Probing Binding and Cellular Activity of Pyrrolidinone and Piperidinone Small Molecules Targeting the Urokinase Receptor

Abstract

The urokinase receptor (uPAR) is a cell-surface protein that is part of an intricate web of transient and tight protein interactions that promote cancer cell invasion and metastasis. Here we evaluate the binding and biological activity of a new class of pyrrolidinone (3) and piperidinone (4) compounds, along with derivatives of previously-identified pyrazole (1) and propylamine (2) compounds. Competition assays revealed that the compounds displaced a fluorescently-labeled peptide (AE147-FAM) with inhibition constant K_i ranging from 6 to 63 μM. Structure-based computational pharmacophore analysis followed by extensive explicit-solvent molecular dynamics simulations and free energy calculations suggested pyrazole-based 1a and piperidinone-based 4 adopt different binding modes, despite their similar two-dimensional structures. In cells, compounds 1b and 1f showed significant inhibition of breast MDA-MB-231 and pancreatic ductal adenocarcinoma (PDAC) cell proliferation, but 4b exhibited no cytotoxicity even at concentrations of 100 μM. 1f impaired MDA-MB-231 invasion, adhesion, and migration in a concentration-dependent manner, while 4b inhibited only invasion. 1f inhibited gelatinase (MMP-9) activity in a concentration-dependent manner, while 4b showed no effect suggesting different mechanisms for inhibition of cell invasion. Signaling studies further highlighted these differences, showing that pyrazole compounds completely inhibited ERK phosphorylation and impaired HIF1α and NF-κB signaling, while pyrrolidinone and piperidinone (3 and 4b) had no effect. Annexin V staining suggested that the effect of pyrazole-based 1f on proliferation was due to cell killing through an apoptotic mechanism.

Introduction

A characteristic of malignant tumors is the ability of a small subpopulation of cells to escape from the primary tumor (1-3). Throughout the metastatic cascade and formation of new blood vessels, there is an intense process of extracellular matrix (ECM) degradation (4). With increased motility and more effective adhesion to the ECM, malignant cells eventually gain access to the vasculature and spread to distant sites to form new colonies (3). These processes require a complex interplay of various cell surface-associated proteins (4,5). Among them, the glycosyl-phosphatidylinositol (GPI) anchored urokinase receptor (uPAR) has been implicated in nearly every step of cancer metastasis including adhesion (6-8), migration (8-10), invasion (8,9,11-14), and angiogenesis (9,11,12). This is attributed to its large number of transient and tight protein-protein interactions with soluble and membrane proteins (15).

uPAR contributes to pericellular proteolysis by binding and sequestering the multi-domain protease urokinase-type plasminogen activator (uPA) to the cell surface (16,17). Most of its reported interactions are transient, such as binding to vitronectin (18), integrins (19), or receptor tyrosine kinases such as the platelet-derived growth factor receptor PDGFR (20). Indirect interactions with EGFR and LRP1 have also been reported (21-23). The former is expected to occur through integrin, while the latter is mediated by PAI-1, which eventually leads to endocytosis and the recycling of uPAR. The weak interactions have only been detected in cellular studies except for the binding of the SMB domain of vitronectin to uPAR, which has been extensively studied through biophysical studies using surface plasmon resonance (SPR) (24). A crystal structure of uPAR in complex with the amino-terminal fragment of ATF (uPA_ATF) shows the presence of a large pocket with hydrophobic characteristic that is located at the uPAR•uPA interface (25,26).

Small molecules that bind to uPAR are expected to modulate its function. We have recently reported three classes of compounds that target uPAR (8,13). An anthraquinone-based compound (IPR-803) was shown to bind at submicromolar affinity (0.2 μM) (27) and inhibited the uPAR•uPA interaction with an IC₅₀ of 10 μM (13). Pyrazole-based compounds displaced a fluorescently-labeled peptide with an IC₅₀ of 30 μM. Cellular studies of these compounds revealed inhibition of breast MDA-MB-231 adhesion, migration invasion, and angiogenesis (28). This range of cellular activity, particularly its inhibition of cell growth, suggested that the compound may likely inhibit other targets in addition to uPAR, making it a potentially useful multi-targeted agent. In vivo, the compound had promising pharmacokinetics and little toxicity (8).

Starting with the structure of the pyrazole- and propylamine-based compounds, we identify a series of derivatives from commercial sources that include molecules with pyrrolidinone and piperidinone core structures. Additional derivatives of the pyrazole-based compounds are designed and synthesized. The four classes of compounds were evaluated in breast MDA-MB-231 and pancreatic ductal adenocarcinoma (PDAC) PANC-1 cells and revealed distinct biological activity despite their two-dimensional structural similarity. Finally, gel zymography, immunoblotting, and flow cytometry studies were carried out to gain further insight into their mechanism of action.

Materials and Methods

Fluorescence polarization

Polarized fluorescence intensities were measured using EnVision® Multilabel Plate Readers (PerkinElmer) with excitation and emission wavelengths of 485 and 530 nm, respectively (8,13). Samples were prepared in black BD Falcon™ 384-well Microplate with a final volume of 50 μL in duplicates. First, the compounds were serially diluted in dimethylsulfoxide (DMSO) and further diluted in 1×PBS buffer with 0.01% Triton X-100 to make final concentrations of 100 μM to 0.046 μM. Triton X-100 was added in the buffer to avoid compound aggregation. 35 μL of the compound solution and 10 μL of PBS containing uPAR was added to the wells and incubated for at least 15 minutes to allow the compound to bind to the protein. Finally, 5 μL of fluorescently-labeled AE147 peptide AE147-FAM (K-S-D-Cha-F-s-k-Y-L-W-S-S-K; Cha: L-b-cyclohexyl-alanine) was added to make a total volume of 50 μL in each well resulting in final uPAR and peptide concentrations of 320 and 100 nM, respectively. The final DMSO concentration was 2%, which had no effect on the binding of the peptide. Controls included wells containing only the peptide and wells containing both protein and peptide each in quadruplicates to ensure the validity of the reaction assay. A unit of millipolarization (mP) was used for calculating percentage inhibition of the compounds. Inhibition constants were measured using the Ki calculator available at http://sw16.im.med.umich.edu/software/calc_ki/.

Reagents

Biotinylated anti-human uPAR antibody (BAF807) was purchased from R&D Systems (Minneapolis, MN). Phospho-p44/42 MAPK (Thr202/Tyr204) rabbit monoclonal antibody (4370) and p44/42 MAPK mouse monoclonal antibody (9107) were from Cell Signaling Technology, Inc. (Danvers, MA). Actin (C-2, sc-8432) antibody was from Santa Cruz Biotechnology Inc. (Santa Cruz, CA).

Cell Culture

MDA-MB-231 and PANC-1 cells were cultured in Dulbecco's Modified Eagle Medium (Cellgro, Manassas, VA). AsPC-1 cells were cultured in RPMI-160 medium. Each medium was supplemented with 10% FBS, 1% penicillin/streptomycin and cells were cultured in a 5% CO₂ atmosphere at 37°C.

Cell Viability

10⁴ cells (MDA-MB-231, PANC-1 or AsPC-1) were plated overnight in 100 μl/well of 96 well plates. Cells were treated with DMSO (control) or compounds at the indicated concentrations for 3 days. Viable cells were quantified by MTT assay at absorbance of 570 nm and 630 nm (reference background) as previously described (8,29).

Adhesion

96-well plates were coated with 15 ng μl^-1 fibronectin (Sigma-Aldrich, St. Louis, MO) at 4 °C overnight, then blocked with 2% BSA in PBS for 1 h, as described previously (8,13,30). After starving with serum-free medium for 4 h, MDA-MB-231 cells (2.5 × 10⁴ cells ml^-1) were suspended in 100 μl of 0.1% FBS DMEM medium with various concentrations of uPAR compounds or DMSO control at 37 °C for 90 min. Medium was then carefully suctioned out from each well. Each well was washed three times with PBS and the number of adherent cells was quantified by MTT assay at 570 nm and 630 nm.

