Kenneth J. Murray

Selective small molecule inhibitors of glycogen synthase kinase-3 modulate glycogen metabolism and gene transcription

BACKGROUND Glycogen synthase kinase-3 (GSK-3) is a serine/threonine protein kinase, the activity of which is inhibited by a variety of extracellular stimuli including insulin, growth factors, cell specification factors and cell adhesion. Consequently, inhibition of GSK-3 activity has been proposed to play a role in the regulation of numerous signalling pathways that elicit pleiotropic cellular responses. This report describes the identification and characterisation of potent and selective small molecule inhibitors of GSK-3. RESULTS SB-216763 and SB-415286 are structurally distinct maleimides that inhibit GSK-3alpha in vitro, with K(i)s of 9 nM and 31 nM respectively, in an ATP competitive manner. These compounds inhibited GSK-3beta with similar potency. However, neither compound significantly inhibited any member of a panel of 24 other protein kinases. Furthermore, treatment of cells with either compound stimulated responses characteristic of extracellular stimuli that are known to inhibit GSK-3 activity. Thus, SB-216763 and SB-415286 stimulated glycogen synthesis in human liver cells and induced expression of a beta-catenin-LEF/TCF regulated reporter gene in HEK293 cells. In both cases, compound treatment was demonstrated to inhibit cellular GSK-3 activity as assessed by activation of glycogen synthase, which is a direct target of this kinase. CONCLUSIONS SB-216763 and SB-415286 are novel, potent and selective cell permeable inhibitors of GSK-3. Therefore, these compounds represent valuable pharmacological tools with which the role of GSK-3 in cellular signalling can be further elucidated. Furthermore, development of similar compounds may be of use therapeutically in disease states associated with elevated GSK-3 activity such as non-insulin dependent diabetes mellitus and neurodegenerative disease.

Relaxation of guinea‐pig trachea by cyclic AMP phosphodiesterase inhibitors and their enhancement by sodium nitroprusside

1 The effects of agents that elevate either cyclic AMP (the phosphodiesterase (PDE) III inhibitor siguazodan, salbutamol) or cyclic GMP (sodium nitroprusside (SNP)) on the relaxant activity of the PDE IV inhibitor, rolipram, were investigated in carbachol (0.1 μm) precontracted guinea‐pig tracheal sheets. 2 Rolipram, siguazodan and SNP caused concentration‐related reductions in tone of tissues precontracted with 0.1 μm carbachol (EC50 values 12.5; 2.73 and 0.35 μm respectively). Whilst the concentration‐response relationship for the PDE III inhibitor, siguazodan, was monophasic that of the PDE IV inhibitor, rolipram, was biphasic. 3 The relaxant activity of rolipram was markedly enhanced in the presence of 10 μm siguazodan (EC50 < 0.01 μm), 0.1 μm salbutamol (EC50 0.03 μm) and 0.3 μm SNP (EC50 0.03 μm). In contrast, the relaxant activity of siguazodan was unaffected by SNP and only modestly enhanced by rolipram (10 μm) and salbutamol (0.1 μm). 4 The relaxant activity of SNP was enhanced by the PDE V inhibitor SK&F 96231 (30 μm: EC50 0.06 μm) and rolipram (30 μm, EC50 0.08 μm) but was unaffected by 30 μm siguazodan. 5 At concentrations up to 10 μm, neither siguazodan nor rolipram elevated tracheal cyclic AMP levels. However, the combination of 10 μm rolipram and siguazodan caused a two fold increase in the cyclic AMP content (from 2.19 to 4.36 pmol cyclic AMP mg−1protein). SNP (0.1–10 μm) failed to produce a significant increase in tracheal cyclic AMP levels. At 0.1 μm the effect of SNP on tracheal cyclic AMP levels was significantly (P < 0.05) increased in the presence of rolipram but not siguadozan. 6 The results indicate that the relaxant effects of rolipram are markedly enhanced by agents that inhibit PDE III activity or elevate cyclic GMP. They support the hypothesis that SNP potentiates the effects of rolipram via the inhibitory action of cyclic GMP on hydrolysis of cyclic AMP by PDE III. The findings also suggest that whilst PDE III may be more significant in regulating basal smooth muscle tone in the absence of any exogenous stimulus to cyclic AMP accumulation, PDE IV activity may be more tightly coupled to the pool of adenylyl cyclase stimulated by β2‐adrenoceptor agonists.

