| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Articles: Therapeutics, Targets, and Development
Characterization of a potent and selective small-molecule inhibitor of the PIM1 kinase
1 Center for Molecular Biology and Gene Therapy, Departments of 2 Biochemistry and Microbiology and 3 Medicine, Loma Linda University School of Medicine, Loma Linda, California; and 4 Plexxikon, Inc., Berkeley, California
Requests for reprints: Michael B. Lilly, Center for Molecular Biology and Gene Therapy, Loma Linda University School of Medicine, 11234 Anderson Street, Loma Linda, CA 92354. Phone: 909-558-8777; Fax: 909-558-0177. E-mail: mlilly{at}llu.edu
Abstract
The pim-1 kinase is a true oncogene that has been implicated in the development of leukemias, lymphomas, and prostate cancer, and is the target of drug development programs. We have used experimental approaches to identify a selective, cell-permeable, small-molecule inhibitor of the pim-1 kinase to foster basic and translational studies of the enzyme. We used an ELISA-based kinase assay to screen a diversity library of potential kinase inhibitors. The flavonol quercetagetin (3,3',4',5,6,7-hydroxyflavone) was identified as a moderately potent, ATP-competitive inhibitor (IC50, 0.34 µmol/L). Resolution of the crystal structure of PIM1 in complex with quercetagetin or two other flavonoids revealed a spectrum of binding poses and hydrogen-bonding patterns in spite of strong similarity of the ligands. Quercetagetin was a highly selective inhibitor of PIM1 compared with PIM2 and seven other serine-threonine kinases. Quercetagetin was able to inhibit PIM1 activity in intact RWPE2 prostate cancer cells in a dose-dependent manner (ED50, 5.5 µmol/L). RWPE2 cells treated with quercetagetin showed pronounced growth inhibition at inhibitor concentrations that blocked PIM1 kinase activity. Furthermore, the ability of quercetagetin to inhibit the growth of other prostate epithelial cell lines varied in proportion to their levels of PIM1 protein. Quercetagetin can function as a moderately potent and selective, cell-permeable inhibitor of the pim-1 kinase, and may be useful for proof-of-concept studies to support the development of clinically useful PIM1 inhibitors. [Mol Cancer Ther 2007;6(1):163–72]
Introduction
The pim family of serine-threonine kinases is composed of three highly homologous genes, pim-1, pim-2, and pim-3. These enzymes are increasingly being recognized as important mediators of survival signals in cancers, stress responses, and neural development (1–6). In addition, these kinases are constitutively expressed in some tumors and function as true oncogenes. Thus, they are of significant interest as targets for therapeutic intervention.
Small-molecule inhibitors are important molecular probes for studying protein kinases. In addition, they may serve as protype therapeutic agents for treating diseases resulting from unregulated kinase activity. Three prior reports have shown that known, promiscuous kinase inhibitors can inhibit PIM1 function in vitro. Jacobs et al. (7) showed that several staurosporine and bisindoyl-maleimide analogues, as well as the morpholino-substituted chromone LY294002, were able to inhibit PIM1 activity in vitro. Subsequently, Fabian et al. (8) presented an interaction map involving 113 kinases and 20 small-molecule kinase inhibitors now under clinical study. Only three inhibitors had detectable binding to (and presumably inhibitory activity against) PIM1—two staurosporine analogues and flavopiridol, a flavonoid undergoing evaluation as an inhibitor of cyclin-dependent kinases. A recent report (9) confirmed the activity of bisindoylmaleimide derivatives as well as some flavonoids in vitro. All of the identified inhibitors either lacked specificity for PIM1 or were only modestly active at low micromolar concentrations, or both. Furthermore, none of these reports showed that the test agents could selectively inhibit PIM1 activity in intact cells.
To further our basic and translational studies of the pim kinases, we have sought to identify small-molecule inhibitors of PIM1. We here report that the flavonol quercetagetin is a selective PIM1 inhibitor with nanomolar potency and can differentially inhibit the kinase in cell-based assays.
