Discovery of Highly Potent Microsomal Prostaglandin E2 Synthase 1

Mar 25, 2013 - *Telephone: 010-62757486. ... In the present study, the model was used for structure-based virtual screen of novel mPGES-1 inhibitors. ...
0 downloads 0 Views 698KB Size
Article pubs.acs.org/jmc

Discovery of Highly Potent Microsomal Prostaglandin E2 Synthase 1 Inhibitors Using the Active Conformation Structural Model and Virtual Screen Shan He, Cong Li, Ying Liu, and Luhua Lai* BNLMS, State Key Laboratory for Structural Chemistry of Unstable and Stable Species, College of Chemistry and Molecular Engineering, Peking University, Beijing 100871, China S Supporting Information *

ABSTRACT: Microsomal prostaglandin E2 synthase 1 (mPGES-1) has been identified as a promising drug target due to its key role in prostaglandin biosynthesis. However, the lack of a well-characterized structure constitutes a great challenge for the development of inhibitors. Recently, we have built a model for the active conformation of mPGES-1. In the present study, the model was used for structure-based virtual screen of novel mPGES-1 inhibitors. Of the 142 compounds tested in the cell-free assay, 10 molecules are highly potent with IC50 values of single digit nanomolar and the strongest inhibition of 1.1 nM. Moreover, nine compounds showed strong activity in the human whole blood (HWB) assay with IC50 values of less than 10 μM. The lead compounds 1 and 2 showed HWB IC50 values of 0.3 and 0.7 μM which are among the most potent mPGES-1 inhibitors reported. These compounds represent new scaffolds for future development of drugs against mPGES-1.



analogues with even higher activity (IC50 of 0.2 and 0.14 μM).8 However, this series of compounds lacks the ability to inhibit mouse or rat mPGES-1, rendering the application of preclinical animal models impossible. To date, no mPGES-1 inhibitor has been reported to be in clinical trials or available on the market. Therefore, the design and discovery of novel mPGES-1 inhibitors with new scaffolds are highly desirable. For the rational design of mPGES-1 inhibitors, the lack of a well-characterized binding site structure constitutes a key challenge. As a membrane protein, the crystal structure of mPGES-1 was determined only recently by electron crystallography to a resolution of 3.5 Å (PDB code 3DWW).9 The mPGES-1 molecule in this structure takes a closed inactive conformation that is not accessible by the substrate PGH2. This makes the straightforward application of the established structure-based virtual screen techniques rather difficult. mPGES-1 belongs to the membrane-associated proteins in eicosanoid and glutathione metabolism (MAPEG) protein family, which also includes microsomal glutathione transferase 1 (MGST-1), MGST-2, MGST-3, leukotriene C4 synthase (LTC4S), and 5-lipoxygenase-activation protein (FLAP).10 In previous studies, Bifulco et al. performed an in silico screen of a small set of molecules based on the MGST-1 structure.11 Hamza et al. built a comparative model of mPGES-1 based on the MGST-1 structure and performed a structure-based virtual

INTRODUCTION Prostaglandin E2 (PGE2), the pivotal prostaglandin produced by most mammalian tissues, regulates multiple biological processes under both normal and pathological conditions.1 Previous evidence has shown that PGE2 plays a key role in the development of various disorders, including inflammatory pain, fever, anorexia, atherosclerosis, stroke, tumorigenesis, and Alzheimer’s disease.2−4 Microsomal prostaglandin E2 synthase 1 (mPGES-1) is the terminal enzyme of the PGE2 production pathway and is not coupled to any downstream enzymes in the enzymatic cascade. Constitutive levels of mPGES-1 are normally low, and it is highly up-regulated by proinflammatory stimuli. Therefore, mPGES-1 constitutes a promising drug target with a low risk of side effects.4 Since the discovery of mPGES-1 in 1999, different types of mPGES-1 inhibitors have been reported, but very few maintain the activity in vivo.5 MK886 [1-[(4-chlorophenyl)methyl]-3[(1,1-dimethylethyl)thio]-α,α-dimethyl-5-(1-methylethyl)-1Hindole-2-propanoic acid] and derivatives potently inhibit mPGES-1 in vitro (IC50 = 3−7 nM). However, because of strong protein binding and poor cell permeability, the potency of the compounds in the cell assay shows large right shifts and no inhibitory effects could be detected in the human whole blood (HWB) assay.6 Recently, 2-(6-chloro-1H-phenanthro[9,10-d]imidazol-2-yl)isophthalonitrile (MF63) has been identified as a selective and orally active mPGES-1 inhibitor.7 It exhibited the highest reported potency in HWB assay (IC50 = 1.3 μM), and further optimization of its structure produced two © 2013 American Chemical Society

Received: December 26, 2012 Published: March 25, 2013 3296

dx.doi.org/10.1021/jm301900x | J. Med. Chem. 2013, 56, 3296−3309

Journal of Medicinal Chemistry

Article

Figure 1. (A) Active conformation structural model of mPGES-1: (three monomers) green, cyan, and gray; (GSH) yellow; (PGH2) blue. (B) Binding mode of 24 (MF63, magenta) to mPGES-1.