OCTIMIBATE, A POTENT NON-PROSTANOID INHIBITOR OF PLATELET AGGREGATION, ACTS VIA THE PROSTACYCLIN RECEPTOR

1 Octimibate, 8‐[(1,4,5‐triphenyl‐1H‐imidazol‐2‐yl)oxy]octanoic acid, is reported to have antithrombotic properties. This is in addition to its antihyperlipidaemic effects which are due to inhibition of acyl‐CoA: cholesterol acyltransferase (ACAT). The aim of this study was to investigate the mechanism of the antithrombotic effect of octimibate, and to determine whether the effects of octimibate are mediated through prostacyclin receptors. 2 In suspensions of washed (plasma‐free) human platelets, octimibate is a potent inhibitor of aggregation; its IC50 is approx. 10 nm for inhibition of aggregation stimulated by several different agonists, including U46619 and ADP. The inhibitory effects of octimibate on aggregation are not competitive with the stimulatory agonist; the maximal response is suppressed but there is no obvious shift in potency of the agonist. In platelet‐rich plasma, octimibate inhibits agonist‐stimulated aggregation with an IC50 of approx. 200 nm. 3 Octimibate also inhibits agonist‐stimulated rises in the cytosolic free calcium concentration, [Ca2+]i, in platelets. Both Ca2+ influx and release from intracellular stores are inhibited. The effects of octimibate on aggregation and [Ca2+]i are typical of agents that act via elevation of adenosine 3′:5′‐cyclic monophosphate (cyclic AMP). Similar effects are seen with forskolin, prostacyclin (PG12) and iloprost (a stable PG12 mimetic). 4 Octimibate increases cyclic AMP concentrations in platelets and increases the cyclic AMP‐dependent protein kinase activity ratio. Octimibate stimulates adenylyl cyclase activity in human platelet membranes, with an EC50 of 200 nm. The maximal achievable activation of adenylyl cyclase by octimibate is 60% of that obtainable with iloprost. Octimibate has no effect on the cyclic GMP‐inhibited phosphodiesterase (phosphodiesterase‐III), which is the major cyclic AMP‐degrading enzyme in human platelets. 5 Octimibate inhibits, apparently competitively, the binding of [3H]‐iloprost (a stable PG12 mimetic) to platelet membranes; the estimated Ki is 150 nm. 6 The platelets of different species show considerable differences in the apparent potency of their inhibition of aggregation by octimibate; platelets from cynomolgus monkeys are 3 fold more sensitive than those from humans, while rat, cat and cow platelets are 50, 100, and 250 fold less sensitive than human platelets. The sensitivity of these different species to iloprost, however, varies over a range of only 10 fold with no obvious difference between primates and non‐primates. 7 Octimibate appears to be a potent agonist (aggregation), or partial agonist (adenylyl cyclase), at prostacyclin receptors and is the first non‐prostanoid agent of this type to be identified. The species differences in relative potency of octimibate and iloprost may reflect the existence of receptor subtypes.

8-(4-Chlorophenyl)thio-cyclic AMP is a potent inhibitor of the cyclic GMP-specific phosphodiesterase (PDE VA)

8-(4-Chlorophenyl)thio-cyclic AMP (8-CPT-cAMP), extensively used as selective activator of cyclic AMP-dependent protein kinase, has been found to be a potent inhibitor of the cyclic GMP-specific phosphodiesterase (PDE VA). Indeed, 8-CPT-cAMP (IC50 = 0.9 microM) inhibited PDE VA with a potency identical to that of zaprinast. 8-CPT-cAMP was also metabolized by PDE VA at a rate half that of cyclic GMP. The cyclic GMP-inhibited phosphodiesterase (PDE III) (IC50 = 24 microM) and the cyclic AMP-specific phosphodiesterase (PDE IV) (IC50 = 25 microM) were also inhibited by 8-CPT-cAMP. In contrast, most of the other cAMP-derivative studies showed little inhibition of any phosphodiesterase isoenzyme. These observations provide further reasons why the mechanism of the physiological effects of 8-CPT-cAMP should be interpreted with caution.