Materials and Methods
Cell Lines and Culture Methods
The prostate epithelial cell lines RWPE1, RWPE2, LNCaP, and PC3 were obtained from the American Type Culture Collection (Manassas, VA) and cultured in the recommended medium. We produced additional pools of RWPE2 prostate cells that overexpressed pim-1 through retroviral transduction. The coding region for the human pim-1 gene was cloned into the pLNCX retroviral vector (Clontech, Mountain View, CA). Infectious viruses were produced in the GP-293 packaging cell line by cotransfection with retroviral backbone plasmids (pLNCX or pLNCX/pim-1) and with pVSV-G, a plasmid that expresses the envelope glycoprotein from vesicular stomatitis virus. Forty-eight hours after transfection, the medium was collected and the virus particles were concentrated as described in the manufacturer's protocol (Clontech). RWPE2 cells were plated at 1 x 105 per 60-mm plate 16 to 18 h before infection. Cells were infected with 5 x 104 viral particles in the presence of 8 µg/mL polybrene. After 6 h of incubation, the virus-containing medium was replaced with fresh medium, and on the next day G418 (400 µg/mL) was added to select infected cells. After 10 days of selection, stable cell pools were established and PIM1 expression was verified by immunoblotting.
For growth-inhibition experiments, cells were plated onto 24-well plates and fixed with formaldehyde at intervals. Cell number was quantified by crystal violet staining (10).
Recombinant pim Kinases and Kinase Assays
We prepared recombinant PIM1 and PIM2 as glutathione S-transferase (GST) fusions in Escherichia coli, as described (11). For the inhibitor screening assays, a solid-phase kinase assay was developed based on our demonstration that PIM1 is a potent kinase for phosphorylating BAD on Ser112 (11, 12). Ninety-six-well flat-bottomed plates were coated overnight at 4°C with recombinant GST-BAD [1 µg/well in HEPES buffer: 136 mmol/L NaCl, 2.6 mmol/L KCl, and 20 mmol/L HEPES (pH 7.5)]. The plates were then blocked for 1 h at room temperature with 10 µg/mL bovine serum albumin in HEPES buffer. The blocking solution was then removed and 5 µL of each inhibitor, dissolved in 50% DMSO, were added to each well. Then, 100 µL of kinase buffer [20 mmol/L MOPS (pH 7.0), 12.5 mmol/L MgCl2, 1 mmol/L MnCl2, 1 mmol/L EGTA, 150 mmol/L NaCl, 10 µmol/L ATP, 1 mmol/L DTT, and 5 mmol/L ß-glycerophosphate] containing 25 ng recombinant GST-PIM1 kinase were added to each well. The final concentration of each inhibitor was 
10 µmol/L. The plate was placed on a gel slab dryer prewarmed to 30°C, and the kinase reaction was allowed to proceed. The reaction was stopped after 60 min by removal of the reaction buffer, followed by the addition of 100 µL of HEPES buffer containing 20 mmol/L EDTA to each well. Phosphorylated GST-BAD was detected by an ELISA reaction, using as first antibody a monoclonal anti–phospho-BAD(S112) antibody (Cell Signaling, Danvers, MA), a secondary goat anti-mouse IgG-peroxidase conjugated antibody (Pierce, Rockford, IL), and Turbo-TMB peroxidase substrate (Pierce). The level of phosphorylated GST-BAD present was proportional to the absorbance at 450 nm.
For quantitative and kinetic studies of inhibitors against various BAD(S112) kinases, a solution phase assay was used. A biotinylated peptide based on the PIM1 phosphorylation site of human BAD was synthesized (GGAGAVEIRSRHSSYPAGTE) and used as the assay substrate. Recombinant GST-PIM1 (25 ng/reaction) was preincubated with various concentrations of inhibitors in the previous kinase buffer (final volume 100 µL). The reaction proceeded by addition of substrate peptide, followed by incubation for 5 min in a 30°C water bath. The reaction was terminated by transferring the mixture to a streptavidin-coated 96-well plate (Pierce) containing 100 µL/well of 40 mmol/L EDTA. The biotinylated peptide substrate was allowed to bind to the plate at room temperature for 10 min. The level of phosphorylation was then determined by ELISA as described above. Curve fitting and enzyme analyses were done using GraphPad Prism version 4.00 for Windows (GraphPad Software, San Diego, CA). For the additional BAD(S112) kinases [PIM2, RSK2 (ribosomal S6 kinase 2), and PKA (cyclic AMP–dependent protein kinase)], reaction components were as described above. As with the PIM1 assays, an ATP concentration of 10 µmol/L was used. Furthermore, with each kinase, linear reaction velocities for the duration of the reaction were confirmed (data not shown).
