Dual targeting of MDM2 with a novel small-molecule inhibitor overcomes TRAIL resistance in cancer

Summary

Most of the conventional MDM2 inhibitors interfere only at MDM2–p53 interaction and confer limited clinical application due to p53-mediated MDM2 transactivation. Here, we report a novel dual-targeting small-molecule inhibitor of MDM2 that overcomes TRAIL resistance in cancer.

Abstract

Mouse double minute 2 (MDM2) protein functionally inactivates the tumor suppressor p53 in human cancer. Conventional MDM2 inhibitors provide limited clinical application as they interfere only with the MDM2–p53 interaction to release p53 from MDM2 sequestration but do not prevent activated p53 from transcriptionally inducing MDM2 expression. Here, we report a rationally synthesized chalcone-based pyrido[ b ]indole, CPI-7c, as a unique small-molecule inhibitor of MDM2, which not only inhibited MDM2–p53 interaction but also promoted MDM2 degradation. CPI-7c bound to both RING and N-terminal domains of MDM2 to promote its ubiquitin-mediated degradation and p53 stabilization. CPI-7c-induced p53 directly recruited to the promoters of DR4 and DR5 genes and enhanced their expression, resulting in sensitization of TNF-related apoptosis-inducing ligand (TRAIL)-resistant cancer cells toward TRAIL-induced apoptosis. Collectively, we identified CPI-7c as a novel small-molecule inhibitor of MDM2 with a unique two-prong mechanism of action that sensitized TRAIL-resistant cancer cells to apoptosis by modulating the MDM2–p53–DR4/DR5 pathway.

Introduction

Mouse double minute 2 (MDM2, also termed HDM2 in humans) oncogene is amplified and overexpressed in a number of human malignancies and its expression often correlates with poor survival of cancer patients ( 1–3 ). The tumor suppressor p53 is known to be the primary target of MDM2-mediated proteasomal degradation ( 4 , 5 ). Structurally, MDM2 protein has two distinct domains with different functional roles. N-terminal domain of MDM2 primarily interacts with p53 ( 6 ), and its C-terminal RING domain has E3 ubiquitin ligase activity that promotes p53 ubiquitination followed by proteasomal degradation ( 7 , 8 ). The major limitation associated with conventional MDM2 inhibitors, such as Nutlin-3, is that they interact only with its N-terminal and interfere with MDM2–p53 interaction but do not exert any effect on its catalytic RING domain leading to accumulation of active MDM2 ( 9–11 ). Moreover, MDM2 and p53 regulate each other mutually through the autoregulatory feedback loop where stabilization and activation of p53 results in MDM2 transactivation and overexpression. On the other hand, MDM2 directly interacts with p53 to inhibit its function ( 4 , 5 , 12 ). Furthermore, MDM2 is stabilized by a closely related protein MDMX (also called MDM4) owing to dimerization mediated by the conserved C-terminal RING domains of both MDM2 and MDMX and this prevents MDM2 autoubiquitination ( 13–15 ). These observations suggest that stabilization of p53 requires a strategy that involves not only interference with the N-terminal domain of MDM2 but also impairment of the MDMX interacting catalytic RING domain so that it can undergo autodegradation. Therefore, ideally total stabilization of p53 should be accomplished by blocking both N-terminal and RING domains of MDM2 followed by MDM2 autodegradation, in order to circumvent p53-mediated MDM2 transactivation.

Various post-translational modifications of p53 are associated with its specific function ( 16 ). For instance, serine 15 phosphorylation of p53 is responsible for uncoupling of MDM2 from p53 and results in p53 stabilization and subsequently apoptosis of target cell ( 17 ). Similarly, functionally diverse p53 regulates cell cycle arrest, DNA repair, senescence and apoptosis by targeting an array of genes such as MDM2 , CDKN1A , BAX , DR4 , DR5 etc ( 1 , 18 ). TNF-related apoptosis-inducing ligand (TRAIL), which acts through its receptors (also known as death receptors) TRAIL-R1 (DR4) and TRAIL-R2 (DR5), is one of the best-known strategies against cancer for its unique ability to induce apoptosis specifically in cancer cells without damaging normal counterpart ( 19–21 ). However, the major obstacle in this approach is the development of TRAIL resistance, particularly due to the paucity of TRAIL receptors DR4 and DR5 on the cell surface of many tumors ( 22–24 ). Furthermore, the promoters of DR4 and DR5 genes have been shown to possess consensus binding sites for p53 ( 25 , 26 ). Therefore, MDM2-induced degradation of p53 may be a key cause of down regulation of DR4 and DR5 and TRAIL resistance.

Anticancer properties of naturally occurring and synthetic chalcones particularly through their ability to inhibit p53–MDM2 interaction are widely reported ( 27 , 28 ). However, their efficacy has been questioned due to the other oncogenic functions of MDM2 protein ( 27–33 ). Subsequently, pyrido[ b ]indole alkaloids have also emerged as anticancer compounds especially due to their ability to directly bind and inhibit MDM2 function ( 34–36 ). Hence, considering the differential but related anticancer effect of naturally occurring and synthetic chalcone moieties, we utilized our natural product-inspired drug-designing approach and synthesized a series of chalcone-based pyrido[ b ]indole derivatives. We reported their potent anticancer activity initially in breast cancer ( 37 ). Here, we provide evidence that one of the compounds, CPI-7c [( E )-1-(3,4-dimethoxyphenyl)-3-(9H-pyrido[3,4- b ]indol-1-yl)prop-2-en-1-one], is a far more potent inhibitor of MDM2 as compared with Nutlin-3 particularly due to its additional ability to interact with RING domain of MDM2. We further provide molecular insight into its anticancer activity and report that CPI-7c restored the expression of death receptors (DR4 and DR5) on the surface of TRAIL-resistant cells by stabilizing and activating p53.