screen using their model.12 In both studies, new mPGES-1 inhibitors have been successfully identified. However, because both screens were mainly based on the structure of MGST-1, their lead compounds only displayed moderate activity (IC50 of 3.2 and 3.5 μM in cell-free assay, respectively), which needs to be further optimized. Recent advances in membrane protein studies and structural biology produced an explosion of crystal structures for MAPEG family proteins including FLAP13 and LTC4S.14 On the basis of these newly disclosed structures, we have built a structural model for the active conformation of mPGES-1 using molecular docking and MD simulation.15,16 By use of this model, a structure-based virtual screen was performed in the present study to identify novel mPGES-1 inhibitors. The inhibitory activities of the selected compounds were tested in cell-free and HWB assays. A series of highly potent novel inhibitors for mPGES-1 were identified.

Figure 2. Virtual screen scheme to identifiy mPGES-1 inhibitors.



(2) Binding conformations of these compounds were evaluated using the pharmacophore-based scoring program PSCORE,19 and 5894 compounds were selected with the highest matching score. (3) The selected compounds were docked into the mPGES1 model using Autodock 4.0,20 and 1046 compounds were selected with the lowest estimated Kd. Then binding conformations of these 1046 compounds were exported and evaluated manually according to the following criteria: (1) forming a hydrogen bond to the catalytic residue Arg126, (2) forming good hydrophobic interaction with both the substrate and cofactor binding sites, (3) containing ring structure at the center region which forms π−π stacking or hydrophobic interaction with Tyr130. After this step, 142 compounds were selected and purchased for experimental testing. Inhibition Testing. The inhibitory activities of the 142 selected compounds were tested in cell-free and HWB assays. In the first screen, all compounds were tested at 10 μM. The mPGES-1 inhibitor 24 (IC50 = 1.3 nM) was used as a positive control, and DMSO (2%, v/v) was used as vehicle control. 21 of the 142 compounds showed significant inhibition of mPGES-1 activity (over 50% inhibition; see Table S1 for ranks of these compounds in each step of the virtual screen).

RESULTS Virtual Screen. The structure of mPGES-1 from our previous modeling work (Figure 1A)15,16 was used to identify potential inhibitors of mPGES-1. The enzyme forms a homotrimer, and only one monomer is active with the open conformation at a time. This active conformation of mPGES-1 has been used to explain the inhibition mechanism of known mPGES-1 inhibitors. mPGES-1 catalyzes the isomerization of PGH2 to PGE2 and requires glutathione (GSH) as an essential cofactor for its activity. Most known inhibitors could bind to the substrate and the cofactor binding sites simultaneously, which was supported by the experimental results (Figure 1B).16 Therefore, in the present virtual screen, the substrate and cofactor binding pockets were used for the docking studies as a whole. Figure 2 shows the three-step virtual screen scheme combining both molecular docking and pharmacophore mapping. This scheme has been successfully used to identify inhibitors of 5-lipoxygenase (5-LOX):17 (1) Compounds in the SPECS library were docked into the substrate and cofactor binding pocket of mPGES-1 using the program DOCK 6.1,18 and 20 000 compounds were selected with the lowest grid score. 3297

dx.doi.org/10.1021/jm301900x | J. Med. Chem. 2013, 56, 3296−3309

Journal of Medicinal Chemistry

Article

Table 1. Inhibition of mPGES-1 by the Active Compounds in Cell-Free and Human Whole Blood (HWB) Assaya

3298

dx.doi.org/10.1021/jm301900x | J. Med. Chem. 2013, 56, 3296−3309

Journal of Medicinal Chemistry

Article

Table 1. continued

3299

dx.doi.org/10.1021/jm301900x | J. Med. Chem. 2013, 56, 3296−3309

Journal of Medicinal Chemistry

Article

Table 1. continued

a

Data shown represent the mean ± SEM (n = 3). 3300

dx.doi.org/10.1021/jm301900x | J. Med. Chem. 2013, 56, 3296−3309

Journal of Medicinal Chemistry

Article

Figure 3. Interaction between mPGES-1 and compounds 1 (A), 4 (B), 7 (C). Competition experiments: (D) effects of cofactor GSH and control compounds GSSG on mPGES-1 activity; (E) compound−cofactor competition experiment; (F) compound−substrate competition experiment. Data shown represent the mean ± SEM (n = 3).