Invasion

Invasion assays were performed using BD Biocoat Matrigel invasion chambers (BD Biosciences, San Jose, CA) as previously described (8,13,31). The undersurface of the inserts was coated with 30 ng μl^-1 of fibronectin at 4 °C overnight. The inserts were equilibrated with 0.5 mL of serum-free medium for 2 h at 37 °C. After 4 h of serum starvation, cells were harvested and 5 × 10⁴ cells in 500 μl medium containing 0.1% FBS and the indicated compounds or 1% DMSO control were plated onto the upper chamber. 500 μL of 10% FBS medium containing the same amount of compounds or DMSO control was added to the lower chamber. After a 16 h incubation at 37°C in 5% CO₂, non-invaded cells were removed from the upper chamber with a cotton swab, and the invaded cells were fixed in methanol for 30 min at room temperature and stained with Hematoxylin Stain Harris Modified Method (Fisher Scientific, Waltham, MA) for 1 h at room temperature. We washed the filters with water 3 times. Filters were air-dried, and the number of invaded cells was counted in ten separate 200 × fields.

Gelatin zymography

MDA-MB-231 cells were treated with compounds in serum free medium for 24 h (8,32,33). The conditioned medium was collected and concentrated by Amicon Ultra centrifugal filter units (Millipore, #UFC 501024), and electrophoresed on 7.5% sodium dodecyl sulfate (SDS) polyacrylamide gels containing 1 mg/mL gelatin. After electrophoresis, the gel was washed twice in 2.5% Triton X-100 for 30 min at room temperature and incubated in buffer that contained 50 mM Tris-HCl (pH 7.6), 200 mM NaCl, 10 mM CaCl₂, and 0.02% Brij 35 at 37 °C for 36 h. Then, the gels were stained with 0.05% Coomassie brilliant blue (CBB) and de-stained with 30% methanol in 10% acetic acid. Areas of gelatinolytic degradation appeared as transparent bands on the blue stained background of the gel. Data were quantified using Li-Cor Odyssey Imaging System (Li-Cor, Lincoln, NE).

Western blot analysis

Six-well plates were coated with 30 ng μl^-1 of fibronectin at 4 °C overnight. 1.5 × 10⁶ serum starved MDA-MB-231 cells were plated onto each well in the presence of DMSO (control) or 50 μM compounds for 30 minutes. Total cell lysates were prepared in standard RIPA extraction buffer containing protease and phosphatase inhibitors (Sigma). Thirty μg of protein was separated by 10% SDS-PAGE and transferred to nitrocellulose membranes (Amersham, Arlington Heights, IL). The membranes were immunoprobed with phospho-p44/42 MAPK (Thr202/Tyr204) rabbit monoclonal antibody (1:2000) or p44/42 MAPK mouse monoclonal antibody (1:2000) at 4°C overnight. Next, membranes were incubated with IRDye 800-conjugated goat anti-mouse IgG (Rockland) or Alexa Fluor 680 goat anti-rabbit IgG (Invitrogen) as secondary antibodies. Bands were detected using Li-Cor Odyssey Imaging System (Li-Cor, Lincoln, NE).

Wound Healing Assay

Confluent cell monolayers in 12-well plates were wounded by scraping with a micropipette tip. The cells were washed and then cultured in complete media containing the compounds. The degree of wound closure was assessed in three randomly chosen regions by measuring the distance between the wound edges just after wounding and after 16 h and 40 h under Nikon Diaphot 300 microscope (13).

Apoptosis Assay

MDA-MB-231 cells were treated with indicated concentrations of 1a, 1f, 1i, 3 and 4 or 1% DMSO control for 24 h and flow cytometry analysis was performed as described previously (8).

Reporter Assay

MDA-MB-231 cells were transfected with each reporter assay. 16h after carrying out the transfection, the medium was changed to complete medium (DMEM containing 10% of fetal bovine serum, 1% NEAA, 100 U/ml Penicillin and 100 μg/ml Streptomycin). After 46 h of transfection, cells were treated with 1f at 25 μM. After 48 h of transfection, the dual-luciferase assay was used to measure the ratio of firefly versus Renilla luciferase units. The fold change was calculated by dividing the normalized luciferase ratio of each treated pathway-focused reporter by the normalized luciferase ratio of the untreated pathway-focused reporter. Experiments were performed in triplicate.

Pharmacophore Hypotheses

Twelve of the highest affinity compounds from pyrazole, pyrrolidinone and piperidinone derivatives (four from each series) were selected to generate pharmacophore hypotheses. Compound structures were prepared with LigPrep to enumerate all possible enantiomers. Three-dimensional conformations of compounds were generated by docking compounds to the uPA binding site on multiple uPAR crystal structures. We have included uPAR crystal structures obtained from the PDB databank both in its complex form (PDB code: 1YWH, 3BT2, 3U73) and in its apo form (PDB code: 3U74). uPAR structures were prepared with Protein Preparation Wizard in the Schrodinger (Portland, OR) modeling suite by removing solvent and ligands, creating disulfide bonds, filling missing side chains, adding hydrogen atoms and optimizing hydrogen network to a pH neutral environment. The docking of compounds to uPAR was carried out using Autodock Vina (34) (version 1.1.2). All optional docking parameters were kept at default values except that the num_modes was reset from 9 to 50 and exhaustiveness from 8 to 100. The docked conformations of the compounds, whose predicted binding affinity (vina score) are stronger than -7.0 kcal/mol, were kept for hypotheses generation.

The Phase program from Schrodinger was employed to generate common pharmacophore hypotheses from docked conformations. The standard features (Acceptor, Donor, Hydrophobic, Negative, Positive, and Aromatic Rings) implemented in the Phase program were used. To find common pharmacophores, a restriction of matching all 12 compounds to a five-site pharmacophore was applied. Further restrictions were applied to match at least two aromatic rings out of a maximum of 3 and 1 acceptor site out of a maximum of 2. After scoring the active compounds, 18 hypotheses survived with the default scoring function. Clustering on the hypotheses reduced the number down to 16. The hypotheses were then visualized in the presence of the uPAR receptor structure. Six hypotheses were selected for further evaluation.

QSAR Model Development

A set of 30 compounds for training and another 30 for testing, whose activity was uniformly distributed from most active to in-active across three different series, were prepared for QSAR model development. With the QSAR Model Development Tools provided in the Phase program, we built atom-based QSAR models using the partial least squared (PLS) regression method applying a set of binary valued variables that encode whether or not ligand atoms occupy various cube-shaped elements of space based on the hypotheses developed above. The QSAR model performance was assessed using a number of metrics, including R² the coefficient of determination for regression, stability (model stability to the change of training set composition), Q² on the test set, and Pearson correlation r on the test set. Assuming that good pharmacophore hypotheses will generate good QSAR model performance, two pharmacophore models (HS1119 and HS1135) were selected from the hypothesis pool with the aforementioned metrics, which were further assessed with extensive molecular dynamics simulations and energy calculations.

Molecular dynamics (MD) simulations and MM-PBSA calculations

One active compound from each of three series of the compound was selected for molecular dynamics simulations. Here we used 1a, 3a and 4c to represent the three series respectively. The compound binding pose that conformed to the corresponding hypothesis was used for setting up the simulation. In total, six MD simulations were carried out on three uPAR-compound complexes in two binding modes based on hypotheses HS1119 and HS1135 respectively. The uPAR crystal structure with PDB code 1YWH was used for all the simulations. The protocol for setting up the MD simulations was described elsewhere (35). Briefly, the simulation of uPAR protein and compound complex solvated by TIP3P (36) water molecules were carried out using AMBER9 (37) MD simulation package with ff99SB (38) protein force field. A deliberate annealing process (39) was employed to equilibrate the solvated structures before production runs were carried out. The pmemd in AMBER was employed for production runs. By assigning different initial velocities, four independent simulation trajectories 12 ns in length were carried out for each uPAR-compound complex. The first 2 ns on each trajectory was discarded for equilibration. MD snapshots were saved every 2 ps, which results in 5,000 structures per trajectory.