Materials and methods

Materials

Compound CPI-7c was synthesized in-house, and procedures of synthesis and nuclear magnetic resonance-based purity profile were published previously ( 37 ). Reagents were purchased from following suppliers: Pierce Magnetic ChIP kit, antibodies for GAPDH and verso one-step reverse transcription–PCR (RT–PCR) kit from Thermo Scientific; human apoptosis proteome profiler kit and neutralizing antibodies for TRAIL-R1 (DR4) and TRAIL-R2 (DR5) from R&D systems; sulforhodamine-B (SRB), TRAIL, crystal violet dye and dimethyl sulfoxide were procured from Sigma–Aldrich; antibodies for p53 (DO1), MDM2, horseradish peroxidase (HRP)-conjugated secondary antibodies from Santacruz; Agarose A/G beads and antibody for DR4 from Millipore; antibodies for DR5, ubiquitin and chromatin immunoprecipitation (ChIP) grade p53(7F5) from Cell Signaling Technologies; phycoerythrin-conjugated antibody for DR4 and DR5 used in the flowcytometry and platinum human TRAIL ELISA kit from e-Biociences; DharmaFECT transfection reagent, control and p53-specific siRNA and MGC Human XIAP Sequence Verified cDNA from Dharmacon; and primers for DR4, DR5, GAPDH used for semiquantitative PCR and that of p53 binding sites in DR4 and DR5 gene sequence were purchased from IDT. Nutlin-3 and ChIP grade antibody for ubiquitin, and HDM2 (catalytic RING domain) peptide were purchased from Enzo Life Sciences. HRP-conjugated secondary antibody against kappa light chain were purchased from Jackson Lab. All chemicals and antibodies were obtained from Sigma unless specified otherwise.

Cell culture

Representative cell lines from different cancer types DLD-1 and SW-620 (colorectal adenocarcinoma), A-549 (lung carcinoma), SKOV-3 (ovarian adenocarcinoma) and MCF-7 (breast adenocarcinoma) and MCF-10A (non-tumorigenic breast epithelial cells) were obtained from American Type Culture Collection (ATCC). HCT-116 [p53 wild type (WT)] and HCT-116 (p53 null) cells were generous gift from Dr Bert Vogelstein, Baltimore, MD. Early passage cells were cultured as monolayers in recommended media supplemented with 10% fetal bovine serum (Invitrogen), 1× Anti-Anti (Invitrogen; containing 100 μg/ml streptomycin, 100U/ml penicillin and 0.25 μg/ml amphotericin B) and maintained in 5% CO ₂ and humidified environment at 37°C. In all treatments, compounds were dissolved in cell culture grade dimethyl sulfoxide at concentration of 50mM. The subconfluent cells were treated with required doses of compounds in all the experiments.

Cell viability assay

A standard colorimetric SRB assay was used for the measurement of cell cytotoxicity ( 37 , 38 ). In brief, 10000–30000 cells (depending on the doubling time of each cell type) were seeded to each well of 96-well plate in 5% serum containing growth medium and incubated overnight to allow for cell attachment. Cells were then treated with test compounds at the required dose and untreated cells receiving the same volume of vehicle containing medium served as control. After 48h of exposure, cells were fixed with ice cold 10% trichloroacetic acid, stained with 0.4% (wt/vol) SRB in 1% acetic acid, washed and air dried. Bound dye was dissolved in 10mM Tris base and absorbance was measured at 510nm on a plate reader (Epoch Microplate Reader, Biotek). The cytotoxic effects of compounds were calculated as percentage inhibition in cell growth as per the formula [100 − (absorbance of compound treated cells/absorbance of untreated cells) × 100].

Crystal violet staining

Crystal violet staining was performed as a visualization method for cell viability. It stains nuclei a deep purple color that aids in their visualization. Here, 10 ⁵ MCF-7 and MCF-10A cells were seeded in a 12-well plate in 5% serum containing growth medium and incubated overnight to allow them to adhere. Cells were treated with either multiple doses of CPI-7c in treatment group or same volume of vehicle containing media in control group. After 48h of treatment, cells were fixed with ice-cold methanol followed by staining with 0.5% crystal violet solution. After 20min of incubation in staining, crystal violet solution was removed, plate was washed thoroughly with water, dried and representative images were taken to monitor cell viability.

Apoptosis protein array

MCF-7 cells were plated at a density of 1×10 ⁶ cells/60mm tissue culture dish in 3ml of culture medium for 24h and then were treated with 10 μM concentration of compound CPI-7c. After 24h of treatment, cell lysates were prepared by homogenization in lysis buffer 17 (R&D systems) and incubated on ice for 30min. Lysates were cleared by centrifugation at 14000 g for 15min at 4°C. The protein concentration of the supernatants was measured using the BCA protein assay kit (Thermo Scientific-Pierce) with bovine serum albumin as a standard. Apoptosis array was done as described previously ( 39 ).

RNA isolation and RT–PCR analysis

Total RNA was prepared using the Pure Link™ RNA Mini kit (Ambion). The cDNA synthesis and PCR were carried out by verso one-step RT–PCR kit (Thermo Scientific) using gene-specific primers as per the manufacturer’s protocol, and detailed procedure is described previously ( 40 ). The oligonucleotide primers used are as follows: for human GAPDH: forward: 5′-GTCAGTGGTGGACCTGACCT-3′, reverse: 5′-AGGGGAGATTCAGTGTGGTG-3′, product size 395bp; for human DR4: forward: 5′-AGAGAGAAGTCCCTGCACCA-3′, reverse: 5′-AGAGAGAAGTCCCTGCACCA-3′, product size 366bp; for human DR5: forward: 5′-TGCAGCCGTAGTCTTGATTG-3′, reverse: 5′-GCACCAAGT CTGCAAAGTCA-3′, product size 389bp; for human MDM2: forward: 5′-CAGTGGCGATTGGAGGGTAG-3′, reverse: 5′-GCTGGAATCTGTGAGGTGGT-3′; product size 130bp.

Western blot analysis

Protein samples were run on 4–15% gradient sodium dodecyl sulfate–polyacrylamide gel (Bio-Rad) and transferred to a polyvinylidene difluoride membrane (Millipore). The membranes were incubated blocking buffer followed by different primary antibodies. Subsequently, membranes were incubated with HRP-linked appropriate secondary antibody. The protein expression was visualized by an enhanced chemiluminescence solution (Immobilon™ western, Millipore) and scanned by gel documentation system (Bio-Rad Chemidoc XRS Plus).

Immunoprecipitation

Immunoprecipitations were performed with 0.5mg of total protein at antibody excess using anti-human MDM2 and anti-human ubiquitin antibodies. Immunocomplexes were captured with protein A/G PLUS-Agarose beads, and bead-bound proteins were subjected to western blot analysis using either anti-MDM2 or anti-p53 antibodies along with 5% input control ( 41 ). To avoid heavy chain immunoglobulin G (IgG) band, we used kappa light chain-specific HRP-conjugated secondary antibody.

Flow cytometry

Surface expression of DR4 and DR5 was measured on MCF-7 cells in 10 µM CPI-7c and vehicle-treated control group of cells. After 24h of treatment, cells were harvested by mild treatment of TrypLE (Invitrogen), washed with phosphate-buffered saline and then each group was split into isotype and test group of tubes. Test group of cells were stained with anti-human phycoerythrin-conjugated DR4 or DR5 antibody and appropriate IgG isotype control antibody. Finally, stained cells were washed with phosphate-buffered saline and acquired by flowcytometry using FACS Calibur (Becton Dickinson) and analyzed by FlowJo software (Treestar).