The IC50 values of these 21 compounds were then determined (Table 1; see Figure S1 for dose−response behavior of screened inhibitors in cell-free assay). Ten compounds had single digit nanomolar IC50 values which were comparable to that of the reference compound 24. The inhibitory effects of these 21 compounds were further tested in an HWB assay (Table 1; see Figure S2 for dose− response behavior of screened inhibitors in the HWB assay). Nine compounds showed strong inhibition activity with IC50 values less than 10 μM. Compounds 1 and 2 were more potent than 24 with IC50 values of 0.3 and 0.7 μM, respectively. Cross-Reactivity. To investigate the selectivity of mPGES-1 inhibitors, the cross-reactivity against other enzymes within the arachidonic acid21 metabolic network was tested for compounds 1−10 (see Table S2 for inhibition activity of compounds 1−10 toward other enzymes within the AA metabolic network). In the AA metabolic network, two separate pathways produce inflammatory mediators. One is initiated by cyclooxygenase (COX) 1 and COX-2 to produce PGH2 which was then catalyzed by mPGES-1 to form PGE2. The other is initiated by 5-LOX and produces leukotrienes (LTs). None of the compounds 1−10 showed inhibitory activity in COX-1/ COX-2 cell-free assay (less than 20% at 100 μM) or 5-LOX/ LTB4 HWB assay (less than 40% at 10 μM). Binding Modes of Test Compounds. The active compounds were predicted to bind to the substrate and cofactor binding pocket according to the Autodock studies (Figure 3). To verify the binding modes of the active compounds, compound−cofactor and compound−substrate competition experiments were performed for compounds 1, 4, and 7: Compound−Cofactor Competitive Experiment. If an inhibitor and the cofactor competitively bind to the same site, at constant inhibitor concentration, increasing cofactor concentration should impair the binding ability of the inhibitor,

therefore decreasing its inhibitory effect. The experimental results supported this hypothesis, as the inhibitory effect of all three compounds decreased from 80% inhibition to below 15% with a 10 times increase in cofactor concentration. To further confirm that the inhibitors’ impaired effect was specifically caused by the competitive binding of the cofactor GSH, we performed a control experiment using glutathione disulfide (GSSG) which is derived from two GSH molecules. The GSSG molecule has properties similar to those of the GSH molecule but cannot bind to the mPGES-1 pocket because of its bigger size. Our enzyme activity assay also revealed that increasing GSSG concentration had no significant effect on the enzyme activity (Figure 3D). We repeated the competition experiment under the same conditions except that the additional GSH was replaced by GSSG. The experimental results revealed that with increased GSSG concentration, no decrease in the inhibitory effect was observed. Therefore, the decrease in the compound− GSH competition experiment was caused by the competitive binding of GSH. This result provided additional support for the compound−cofactor competitive binding mode. Compound−Substrate Competitive Experiment. Similar to the compound−cofactor competition experiment, if an inhibitor and substrate competitively bind to the same site, at constant inhibitor concentration, increasing the substrate concentration should reduce the effect of the inhibitor. The experimental observations confirmed this hypothesis, as the inhibitory effect of all three compounds decreased from above 60% to nearly 0% with increased substrate concentration. This result also supported the compound−substrate competitive binding mode. SAR of 1 and 2 Derivatives. For the lead compounds 1 and 2, SAR studies were undertaken with various truncated analogues to verify the pharmacophore required for mPGES1 binding. We searched SciFinder for analogues (chemical structure search by substructure and manually select the 3301

dx.doi.org/10.1021/jm301900x | J. Med. Chem. 2013, 56, 3296−3309

Journal of Medicinal Chemistry

Article

Table 2. SAR of Compound 1 Derivativesa

3302

dx.doi.org/10.1021/jm301900x | J. Med. Chem. 2013, 56, 3296−3309

Journal of Medicinal Chemistry

Article

Table 2. continued

a

Data shown represent the mean ± SEM (n = 3).

compounds with different substituents), and the inhibitory activities of these compounds were tested in a cell-free assay (Tables 2 and 3; see Figures S3 and S4 for dose−response behavior of these compounds in cell-free assay). Compound 1 Derivatives. Compound 1 can be divided into three parts (Figure 4). According to the docking studies, they form the following interactions with the protein: (1) The R1 region was located in the mPGES-1 cofactor binding site and formed mainly hydrophobic interactions with the protein. That explained the decreasing trend in potency from 1_1 to 1_4, from 1_5 to 1_6, and from 1_7 to 1_12. Furthermore, rigid and slender substituent groups were needed to fit the pocket shape. Replacement of the fluorene group by carbon chain (in 1_3) or adamantine (in 1_11) resulted in a significant decrease in inhibitory activity. (2) The mPGES-1 pocket had a dumbbell shape (Figure 1), and the R2 region was located at the narrow passage that connects the cofactor and substrate binding sites.