To further assess the specificity of quercetagetin as a PIM1 inhibitor, its activity against a panel of serine-threonine kinases was also studied through a commercial kinase inhibitor profiling service (KinaseProfiler; Upstate Biotechnology, Charlottesville, VA). All KinaseProfiler assays were conducted using 10 µmol/L ATP concentrations.
Small-Molecule Library Screening
We obtained a library of 1,200 compounds that had structural affinity to known kinase inhibitors (TimTec, Inc., Newark, DE). The entire library was screened once with our solid-phase ELISA kinase assay, with each compound at 10 µmol/L concentration. Positive hits were rescreened at the same concentration. Compounds that had reproducible activity at 10 µmol/L were then screened at a range of concentrations from 0.001 to 300 µmol/L. Additional flavonoids were purchased from Indofine Chemicals (Hillsborough, NJ) and were tested in a similar protocol.
Measurement of PIM1 Kinase Activity in Cells
RWPE2 cell pools, stably infected with empty retrovirus or pim-1–encoding retrovirus, were seeded in six-well plates at 5 x 105 cells per well. After 18 h, the normal supplemented keratinocyte medium was removed and replaced with supplement-free keratinocyte medium. Cells were then incubated for an additional 20 h. Quercetagetin, or an equivalent volume of DMSO, was added to the cells 3 h before the end of the starvation period. At the conclusion of the starvation period, the cells were washed twice with PBS and subsequently lysed in a denaturing buffer with protease, phosphatase inhibitors. The lysates were normalized by total protein content (BCA protein assay, Pierce), then analyzed by immunoblotting with the following antibodies: monoclonal anti-PIM1 (Santa Cruz Biotechnologies, Santa Cruz, CA); monoclonal anti–ß-actin (Sigma, St. Louis, MO); monoclonal anti-BAD (Transduction Laboratories, Franklin Lakes, NJ); and monoclonal anti–phospho-BAD(S112), polyclonal anti–phospho-AKT(S473), and anti-AKT (all from Cell Signaling).
Cloning, Expression, Purification, and Crystallization of PIM1
The production, purification, and characterization of recombinant 6His-tagged PIM1 proteins for crystallography have been described previously (13). To obtain cocrystals of complexes of the protein with ligands, the protein solution was initially mixed with the compound (dissolved in DMSO) at a final compound concentration of 1 mmol/L and then set up for crystallization. The protein was crystallized by a sitting-drop, vapor-diffusion experiment in which equal volumes of protein (10–15 mg/mL concentration) and reservoir solution [0.4–0.9 mol/L sodium acetate, 0.1 mol/L imidazole (pH 6.5)] were mixed and allowed to equilibrate against the reservoir at 4°C. The crystals routinely grew to a size of 200 x 200 x 800 µm in 2 to 3 days.
Structure Determination
X-ray diffraction data were collected at Advanced Light Source (Berkeley, CA). All data were processed and reduced with MOSFLM and scaled with SCALA of the CCP4 suite of programs using the software ELVES. The space group of all crystals was determined to be P65, with the cell axes being approximately 99, 99, and 80, and one protein monomer being present in the asymmetrical unit. All structures were determined by molecular replacement using the apo PIM1 structure (1YWV; ref. 13) as a model, and refined by CNX and REFMAC5. Crystallographic statistics are reported in Supplementary Table S1.5 The coordinates and structure factors for the structures have been deposited with the RCSB Protein Data Bank (accession codes 2O63, 2O64, 2O65).