Receptor neutralization experiments

For the functional neutralization of TRAIL-R1 (DR4) and TRAIL-R2 (DR5) present on cell surface, cells were pretreated with 1 µg/ml of either anti-DR4 or anti-DR5 or both for 2h followed by treatment with 2.5 µM of CPI-7c. After 48h of incubation, SRB assay was performed to analyze corresponding rescue in CPI-7c-induced cellular cytotoxicity.

Enzyme-linked immunosorbent assay

Highly concentrated solution of culture supernatant and total cell lysate of control and either 2.5 or 5 µM CPI-7c-treated MCF-7 cells were taken, and enzyme-linked immunosorbent assay was performed as per manufacturer’s protocol. Recombinant TRAIL was used as positive control, and a standard curve was prepared for quantification of TRAIL level in the test samples.

siRNA knockdown experiments

MCF-7 cells were seeded on a six-well cell culture plate and allowed to grow up to 40% confluent monolayer of cells followed by addition of 50nM of p53-targeting siRNA or non-targeting control siRNA along with 6 μl of DharmaFECT transfection reagent in antibiotic-free medium as per manufacturer’s protocol. Cells were harvested for protein extraction and estimation followed by p53 knockdown validation using western blotting.

XIAP overexpression experiments

Human XIAP cDNA cloned in pCMV sport plasmid cloning vector was isolated from supplied bacterial glycerol stocks. For overexpression, 500000 of MCF-7 cells were seeded on a six-well cell culture plate and allowed to grow overnight followed by addition of 2 µg/ml of pCMV-XIAP plasmid or pCMV empty vector along with 4 μl of DharmaFECT transfection reagent in antibiotic-free medium as per manufacturer’s protocol. After 48h of addition, cells were harvested for protein extraction and estimation followed by XIAP overexpression validation using western blot analysis.

ChIP assay

ChIP assay was conducted using the ChIP assay kit purchased from Thermo Scientific following the manufacturer’s instruction. In brief, MCF-7 cells were treated with differential dose of CPI-7c. After 24h of treatment, genomic DNA and protein were cross-linked by addition of formaldehyde (1% final concentration) directly into the culture medium and incubated for 10min at 37°C. Cells then were collected and lysed in 200 μl of membrane extraction buffer containing protease inhibitor cocktail followed by 20U of MNase treatment in digestion buffer to obtain chromatin fragments. Cells were sonicated to generate DNA fragments of 100–500bp long. After centrifugation, the cleared supernatant was diluted 10-fold with IP buffer and 10 µl of it was kept as input control and rest is incubated at 4°C overnight with anti-RNA polymerase II antibody (Pierce) as positive control, anti-p53 monoclonal antibody (Cell Signaling Technology) as test for different groups and mouse IgG2 isotype antibody (Pierce) as negative control. Immune complexes were precipitated, washed and eluted as recommended. After DNA-protein cross-linkages were reversed by heating at 65°C for 4h, DNA was extracted in phenol/chloroform, precipitated with ethanol and resuspended in 50 µl of elution of buffer (pH 8.0). Each sample at the same volume was used as a template for PCR amplification of fragments containing the potential p53 binding site on immunoprecipitated chromatin, using specific primers as follows for DR4 and DR5 , respectively: DR4 forward: 5′-CTCGAGAAGTTTGTCGTCGTCGGGGT-3′; DR4 reverse: 5′-GAGCTCCCGTTCTTCCTCCGACTC-3′; DR5 forward: 5′-CTCGAGGTCCTGCTGTTGGTGAGT-3′; and DR5 reverse: 5′-GAGCTCGGG AATTTACACCAAGTGGAG-3′ ( 26 ).

Confocal microscopy

Vehicle control and treated MCF-7 cells were fixed with ice-cold pure methanol for 10min at −20°C followed by blocking with 2% bovine serum albumin for 1h at room temperature. After overnight primary antibody (anti-p53 and anti-MDM2) incubation, cells were washed twice with phosphate-buffered saline and incubated with fluorescent-conjugated secondary antibodies at room temperature for 1h, followed by 4′,6-diamidino-2-phenylindole staining for 5min at room temperature. After washing, cells were mounted with anti-fade mounting medium on glass slides and viewed under an inverted confocal laser scanning microscope (Ziess Meta 510 LSM; Carl Zeiss, Jena, Germany). Plan Apochromat 63X/1.4 NA Oil DIC objective lens was used for imaging and data collection. Appropriate excitation lines, excitation and emission filters were used for imaging.

Molecular modeling

Initially, the binding domains study of MDM2–p53 complex was done by using co-crystal complex structure with PDB id: 1YCR, which was retrieved from PDB database ( 42 , 43 ). The crucial docking studies were undertaken with more recent crystal structures of MDM2 for N-terminal (PDB id: 4OBA) and C-terminal (PDB id: 2VJF) by using Patchdock software ( 42 , 44 ). All heteroatoms and water molecules are removed from the PDB file, and bond angle and bond length were optimized by energy minimization using GROMOS 96 force field inbuilt in GROMACS 4.5.4 package ( 45 ). After optimizing the bond length and bond angles of N-terminal MDM2 structure (4OBA), bound inhibitor was removed and the grid was set over key residues L54, G58, Y67, V93, H96, I99, Y100 and Y103 (assuming as binding site residues as reported in earlier study), whereas for C-terminal fragment of MDM2 (2VJF), grid was set over the full-length structure of C-terminal MDM2. The final docking studies were performed on the defined grid of both structure with two inhibitors CPI-7c and Nutlin-3. The clustering root mean square deviation size was set as 4 Å and for each case and top 20 docked complexes were selected and analyzed. The most suitable complex was selected based on best docking score and best-fitting ligand pose in the binding cavity. The final selected MDM2 and inhibitor complex were subjected to energy minimization using GROMOS 96 43B1 force field parameters inbuilt in GROMACS 4.5.4 package. Further, the detailed binding interaction analysis of protein inhibitors and MDM2–p53 (1YCR) was carried out by using PyMOL v0.99. The three-dimensional structure views and pictures were generated by PyMOL v0.99.