Therefore, adding 2 methyl group to this narrow region from 1_8 to 1_13 resulted in a 4-fold loss in potency. Furthermore, the length of this linker region is essential for the potency, as lengthening (in 1_14) or shortening (in 1_15) the linker resulted in a remarkable decrease in activity. 3) The R3 region was located in the mPGES-1 substrate binding site which was also dominated by hydrophobic interactions. That explained the potency reducing from 1_1 to 1_17 and from 1_9 to 1_19. Furthermore, replacement of the triazine ring (in 1_9) by an imidazole ring (in 1_18) resulted in a 4-fold loss in potency, implying the importance of the π−π stacking interaction formed between the ring substituents and the protein (Tyr130). Compound 2 Derivatives. Compound 2 can be divided into three parts (Figure 5). According to the docking studies, they form the following interactions with the protein: 3303

dx.doi.org/10.1021/jm301900x | J. Med. Chem. 2013, 56, 3296−3309

Journal of Medicinal Chemistry

Article

Table 3. SAR of Compound 2 Derivativesa

3304

dx.doi.org/10.1021/jm301900x | J. Med. Chem. 2013, 56, 3296−3309

Journal of Medicinal Chemistry

Article

Table 3. continued

a

Data shown represent the mean ± SEM (n = 3).

Figure 4. Predicted binding mode of compound 1 to mPGES-1. (A) Compound 1 can be divided into three parts: R1 (green), R2 (red), and R3 (blue). (B) Pharmacophore features contributed to the ligand binding: aromatic ring (yellow), hydrogen bond acceptor (red), and hydrophobic centers (green).

while shortening the linker (2_15 to 2_17) resulted in a significant decrease in potency until completely lost. Among all the derivatives, no compound exhibited stronger inhibitory activity than the lead compound 1 or 2. However, with the truncated analogues, pharmacophore and interactions which are important for mPGES-1 inhibitors were verified and the SAR results can be used for future optimization of mPGES1 inhibitors. For the 21 active compounds, we also searched SciFinder for substances with similar structures in related studies (chemical structure search with similarity larger than 85% and being included in references). The search results will be discussed in the following section.

(1) The R1 region was located in the mPGES-1 cofactor binding site. The nitro group formed a hydrogen bond to the protein (Arg38 and Arg70). Therefore, removal of the nitro group from 2 to 2_3 resulted in a significant loss of potency. Replacement of the nitro group by other groups including chlorine (in 2_1), methoxy group (in 2_2), and ethyl acetate (in 2_4) exhibited similar loss of inhibitory activity. 2) The R2 region was located at the narrow passage of the protein. Shortening the linker from 2_5 to 2_6 and from 2_7 to 2_8 resulted in a 2-fold loss in potency. Adding sterically large substituents (from 2_3 to 2_9 and 2_10) to the central triazole ring also led to a great potency decrease. 3) The R3 region was located in the mPGES-1 substrate binding site. The compounds with longer linker (2_11 to 2_14) maintained relatively strong inhibitory activity,



DISCUSSION In order to identify mPGES-1 inhibitors with novel scaffolds, structure-based virtual screen was performed based on our recently developed structural model of mPGES-1. This enzyme 3305

dx.doi.org/10.1021/jm301900x | J. Med. Chem. 2013, 56, 3296−3309

Journal of Medicinal Chemistry

Article

Figure 5. Predicted binding mode of compound 2 to mPGES-1. (A) Compound 2 can be divided into three parts: R1 (green), R2 (red), and R3 (blue). (B) Pharmacophore features contributed to the ligand binding: aromatic ring (yellow), hydrogen bond acceptors (red), and hydrophobic centers (green).

The active compounds also possessed high selectivity with no cross-reactivity against other enzymes within the AA metabolic network including COX-1 and COX-2. In the AA metabolic network, the biosynthesis of the inflammatory mediator PGE2 requires two steps: AA is converted to PGH2 by COX-1/COX2, and then PGH2 is converted to PGE2 by mPGES-1. The currently available nonsteroidal anti-inflammatory drugs (NSAIDs) inhibit COX and prevent the production of all prostaglandins downstream of PGH2, which may cause certain types of side effects (e.g., stomach damage).22 Unlike COX inhibition, selective mPGES-1 inhibitors only block the production of PGE2 with low risk of side effects. The active compounds 1−10 we identified showed no inhibitory activity against COX-1 or COX-2, therefore representing promising candidates for the development of future anti-inflammatory drugs. For the 21 active compounds, we searched SciFinder for substances with similar structures in related studies. Most of the compounds have unique structures not published before except for compounds 11 and 6 (Figure 7). Compounds 22 and 11 share 86% similarity. Schuster et al. performed ligand-based virtual screen using the same SPECS database23 before the ligand binding structures of mPGES-1 were available. Their ligand-based pharmacophore model shared a number of similarities with our receptor-based pharmacophore model. Both of the models contained one aromatic ring, one hydrogen bond acceptor (or one negatively ionizable group), and two hydrophobic centers. They selected compound 22 for biological testing and found that 22 showed 10% inhibition at 10 μM in a cell-free assay. Compared to that, compound 11 has cell-free IC50 value of 0.17 μM. The key difference between these two compounds is the substituent position of the carboxyl group (−OCH2COOH). This group is important for inhibitor binding by forming two hydrogen bonds to the catalytic residue Arg126 according to the docking studies. Therefore, changing its substituent position may strongly reduce its inhibitory activity. Compounds 23 and 6 share 86% similarity. Compound 23 was reported in a patent literature as inhibitor of MAPEG family for treatment of inflammation and exhibited 50% inhibition of mPGES-1 at 10 μM.21 Compared to that, compound 6 has a cell-free IC50 value of 4.6 nM. The nitro group that forms two hydrogen bonds to Arg38 and Arg70