Results
Screening of a Chemical Library with Structural Affinity to Known Kinase Inhibitors
As an initial approach to the identification of PIM1 inhibitors, we screened a library of small molecules whose structures were similar to those of known kinase inhibitors. Of the seven compounds that had reproducible inhibitory activity at 10 µmol/L, six were flavonoids [quercetin, luteolin, kaempferol, 7-hydroxyflavone, (S)-5,7-dihydroxy-8-(3-methylbut-2-ene)flavanone, and (R)-5,7-dihydroxyflavanone]. These compounds exhibited a range of inhibitory potencies (as IC50) from 1.1 to 60 µmol/L. Thirty-seven other flavonoids failed to show detectable inhibitory activity at 10 µmol/L. These inactive compounds were characterized in most cases by bulky (charged or uncharged) groups at the 3, 3', 4', or 7 positions; lack of at least two hydrogen bond donors on the A or C rings; presence of glycoside linkages; or failure of all rings to adopt a planar conformation.
The most active compound in the chemical library was the flavonol quercetin (IC50, 1.1 µmol/L), a known inhibitor of kinases and many other enzymes (14–19). Furthermore, six of the seven compounds with reproducible activity at 10 µmol/L were flavonoids. Hence, we screened additional flavonoids to identify molecules with inhibitory activity against the PIM1 kinase (Fig. 1 ). The most active molecule was the flavonol quercetagetin (IC50, 0.34 µmol/L). The four flavonoids with the highest inhibitory activity were characterized by the presence of five to six –OH groups distributed between the A and B rings. In comparison, the hydroxyl groups on the B ring seemed to be more critical for the activity of the compounds than those on the A ring, as compounds with an unsubstituted B ring showed greatly reduced activity. Finally, a hydrophobic substituent at the 8 position was tolerated.
|
|
; Fig. 2
). At the lower concentration, the selectivity of quercetagetin was most apparent. In the presence of 1 µmol/L inhibitor, PIM1 activity was inhibited by 92%. In contrast, the activity of the other kinases was inhibited by only 0% to 41%. In aggregate, these studies established that quercetagetin was a severalfold more potent inhibitor for pim-1 kinase than for seven other serine-threonine kinases. In addition, quercetagetin was completely inactive against the c-abl tyrosine kinase when tested at the 200 µmol/L concentration (data not shown).
|
|
The binding pose of quercetagetin in PIM1 (Fig. 3A) closely resembles that of quercetin in phosphatidylinositol 3-kinase 
(1E8W; ref. 22) and that of fisetin in CDK6 (1XO2; ref. 23), designated here as orientation I. As seen in the two earlier structures (Fig. 3D and E), the 3-OH of the quercetagetin (Fig. 3A) makes a canonical hydrogen bond with backbone carbonyl oxygen of the hinge residue Glu121. In addition, the B ring of quercetagetin binds deep inside the PIM1 ATP-binding pocket, with the 4'-hydroxyl group hydrogen-bonded to the side chains of two highly conserved residues, Lys67 and Glu89. However, significant differences were also observed between the current structure and the two reported structures. In both 1E8W and 1XO2, the 4-keto group of the chromenone core of the compound formed a hydrogen bond with the same hinge amide nitrogen [Val882 in phosphatidylinositol 3-kinase (Fig. 3D) and Val101 in CDK6 (Fig. 3E)]. However, there is no direct interaction between the 4-keto group of quercetagetin and the amide nitrogen of the corresponding residue Pro123 in PIM1 because proline is incapable of acting as a hydrogen bond donor. Instead, the 4-keto group of quercetagetin makes close contact with the backbone C
of Arg122 (3.4 Å). It is not clear whether this interaction makes a positive contribution to the binding of quercetagetin to PIM1.
The B ring of quercetagetin binds deep inside the PIM1 ATP-binding pocket. The 4'-hydroxyl group forms hydrogen bonds with both Lys67 and Glu89, two of the most conserved residues in kinases. As has been noted, satisfying the hydrogen bonding requirements at this region is one of the determining features of binding of compounds to PIM1 (13).