Cloning of human MDM2

Human MDM2 gene (5–125) construct was synthesized from pIDTSmart between NcoI and XhoI restriction site. The synthesized gene fragments were digested with NcoI and XhoI and then ligated into the pET-21d(+) vector (Novagen) digested with the same enzymes. Competent Escherichia coli DH5-α cells were transformed with the plasmid constructs and screened for positive clones. The DNA sequencing of pIDTSmart- MDM2 (5–125) construct confirmed the homogeneity of the sequence.

Preparation of recombinant human MDM2 (5–125) protein (NT MDM2 )

Plasmid constructs expressing MDM2 (5–125) construct (NT MDM2 ) with the C-terminal histidine tag was transformed into the BL21 (DE3) E.coli cells. A single colony from transformed plate was inoculated into 5ml of LB broth (Hi-media) having ampicillin at a concentration of 100 μg/ml and allowed to grow overnight at 37°C. It was then subcultured in 400ml of LB broth containing ampicillin and allowed to grow at 37°C until A₆₀₀ ~0.6 was achieved and further induced at 20°C with 0.5mM isopropyl-1-thio-β-D-galactopyranoside. Induced cultures were further grown for 8h with shaking at the same temperature. The cells were harvested at 8000 r.p.m. for 5min, and the resultant pellet was stored at −70°C until further use. The histidine-tagged proteins were purified by resuspending the cells in lysis buffer containing 20mM Tris–HCl, 0.5M NaCl and 10% glycerol (pH 8) and disrupted using a probe-type ultrasonicator followed by centrifugation at 12000 r.p.m. for 30min at 4°C. The cell lysates for His-tagged proteins were loaded onto a nickel-nitrilotriacetic acid column pre-equilibrated with buffer containing 20mM Tris–Cl, pH 8, and 0.5M NaCl and washed with the same buffer followed by 20 and 30mM imidazole buffer. Protein was finally eluted with 300mM imidazole buffer and dialyzed against buffer containing 20mM Tris–Cl, pH 8, and 0.15M NaCl. The protein was dialyzed, concentrated and finally purified by passing through the size exclusion column pre-equilibrated with buffer containing 20mM Tris–Cl, pH 8, and 0.15M NaCl. The purity of recombinant proteins NTMDM2 with ~14kDa was checked by sodium dodecyl sulfate–polyacrylamide gel electrophoresis, and histidine tag was confirmed by western blotting using anti-His antibody.

Fluorescence quenching measurements

The fluorescence intensities were recorded with a Perkin Elmer Life Sciences LS 50B spectrofluorimeter in a 5mm path-length quartz cell at 25°C, using 15nm excitation and 12nm emission slit widths. For studying intrinsic fluorescence of proteins NTMDM2 and RdMDM2 (human recombinant GST tagged catalytic RING domain, Cat. No. BML-UW0200-0025, Enzo Life Sciences), excitation wavelength of 280nm was used and the spectra were recorded between 300 and 400nm. The λem _max for NTMDM2 and RdMDM2 were 346 and 340nm, respectively. An aliquot of protein NTMDM2 and RdMDM2 with final concentration of X and Y was titrated with the ligands Nutlin-3 (Cat. No. 430-128-M025, Enzo Life Sciences) and test compound at 25°C. For fluorescence quenching experiment, the concentration range of 5–100nM was used for both Nutlin-3 and test compound against NTMDM2 as well as range of 5–100 and 2.5–70nM was used for Nutlin-3 and test compound against RdMDM2, respectively, to the final volume. To eliminate the background effect on NTMDM2 and RdMDM2 fluorescence quenching values, the fluorescence emission intensities obtained through titrating equal amount of buffer without quencher and fluorescence intensity value obtained through GST protein in the case of RdMDM2 were subtracted from the fluorescence intensity values obtained for NTMDM2 and RdMDM2 quenching. Each measurement was repeated in triplicate, and the mean and standard deviation were calculated by Prism 3 software. Dissociation equilibrium constant ( K_d ) values were determined from data fitted to a single exponential hyperbolic equation, by using the PRISM 3 non-linear regression tool (GraphPad, San Diego, CA).

Statistical analysis

Statistical evaluation for data analysis was determined by Student’s t -test. Differences with P < 0.05 were considered statistically significant.

Results

CPI-7c activates p53 and upregulates the expression of death receptors DR4 and DR5

Previously, we have reported the synthesis and preliminary anticancer properties of chalcone-based pyrido[ b ]indoles ( 37 ), in which CPI-7c was found to be the most potent one. In our current endeavor, we first revalidated the cytotoxic ability of CPI-7c on representative cell lines of breast (MCF-7), colon (DLD-1) and ovarian (SKOV-3) cancer types in four different doses (1.25, 2.5, 5 and 10 µM) by SRB assay. We observed significant growth inhibition of CPI-7c in MCF-7 and DLD-1 cells; however, it was less effective on SKOV-3 cells ( Figure 1A ), indicating some selectivity in the target engagement, which is probably partially impaired in SKOV-3 cells. We also confirmed its apoptotic potential in breast cancer (MCF-7) cells by observing the induction of Annexin V staining in treated cells compared with vehicle-treated cells ( Figure 1B ). To understand in vitro safety efficacy of CPI-7c, we evaluated cytotoxic effects of CPI-7c in non-tumorigenic (MCF-10A) versus tumorigenic (MCF-7) cells by crystal violet staining of cells after 48h treatment. Interestingly, we observed that in spite of its higher cytotoxicity against tumorigenic MCF-7 cells, CPI-7c was found to be minimally effective in non-tumorigenic MCF-10A cells ( Figure 1C ). To decipher the potential target of CPI-7c in modulating apoptotic pathway, we made use of human apoptosis array from R&D Systems ( 39 , 46 ). This array platform simultaneously assessed the expression of 35 different apoptosis-related proteins spanned around both intrinsic and extrinsic pathways. Overnight grown MCF-7 cells were treated with either 10 µM of CPI-7c in test or with required amount of dimethyl sulfoxide in the vehicle control group for 24h, respective protein lysates were prepared and apoptosis antibody array were performed in both groups. Our results showed that a number of apoptotic signaling proteins were modulated following treatment of CPI-7c compared with control and shown along with reference name corresponding to each dot ( Figure 1D and Supplementary Figure 1 , available at Carcinogenesis Online). On the basis of the densitometry analysis of corresponding dots, a heat map was prepared for change in all 35 apoptosis-related proteins ( Figure 1F ). We observed marked increase in p53 phosphorylation at serine 15, serine 46 and serine 392 residues; strong upregulation of DR4 and DR5 proteins; and downregulation of anti-apoptotic protein XIAP ( Figure 1E ). However, there were no significant changes in the expression of Bax, Bak, Bcl2 or cytochrome C, implying the selective involvement of extrinsic pathway of apoptosis over intrinsic pathway ( Figure 1D ). Altogether, results obtained in the apoptosis protein array and respective heat map indicate that CPI-7c is able to induce apoptosis in MCF-7 cells by inducing and activating pro-apoptotic proteins p53, DR4 and DR5 and inhibiting XIAP.