Comparison of the effect of isobutylmethylxanthine and phosphodiesterase-selective inhibitors on cAMP levels in SH-SY5Y neuroblastoma cells

A comparison of the effects of various phosphodiesterase (PDE) inhibitors upon cellular cAMP levels was undertaken in human neuroblastoma SH-SY5Y cells. When inhibitors such as rolipram and Ro 20 1724 (selective for the low Km cAMP-specific PDE) were used, cAMP levels were seen to rise dramatically under basal (< or = 60 fold) or forskolin-stimulated (< or = 200 fold) conditions. However, the non-selective PDE inhibitor isobutylmethylxanthine (IBMX) was 7-18% as effective as these other agents even at 1 mM. The poor efficacy of IBMX was not attributable to concomitant increases in cGMP, to alterations in cAMP egress or to a lack of sensitivity of the cellular PDEs to IBMX inhibition. In additivity experiments, IBMX potently and rapidly reduced cAMP that had accumulated after rolipram treatment. The fact that the agonist 2-chloroadenosine can enhance cAMP accumulation in these cells, and that cAMP elevated by rolipram or forskolin can be reduced by adenosine deaminase and theophylline suggest that cell-derived adenosine enhances cAMP in these cells in an autocrine fashion. Since IBMX is an adenosine receptor antagonist, it is suggested that its blockade of endogenous adenosine effects is at least partly responsible for its poor response when compared to other PDE inhibitors which are weaker adenosine receptor antagonists. These results forewarn against assuming that similar levels of cAMP accumulate after application of PDE inhibitors in these cells.

The effect of SK&F 95654, a novel phosphodiesterase inhibitor, on cardiovascular, respiratory and platelet function.

1 SK&F 95654 inhibited the guanosine 3′:5′‐cyclic monophosphate (cyclic GMP)‐inhibited phosphodiesterase (cGI‐PDE) with an IC50 value of 0.7 μm. The IC50 values were greater than 100 μm for the other four phosphodiesterase isoenzymes tested. The R‐enantiomer of SK&F 95654 (IC50 = 0.35 μm) was a more potent inhibitor of cGI‐PDE than was the S‐enantiomer (IC50 = 5.3 μm). 2 In the guinea‐pig working heart, SK&F 95654 produced a positive inotropic response without altering heart rate. 3 Oral administration of SK&F 95654 to conscious dogs caused dose‐dependent increases in left ventricular dp/dtmax in the range 10–50 μg kg−1. These positive inotropic responses were maintained for 3 h without simultaneous changes in heart rate or blood pressure. The peak effects on left ventricular dp/dtmax were similar for orally and intravenously administered compound, indicating good oral bioavailability. 4 SK&F 95654 caused a potent inhibition of U46619‐induced aggregation in both a human washed platelet suspension (WPS) (IC50 = 70 nm) and in human platelet‐rich plasma (PRP) (IC50 = 60 nm), indicating that the compound shows negligible plasma binding. 5 The R‐enantiomer of SK&F 95654 was twenty fold more potent as an inhibitor of platelet aggregation than was the S‐enantiomer. The similarity of this ratio to that obtained on the cGI‐PDE suggests that SK&F 95654 inhibits platelet aggregation via its effects on cGI‐PDE. This was also indicated by studies which showed that SK&F 95654 increased adenosine 3′:5′‐cyclic monophosphate (cyclic AMP) levels and activated cyclic AMP‐dependent protein kinase in human platelets. 6 Collagen‐induced aggregation of rat PRP was also inhibited by SK&F 95654 (IC50 = 65 nm). The effects of SK&F 95654, administered intravenously, on ex vivo platelet aggregation were studied in the conscious rat. At 1 mg kg−1, SK&F 95654 inhibited aggregation for at least 4 h post dose and was more potent than the two other cGI‐PDE inhibitors studied (siguazodan and SK&F 94120). 7 In contrast to its potent effects on heart and platelets, SK&F 95654 caused only a modest relaxation of histamine‐ or U46619‐induced bronchoconstriction in the anaesthetized, ventilated guinea‐pig. 8 Taken together, these results indicate that SK&F 95654 may be a suitable agent for the treatment of congestive heart failure.

Conformation of the cytoplasmic domain of phospholamban by NMR and CD

Nuclear magnetic resonance (NMR) and circular dichroism (CD) spectroscopy have been used to characterize the conformation of the putative cytoplasmic domain of phospholamban (PLB), an oligomeric membrane-bound protein which regulates the activity of the cardiac sarcoplasmic reticulum Ca(2+)-dependent ATPase. In aqueous solution the 25-residue peptide adopts a number of rapidly interconverting conformers with no secondary structural type obviously predominating. However, in trifluoroethanol (TFE) the conformation, while still highly dynamic, is characterized by a high proportion of helical structures. Evidence for this is provided by alpha CH chemical shifts and low NH chemical shift temperature coefficients, small NH-alpha CH intraresidue scalar coupling constants, a substantial number of distinctive interresidue nuclear Overhauser effects (NOEs) [dNN(i, i + 1), d alpha N(i, i + 3), d alpha beta(i, i + 3) and d alpha N(i, i + 4)] and characteristic CD bands at 190 (positive), 206 (negative) and 222 nm (negative). The helicity is interrupted around Pro-21. The activity of PLB is regulated by phosphorylation at either Ser-16 or Thr-17. CD shows that phosphorylation at Ser-16 by the cAMP-activated protein kinase causes about an 11% decrease in alpha-helical content in TFE.