undergoes conformation transition between the open and the closed conformations, and its crystal structure was determined in a closed inactive conformation that cannot be used directly for virtual screen studies. Previous screen studies were mainly based on the structure of MGST-1, and only moderate inhibitors were identified.11,12 Using molecular docking and MD simulation, we have built a structural model for the active conformation of mPGES-1. In the model, the substrate PGH2 and cofactor GSH bind in a position suitable for enzymatic reaction.15 The substrate and cofactor binding site also constitute a promising site for the inhibitor binding. Virtual screen using this model successfully identified mPGES-1 inhibitors with a success rate of about 15% (21 out of 142 selected compounds with cell-free IC50 less than 10 μM). For the 21 active compounds, the correlation coefficient between experimental pIC50 and Autodock predicted pKd reached 0.63 (Figure 6). Furthermore, the predicted binding modes of the

Figure 6. Correlation analysis between experimental pIC50 and calculated pKd.

active compounds were supported by the compound−cofactor and compound−substrate competition experiment. These results validated our model and its effectiveness in virtual screen. Ten compounds showed strong inhibition activity in the cell-free assay with IC50 values less than 10 nM, and two compounds exhibited HWB IC50 values of 0.3 and 0.7 μM which are among the most potent mPGES-1 inhibitors reported. 3306

dx.doi.org/10.1021/jm301900x | J. Med. Chem. 2013, 56, 3296−3309

Journal of Medicinal Chemistry

Article

Figure 7. (A) Chemical structures of compound 11 and similar compound 22 (left) and binding mode of compound 11 (right). (B) Chemical structures of compound 6 and similar compound 23 (left) and binding mode of compound 6 (right). these compounds). Other reagents were from Sigma Aldrich unless indicated otherwise. Molecular Cloning and Protein Expression of mPGES-1. The mPGES-1 coding region was PCR-amplified from the plasmid DNA and ligated into pET28a(+) vector and transformed to the Rosetta(DE3) strain of Escherichia coli. Recombinant cells were cultivated at 37 °C until the OD600 reached 1.0.15 Then mPGES-1 expression was induced and the cells were grown for another 12 h at 25 °C. The cells were harvested by centrifugation at 5000g for 20 min at 4 °C and broken by sonication. Insoluble material was separated by centrifugation at 12000g for 30 min at 4 °C. The supernatant was then ultracentrifuged at 174000g for 1 h. The membrane pellet was washed and resuspended in solubilizing buffer. Protein concentrations were determined using a bicinchoninic acid protein assay kit (Biomed). Inhibition Assay of mPGES-1 in Cell-Free Systems. The enzyme activity was measured by assessment of PGH2 conversion to PGE2. Briefly, PGH2 was added to each well of a 96-well plate and the reaction was started by adding a microsome sample. After reaction at 4 °C for 1 min, the reaction was terminated by adding stop solution. PGE2 in the reaction mixture was quantified using the PGE2 EIA kit (Cayman). For the IC50 determinations, 2.5 mM cofactor and 17 μM PGH2 were used. Enzyme samples were preincubated with the test compounds for 15 min at 4 °C. For the first screen, compounds that may perturb the assay (having absorption at 412 nm) were deleted. Measuring PGE2 Formation in Human Whole Blood. Fresh human blood was obtained from healthy volunteers who did not received NSAIDs for at least 14 days. Coagulation was prevented using blood collection tubes with heparin (VACUETTE), and 0.1 mL of blood was immediately aliquoted into each well of a 96-well plate. Blood samples were preincubated with either vehicle (DMSO) or test compounds for 15 min at 37 °C. To assay PGE2 formation, LPS (Sigma) was added to the blood with a final concentration of 100 μg/ mL and the blood samples were incubated for 24 h at 37 °C to induce PGE2 formation. The incubation was terminated by centrifugation at 3000g for 5 min at 4 °C to obtain plasma. PGE2 in the plasma was quantified using the PGE2 EIA kit (Cayman). Inhibition Assay of COX-1 and COX-2 in Cell-Free Systems. The enzyme activities of COXs were determined spectrophotometri-

contributes to the inhibitor binding. Therefore, changing its substituent position as in compound 23 weakens its inhibitory activity. For both compounds 22 and 23, although their activities against mPGES-1 have been identified, their inhibition mechanisms remain to be fully disclosed. Our active conformation structural model can give a good explanation for the binding modes of mPGES-1 inhibitors and provide guidance for their optimization. On the basis of the model, compounds 11 and 6 with similar structure but significantly improved activity were identified through structure-based virtual screen. This result supported that the active conformation structural model can be used for future rational design and optimization of mPGES-1 inhibitors. In conclusion, based on the structural model for the active conformation of mPGES-1, novel inhibitors with new scaffolds have been identified through structure-based virtual screen. The active compounds showed strong inhibition activity in both the cell-free and HWB assays and high selectivity with no crossactivity against COX-1 or COX-2. Furthermore, compounds 1 and 2 showed HWB IC50 values of 0.3 and 0.7 μM which are among the most potent published mPGES-1 inhibitors in literature. The compounds reported here represent promising inhibitors for the development of next generation antiinflammatory drugs.



