Diabetes In Press, published online June 11, 2007 Characterisation of a protein kinase B inhibitor in vitro and in insulin treated liver cells. Received for publication 17 March 2007 and accepted in revised form 5 June 2007. Additional information for this article can be viewed in an online appendix at http://diabetes.diabetesjournals.org Lisa Logie, Antonio J. Ruiz-Alcaraz, Michael Keane*$, Yvonne L. Woods*, Jennifer Bain*, Rudolfo Marquez$, Dario R. Alessi* and Calum Sutherland. Division of Pathology and Neuroscience, *MRC Protein Phosphorylation Unit, $Division of Biological Chemistry, University of Dundee, Dundee, Scotland. Corresponding author Dr Calum Sutherland Pathology and Neurosciences University of Dundee, Ninewells Hospital Dundee, Scotland, UK, DD1 9SY email: [email protected] Word Count: 3978, Tables: 2, Figures: 6. Keywords: PKB, AKT, PEPCK, insulin, Akti-1/2, diabetes. Abbreviations: PKB, protein kinase B; G6Pase, glucose-6-phosphatase; PEPCK, phosphoenolpyruvate carboxykinase; IGFBP1, IGF binding protein-1; TIRE, thymine-rich insulin response element; PRAS40, proline rich akt substrate of 40kDa; Akti, Akt inhibitor. Running Title: PKB inhibition and insulin action. Copyright American Diabetes Association, Inc., 2007 Abstract: Objective: Abnormal expression of the hepatic gluconeogenic genes (glucose-6-phosphatase (G6Pase) and phosphoenolpyruvate carboxykinase (PEPCK)) contributes to hyperglycemia. These genes are repressed by insulin but this process is defective in diabetes. Protein kinase B (PKB) is implicated in this action of insulin. An inhibitor of PKB, termed Akti-1/2, was recently reported, however the specificity and efficacy against insulin induced PKB was not reported. Our aim was to characterise the specificity and efficacy of Akti-1/2 in cells exposed to insulin and then establish whether inhibition of PKB is sufficient to prevent regulation of hepatic gene expression by insulin. Design: Akti-1/2 was assayed against seventy kinases in vitro, and its ability to block PKB activation in cells exposed to insulin fully characterised. Results: Akti-1/2 exhibits high selectivity toward PKBα and PKBβ. Complete inhibition of PKB activity is achieved in liver cells incubated with 1 to 10µM Akti-1/2 and this blocks insulin regulation of PEPCK and G6Pase expression. Our data demonstrates that only 5-10% of maximal insulin induced PKB is required to fully repress PEPCK and G6Pase expression. Finally, we demonstrate reduced insulin sensitivity of these gene promoters in cells exposed to sub-maximal concentrations of Akti-1/2, however full repression of the genes can still be achieved by high concentrations of insulin. Conclusion: This work establishes the requirement for PKB activity in the insulin regulation of PEPCK, G6Pase and a third insulin regulated gene, IGFBP1, suggests a high degree of functional reserve and identifies Akti-1/2 as a useful tool to delineate PKB function in the liver. Protein kinase B (PKB) is a member of the induce insulin responsive DNA sequences AGC family of protein kinases (1-3). In within these gene promoters (24-27). This mammals there are three isoforms (PKBα, data suggests that insulin turns off these gene PKBβ and PKBγ) (1). PKB is activated promoters by activating the PI 3-kinase- following induction of phosphatidylinositol PDK1-PKB pathway to inhibit FOXO. (PI) 3-kinase activity and the resultant However the importance of PKB and/or generation of the lipid second messengers PI FOXO in the regulation of these genes has 3,4,5 trisphosphate (PIP3) and PI 3,4 been questioned (18; 28). For example, bisphosphate (4). These lipids bind to the PH overexpression of dominant negative PKB domain of PKB, altering its conformation and does not block insulin action on PEPCK (29) permitting access to upstream protein kinases and G6Pase (22), inhibitors of mTOR will (5). Phosphoinositide-dependent protein block insulin regulation of IGFBP1 but not kinase (PDK)-1 phosphorylates PKB at Thr- PKB or FOXO (21; 26), while inhibitors of 308 (6) and a second phosphorylation (at Ser- GSK3 (also downstream of PKB) will inhibit 473) occurs through the action of an these genes without regulating FOXO activity alternative kinase, such as the rapamycin (30; 31). It is also suggested that insulin can insensitive mTOR, TORC2 (7). Therefore regulate FOXO activity independently of most growth factors, including PDGF, EGF PKB (32). Finally, a base by base mutational and insulin, that are potent activators of PI 3- analysis of the insulin responsive promoter kinase, also strongly induce PKB in cells. sequences from PEPCK and IGFBP1 showed that the DNA sequence required for the One of the first substrates of PKB to be response to insulin is not identical to that characterised was GSK3, as part of the insulin required for response to FOXO (33). All of signalling pathway that regulates glycogen this data suggests that although PKB and metabolism (8). Since then multiple potential FOXO can regulate these genes the inhibition substrates of PKB have been proposed of FOXO by PKB may not be absolutely including the pro-apoptotic protein Bad (9; required for their response to insulin. 