The binding energy of compounds to uPAR protein was calculated with MM-PBSA approach (40). The conformational stability of the bound compound during simulations was analyzed using RMSD from its initial binding mode. In total, 600 snapshots were extracted evenly from the production trajectories that kept on the initial binding mode and subject to MM-PBSA energy analysis. The MM-PBSA Perl scripts in Amber9 were employed to determine the binding energy.

General

All chemicals were purchased from either Aldrich or Acros and used as received. Column chromatography was carried out with silica gel (25-63 μ and used as received). ¹H and ¹³C NMR were recorded in CDCl₃ or d₄-Methanol on a Bruker 500 MHz spectrometer. Chemical shifts are reported in ppm using either residual CHCl₃ or MeOH as internal references.

1-(tert-butoxycarbonyl)piperidine-4-carboxylic acid (6)

To a stirred solution of isonipecotic acid (77.4 mmol, 10.0 g) and potassium carbonate (154.8 mmol, 21.4 g) in water (150 mL) at 0 °C was added dropwise a solution of di-t-butyldicarbonate (77.4 mmol, 16.9 g) in THF (150 mL). The reaction mixture was gradually warmed to r.t. and stirred overnight. The solvents were evaporated and the residue was dissolved in DCM. DCM layer was washed with 1N HCl (3 × 100 mL), water, dried over sodium sulfate, and concentrated in vacuo to give pure 6 (13.03 g, 75%) as a white powder. ¹H NMR (500 MHz, CDCl₃) δ 4.02 (br s, 2H), 2.85 (t, J = 11.5 Hz, 2H), 2.49 (m, 1H), 1.90 (d, J = 11.5 Hz, 2 H), 1.65 (m, 2H), 1.45 (s, 9H); ¹³C NMR (126 MHz, CDCl₃) δ 180.1, 154.7, 79.7, 40.7, 28.3, 27.6. HRMS m/z calcd for C₁₁H₁₈NO₄ [M-H]^-: 228.1241, found 228.1240.

tert-butyl 4-(2,2-dimethyl-4,6-dioxo-1,3-dioxane-5-carbonyl)piperidine-1-carboxylate (7)

To a stirred soluion of 6 (56.8 mmol, 13.03 g) and DMAP (5.68 mmol, 694 mg) in DCM (10 mL) at 0 °C were added DCC (62.5 mmol, 12.9 g) and 2,2-dimethyl-1,3-dioxane-4,6-dione (62.5 mmol, 9.00 g) sequentially. The reaction mixture was gradually warmed to r.t. and stirred overnight. Reaction was filtered over filter paper and washed with DCM. The resultant orange solution was concentrated in vacuo. Product was not isolated.

tert-butyl 4-(3-ethoxy-3-oxopropanoyl)piperidine-1-carboxylate (8)

To 7 was added abs. ethanol (200 mL) and the solution was refluxed for 48 h. The solution was concentrated in vacuo and purified by flash chromatography (DCM) to give 8 as a reddish oil (14.36 g, 85%). ¹H NMR (500 MHz, CDCl₃) δ 12.09 (s, 0.14H, enol OH), 4.89 (s, 0.14H enol C-H), 4.13 (q, J = 7.0 Hz, 2H), 4.10-3.96 (m, 2H), 3.42 (s, 2H), 2.81-2.67 (m, 2H), 2.62-2.52 (m, 1H), 1.85-1.71 (m, 2H), 1.55-1.43 (m, 2H), 1.39 (s, 9H), 1.21 (t, J = 7.0 Hz, 3H); ¹³C NMR (126 MHz, CDCl₃) δ 204.0, 180.2 (enol), 172.7 (enol), 167.0, 154.4, 87.52, 79.51, 61.3, 48.5, 47.1, 28.2, 27.1, 13.9; R_f = 0.2 (DCM). HRMS m/z calcd for C₁₅H₂₆NO₅ [M+H]⁺: 300.1805, found 300.1808.

(E)-tert-butyl 4-(3-ethoxy-2-(ethoxycarbonyl)acryloyl)piperidine-1-carboxylate (9)

Under argon, 8 (47.9 mmol, 14.36 g), triethyl orthoformate (143.7 mmol, 24 mL), and acetic anhydride (95.8, 9 mL) were mixed and refluxed at 100 °C for 48 h. Low-boiling impurities were evaporated off and the crude product was purified by flash chromatography (DCM) to give 9, as a yellowish oil (14.92 g, 88%). ¹H NMR (500 MHz, CDCl₃) δ 7.59 (s, 0.54 H, minor), 7.52 (s, 1H, major), 4.24 (q, J = 7.1 Hz, 2H), 4.21-4.09 (m, 7H), 4.08-3.92 (m, 4H), 3.09-3.01 (m, 0.58H, minor), 2.95-2.87 (m, 1H, major), 2.83-2.67 (m, 4H), 1.85-1.67 (m, 4 H), 1.58-1.47 (m, 4H), 1.41 (s, 19H), 1.37-1.26 (m, 9H), 1.23 (t, J = 7.1 Hz, 6H); ¹³C NMR (126 MHz, CDCl₃) δ major isomer: 201.7, 165.7, 165.3, 162.3, 154.61, 112.6, 79.3. 72.2, 60.5, 48.0, 45.4, 28.3, 27.2, 15.2, 14.2; minor isomer: 199.6, 165.2, 154.59, 112.9, 72.7, 60.7, 28.0, 15.1, 14.1. HRMS m/z calcd for C₁₈H₃₀NO₆ [M+H]⁺: 356.2068, found 356.2067.

tert-butyl 4-(1-(3,4-dimethylphenyl)-4-(ethoxycarbonyl)-1H-pyrazol-5-yl)piperidine-1-carboxylate (10)

3,4-dimethylphenyl hydrazine was prepared from its HCl salt by washing with sat. sodium bicarbonate solution and extracting with DCM. DCM was removed in vacuo. To a stirred solution of free hydrazine (1.1 eq) in abs. ethanol (0.3 M) was added 9 (1.0 eq) in abs. ethanol (0.1 M). The reaction was refluxed at 100 °C for 48 h. Ethanol was removed in vacuo and the crude reddish-brown residue was purified by flash chromatography (1% MeOH/DCM) to give 10 (4.76 g, 82%) as a reddish-brown oil. ¹H NMR (500 MHz, CDCl₃) δ 7.95 (s, 1H), 7.17 (d, J = 10.0 Hz, 1H), 7.08 (s, J = 10.0 Hz, 1H), 6.96 (d, J = 10.0 Hz, 1H), 4.25 (q, J = 7.1 Hz, 2H), 4.13-4.01 (m, 2H), 3.09-3.02 (m, 1H), 2.61-2.45 (m, 1H), 2.28 (s, 3H), 2.26 (s, 3H), 2.25-2.15 (m, 2H), 1.84-1.71 (m, 1H), 1.51 (app d, 2H), 1.39 (s, 9H), 1.31 (t, J = 7.0 Hz, 3H);

5-(1-(tert-butoxycarbonyl)piperidin-4-yl)-1-(3,4-dimethylphenyl)-1H-pyrazole-4-carboxylic acid (11)