Figure 1.

CPI-7c is potently cytotoxic to cancer cells and promotes DR4, DR5 upregulation and p53 activation. In panel ( A ), breast (MCF-7), colon (DLD-1) and ovary (SKOV-3) cancer cells were treated with four different doses (1.25, 2.5, 5 and 10 µM) of CPI-7c for 48h, and percentage growth inhibition was measured by SRB assay. In panel ( B ), MCF-7 cells were treated with 10 µM of CPI-7c for 24h and analyzed by flowcytometry for the Annexin V positivity. Panel ( C ) represents crystal violet staining in MCF-7 and MCF-10A cells after 48h treatment of vehicle, 1.0, 2.5 and 5.0 µM of CPI-7c. In ( D ), MCF-7 cells were treated with (10 µM) of CPI-7c for 24h, harvested for protein extraction and analyzed for the expression of apoptotic genes using human proteome profiler apoptosis array. Chemiluminescent images for the expression of 35 different apoptosis-related genes with positive and negative controls in duplicates for vehicle and CPI-7c-treated cells were shown. In ( E ), histogram representations of comparative pixel densities of significantly changed proteins were shown for control and treated cells. Panel ( F ) shows the comparative heat map image of these 35 genes based on pixel density of corresponding dots. Panel (F) represents the pixel intensity histogram of markedly altered target proteins.

CPI-7c transcriptionally upregulates the expression of death receptors on the surface of TRAIL-resistant cells and sensitizes them for TRAIL-induced apoptosis

For further validation of upregulation of death receptors and phospho-p53 in our array experiment, we treated MCF-7 cells with two different doses of CPI-7c (5, 10 µM) and vehicle for 24h and whole-cell protein lysates were analyzed for the expression of DR4, DR5 and p53. We observed a dose-dependent increase of DR4 and phospho-p53 ^Ser15 protein level compared with vehicle-treated cells, whereas CPI-7c was found to be most effective in inducing DR5, p53 and phospho-p53 ^Ser392 at 10 µM dose ( Figure 2A and B ). Phospho-p53 ^S392 acts as a transcription factor for many target genes such as DR4, DR5 and MDM2, whereas Phospho-p53 ^S15 is reported to be present in MDM2 unbound form ( 26 ). Further, we assessed transcriptional changes in death receptor genes in response to CPI-7c treatment. Accordingly, we performed quantitative RT–PCR analysis for DR4 and DR5 mRNA expression in CPI-7c (10 µM) treated MCF-7 cells and found that it robustly augmented the mRNA expression of DR4 and DR5 genes compared with vehicle-treated cells ( Figure 2C ). We also validated the surface expression of DR4 and DR5 protein using flowcytometry and observed that CPI-7c not only transcriptionally upregulated the expression of DR4 and DR5 but also promoted their surface expression as observed in FACS histogram overlays ( Figure 2D and E ). As MCF-7 cells are TRAIL resistant ( 47 ), we reasoned that CPI-7c might be effective in inducing death receptors in other TRAIL-resistant cells. We used two TRAIL-resistant cell lines A-549 (lung carcinoma) and SW-620 (colon adenocarcinoma) ( 29 , 48 , 49 ) and noted similar increase in the expression of DR4 and DR5 proteins after CPI-7c treatment as observed by western blot analysis ( Figure 2F , upper and lower panels). To confirm the functional involvement of CPI-7c-induced death receptor upregulation in execution of cancer cell apoptosis, we utilized neutralizing antibodies against DR4 and DR5 to block their downstream signals. Pretreatment of both (DR4 and DR5) neutralizing antibodies together but not alone remarkably rescued CPI-7c-induced apoptosis of cancer cells indicating their pivotal role for CPI-7c’s cytotoxic effects ( Figure 2G ). Next, we assessed the effect of CPI-7c on synthesis and release of DR4 and DR5 ligand TRAIL by using enzyme-linked immunosorbent assay. Here, we took concentrated culture media supernatant and cell lysate of control and treated cells and found that compared with respective controls, there was no significant change in TRAIL level either in supernatant or in cell lysate before and after treatment ( Figure 2H ). Basal expression of death receptors is very low in all these three cell lines, may contribute to their non-responsiveness to TRAIL. Therefore, we restored the expression of death receptors using CPI-7c treatment and assessed whether it resulted in sensitization to TRAIL-induced apoptosis. We performed the functional validation of death receptor upregulation by treating MCF-7, A-549 and SW-620 cells with TRAIL (10 and 50nM), and CPI-7c (5 µM in SW-620/A-549 and 2.5 µM in MCF-7) alone and in combinations for 48h and determined cellular cytotoxicity by SRB assay. Consistent with previous findings ( 22 , 50–52 ), TRAIL marginally induced apoptosis in these cells. However, a synergistic increase in TRAIL-mediated cellular cytotoxicity was noted in all three cell lines when they were pretreated with a lower dose of CPI-7c for 12h followed by TRAIL treatment ( Figure 2I ). These findings indicate that CPI-7c-sensitized TRAIL-resistant cancer cells to apoptosis by TRAIL via restoring the functional expression of death receptors. As XIAP was found to be downregulated after CPI-7c treatment, we overexpressed XIAP in MCF-7 cells by transient transfection and evaluated its contribution in regulating CPI-7c-mediated apoptosis, but we did not observe any significant change in the rescue of apoptosis ( Supplementary Figure 2 , available at Carcinogenesis Online).

Figure 2.