When compared with quercetagetin, the chromenone core of myricetin (Fig. 3B) and 5,7,3',4',5'-pentahydroxyfavone (Fig. 3C) has flipped 180° in PIM1 such that the B ring is now oriented toward the entrance of the ATP pocket. A possible explanation for adopting this orientation is that the interior of the ATP pocket cannot accommodate the B ring with three hydroxyl substitutions. Although they bind in the same orientation, there are important differences between the binding poses of the two compounds, which can be attributed to the presence or absence of the 3-hydroxyl group. The 3-hydroxyl group in myricetin still makes a hydrogen bond with the carbonyl oxygen of Glu121, despite the difference in binding orientation. Because of the adjacent 4-keto group, the 3-hydroxyl is likely to be most acidic of all the hydroxyl groups in the compound, and, as a result, it dictates the overall positioning of the compound. Another interaction that may contribute to the observed binding pose is a hydrogen bond between the 3'-hydroxyl group of myricetin and the carbonyl oxygen of Pro123 (Fig. 3B). The importance of the 3-hydroxyl group is evident. The second compound, 5,7,3',4',5'-pentahydroxyflavone, lacking such a group, makes no direct interaction with the hinge region.
Quercetagetin Inhibits PIM1 Kinase Activity in Intact Cells
To determine if quercetagetin could act as a cell-permeable PIM1 inhibitor, we examined the activity of the flavonol in RWPE2 prostate cancer cells. We studied the phosphorylation of endogenous BAD on Ser112, under conditions of growth factor starvation, as an indicator of intracellular PIM1 activity (Fig. 4
).
|
Quercetagetin Treatment Reproduces a Known pim-1 Knockdown Phenotype
If quercetagetin acts as a true PIM1 inhibitor, then it should reproduce a pim-1–dependent phenotype in the target cells. We have shown that PIM1 inhibition by genetic means (small interfering RNA) inhibits the proliferation of RWPE1 and RWPE2 cells (Supplementary Fig. S1).5 We therefore determined if quercetagetin could reproduce this phenotype. RWPE2 cells were treated with quercetagetin for up to 72 h (Fig. 5A
). Marked dose-dependent growth inhibition was apparent by 24 h, leading to persistent growth arrest thereafter. Quercetagetin reproduced this pim-1–dependent phenotype at a drug concentration that inhibited the enzyme in cells (ED50, 3.8 µmol/L; Fig. 5B). Similar results were seen in RWPE1 cells (data not shown). Apoptotic cells, showing cytoplasmic blebbing and detachment, were rare, but dividing cells virtually disappeared in cultures treated with quercetagetin at 6.25 µmol/L or higher concentrations (data not shown). DNA histograms obtained at 24 h after the addition of quercetagetin (6.25 µmol/L) or DMSO vehicle were very similar (Fig. 5C). Neither showed a <2n population suggestive of apoptosis. There was a slight increase in the proportion of cycling cells (S + G2-M) in the drug-treated samples.
|
12.5 µmol/L. Thus, the ability of the flavonol to inhibit proliferation was proportional to the amount of PIM1 protein in the target cells, particularly at lower drug concentrations. Although other interpretations are possible, these data support our observation that quercetagetin can act as a PIM1 inhibitor.
|
The development of clinically useful small-molecule kinase inhibitors has been a seminal event in the world of oncology. Flavonoids were among the early scaffold structures identified as potential kinase inhibitors. However, although many flavones, isoflavones, and flavonols have been shown to regulate the activity of kinases in cell-based assays, fewer data exist to show that these molecules can directly bind and inhibit kinase targets both in vitro and in cells. It is clear that some flavonoids are ATP-competitive ligands for both tyrosine and serine-threonine kinases, as well as other ATP-binding enzymes. The flavonol quercetin is one such ligand, and its ability to directly bind to ATP-binding enzymes has been well shown. At low-micromolar concentrations, it directly binds and inhibits such diverse enzymes as the phosphatidylinositol 3-kinase (14), the epidermal growth factor receptor tyrosine kinase (15), retroviral reverse transcriptases (16), DNA gyrases (17), phosphodiesterases (18), and thioredoxin reductase (19). Other direct flavonoid inhibitors have been described for RSK2 kinase (24), mitogen-activated protein/extracellular signal-regulated kinase 1 (25), and several cyclin-dependent kinases (23, 26–28). One such ligand, flavopiridol, has already entered clinical trials for the treatment of cancer. Others, such as PD98059, are familiar laboratory reagents for inhibition of kinase pathways. We now show, by means of crystallography, that quercetagetin is a direct ligand for the ATP-binding pocket of PIM1 kinase (Fig. 3).
