.S. Manning A.M. Anderson D.W. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 13681-13686Crossref PubMed Scopus (2244) Google Scholar). One approach to further test the current theory would be to determine whether PMA and TNFα induce IRS1 Ser-307 phosphorylation and inhibit insulin signaling in JNK-deficient cells.