10), the TSC2 gene product (11), the Rab- GAP AS160 (12), PRAS40 (13) and the key Interestingly, PKBβ KO mice exhibit reduced forkhead transcription factor sub-family, insulin sensitivity and can develop diabetes FOXO. PKB phosphorylates FOXO on and lipoatrophy (34), while PKBα knockout several residues promoting its inactivation (KO) animals are smaller, have reduced and exclusion from the nucleus (14-16). A lifespan and defective adipogenesis (35). growing number of insulin-inhibited genes are PKBγ KO mice have a reduced brain size, proposed to be targets of FOXO. These possibly due to its relatively high expression include Glucose-6-Phosphatase (G6Pase), in this tissue (36). These data suggest that Phosphoenolpyruvate carboxyinase (PEPCK) PKBβ is the most likely link between PI 3- and the IGF-binding protein-1 (IGFBP-1) kinase and the gluconeogenic genes PEPCK (17). All three genes are completely repressed and G6Pase, although insulin regulation of in liver cells exposed to insulin (18) or in these genes was not characterised in this intact liver following feeding (19). This gene model (34). In contrast, PKBα is the major regulation requires PI 3-kinase (20-22) and insulin activated PKB isoform detectable in PDK1 (19) activity, and can be recapitulated hepatocytes (37). by overexpression of active PKB (23). Meanwhile, overexpression of FOXO will Recently, selective non-ATP competitive and phospho TSC2 and PRAS40 antibodies inhibitors of PKB were reported (38; 39). were generated in the Division of Signal Compound 16h (now termed Akti-1/2) Transduction and Therapy, University of potently inhibits PKBα and PKBβ in vitro Dundee. Anti-β-Actin was purchased from and in cells, with relatively poor inhibition of Sigma-Aldrich, Inc. (St Louis, MO, U.S.A.). PKBγ (39). Akti-1/2 also prevents phosphorylation at Ser308 by PDK1 in vitro Akti-1/2 compound Synthesis and at Ser308 and Thr473 in cells without The dual PKB α/β inhibitor (compound 16h) inhibiting PDK1 activity (38). Interaction of was synthesised in-house through a Akti-1/2 with the PH domain of PKB prevents modification of the procedure of Lindsley and the conformational change required for co-workers (39). Purity was established as phosphorylation by upstream kinases (38). >98% by 1H-NMR and LCMS. This Since isolated hepatocytes contain very little, compound (termed Akti-1/2 or AKT inhibitor if any, PKBγ (37), we hypothesised that Akti- VIII in the Calbiochem catalogue) is a 1/2 would produce acute and complete derivative of Akti-1/2a (a compound inhibition of PKB activity in an insulin originally termed Akti-1/2 in (38)), and sensitive liver cell line (HL1c) allowing exhibits increased potency. careful analysis of the requirement for PKB in processes such as insulin regulation of gene Kinase Screen expression. Details of the kinase selectivity screens have been provided previously (40) and complete Research Design and Methods assay conditions are provided in Materials supplementary data. Briefly, all assays (25.5 Actrapid (human insulin) was from Novo µl, 21 oC, 30 min) were performed using a Nordisk A/S (Bagsværd, Denmark). 2x Biomek 2000 Laboratory Automation Universal PCR Master Mix, No AmpErase Workstation in a 96-well format (Beckman UNG from Applied Biosystems, CompleteTM Instruments, Palo Alto, CA, USA). Reactions protease inhibitor cocktail tablets from Roche, contained 5-20 mU of purified kinase along and 8-(4-chloro-phenylthio)-cAMP from with substrate peptide or protein and were Calbiochem. All primers and probes were initiated by addition of 10mM MgAcetate and synthesized and purified by Sigma-Aldrich. either 5, 20 or 50 µM ATP ([γ-33P]ATP, 800 All other chemicals were of the highest grade cpm/pmol). obtainable. Cell Culture Antibodies HL1c rat hepatoma cells were maintained in PKB isoform specific antibodies have been 1g/L glucose containing Dulbeccos Modified characterised elsewhere (37). Antibodies to Eagles Medium supplemented with 5% FCS phospho-PKB (Thr308), phospho-PKB and 100U/ml penicillin and 100ug/ml (Ser473), phospho-FOXO1 (Ser256), streptomycin. phospho-p42/p44 MAPK (Thr202/Tyr204), phospho-S6 ribosomal protein (Ser240/244) RNA Isolation and Real-Time quantitative were from Cell Signaling Technology reverse transcriptase PCR (Beverly, MA, U.S.A.), anti-PKB was from Following hormone treatments total cellular Upstate Biotechnologies (Lake Placid, NY, RNA was extracted from HL1c cells using U.S.A.), FOXO antibody was from Dr TRIreagentTM (Sigma) according to the Graham Rena, University of Dundee and total manufacturer’s instructions. cDNA was synthesized from total cellular RNA using primary antibodies at 4 °C overnight before SuperscriptTM II reverse transcriptase kit incubation for 1 h at room temperature with (Invitrogen). Real-time PCR analysis was the appropriate secondary antibody. Protein carried out using ABIPrism7700 sequence bands were visualized using ECL kit detector (Applied Biosystems). PCR reactions (Amersham Biosciences, Inc.).