CPI-7c upregulates the expression of DR4 and DR5 in TRAIL-resistant cells and sensitizes them toward TRAIL-induced apoptosis. In ( A ) and ( B ), MCF-7 cells were treated with different (5 and 10 µM) doses of CPI-7c for 24h, and cells were harvested and immunoblotted for DR4, DR5, p53, GAPDH, phospho-p53 ^Ser392 and phospho-p53 ^Ser15 . In ( C ), MCF-7 cells were treated with 10 µM of CPI-7c for 24h and subjected to mRNA isolation for RT–PCR analysis. Respective gel images of DR4, DR5 and GAPDH were shown. In ( D ) and ( E ), MCF-7 cells were treated with 10 µM of CPI-7c for 24h, and flowcytometry was performed to assess the surface expression of DR4 and DR5, which are shown in (D) and (E), respectively. In ( F ), SW-620 and A-549 were treated with different (5 and 10 µM) doses of CPI-7c for 24h, and cells were harvested for protein extraction followed by western blot analysis for the expression of DR4 and DR5. Respective immunoblot images were documented. In ( G ), percentage cell survival as determined by SRB assay in 1 µg of anti-DR4 or/and anti-DR5 pretreated (2h) MCF-7 cells followed by 42h treatment with 2.5 µM of CPI-7c were shown in bar graph. Columns, average of triplicate readings of samples; error bars, ±SD. * P < 0.05, compared with vehicle-treated cells, whereas ^#P < 0.05, compared with only CPI-7c-treated cells. In ( H ), changes in TRAIL synthesis and secretion were determined by enzyme-linked immunosorbent assay after vehicle, 2.5 and 5 µM of CPI-7c treatment in MCF-7 cells and represented in bar graph. Columns, average of triplicate readings of samples; error bars, ±SD. All the data presented from (A) to (H) are representative of at least three independent experiments. Panel ( I ) represents synergistic increase in TRAIL-mediated cytotoxic effects in SW-620, A-549 and MCF-7 cells as observed by SRB assay in response to 12h pretreatment of 5 µM dose of CPI-7c in SW-620/A-549 and 2.5 µM dose in MCF-7 cells followed by treatment with 10 and 50ng/ml of TRAIL for 36h. Results are representative of three independent experiments. Columns, average of triplicate readings of samples; error bars, ±SD. * P < 0.05, compared with only 10ng/ml TRAIL-treated cells, whereas ^#P < 0.05, compared with only 50ng/ml TRAIL-treated cells.

CPI-7c-induced cytotoxicity and upregulation of DR4 and DR5 is linked to the recruitment of activated p53 into their promoters

As CPI-7c activated p53 and the promoters of DR4 and DR5 genes have consensus binding motifs for p53, we examined whether p53 was involved in CPI-7c-induced death receptor upregulation. To test this, we utilized p53-specific siRNA to knockdown p53 in MCF-7 cells. After confirming >80% knockdown efficiency, we treated control and knockdown cells with different doses of CPI-7c for 24h and determined the expression of DR4 and DR5 by western blot analysis. As shown in Figure 3A (left panel), CPI-7c-induced DR4 and DR5 upregulation was markedly impaired in p53 knockdown cells compared with control siRNA transfected cells. Moreover, the cytotoxic effects of CPI-7c on breast cancer cells were decreased after p53 knockdown. These findings indicated the requirement of p53 in regulating the expression of death receptors and the cytotoxic effects of CPI-7c ( Figure 3A , right-hand panel).

Figure 3.

CPI-7c-induced cytotoxicity is partially p53 dependent, and p53 is being recruited in the regulatory sequence of DR4 and DR5 genes to promote their expression. In ( A ), MCF-7 cells were transfected with either control or p53 siRNA (50nM) for 24h followed by the treatment with 5 and 10 µM of CPI-7c for another 24h, and photomicrograph (A, right panel) were taken before harvesting the cells for expression analysis of DR4, DR5 and p53. Immunoblot images were documented in (A) left panel. In panel ( B ), MCF-7 cells were treated with 5 and 10 µM of CPI-7c for 24h, and cells were harvested for chromatin purification followed by immunoprecipitation with either anti-RNA polymerase II or IgG or anti-p53 binding proteins and finally subjected to PCR amplification for p53 binding sites in DR4 and DR5 genes. In ( C ), p53 WT and null HCT-116 cells were treated with 2.5, 5 and 10 µM of CPI-7c for 48h, and cytotoxicity was assessed by SRB assay. Results are representative of three independent experiments. Columns, average of triplicate readings of samples; error bars, ±SD. * P < 0.01. Panels ( D ) and ( E ) represent the western blot analysis of change in p53, MDM2 (D) and DR4/DR5 (E) in p53 WT and null HCT-116 after 24h treatment of 5 and 10 µM of CPI-7c. Right-hand panel of (D) shows pixel intensity graph of change in MDM2.

Next to confirm the functional significance of p53 in expression of DR4 and DR5 genes, we performed the ChIP assays for the recruitment of p53 to its consensus binding site in these genes, in the presence or absence of CPI-7c treatment. Using the PCR-based amplification of p53 recruitment site-specific primers and DNA obtained from p53 pull down chromatin, we noted a significant increase in p53 recruitment to the death receptor genes following CPI-7c treatment relative to vehicle control ( Figure 3B ). Collectively, our findings suggest that p53 plays a major role in CPI-7c-mediated cancer cell cytotoxicity by recruitment to the promoters of DR4 and DR5 to upregulate death receptor expression.

To further confirm the p53 dependency of CPI-7c-mediated effects, we used p53 WT or p53 null HCT-116 colon cancer cells and perform cell viability assay. Here, we found that at lower doses CPI-7c was less effective in p53 null cells but at higher dose, CPI-7c exhibited comparable cytotoxicity in both p53 WT and p53 null cells ( Figure 3C ). These observations indicate that the presence of p53 may be a determining factor underlying the cytotoxic effect of the lower dose of CPI-7c. However, p53 was dispensable at the higher dose of CPI-7c.

Oncoprotein MDM2 is an ubiquitin ligase that binds to p53 via its N-terminal domain and promotes its proteasomal degradation in order to impair its tumor suppressive function ( 1 ). To obtain further insights into the action of CPI-7c, we focused on its effect on MDM2. We performed western blot analysis to determine the expression of p53 and MDM2 in HCT-116-p53-WT and HCT-116-p53-null cells following CPI-7c treatment. We noted a dose-dependent increase of MDM2 and p53 level in p53 WT HCT-116 cells. On the other hand, in p53 null HCT-116 cells, we observed a trend in MDM2 reduction after CPI-7c treatment as compared with vehicle treatment ( Figure 3D ). These results raised the possibility that although CPI-7c is a MDM2 inhibitor, its effect may be masked by the p53-mediated transactivation of MDM2 in p53 WT cells but at the same dose of CPI-7c used in cells lacking p53 function, MDM2 degradation was uncovered. To further confirm p53 dependency of CPI-7c-induced death receptor induction, we again utilized p53 WT and null cells. As shown in Figure 3E , we observed that CPI-7c was more effective in inducing death receptors in p53 WT HCT-116 as compared with p53 null HCT-116 cells.