To a stirred solution of 10 (1.0 eq) in 95% ethanol (0.1 M) was added a 2.0 M NaOH solution (4.0 eq). The reaction mixture was refluxed at 70 °C for 20 h. Ethanol was removed in vacuo and the resulting solid was acidified to pH 2 at 0 °C using 1M HCl. The reddish-brown solid was filtered off and washed with cold water to give 11 (3.57 g, 80%) as a tan solid. ¹H NMR (500 MHz, CDCl₃) δ 8.06 (s, 1H), 7.29-7.22 (m, 1H), 7.12 (s, 1H), 7.03 (d, J = 8.0 Hz, 1H), 4.27-4.01 (m, 2H), 3.12 (app t, J = 12.0 Hz, 1H), 2.69-2.55 (m, 2H), 2.34 (s, 3H), 2.32 (s, 3H), 2.29-2.20 (m, 2H), 1.57-1.54 (m, 2H), 1.46 (s, 9H); ¹³C NMR (126 MHz, CDCl₃) δ 168.1, 154.9, 150.6, 143.5, 138.3, 138.1, 136.8, 130.2, 127.5, 123.6, 111.1, 79.6, 35.1, 28.5, 28.4, 19.8, 19.6. HRMS m/z calcd for C₂₂H₃₀N₃O₄ [M+H]⁺: 400.2231, found 400.2246.

tert-butyl 4-(4-((3,5-dimethylphenyl)carbamoyl)-1-(4-isopropylphenyl)-1H-pyrazol-5-yl)piperidine-1-carboxylate – (12e-i)

To a stirred solution of 11 (1.0 eq) and DMAP (0.1 eq) in DCM (0.1 M) at 0 °C was added DCC (1.1 eq) and aniline (1.0 eq) sequentially. The reaction was gradually warmed to r.t., and stirred for 48 h. DCM was removed in vacuo and the crude product was isolated and flash columned with 30% ethyl acetate/hexanes to remove unreacted starting material. The isolated semi-pure material was used immediately in the next step.

1-(3,4-dimethylphenyl)-N-(3,5-dimethylphenyl)-5-(piperidin-4-yl)-1H-pyrazole-4-carboxamide - (5d-h)

To a stirred solution of 12e-i (1.0 eq) in DCM (0.5 M) at 0 °C was added TFA (0.5 M). The reaction mixture was warmed to r.t. and stirred for 1 h. The solvents were removed in vacuo. The organic residue was re-dissolved in DCM. The organic layer was washed with sat. sodium bicarbonate, brine, and dried over MgSO₄. The solvent was removed in vacuo to yield 1e-i. (In some cases purification by flash chromatography was employed using a solvent system of 10% (10% NH₄OH/MeOH)/DCM).

1-(3,4-dimethylphenyl)-N-(3-methoxy-5-(trifluoromethyl)phenyl)-5-(piperidin-4-yl)-1H-pyrazole-4-carboxamide - (1e)

(42 mg, 36%); ¹H NMR (500 MHz, CDCl₃) δ 8.26 (s, 1H), 7.98 (s, 1H), 7.46 (s, 1H), 7.41 (s, 1H), 7.22 (d, J = 8.0 Hz, 1H), 7.08 (m, 1H), 7.00 (d, J = 8.0 Hz, 1H), 6.85 (s, 1H), 4.86 (br s, 2H), 3.81 (s, 2H), 3.25 (app d, J = 12.5 Hz, 2H), 3.09 (m, 1H), 2.64 (app t, J = 11.0 Hz, 2H), 2.45 (m, 2H), 2.32 (s, 3H), 2.29 (s, 3H), 2.29 (m, 2H), 1.67 (app d, J = 12.5 Hz, 2H); ¹³C NMR (126 MHz, CDCl₃) δ 161.8, 160.3, 148.6, 139.8, 138.9, 138.6, 138.3, 136.6, 130.4, 127.3, 123.5, 115.1, 55.6, 45.1, 34.1, 28.2, 19.8, 19.5. HRMS m/z calcd for C₂₅H₂₈F₃N₄O₂ [M+H]⁺: 473.2159, found 473.2145.

N-(2-(benzyloxy)phenyl)-1-(3,4-dimethylphenyl)-5-(piperidin-4-yl)-1H-pyrazole-4-carboxamide - (1f)

(51 mg, 41%); ¹H NMR (500 MHz, CDCl₃) δ 8.50 (d, J = 7.0 Hz, 1H), 8.31 (s, 1H), 7.69 (s, 1H), 7.45-7.40 (m, 4H), 7.36 (m, 1H), 7.24 (d, J = 8.0 Hz, 1H), 7.13 (s, 1H), 7.07-6.99 (m, 4H), 5.15 (s, 2H), 3.16-3.06 (m, 3H), 2.54 (app t, J = 11.0 Hz, 2H), 2.34 (s, 3H), 2.32 (s, 3H), 2.30 (m , 2H), 2.03 (br s, 1H), 1.60 (app d, J = 12.5 Hz, 2H); ¹³C NMR (126 MHz, CDCl₃) δ 161.5, 149.6, 147.2, 138.6, 138.1, 137.9, 137.2, 136.4, 130.1, 128.8, 128.5, 128.4, 127.6, 127.5, 123.7, 123.5, 121.6, 119.8, 115.8, 111.7, 71.1, 46.7, 35.3, 31.6, 30.6, 29.0, 22.6, 19.8, 19.6, 14.1. HRMS m/z calcd for C₃₀H₃₃N₄O₂ [M+H]⁺: 481.2598, found 481.2596.

1-(3,4-dimethylphenyl)-N-(4-fluorophenethyl)-5-(piperidin-4-yl)-1H-pyrazole-4-carboxamide - (1g)

(61 mg, 58%); ¹H NMR (500 MHz, CDCl₃) δ 7.59 (s, 1H), 7.26-7.16 (m, 3H), 7.09 (s, 1H), 7.02-6.97 (m, 3H), 5.99 (m, 1H), 3.63 (dd, J = 12.5, 6.5 Hz, 2H), 3.13-3.05 (m, 3H), 2.89 (t, J =7.0 Hz, 2H), 2.56 (app t, J =11.0 Hz, 2H), 2.33 (s, 3H), 2.30 (s, 3H), 2.28 (m, 2H), 1.58 (app d, J =12.5 Hz, 2H); ¹³C NMR (126 MHz, CDCl₃) δ 163.7, 162.6, 160.7, 148.5, 138.5, 138.1, 138.0, 137.0, 134.6, 130.2, 127.4, 123.6, 115.5, 115.3, 115.1, 46.1, 40.7, 35.0, 34.7, 30.0, 19.8, 19.5. HRMS m/z calcd for C₂₅H₃₀FN₄O [M+H]⁺: 421.2398, found 421.2386.

1-(3,4-dimethylphenyl)-N-(4-isopropylphenyl)-5-(piperidin-4-yl)-1H-pyrazole-4-carboxamide - (1h)

(60 mg, 58%); ¹H NMR (500 MHz, CDCl₃) δ 7.87 (s, 1H), 7.82 (s, 1H), 7.23-7.17 (m, 4H), 7.11 (s, 1H), 7.04-7.02 (m, 1H), 3.12-3.05(m, 4H), 2.88 (m, 1H), 2.53 (app t, J = 11.0 Hz, 2H), 2.33 (s, 3H), 2.30 (s, 3H), 2.28 (m, 2H), 1.60 (app d, J = 12.5 Hz, 2H), 1.22 (d, J = 7.0 Hz, 6H); ¹³C NMR (126 MHz, CDCl₃) δ 161.9, 149.2, 145.0, 138.7, 138.1, 138.0, 137.0, 135.6, 130.1, 127.5, 126.8, 123.6, 120.6, 115.5, 46.3, 34.9, 33.5, 30.2, 24.0, 19.8, 19.5. HRMS m/z calcd for C₂₆H₃₃N₄O [M+H]⁺: 417.2649, found 417.2637.