EXPERIMENTAL SECTION

Materials. The plasmid DNA harboring the full-length cDNA of the PIG12 gene was from Addgene (Addgene plasmid 16506: pBKCMV Pig12). PGH2 and PGE2 EIA kits were from Cayman Chemical. 1-Hexadecanoyl-2-(9Z-octadecenoyl)-sn-glycero-3-phosphocholine (POPC) was from Avanti Polar Lipids. The selected compounds were from the SPECS, ChemDiv, ChemBridge, Enamine, and Aldrich with purity of more than 90% and for most compounds greater than 95% (confirmed by the supplier, using NMR, LS-MS, or both; data are available through the Web site; see Table S3 for the order numbers of 3307

dx.doi.org/10.1021/jm301900x | J. Med. Chem. 2013, 56, 3296−3309

Journal of Medicinal Chemistry

Article

50 000 000; maximum number of generations, 270 000; grid box, 48 × 56 × 56 points. The final docking result was selected as the structure with the lowest energy in the largest cluster.

cally by oxidation of TMPD during the conversion of PGG2 to PGH2. Inhibition activities were measured as described.24 Measuring LTB4 Formation in Human Whole Blood. To assay LTB4 formation, calcium ionophore A23187 (Sigma) was added to the blood with a final concentration of 20 μg/mL, and the blood samples were incubated for 0.5 h at 37 °C to induce LTB4 formation. LTB4 formation in human whole blood was measured as described.17 Competition Experiments. In order to investigate the binding modes of the active compounds, compound−cofactor competition experiments and compound−substrate competition experiments were performed as follows:16 Compound−Cofactor Competition Experiment. Prior to the activity assay, the mPGES-1 enzyme sample was copreincubated with test compound and cofactor for 15 min at 4 °C. The compound concentration was kept constant while gradually increasing the cofactor concentration from 2.5 to 30 mM. The concentrations of test compounds were selected as the concentrations that inhibited the mPGES-1 activity 80% when the cofactor concentration was 2.5 mM. The substrate PGH2 concentration was kept constant as 17 μM. A control experiment was performed under the same conditions except that the GSH was replaced by GSSG with the same final mass-volume percentage. To ensure the enzyme activity of mPGES-1 (mPGES-1 requires GSH as an essential cofactor for its activity), 2.5 mM GSH was always present in the assay buffer. For calculation of the percentage of inhibition in Figure 3E, the activity curve shown in Figure 3D (the upper curve for GSH) was set as 0% inhibition. Compound−Substrate Competition Experiment. As the enzyme sample could not be preincubated with the substrate, it was preincubated with the test compound for 5 min at 4 °C. Then PGH2 was added to start the reaction. The inhibitor concentration was kept constant while gradually increasing the substrate concentration from 2.8 to 200 μM. As we need to determine the effect of compounds on the mPGES-1 activity at a lower substrate concentration (2.8 μM), the concentrations of compounds were selected as the concentrations that inhibited mPGES-1 activity by 40% when substrate concentration was 17 μM. The cofactor GSH concentration was kept constant as 2.5 mM. Molecular Docking. The SPECS compound database (November 2009 version for 10 mg, 201 007 compounds) was used for the virtual screen. The 3D structures of the compounds were built with the LigPrep25 program of the Schrö dinger software (using default settings). A 3D structure library containing 197 211 compounds was established (generation of 3D structure failed for 3796 compounds). For the DOCK step in virtual screen, rigid body docking was performed with default parameters using the program DOCK 6.1.18 For the PSCORE step in virtual screen, a receptor-based pharmacophore model was generated using Pocket v.2.26 Pocket v.2 is able to derive a pharmacophore model directly from a given protein−ligand complex structure without human intervention. Our model of mPGES-1 in PDB format and the most potent hit from MK886 series (named 30 in the original literature6) in Mol2 format were used as input files for Pocket v.2 to derive the pharmacophore model. The key features in the model were automatically reduced to a reasonable number. Then the binding conformations of the compounds were evaluated by PSCORE,19 an in-house program developed in our laboratory to check whether the binding conformations of compounds accommodate the pharmacophore model. To validate the screening efficiency of the PSCORE step, an mPGES-1 inhibitor test set that was developed by us before15 was screened against the pharmacophore model. The test set contained 96 mPGES-1 inhibitors and 21 compounds that have been biologically tested to be inactive. The result showed that the pharmacophore model has a good retrieval of active compounds (actives hit rate, 81.8%) and sufficient selectivity against inactive ones (inactives hit rate, 27.3%). For the Autodock step in virtual screen, flexible ligands and rigid receptor docking was performed using the program Autodock 420 with following parameters: empirical free-energy function and the GALS algorithm (genetic algorithm with local search); number of runs, 50; number of individuals, 300; maximum number of energy evaluations,