CPI-7c promotes ubiquitination and degradation of MDM2 by directly binding to its RING and N-terminal domains to stabilize and activate p53

We next compared the effect of CPI-7c with MDM2 inhibitor Nutlin-3. We treated MCF-7 cells with different doses of CPI-7c and Nutlin-3 for 24h and analyzed the levels of MDM2, global ubiquitination and p53 by western blot analysis. At 5 and 10 μM doses, both CPI-7c and Nutlin-3 increased the level of MDM2. However, unlike Nutlin-3, at 15 and 20 µM, CPI-7c reduced the level of MDM2 and increases global ubiquitination level ( Figure 4A ). This indicated that CPI-7c might have significantly increased the ubiquitin-mediated proteasomal degradation of MDM2. On the other hand, p53 expression was consistently elevated. Similarly, confocal studies indicated that CPI-7c promoted MDM2 degradation and p53 stabilization, but Nutlin-3 was unable to degrade MDM2 ( Figure 4B ). To check the transcriptional regulation of MDM2 in response to CPI-7c and Nutlin-3, we performed RT–PCR analysis in MCF-7 cells, treated with vehicle and 10 and 20 µM doses of both CPI-7c and Nutlin-3. Interestingly, in contrast to CPI-7c (20 µM) mediated expression of MDM2 protein downregulation, here, we observed dose-dependent increase in the mRNA expression of MDM2 gene in both the cases of CPI-7c and Nutlin-3 treatment compared with vehicle-treated cells ( Figure 4C ). Altogether, these findings indicate that CPI-7c not only stabilizes and activates p53 but also promotes degradation of MDM2. Thus, CPI-7c functions by a distinct mechanism of action as compared with Nutlin-3, which did not exhibit MDM2 degradation.

Figure 4.

CPI-7c efficiently degrades MDM2 but leads to stabilization of p53 and induces global ubiquitination response. In ( A ), MCF-7 cells were treated with 5, 10, 15 and 20 µM CPI-7c and Nutlin-3, and expression levels of MDM2, ubiquitin and p53 were assessed by using immunoblotting and shown in the image. In panel ( B ), MCF-7 cells were grown in coverslips and treated with either vehicle or 10 and 20 µM of CPI-7c and Nutlin-3, respectively, for 24h and subjected to immunofluorescence staining and analyzed by confocal microscope. Merged confocal photographs represent the superimposition of green (p53) and red (MDM2) images, and the magnified area of the box was shown in inset pictures. Scale bar, 20 µm. Representative of three independent experiments. In panel ( C ), RT–PCR analysis shows the mRNA expression MDM2 and GAPDH genes in response to 10 and 20 µM doses of CPI-7c and Nutlin-3 treatment.

Generally, ubiquitination of MDM2 is promoted by its own catalytic RING domain when it is unbound with MDMX ( 13 , 53 ). On the other hand, the N-terminal domain of MDM2 that offers a binding site for Nutlin-3 and other chalcones is involved in the interaction with p53 ( 9 , 27 , 34 ). On the basis of the above-mentioned findings, we performed in silico binding studies and analyzed the binding interactions of CPI-7c and Nutlin-3 at C-terminal RING domain of MDM2 (2VJF). We observed that CPI-7c interacts with RING domain of MDM2 by making hydrogen bonds with N433, K446 and L458 and non-bonding interactions (including hydrophobic, ionic-π, π–π, aromatic-π and Van der Wall interactions) with surrounding residues ( Figure 5A ). Particularly, pyrido[ b ]indole moiety of CPI-7c forms a hydrogen bond with L458, whereas chalcone part makes hydrogen bond with N433 and K446 ( Figure 5C ). In contrast, Nutlin-3 bound weakly to the C-terminus of MDM2 and did not form a hydrogen bond ( Figure 5B ). Accordingly, CPI-7c may interact at the C-terminus dimer-forming interface of MDM2 (RING domain) that is known to associate with MDMX. CPI-7c aligns with the C-terminus dimer stabilizing residues (L430, A434 and I435) of MDM2 ( Figure 5 ). More importantly, CPI-7c forms non-bonding interactions with residue Y489 that plays a key role in E3 ubiquitin ligase activity of MDM2 ( 8 ).

Figure 5.

In silico binding studies of CPI-7c with C-terminal RING and N-terminal domains of MDM2. Panel ( A ) shows the binding of CPI-7c in the C-terminal RING domain and N-terminal p53 binding groove of MDM2. Here, CPI-7c is represented by pink color stick and MDM2–p53 complex is shown by electrostatic surface representation. Lower panel of (A) shows ribbon model for the possible binding of CPI-7c with the RING and N-terminal domains, where dotted lines represent the hydrogen bonds with respective residues. In ( B ), ribbon model of predicted Nutlin-3 binding in the RING and N-terminal domains is shown. In panel ( C ), table shows the structure of inhibitors, position and number of predicted hydrogen bonds to be formed with MDM2.

We also performed the CPI-7c binding studies with N-terminal domain of MDM2 and found that CPI-7c occupied the hydrophobic binding pocket of MDM2–p53 complex formed by Phe19, Trp23 and Leu26 and that it formed two hydrogen bonds each with His96 and Val93 of N-terminal MDM2. The pyrido[ b ]indole group of CPI-7c bound in the hydrophobic grooves created by Leu54, Leu57, Gly58, Ile99, Tyr100, Ile103 and Tyr104 of MDM2 that were occupied by Phe19 of p53. The chalcone group of CPI-7c interacted with His96 and Val93 through hydrogen bond. The hydrogen bond acceptor group (O=C–) of CPI-7c interacted with hydrogen bond donor group of His96 (HN–) and Val93 (HO–). We also performed similar docking studies between N-terminal sequence of MDM2 (4OBA) and MDM2 inhibitor Nutlin-3. Altogether, these in silico studies revealed that Nutlin-3 aligns in the p53 binding groove of MDM2 but interacts distinctly compared with CPI-7c. Nutlin-3 forms two hydrogen bonds with H96 and non-bonding interactions with surrounding residues. Thus, our molecular modeling studies provided a clear distinction of the binding affinity of Nutlin-3 and CPI-7c to the N-terminal and RING domains of MDM2 and indicated that CPI-7c is a novel inhibitor of MDM2 ( Figure 5 ).