1-(3,4-dimethylphenyl)-5-(piperidin-4-yl)-N-(4-propylphenyl)-1H-pyrazole-4-carboxamide - (1i)

(68 mg, 65%); ¹H NMR (500 MHz, CDCl₃) δ 7.85 (s, 1H), 7.81 (s, 1H), 7.44 (m, 2H), 7.22 (d, J = 8.0 Hz, 1H), 7.16-7.10 (m, 3H), 7.05-7.03 (m, 1H), 3.13-3.02 (m, 3H), 2.57-2.47 (m, 4H), 2.33 (s, 3H), 2.30 (s, 3H), 2.29 (m, 2H), 1.66 (br s, 1H), 1.61-1.55 (m, 4H), 0.93 (t, J = 7.0 Hz, 3H); ¹³C NMR (126 MHz, CDCl₃) δ 161.9, 149.5, 138.8, 138.7, 138.0, 137.9, 137.1, 135.5, 130.1, 128.9, 127.5, 123.6, 120.4, 115.5, 46.7, 37.4, 35.2, 30.8, 24.5, 19.8, 19.5, 13.7. HRMS m/z calcd for C₂₆H₃₃N₄O [M+H]⁺: 417.2649, found 417.2646.

4-methoxy-N-(4-methoxybenzylidene)aniline - (13)

In a round bottom flask with p-anisidine (1.0 g, 8 mmol) was added dichloromethane (40 mL), anhydrous magnesium sulfate (6 g), and followed by p-anisaldehyde (0.97 mL, 8 mmol). The reaction was stirred for 72 h and was then filtered and the solvent was removed in vacuo to give 1.92 g (99%) of an off-white solid. ¹HNMR (500 MHz, CDCl₃) δ 8.40 (s, 1H), 7.83 (d, J = 8.5 Hz, 2H), 7.21 (d, J = 8.5 Hz, 2 H), 6.97 (d, J = 9.0 Hz, 2H), 6.92 (d, J = 9.0 Hz, 2H), 3.86 (s, 3H), 3.83 (s, 3H).

1,2-bis(4-methoxyphenyl)-5-oxopyrrolidine-3-carboxylic acid - (14)

A solution of imine 13 (200 mg, 0.83 mmol) and succinic anhydride (200 mg, 0.83 mmol) in xylenes (1 mL) was refluxed for 24 h. The reaction was slowly cooled to ambient temperature and extracted with saturated sodium bicarbonate (3x). The aqueous extracts were washed with hexane and then made acidic by addition of concentrated HCl resulting in a white precipitate that was extracted into ether. The ether extract was dried and removed in vacuo to give a yellow-white powder (181 mg, 64%). ¹HNMR (500 MHz, CDCl₃) δ 7.21 (d, J = 9.0 Hz, 2H), 7.16 (d, J = 9.0 Hz, 2H), 6.83 (d, J = 9.0 Hz, 2H), 6.77 (d, J = 9.0 Hz, 2H), 5.38 (d, J = 5.0 Hz, 1H), 3.76 (s, 3H), 3.72 (s, 3H), 3.20-3.14 (m, 1H), 3.05 (dd, J = 17.5, 9.5 Hz, 1H), 2.96 (dd, J = 17.5, 7.0 Hz, 1H).

N-(2-ethoxyphenyl)-1,2-bis(4-methoxyphenyl)-5-oxopyrrolidine-3-carboxamide - (3)

To a stirred solution of oxo-pyrrolidine carboxylic acid 14 (181 mg, 0.53 mmol) and DMAP (6 mg, 0.053 mmol) in DCM (5 mL) at 0 °C was added DCC (120 mg, 0.58 mmol) followed by o-phenetidine (0.069 mL, 0.53 mmol). The reaction was warmed to ambient temperature and stirred for 3 days. The reaction was filtered over Celite and the solvent was removed in vacuo. The crude solid was purified by flash chromatography eluting with 60% ethyl acetate/hexane to give 146 mg (70%), as a white solid. ¹HNMR (500 MHz, CDCl₃) δ 8.38 (d, J = 8.0 Hz, 1H), 7.58 (s, 1H), 7.21-7.15 (m, 4H), 7.06-7.01 (m, 1H), 7.00 – 6.93 (m, 1H), 6.84-6.79 (m, 3H), 6.79-6.74 (m, 2H), 5.34 (d, J = 7.0 Hz, 1H), 4.04-3.94 (m, 2H), 3.75 (s, 3H), 3.71 (s, 3H), 3.24-3.11 (m, 2H), 3.00-2.93 (m, 1H), 1.26 (t, J = 7.0 Hz, 3H); ¹³CNMR (126 MHz, CDCl₃) δ 172.3, 168.7, 159.4, 157.3, 147.0, 131.4, 130.1, 128.1, 127.1, 125.4, 124.2, 120.9, 119.9, 114.4, 114.0, 110.9, 66.6, 64.0, 55.2, 55.1, 50.3, 34.8, 14.5. HRMS m/z calcd for C₂₇H₂₉N₂O₅ [M+H]⁺: 461.2071, found 461.2080.

N-(4-fluorobenzylidene)-4-methoxyaniline - (15)

In a round bottom flask with p-anisidine (1.0 g, 8 mmol) was added dichloromethane (40 mL), anhydrous magnesium sulfate (6 g), and followed by 4-fluorobenzaldehyde (0.86 mL, 8 mmol). The reaction was stirred for 72 h and was then filtered and the solvent was removed in vacuo to give 1.91 g (99%) of an off-white solid. ¹HNMR (500 MHz, CDCl₃) δ 8.44 (s, 1H), 7.88 (dd, J = 8.5, 5.5 Hz, 2H), 7.23 (d, J = 8.5 Hz, 2H), 7.15 (t, J = 9.0 Hz, 2H), 6.94 (d, J = 9.0 Hz, 2H), 3.84 (s, 3H).

anti-2-(4-fluorophenyl)-1-(4-methoxyphenyl)-6-oxopiperidine-3-carboxylic acid - (16)

A solution of imine 15 (910 mg, 3.97 mmol) and glutaric anhydride (453 mg, 3.97 mmol) in xylenes (5 mL) was refluxed for 72 h. The reaction was slowly cooled to ambient temperature and extracted with saturated sodium bicarbonate (3x). The aqueous extracts were washed with hexane and then made acidic by addition of concentrated HCl resulting in a white precipitate (457 mg, 34%). ¹HNMR (500 MHz, CDCl₃) δ 7.23-7.19 (m, 2H), 7.07 (d, J = 9.0 Hz, 2H), 7.01-6.96 (m, 2H), 6.74 (d, J= 9.0 Hz, 2H), 5.37 (d, J = 3.5 Hz, 1H), 3.70 (s, 3H), 2.84 (q, J = 9.5, 4.5 Hz, 1H), 2.78-2.64 (m, 2H), 2.40 (t, J = 7.0 Hz, 1H), 2.36 (t, J = 7.0 Hz, 1H), 2.23-2.16 (m , 1H), 2.10-2.00 (m, 1H).

anti-N-(2,4-dimethoxyphenyl)-2-(4-fluorophenyl)-1-(4-methoxyphenyl)-6-oxopiperidine-3-carboxamide - (4)

To a stirred solution of oxo-pyrrolidine carboxylic acid 16 (457 mg, 1.33 mmol), EDC (255 mg, 1.33 mmol), and HOBT (54 mg, 0.40 mmol) in DCM (7 mL) at 0 °C was added dropwise 2, 4-dimethoxyaniline (0.189 mL, 1.33 mmol). The reaction was warmed to ambient temperature and stirred for 2 days. The mixture was evaporated and partitioned between 1M HCl and ethyl acetate. The organic layer was washed with saturated sodium bicarbonate, brine, dried over anhydrous sodium sulfate, and evaporated. The crude solid was purified by flash chromatography eluting with ethyl acetate to give 444 mg (70%) of a pinkish solid after trituration with hexanes. ¹HNMR (500 MHz, CDCl₃) δ 8.10 (d, J = 9.0 Hz, 1H), 7.30 (s, 1H), 7.21-7.15 (m, 2H), 7.02 (d, J = 9.0 Hz, 2H), 6.93 (t, J = 9.0 Hz, 2H), 6.73 (d, J = 9.0 Hz, 2H), 6.45 (d, J = 9.0 Hz, 1H), 6.40 (d, J = 2.5 Hz, 1H), 5.26 (d, J = 7.5 Hz, 1H), 3.77 (s, 3H), 3.73 (s, 3H), 3.70 (s, 3H), 2.90-2.71 (m, 3H), 2.40-2.04 (m, 2H); ¹³CNMR (126 MHz, CDCl₃) δ 169.9, 168.9, 161.1, 158.0, 156.8, 149.3, 136.0, 133.9, 129.1, 129.0, 128.9, 121.02, 120.3, 115.7, 115.5, 114.1, 103.7, 98.5, 66.1, 55.6, 55.5, 55.2, 51.1, 31.5, 31.2, 23.2, 22.6, 14.1. HRMS m/z calcd for C₂₇H₂₈FN₂O₅ [M+H]⁺: 479.1977, found 479.1987.