ASSOCIATED CONTENT

S Supporting Information *

Ranks of identified hits (active in cell-free assay) in virtual screen, inhibition activity of compounds 1−10 toward other enzymes within the AA metabolic network, order numbers for 1 and 2 derivatives, and dose−response behavior of compounds. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*Telephone: 010-62757486. Fax: (+86)10-62751725. E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This research was supported in part by the Ministry of Science and Technology of China (Grants 2009CB918500 and 2012AA020308) and the National Natural Science Foundation of China (Grants 90913021, 11021463, and 20873003).



ABBREVIATIONS USED mPGES-1, microsomal prostaglandin E2 synthase 1; HWB, human whole blood; AA, arachidonic acid; PGE2, prostaglandin E2; MAPEG, membrane-associated proteins in eicosanoid and glutathione metabolism; MGST-1, microsomal glutathione transferase 1; LTC4S, leukotriene C4 synthase; FLAP, 5lipoxygenase-activation protein; POPC, 1-hexadecanoyl-2-(9Zoctadecenoyl)-sn-glycero-3-phosphocholine; GSH, glutathione; COX, cyclooxygenase; LT, leukotriene; NSAID, nonsteroidal anti-inflammatory drug



REFERENCES

(1) Chang, H. H.; Meuillet, E. J. Identification and development of mPGES-1 inhibitors: Where we are at? Future Med. Chem. 2011, 3, 1909−1934. (2) Hara, S.; Kamei, D.; Sasaki, Y.; Tanemoto, A.; Nakatani, Y.; Murakami, M. Prostaglandin E synthases: understanding their pathophysiological roles through mouse genetic models. Biochimie 2010, 92, 651−659. (3) Samuelsson, B.; Morgenstern, R.; Jakobsson, P. J. Membrane prostaglandin E synthase-1: a novel therapeutic target. Pharmacol. Rev. 2007, 59, 207−224. (4) Satoh, K.; Nagano, Y.; Shimomura, C.; Suzuki, N.; Saeki, Y.; Yokota, H. Expression of prostaglandin E synthase mRNA is induced in beta-amyloid treated rat astrocytes. Neurosci. Lett. 2000, 283, 221− 223. (5) Friesen, R. W.; Mancini, J. A. Microsomal prostaglandin E2 synthase-1 (mPGES-1): a novel anti-inflammatory therapeutic target. J. Med. Chem. 2008, 51, 4059−4067. (6) Riendeau, D.; Aspiotis, R.; Ethier, D.; Gareau, Y.; Grimm, E. L.; Guay, J.; Guiral, S.; Juteau, H.; Mancini, J. A.; Methot, N.; Rubin, J.; Friesen, R. W. Inhibitors of the inducible microsomal prostaglandin E2 synthase (mPGES-1) derived from MK-886. Bioorg. Med. Chem. Lett. 2005, 15, 3352−3355. (7) Xu, D.; Rowland, S. E.; Clark, P.; Giroux, A.; Cote, B.; Guiral, S.; Salem, M.; Ducharme, Y.; Friesen, R. W.; Methot, N.; Mancini, J.; Audoly, L.; Riendeau, D. MF63 [2-(6-chloro-1H-phenanthro[9,10d]imidazol-2-yl)-isophthalonitrile], a selective microsomal prostaglan3308