Specificity is always a concern with ATP pocket ligands. There are probably no absolutely selective inhibitors for a kinase but rather ligands that show a spectrum of affinities for their various targets. We have shown that quercetagetin is severalfold more active against PIM1 than against eight other serine-threonine kinases and a tyrosine kinase, either with in vitro assays or in cell cultures. Interestingly, quercetagetin showed 10-fold more selectivity for PIM1 than for the homologous PIM2 kinase (sequence identity 56%). The ATP-binding pockets of these two kinases are identical with the exception of three residues along the edge of the PIM1 ATP-binding pocket—Ser54 (Ala58 in PIM2), Glu124 (Leu120 in PIM2), and Val126 (Ala122 in PIM2). Val126 of PIM1 makes direct van der Waal's contact with the A ring of quercetagetin (Fig. 3A). Loss of such a contact due to the Val-to-Ala substitution is likely a contributing factor to the reduced activity of the compound in PIM2. The other residues are located close to the hinge Arg122 (Arg118 in PIM2). The polar side chains of Ser54 and Glu124 can form hydrogen bonds with Arg122, thus affecting its conformation. Substitutions of these residues to hydrophobic amino acids in PIM2 will change the local environment (Fig. 3A).
The only large-scale examination of the specificity of flavonoid kinase inhibitors was reported recently by Fabian et al. (8). This investigation used a competitive binding assay to predict the inhibitor potency and specificity of the test agents. Flavopiridol was tested for binding affinity to 119 kinases. Twenty-three kinases bound flavopiridol under the test conditions, with binding constants ranging from 0.019 to 6.6 µmol/L. Interestingly, the tested cyclin-dependent kinases bound flavopiridol less well than did calcium/calmodulin–dependent protein kinase kinase I. These data suggest that cyclin-dependent kinases may not be the only kinases inhibited in cells by flavopiridol. Both PIM1 and PIM2 were among the bound kinases, with binding constants of 0.52 and 0.65 µmol/L, respectively. Although there is no absolute correlation between binding constants and enzymatic activity, flavopiridol could conceivably inhibit the activity of both PIM1 and PIM2 in test systems. Because quercetagetin has not been tested against a large number of other kinases, we cannot predict what other enzymes would be perturbed by this flavonoid. It is likely, however, that its spectrum of selectivity will be substantially different from that of flavopiridol. Quercetagetin showed clear preference for inhibiting PIM1 over PIM2, whereas flavopiridol did not. Furthermore quercetagetin inhibited the activity of the Aurora-A kinase (IC50, 4 µmol/L), a kinase that did not bind flavopiridol (8). The substantial homology between Aurora-A kinase and PIM1 kinase likely contributed to the low-level inhibitory activity of quercetagetin for the former; Aurora-A and PIM1 are 29% identical over their entire kinase domains; and the ATP binding pockets have 68% conserved amino acids.
An earlier, smaller-scale study looked at the effect of the flavonol quercetin on the in vitro kinase activity of 25 kinases, none of which were pim family kinases (29). At the tested concentration (20 µmol/L), quercetin inhibited the enzymatic activity of eight of the kinases. The propensity of this flavonol to form aggregates in aqueous solution has been advanced as an explanation for its widespread enzyme-inhibitory activity in vitro (30). We have not detected quercetagetin aggregates at concentrations of <10 µmol/L in aqueous solution, using a light-scattering assay (data not shown). Thus, we feel that this artifact does not account for the ability of this flavonol to inhibit PIM1 at nanomolar concentrations.