Pulmonary effects of type V cyclic GMP specific phosphodiesterase inhibition in the anaesthetized guinea-pig

1 We have investigated the bronchodilator potential of type V phosphodiesterase (PDE V) inhibitors in anaesthetized ventilated guinea‐pigs using the potent and selective PDE V inhibitor, SK&F 96231. We have compared its activity to that of salbutamol, the PDE III inhibitors, siguazodan and SK&F 95654 and to the PDE IV inhibitor rolipram. 2 Administered as an i.v. infusion SK&F 96231 (0.6 and 1 mg kg−1min−1, i.v.) caused a slowly developing inhibition of histamine (100 nmol kg−1, i.v.)‐induced bronchoconstriction and elevated tracheal cyclic GMP levels in the anaesthetized guinea‐pig. SK&F 96231 (0.1 and 0.3 mg kg−1min−1, i.v.) was without effect on histamine‐induced bronchoconstriction. In the presence of a sub‐threshold infusion of SNP (0.1 μmol kg−1min−1, i.v.) there was a marked enhancement of SK&F 96231‐induced inhibition of histamine responses such that at infusion rates that were ineffective alone, SK&F 96231 caused a > 50% inhibition of histamine responses. The stimulation of tracheal cyclic GMP accumulation by SK&F 96231 was also potentiated. 3 Administered directly into the airway, SK&F 96231 (300 μg in 5 mg lactose carrier) was largely without effect on histamine‐induced bronchoconstriction (4.9 ± 1.9% inhibition). In the presence of SNP (0.1 μmol kg−1min−1, i.v.) or isosorbide dinitrate (200 μg administered by insufflation into the trachea) there was a marked potentiation of the inhibitory activity of SK&F 96231 (40 ± 4% and 62 ± 1.8% respectively). 4 Salbutamol and rolipram (3–300 μg by insufflation) caused a dose‐related inhibition of histamine responses with a maximum of 91 ± 2% and 59 ± 10% respectively. The PDE III inhibitor, siguazodan, was without effect on histamine responses but they were reduced (27.7 ± 4.8% at 300 μg) by SK&F 95654. There was a marked enhancement of the inhibitory activity of rolipram in the presence of SK&F 95654. 5 We conclude that SK&F 96231 has weak anti‐spasmogenic activity in the guinea‐pig in vivo, we suggest that this is primarily a consequence of a low endogenous guanylate cyclase activity in the airway. The potentiation of the anti‐spasmogenic activity of SK&F 96231 by SNP suggests that a combination of PDE V inhibitor and guanylate cyclase agonist might provide significant bronchodilator activity. 6 We have established that PDE IV inhibitors are bronchodilators when administered directly into the airway of anaesthetized guinea‐pigs but that PDE III inhibitors are only weakly active. The marked enhancement of the inhibitory activity of rolipram by the PDE III inhibitor, SK&F 95654, indicates that inhibitors of both PDE III and PDE IV might offer greater potential as bronchodilators than inhibitors of either isoenzyme alone.

Chapter 26. Protein Kinases and Phosphatases: Structural Biology and Synthetic Inhibitors

Publisher Summary The reversible covalent phosphorylation of proteins was discovered in 1955 and now represents a well established mechanism for the regulation of the biological function. A large number of proteins are known to exist in phospho- and dephospho-forms; the exact effect of phosphorylation varies from protein to protein, but changes in enzyme kinetics and ligand binding are common responses. Ultimately, the level of phosphoryation of any protein is determined by the activity of two classes of enzyme: the protein kinases that catalyse the transfer of the terminal phosphoryl moiety of adenosine triphosphate (ATP) or very occasionally guanosine triphosphate (GTP) to the acceptor amino acid and the phosphoprotein phosphatases that remove the phosphate. It is common for a protein to be acted on by a number of protein kinases, either at different sites or at the same amino acid residue. Largely, as a result of molecular biology, the number of known protein kinases and phosphatases are rapidly increasing and the information, regarding their cellular function, is also being generated quickly. There is a 230 amino acid catalytic domain common to all protein tyrosine phosphatases (PTPs) and, in contrast to protein kinases there is no homology with PS/TPs. Two features of protein kinases relevant to medicinal chemistry, the crystal structure, and synthetic inhibitor molecules and a brief overview of phosphatases are discussed in the chapter.