dx.doi.org/10.1021/jm301900x | J. Med. Chem. 2013, 56, 3296−3309

Journal of Medicinal Chemistry

Article

din E synthase-1 inhibitor, relieves pyresis and pain in preclinical models of inflammation. J. Pharmacol. Exp. Ther. 2008, 326, 754−763. (8) Giroux, A.; Boulet, L.; Brideau, C.; Chau, A.; Claveau, D.; Cote, B.; Ethier, D.; Frenette, R.; Gagnon, M.; Guay, J.; Guiral, S.; Mancini, J.; Martins, E.; Masse, F.; Methot, N.; Riendeau, D.; Rubin, J.; Xu, D.; Yu, H.; Ducharme, Y.; Friesen, R. W. Discovery of disubstituted phenanthrene imidazoles as potent, selective and orally active mPGES1 inhibitors. Bioorg. Med. Chem. Lett. 2009, 19, 5837−5841. (9) Jegerschold, C.; Pawelzik, S. C.; Purhonen, P.; Bhakat, P.; Gheorghe, K. R.; Gyobu, N.; Mitsuoka, K.; Morgenstern, R.; Jakobsson, P. J.; Hebert, H. Structural basis for induced formation of the inflammatory mediator prostaglandin E2. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 11110−11115. (10) Jakobsson, P. J.; Morgenstern, R.; Mancini, J.; Ford-Hutchinson, A.; Persson, B. Membrane-associated proteins in eicosanoid and glutathione metabolism (MAPEG). A widespread protein superfamily. Am. J. Respir. Crit. Care Med. 2000, 161, S20−S24. (11) De Simone, R.; Chini, M. G.; Bruno, I.; Riccio, R.; Mueller, D.; Werz, O.; Bifulco, G. Structure-based discovery of inhibitors of microsomal prostaglandin E2 synthase-1, 5-lipoxygenase and 5lipoxygenase-activating protein: promising hits for the development of new anti-inflammatory agents. J. Med. Chem. 2011, 54, 1565−1575. (12) Hamza, A.; Zhao, X.; Tong, M.; Tai, H. H.; Zhan, C. G. Novel human mPGES-1 inhibitors identified through structure-based virtual screening. Bioorg. Med. Chem. 2011, 19, 6077−6086. (13) Ferguson, A. D.; McKeever, B. M.; Xu, S.; Wisniewski, D.; Miller, D. K.; Yamin, T. T.; Spencer, R. H.; Chu, L.; Ujjainwalla, F.; Cunningham, B. R.; Evans, J. F.; Becker, J. W. Crystal structure of inhibitor-bound human 5-lipoxygenase-activating protein. Science 2007, 317, 510−512. (14) Martinez Molina, D.; Wetterholm, A.; Kohl, A.; McCarthy, A. A.; Niegowski, D.; Ohlson, E.; Hammarberg, T.; Eshaghi, S.; Haeggstrom, J. Z.; Nordlund, P. Structural basis for synthesis of inflammatory mediators by human leukotriene C4 synthase. Nature 2007, 448, 613−616. (15) He, S.; Wu, Y.; Yu, D.; Lai, L. Microsomal prostaglandin E synthase-1 exhibits one-third-of-the-sites reactivity. Biochem. J. 2011, 440, 13−21. (16) He, S.; Lai, L. Molecular docking and competitive binding study discovered different binding modes of microsomal prostaglandin E synthase-1 inhibitors. J. Chem. Inf. Model. 2011, 51, 3254−3261. (17) Wu, Y.; He, C.; Gao, Y.; He, S.; Liu, Y.; Lai, L. Dynamic modeling of human 5-lipoxygenase−inhibitor interactions helps to discover novel inhibitors. J. Med. Chem. 2012, 55, 2597−2605. (18) Lang, P. T.; Brozell, S. R.; Mukherjee, S.; Pettersen, E. F.; Meng, E. C.; Thomas, V.; Rizzo, R. C.; Case, D. A.; James, T. L.; Kuntz, I. D. DOCK 6: combining techniques to model RNA-small molecule complexes. RNA 2009, 15, 1219−1230. (19) Wei, D.; Jiang, X.; Zhou, L.; Chen, J.; Chen, Z.; He, C.; Yang, K.; Liu, Y.; Pei, J.; Lai, L. Discovery of multitarget inhibitors by combining molecular docking with common pharmacophore matching. J. Med. Chem. 2008, 51, 7882−7888. (20) Morris, G. M.; Goodsell, D. S.; Halliday, R. S.; Huey, R.; Hart, W. E.; Belew, R. K.; Olson, A. J. Automated docking using a Lamarckian genetic algorithm and an empirical binding free energy function. J. Comput. Chem. 1998, 19, 1639−1662. (21) Pelcman, B. O., K.; Schaal, W.; Kalvins, I.; Katkevics, M.; Ozola, V.; Suna, E. Preparation of Benzoxazoles as Inhibitors of MAPEG Family for Treatment of Inflammation. WO Patent 2007042816 A1, 2007. (22) Cryer, B.; Kimmey, M. B. Gastrointestinal side effects of nonsteroidal anti-inflammatory drugs. Am. J. Med. 1998, 105, 20S− 30S. (23) Waltenberger, B.; Wiechmann, K.; Bauer, J.; Markt, P.; Noha, S. M.; Wolber, G.; Rollinger, J. M.; Werz, O.; Schuster, D.; Stuppner, H. Pharmacophore modeling and virtual screening for novel acidic inhibitors of microsomal prostaglandin E(2) synthase-1 (mPGES-1). J. Med. Chem. 2011, 54, 3163−3174.

(24) Chen, Z.; Wu, Y.; Liu, Y.; Yang, S.; Chen, Y.; Lai, L. Discovery of dual target inhibitors against cyclooxygenases and leukotriene A4 hydrolyase. J. Med. Chem. 2011, 54, 3650−3660. (25) LigPrep, version 2.1; Schrödinger, LLC: New York, NY, 2005. (26) Chen, J.; Lai, L. Pocket v.2: further developments on receptorbased pharmacophore modeling. J. Chem. Inf. Model. 2006, 46, 2684− 2691.

3309

dx.doi.org/10.1021/jm301900x | J. Med. Chem. 2013, 56, 3296−3309