Because of the potential ambiguities that may accompany the use of small-molecule kinase inhibitors, a series of standards have been proposed for their use (29). To validate the results, it is desirable to show that the effects of an inhibitor disappear when a drug-resistant mutant of the protein kinase is overexpressed. Although convincing, this standard often fails due to the lack of an identified mutant with the desired properties. No such mutant has been identified for any of the pim kinases. Another potential standard is to show that the cellular effect of the drug occurs at the same concentrations that prevents the phosphorylation of an authentic physiologic substrate of the protein kinase. We have seen in these studies that half-maximal growth inhibition of prostate cancer cells occurred at a drug concentration (3.8 µmol/L) that approximated the IC50 for PIM1 enzyme inhibition in cells (5.5 µmol/L). Furthermore, the selectivity for prostate cancer growth inhibition, in proportion to endogenous PIM1 levels, was greatest at 6.25 µmol/L. Higher concentrations suppressed growth more, but the relationship to endogenous PIM1 levels was obscured. These data suggest that, at relatively low concentrations (perhaps 5–10 µmol/L), the growth-inhibitory effects of quercetagetin likely involve PIM1 antagonism. A third standard is to observe the same effect with at least two structurally unrelated inhibitors of the protein kinase. Previously described inhibitors of pim kinases are either less active or less specific flavonoids (7, 9), the same structural class as quercetagetin, or staurosporine analogues (8, 9, 21). We therefore used small interfering RNA as a genetic means to identify a pim-1–dependent phenotype. Proliferation of prostate cells was suppressed with both the genetic and chemical inhibitors of PIM1 activity. These data show that quercetagetin is an authentic small-molecule inhibitor of PIM1 kinase.
The crystal structures of PIM1 complexed with quercetagetin, myricetin, and 5,7,3',4',5'-pentahydroxyflavonone show that flavonoids bind to PIM1 in two distinct orientations. Although interesting, this is not a surprising observation, as flavones have shown a variety of binding modes in kinases (9, 22, 23, 26–28). An examination of the intermolecular interactions of each flavonoid with PIM1 does not clearly reveal why one orientation was adopted over the other. However, it is possible that the presence of three hydroxyl groups on the B ring of myricetin and 5,7,3',4',5'-pentahydroxyflavone discourages these two flavonoids from adopting the binding orientation observed for quercetagetin. The hydrophobic side chain of Leu120, which extends into the ATP pocket in the same region occupied by the B ring of quercetagetin (Fig. 3A), may be incompatible with the 5' hydroxyl group of myricetin and 5,7,3',4',5'-pentahydroxyflavone.
Both pim-1 and pim-2 can phosphorylate 4EBP-1, a regulator of protein translation (31, 32). Rapamycin was unable to block this effect. These data suggest that pim kinases may function in a parallel pathway to the phosphatidylinositol 3-kinase/AKT/mammalian target of rapamycin cascade to regulate and support protein synthesis under stress conditions. Because AKT-1 and PIM2 function cooperatively to induce lymphoma formation in transgenic mice (6), it may be necessary to target both pathways for effective antitumor effects. Several prototype AKT inhibitors have been described (33, 34). Our identification of quercetagetin as a PIM1 inhibitor provides a tool for tissue culture studies to investigate this hypothesis. Under the tested conditions, we found no evidence that quercetagetin inhibited the phosphorylation of AKT on Ser473. Thus, it may be possible to combine inhibitors of these kinases to detect additive or synergistic effects resulting from the blockade of the two kinase pathways.
Footnotes
Grant support: Department of Defense, Congressionally Directed Medical Research Program, Award W81XWH-04-1-0887.
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.
5 Supplementary data for this article are available at Molecular Cancer Therapeutics Online (http://mct.aacrjournals.org/). 
Received 7/10/06; revised 10/10/06; accepted 11/17/06.
References
This article has been cited by other articles:
|
|
 |
|
  J. R. Gledhill, M. G. Montgomery, A. G. W. Leslie, and J. E. Walker Mechanism of inhibition of bovine F1-ATPase by resveratrol and related polyphenols PNAS, August 21, 2007; 104(34): 13632 - 13637. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