Results

Synthesis

The synthesis of N-1 substituted pyrazoles 1e-i (Scheme 2) followed the route reported previously (41). The pyrazole core was prepared by condensation of 1,3-dicarbonyl enol ethers with hydrazines. Thus, commercially available isonipecotic acid was protected to the N-Boc version 6, which was converted to known β-keto ester 8 through a simple two-step sequence (coupling of Meldrum's acid followed by ethanolysis) (42). With the requisite β-keto ester 8 in hand, the formation of the pyrazole core was explored with commercially available hydrazine. Subsequent hydrolysis of 10, the amide bond formation of 12 with various anilines, and following the removal of N-Boc gave the desired N-1 substituted pyrazoles 1e-i.

Scheme 2.

The synthesis of substituted pyrrolidin-2-one 3 and piperidin-2-one 4 are summarized in Schemes 3 and 4, respectively. First, Schiff bases 13 and 15 were obtained with the appropriate aldehyde and amine in the presence of anhydrous MgSO₄ (43). The subsequent condensation of Schiff base with succinic anhydride or glutaric anhydride in refluxing xylenes gave the pyrrolidin-2-one 14 and piperidin-2-one 16 carboxylic acids, respectively (44). Amide formation was carried out by EDC coupling in the case of 16 and DCC coupling for 14 to give the final substituted pyrrolidin-2-one 3 and piperidin-2-one 4.

Scheme 3.

Scheme 4.

uPAR binding activity

Previously, we had reported the use of virtual screening to identify compounds that bind to uPAR (8). Docking of 300,000 compounds to the uPAR crystal structure was focused to a pocket located at the interface of the protein-protein complex between uPAR and its serine protease ligand urokinase-type plasminogen activator (uPA). Among the active compounds that had emerged were pyrazole 1 (IPR-69) and propylamine 2. Among the top ranking candidates, two compounds shared structural similarity to 1. These include pyrrolidinone 3 and piperidinone 4 (Scheme 1). Here, for the first time, we characterize the binding and cellular activity of these compounds along with their derivatives.

Scheme 1.

A similarity search of the ZINC chemical library was carried out using the chemical structure 1-4. Derivatives were acquired and tested for activity at an initial concentration of 50 μM using a fluorescence polarization (FP) assay (Fig. 1A). The FP assay probe consisted of labeling an α-helical peptide (AE147) with fluorescein (AE147-FAM). The labeled-peptide was shown to bind to uPAR with a K_D of 499 nM. A concentration-dependent study was performed for several derivatives that were active (Fig. 1B). The structure of these derivatives is shown in Tables 1-4. For 1, four derivatives were selected, namely 1a, 1b, 1c, and 1d (Table 1). Their K_i ranged from 10 to 27 μM. The replacement of the para-isopropopyl group of the parent compound with a dimethyl substituted phenyl ring in 1a and 1b enhanced the affinity of the compound from the parent compound suggesting a better fit into the binding pockets of uPAR. This is reflected in 1c and 1d whereby a hydrogen or a chlorine at the meta-position of R₃, also led to inhibition constants that were higher than the parent compound 1. Starting with the structure of 1a, a structure-based approach was followed to design and synthesize additional derivatives (1f-1i). Four compounds were more active than the parent at single concentration and were studied in a concentration-dependent manner (Table 1). The highest affinity derivative was 1e (K_i = 17 μM), whereas the lowest affinity derivative was 1g (K_i = 63.3 μM) (Table 1).

Figure 1. Fluorescence polarization.

(A)Fluorescence polarization screening of 393 compounds that are derivatives of 1-4. Compounds were screened against uPAR (320 nM) and AE147-FAM peptide (100 nM) at 50 μM. (B) Concentration-dependent fluorescence polarization study of the top 22 compounds.

Table 1. Derivatives of 1.

	R1	R2	R3	R4	Ki(μM)
1a(IPR-81)					21.9
1b (IPR-82)					18.9
1c(IPR-361)					27.0
1d (IPR-380)					10.3
1e(IPR-735)					51.2
1f(IPR-737I					17.2
1g (IPR-743)					28.3
1h (IPR-744)					63.3
1i (IPR-745)					31.9

Table 4. Derivatives of 4.

Compound	R1	R2	R3	Ki(μM)
4 (IPR-84)				18.9
4a(IPR-89)				18.9
4b (IPR-108)				24.1
4c (IPR-949)				5.7
4d (IPR-968)				19.5
4e (IPR-974)				16.0
4f (IPR-1036)				12.6
4g (IPR-1046)				10.9

The process was repeated for compounds 2-4. Similarity search identified derivatives that were tested at a single concentration and follow-up studies in a concentration-dependent manner were done for compounds that were more active than the parent. Three derivatives were identified for 2, namely 2a, 2b, and 2c, which had inhibition constants of 19.3, 33.8, and 13.9 μM respectively (Table 2). For 3, two compounds were identified with higher affinity, namely 3a and 3b having K_i values of 10.3 and 11.4 μM respectively (Table 3). For 4, we found seven active derivatives with binding affinities ranging from 5.7 to 24.1 μM. The highest affinity compound was 4c, which has a phenyl rather than a benzyl moiety at R₃. When two methoxy groups are introduced at R₂ to generate 4b, the affinity is significantly reduced. A methoxy group at the meta-position is responsible for the reduced affinity since 4e and 4f have a methoxy moiety at the ortho and para positions with K_i values that are less than 16 and 12.6 μM, respectively (Table 4).

Table 2. Derivatives of 2.

Compound	R1	R2	R3	Ki(μM)
2a (IPR-8)				19.3
2b (IPR-21)				33.8
2c (IPR-22)				13.9

Table 3. Derivatives of 3.

Compound	R1	R2	R3	IC50(μM)	Ki(μM)
3 (IPR-99)				ND	ND
3a (IPR-979)				18	10.3
3b(IPR-994)				20	11.4

Computational studies

Two binding modes were selected from the pharmacophore and QSAR modeling studies (Fig. 2A and 2B). We refer to these two modes as hypothesis (HS1119) and hypothesis (HS1135). In HS1119, the R₂ group of the pyrazole-, piperidinone-, and pyrrolidinone compounds are ensconced in a pocket in the uPA-binding hydrophobic pocket on uPAR (Fig. 2C). In HS1135, the R₂ is instead directed away from the protein toward the solvent (Fig. 2D). Rigorous computational studies were conducted to determine the stability and binding strength of the compounds in both binding modes. Explicit-solvent MD simulations for each binding mode were carried out for three compounds, 1b, 3a, and 4c. Analysis of the 12 ns trajectories revealed that 1b was highly unstable in binding mode HS1119 as evidenced by the RMSD of the compound that increases to 4 Å for three of the 12 ns trajectories and 8 Å for one of the trajectories (Fig. 3A and 3C). The compound was more stable in binding mode HS1135 (Fig. 3B and 3D). 3a on the other hand, showed significantly greater stability in the HS1119 binding pose with an RMSD that remained within 3 Å (Fig. 3E and 3G) in contrast to the HS1135 binding pose that exhibited an RMSD greater than 6 Å for two of the trajectories (Fig. 3F and 3H). A similar observation was made for piperidinone-based 4c, which remained remarkably stable over the course of the simulations, showing RMSDs within 1.5 Å for binding pose HS1119 (Fig. 3I and 3K) compared to the significantly larger RMSDs observed for HS1135 (Fig. 3J and 3L).

Figure 2. Compound binding modes.

Hypothesis HS1119 (A,C) and HS1135 (B, D) that emerged from pharmacophore and QSAR modeling. Features were rendered as orange rings (aromatic ring), green spheres (hydrophobic), blue spheres (donor), and pink spheres (acceptor). All active compounds (shown in sticks) were aligned to pharmacophore hypothesis in A and B. The pharmacophore along with its reference compound was shown in the pocket of uPAR, which is shown as ribbon (C,D). The residues making contact with the reference compound were rendered in thinner capped-sticks.

Figure 3. Compound binding mode stability.

Root-mean-square deviation (RMSD) of compound versus time for 1b (A-B), 3a (E-F), and 4c (I-J), and corresponding illustration of the motion of the compound for 1b (C-D), 3a (G-H), and 4c (K-L). RMSDs are determined for the compound relative to the initial docked structure. Illustrations of superimposed snapshots collected over the course of the trajectories are shown in ribbon representation for uPAR and in capped-sticks for compound (C, N, O, Cl and Br are shown in white, blue, red, pink and cyan, respectively). Carbon atoms are shown in yellow for the initial docked structure.

Free energy calculations using protein-compound structures collected from the MD simulations were carried out using the MM-PBSA approach (Table 5). The MM-PBSA free energy ΔG_PBSA is obtained by subtracting the free energy of the protein-compound complex from protein and compound for each snapshot collected from the MD simulation. The free energy consists of an internal energy term consisting of van der Waals and Coulomb potential energy (ΔE_PBELE), a solvation energy term consists of electrostatic and non-polar components (ΔE_PBTOT), and finally an entropy term (TΔS_NM) (Table 1). For 1a, ΔG_PBSA was only determined only for HS1135, as the compound was unstable in the HS1119 binding mode. In contrast, piperidinone- and pyrrolidinone-based 3a and 4c trajectories were stable in the HS1119 binding mode and resulted in ΔG_PBSA of -20 and -25 kcal/mol, respectively. The trajectories for HS1135 binding mode resulted in binding less stable ΔG_PBSA, namely -13, -11, and -21 kcal/mol for 1a, 3a and 4c, respectively. The binding free energy for 3a and 4c was significantly more stable for the HS1119 binding mode, in contrast to the pyrazole-based 1a, which prefers the HS1135 binding mode.

Table 5. Binding Free Energy and its Components from Computational Studies.

Compound	Hypothesis	ΔEVDW	ΔEPBEIE	ΔEGBELE	ΔESA	ΔEPBTOT	ΔEGBTOT	TΔSNM	ΔGPBSA	ΔGGB
1a	H1119	unstable
	H1135	-49.3±0.2	20.0±0.2	18.2±0.1	-5.6±0.0	-34.9±0.2	-36.7±0.2	-23.4±0.6	-11.5±0.6	-13.3±0.6
3a	H1119	-54.2±0.2	18.6±0.3	18.4±0.1	-6.9±0.0	-42.5±0.2	-42.7±0.2	-22.4±0.6	-20.1±0.6	-20.3±0.6
	H1135	-43.6±0.2	22.1±0.2	18.9±0.1	-6.2±0.0	-27.7±0.2	-30.9±0.2	-19.9±0.6	-7.8±0.6	-11.0±0.6
4c	H1119	-55.7±0.1	15.0±0.2	15.3±0.1	-6.6±0.0	-47.4±0.2	-47.0±0.2	-22.0±0.6	-25.4±0.6	-25.0±0.6
	H1135	-53.8±0.1	17.4±0.1	15.6±0.1	-6.6±0.0	-43.0±0.2	-44.8±0.2	-23.5±0.6	-19.5±0.6	-21.3±0.6

ΔE_VDW, van der Waals potential energy; ΔE_PBElE, electrostatic contributions to the binding energy, of which the polar solvent contributions were calculated with Poisson-Boltzmann equation; ΔE_GBELE, electrostatic contributions to the binding energy, of which the polar solvent contributions were calculated with Generalized Born equation; ΔE_SA, nonpolar solvent contribution to solvation free energy; ΔE_PBTOT, the sum of ΔE_VDW, ΔEI_PBELE and ΔE_SA; ΔE_GBTOT, the sum of ΔE_VDW, ΔE_GBELE and ΔE_SA; TΔS_NM, entropy calculated with normal mode analysis; ΔG_PB, the calculated free energy of binding using PB model; ΔG_GB, the calculated free energy of binding using GB model.

Effect of compounds on cell proliferation

We tested the cytotoxicity of several compounds using MTT as previously described (8,29). A concentration-dependent MTT study in MDA-MB-231 cells gave IC₅₀ values of 31, 13, and 25 μM for 1b, 2c, and 1f, respectively (Fig. 4). 1f had IC₅₀ values of 14 and 22 μM in AsPC-1 and PANC-1 cell lines respectively. In comparison, 4b showed no cytotoxicity with 70% cell viability even at 100 μM (Fig. 4).

Figure 4. Cytotoxicity and cell viability studies.

Effects of 1b, 1f, 2c, 4b on MDA-MB-231, AsPC-1 and PANC-1 cell viability as assessed by MTT assay.

To get insight into the cell killing mechanism, a flow cytometry analysis was performed with Annexin V-FITC and PI staining for 1a, 1f, 1i, 3, and 4 (Table 6 and Fig. 5). The level of apoptosis and necrosis in MDA-MB-231 cells was assessed after exposure to increasing concentrations of compound for 24 hours as a percentage of Annexin V-positive/PI-negative cells (apoptotic) and Annexin V-positive/PI-positive cells (necrotic) respectively (Table 5 and Fig. 5). The data reveal that only the pyrazole compounds cause apoptosis, with 1f and 1i exhibiting the most pronounced effects. Control DMSO-treated cells showed about 3% apoptotic and 6% necrotic cells. At concentrations of 1, 10, 25 and 50 μM, 1f resulted in 3, 45, 76 and 61% apoptotic cells and 7, 15, 8 and 21% necrotic cells respectively (Fig. 5F). For compound 1i, 10, 25 and 50 μM resulted in 11, 6, 46, and 86% apoptotic cells, and 7, 9, 10, and 12% necrotic cells, respectively. Finally, 1a exhibited less significant apoptosis with nearly 39 and 32% of cells undergoing apoptosis and necrosis at 50 μM, respectively.

Table 6. Mechanism of Compound Cell Killing Studied by Flow Cytometry.

		DMSO	1 μM	10 μM	25 μM	50 μM
1a	Percentage apoptosis	2	1	5	16	39
	Percentage necrosis	6	5	12	17	32
1f	Percentage apoptosis	3	3	45	76	61
	Percentage necrosis	6	7	15	8	21
1i	Percentage apoptosis	2	11	6	46	86
	Percentage necrosis	6	7	9	10	12
3	Percentage apoptosis	2	2	1	2	4
	Percentage necrosis	6	6	3	8	7
4	Percentage apoptosis	2	4	2	3	3
	Percentage necrosis	6	3	14	6	5

Figure 5. Flow cytometer analysis using Annexin V-FITC and PI staining.

MDA-MB-231 cells were treated for 24 hr with 1f, 1a, 1i, 3 or 4 and analyzed for apoptosis and necrosis by flow cytometry.

Western blot analysis was performed test for the effects of compounds on MAPK signaling, given that the pathway regulates cell proliferation and survival. Western blot analysis was carried out following plating of serum starved cells onto fibronectin coated culture plates for 30 min in the presence of 50 μM compounds. Results show that pyrazole-based 1f, 1i and 1a significantly impaired p44/42 MAPK phosphorylation. Propylamine-based 2c exhibited moderate inhibition. In contrast, 4, 4b and 3 showed no inhibition of p44/42 MAPK phosphorylation (Fig. 6A), consistent with their complete lack of cytotoxicity. Luciferase reporter assays on MDA-MB-231 cells revealed that NF-κB, HIF1α and ERK were active even at basal level (Fig. 6B). 1f impaired all three pathways. The effect on HIF1α is consistent with inhibition of ERK1/2 phosphorylation, which is known to control activity of HIF1α. The effect on NF-κB can also be attributed to inhibition of ERK phosphorylation. In both cases, however, it is possible that the effects are ERK-independent.

Figure 6. Effect on signaling.

(A)Western immunoblotting showing the effects of compounds at 50 μM concentration on phospho-p44/42 MAPK and total p44/42 MAPK in MDA-MB-231 cell lysates prepared as described in Materials and Methods. Actin is shown as loading control. (B) Luciferase reporter assay testing NF-κB, Hypoxia, MAPK/ERK transcription of 1f on MDA-MB-231 cells.

Effect on cell invasion, migration and adhesion

To test the effects of compounds on MDA-MB-231 invasion, the Matrigel coated Boyden chamber was used as previously described (13) (Fig. 7). The compounds were initially tested at two concentrations (Fig. 7A). First, at 10 μM concentration, 1b, 1f, 2b, and 2c, and 4b. Each compound inhibited invasion with the pyrazole (1) and propylamine (2) derivatives exhibiting highest activity, while the piperidinone (4) and pyrolidinone (3) showing milder effects. At 25 μM 1f, 2b and 2c completely inhibited invasion, while compounds 1b, 4b and 3 showed moderate inhibition of invasion (Fig. 7A). The effects of these compounds on migration were also tested at 25 μM using wound healing. Compounds 1f and 2c inhibited cell migration by 70 and 50%, respectively (Fig. 7C and 7D). In contrast, 1b, 3, and 4b had no effect on migration (Fig. 7C and 7D). Cell viability was studied using MTT at 16 h within a similar time frame of the invasion and migration studies (Fig. 7C). At 10 μM, most compounds did not exhibit any cytotoxicity except for 1f, 2b and 2c, inhibit cell growth by 20, 60 and 80%, respectively. Hence, the effects of invasion observed at 10 μM for 1b, 1f, 3, and 4b are unlikely due to cell toxicity. At 25 μM, only the pyrrolidinone 3 and piperidone 4b were not cytotoxic, confirming that the 50% inhibition of cell invasion at this concentration is unlikely due to effects on cell viability.

Figure 7. Invasion and migration studies.

(A)Boyden chamber apparatus (with Matrigel) is used to assess the effects of compounds at 10 μM and 25 μM concentrations on MDA-MB-231 invasion. (B) Cell viability measured by MTT after 16 hours within the time frame of the invasion studies. (C) Effects of compounds at 25 μM concentration on MDA-MB-231 migration as assessed by wound healing assay. (D) Snapshots taken over the course of the wound healing experiment to illustrate the effect of the compounds on migration.

We used a Matrigel-coated Boyden chamber apparatus in a concentration-dependent study to determine the effect of 1f and 4b on MDA-MB-231 invasion in a concentration-dependent manner. 1f inhibited invasion in a concentration-dependent manner with an estimated IC₅₀ less than 10 μM (Fig. 8A). At 25 μM, a complete inhibition of invasion was observed. To test whether the inhibition of invasion by 1f is dependent on the degradation of ECM by matrix metalloproteinases (MMPs), we performed gelatin zymography as previously described (13,29). 1f inhibits MMP-9 activity in a concentration-dependent manner with an IC₅₀ of 28 μM (Fig. 8B). Similarly at 25 μM, a complete inhibition of migration is observed with 1f as assessed by wound healing (Fig. 7B). We then tested the effects of the piperidinone derivative 4b in cancer cell invasion and metastasis. Similar to pyrazole-based compound 1f the piperidinone-based 4b impaired invasion in a concentration-dependent manner, with 50% inhibition observed at 25 μM (Fig. 7A). Importantly 4b showed no cytotoxicity in cell viability studies conducted in parallel with invasion studies. Furthermore, unlike 1f, 4b showed no inhibition of MMP-9 activity (Fig. 8B) suggesting that the effects of 4b on invasion are MMP-9 independent.

Figure 8. Invasion, MMP activity and adhesion studies.

(A)Boyden chamber apparatus (with Matrigel) is used to assess the effect of 1f and 4b on MDA-MB-231 invasion in a concentration dependent manner. (B) Effect of 1f and 4b on MDA-MB-231 matrix metalloproteinase (MMP-9) activity using gelatin zymography. (C) Effect of 1f and 4b on MDA-MB-231 cell adhesion to fibronectin.

Next, we tested the effect of the 1f on the integrin-mediated process of cell adhesion (18,46-48). MDA-MB-231 cells were added to wells pre-coated with fibronectin. In the presence of 1f, a concentration-dependent impairment of cell adhesion is observed with an estimated IC₅₀ of 25 μM (Fig. 8C). In contrast to the pyrazole-based compound 1f, the piperidinone derivative 4b showed no impairment of cell adhesion (Fig. 8C) consistent on the lack of effect on migration (Fig. 7B).

Discussion

Previously, we had reported the discovery of anthraquinone-based compound (13,32,49) and pyrazole-based compound (8) that bind to uPAR and characterized their activity both in vitro and in vivo. In this study, we report the design, synthesis and biological characterization of a series of derivatives of the pyrazole-based 1, piperidinone 4 and pyrrolidinone 3 cores that were shown to bind to uPAR. Using fluorescence polarization (FP), we measured the inhibition constants of these derivatives and their K_i ranged from 6 to 63 μM. Computational studies revealed that the flat aromatic ring of the pyrazole-based compounds preferred a different binding mode to uPAR in comparison to 3 and 4. The chiral centers in the latter compounds direct the substituents bound to the central core into different pockets, resulting in better complementarity with the protein.

The difference in the structure of 1 and 4 is reflected in their dramatically different cellular activity. In cell-based assays the pyrazole-containing compound such as 1f showed significantly greater cytotoxicity. Flow cytometry analysis revealed that this compound also caused significant apoptosis in MDA-MB-231 cells. In contrast, the piperidinone compound, 4b was not cytotoxic even up to 100 μM. Furthermore, 1f impaired MDA-MB-231 cell invasion, adhesion, and migration in a concentration-dependent manner, whereas 4b inhibited invasion but had no effect on adhesion or migration. The effects of 1f on adhesion and migration suggest that the pyrazole-based 1f likely inhibits the interactions of uPAR with integrins. On the other hand, the lack of effects of the piperidinone compound on cell migration and adhesion suggest that it is unlikely to perturb the uPAR-integrin association. Gelatin zymography shows that 1f impaired gelatinase (MMP-9) activity in a concentration-dependent manner, whereas the piperidinone derivative did not, indicating that these compounds also differ in their anti-invasion mechanism. It is well documented that binding of uPA to uPAR is critical for its activation, and this triggers the proteolytic degradation of the ECM (50,51). Activated uPA converts the extracellular zymogen plasminogen into the active serine protease plasmin, which degrades the ECM components including fibrinogen, fibronectin and vitronectin and activates several matrix metalloproteases (MMPs) (52-55).

Finally, our signaling studies revealed that the pyrazole-based compounds such as 1f, 1i and 1a significantly inhibit MAPK signaling whereas the piperidinone compounds (4 and 4b) and the pyrrolidinone (3) had no effect. The inhibition of MAPK phosphorylation may explain the broad cellular activity exhibited by 1f that includes inhibition of cell growth, migration, invasion and adhesion, as well as apoptosis. The compound also impaired HIF1 α and NF-κB signaling. These effects may be due to additional unknown off-targets. Future studies will utilize these compounds as valuable leads to design derivatives with higher affinity and to unravel the protein-protein interactions of uPAR.