We also carried out pull down experiments using CPI-7c-treated and Nutlin-3-treated MCF-7 cells and determined the ubiquitin-associated MDM2 levels by western blot analysis. Interestingly, we found that ubiquitinated MDM2 level was markedly higher in CPI-7c-treated cells as compared with vehicle-treated cells. There was no change in ubiquitinated MDM2 in Nutlin-3-treated cells as compared with control, although the total MDM2 level was increased in the input control ( Figure 6A ). To confirm this observation, immunoprecipitation of MDM2 was performed using the protein lysates of CPI-7c-treated or Nutlin-3-treated MCF-7 cells, and p53 was analyzed by western blot analysis. CPI-7c and Nutlin-3 increased the p53 level compared with vehicle control, but there was no increase in MDM2-associated p53, indicating that the enhanced level of p53 in the input was dissociated from MDM2 ( Figure 6B ). On the basis of these findings, we sought to address the direct binding of CPI-7c in the RING domain and in its N-terminal domain along with Nutlin-3. Toward this objective, we cloned and purified the N-terminal domain peptide of MDM2 using His-tag purification system. We further procured the catalytically active RING domain peptide of MDM2. Finally, we performed the fluorescence quenching measurement assays separately with N-terminal peptide and with the RING domain peptide by using varying concentration of CPI-7c and Nutlin-3. CPI-7c efficiently bound with the RING and N-terminal domains of MDM2, whereas Nutlin-3 interacted only with the N-terminal domain ( Figure 6C ). Together, our data suggest that CPI-7c is a unique MDM2 inhibitor that has the ability to release p53 from MDM2 sequestration resulting in p53 stabilization and also cause degradation of MDM2 by interacting with its both N-terminal and RING domains.

Figure 6.

CPI-7c physically binds to RING and N-terminal domains of MDM2 and selectively induces ubiquitination of MDM2 as well as interferes at p53–MDM2 interaction. MCF-7 cells were treated with 10 µM dose of CPI-7c and Nutlin-3 for 24h, and protein lysates were prepared for immunoprecipitation using human anti-ubiquitin. Ubiqutinated MDM2 level was measured using western blotting of anti-MDM2 antibody with proper input control and shown in panel ( A ). Similarly, ( B ) represents the level of associated p53 when lysate was immunoprecipitated using anti-MDM2 antibody. Panel ( C ) represents standard scatchard exponential hyperbolic binding curve of direct physical binding of CPI-7c and Nutlin-3 with N-terminal peptide and RING domain peptide of MDM2 in left and right panels, respectively, along with respective dissociation equilibrium constant ( K_d ) values.

Discussion

Multiple reports have indicated that the overexpression of MDM2 and resultant loss of p53 function is directly associated with tumor development and that reactivation of p53 by dampening MDM2 is a promising approach for cancer therapeutics ( 9 , 54 , 55 ). There is increasing interest in exploring the potential small-molecule inhibitors of MDM2 as novel anticancer agents ( 11 ). In our pursuit to develop natural compound-inspired anticancer agents, we discovered a series of chalcone-based pyrido[ b ]indoles that were found to be potently active against breast cancer cells ( 37 ). The present study provides mechanistic insights into CPI-7c, the most potent anticancer molecule of the chalcone-based pyrido[ b ]indole series. CPI-7c directly inhibited MDM2 oncoprotein by a novel mechanism of action and sensitized TRAIL-resistant tumor cells to apoptosis by regulating the MDM2–p53–DR4/DR5 pathway. Unlike earlier reports speculating that chalcones may interfere with p53–MDM2 interactions ( 27 , 28 ), our studies provide evidence that CPI-7c acts as a direct MDM2 inhibitor. Though not tested so far but there would be a good possibility of our molecule to be pharmacokinetically stable and in vivo active as the pharmacophore components of CPI-7c such as chalcone and pyridoindoles are known to be safe, stable and in vivo active ( 10 , 56 ).

MDM2 inhibitors described so far comprise peptide inhibitors ( 55 ) and small molecules such as Nutlin-3 ( 9 ) and RITA ( 57 ). These inhibitors act through the N-terminal domain of MDM2 and activate p53. MDM2 E3 ligase inhibitors, such as the HLI series of compounds, have been identified by in vitro screens, but they lack specificity toward MDM2 ( 58 ). There are also reports about MDM2 inhibitors of spiro-oxindole series, such as MI-319 and MI-219, which partially ubiquitinate MDM2 in addition to activating p53 ( 59 , 60 ). Other MDM2 E3 ligase inhibitors revealed by in vitro ubiquitination assays, such as sempervirine and lissochlinidine B, do not exert any effect on the N-terminal domain ( 61 , 62 ). More recently, molecules containing pyrido[ b ]indole or β-carboline moieties such as SP-141 ( 10 ) and MEL23 and MEL24 ( 36 ) were discovered as unique class of MDM2 inhibitors having potent E3 ligase diminishing activity. Interestingly, none of these inhibitors have dual MDM2 inhibitory activities such as CPI-7c, which has two-prong mechanism of action for preventing p53–MDM2 interactions as well as MDM2 degradation by promoting its ubiquitination. At first glance, this dual mechanism of action of CPI-7c seems consistent with its structure, as CPI-7c contains chalcone and pyrido[ b ]indole moieties that exhibit interaction with the N-terminal and RING domains of MDM2, respectively. On the other hand, our in silico studies indicated that the RING domain of MDM2 is involved in forming hydrogen bonds with both chalcone and pyrido[ b ]indole moieties.

The expected outcome of the MDM2 inhibitors is p53 stabilization and activation of downstream target genes that may lead to cancer cell apoptosis. Unfortunately, a major limitation of this process is p53-mediated MDM2 transactivation. Consistent with previous reports ( 36 , 63 ), we observed robust upregulation of MDM2 upon treatment even with high doses of Nutlin-3. Although p53 is effectively stabilized in response to Nutlin-3, p53-independent MDM2 oncogenic effects may be a major shortcoming. Another important aspect is the complete or partial p53 independence of human malignancies as it has been found to be mutated or absent in many solid tumors ( 64 ). Keeping this limitation in mind especially in case of p53–MDM2 dissociation inhibitors, lot of effort has been recently focused in generating MDM2 E3 ligase inhibitors that can function in the absence of p53 ( 10 , 36 ); however, their efficacy in WT p53 situation is questionable because of p53-mediated transactivation of MDM2. Our immunoprecipitation and direct binding assay experiments strongly suggest that our CPI-7c is unique as it has functions as a double-edged sword to work effectively in both p53-dependent and p53-independent conditions.

We discovered that the most important outcome of CPI-7c-mediated activation of p53 is the ability to surmount TRAIL resistance in cancer cells by restoring the expression of TRAIL receptors DR4 and DR5 on cell surface. TRAIL holds promise as a potential frontline therapeutic in the tumors harboring WT p53 function ( 20 , 21 , 65 ). However, a paucity of death receptors often significantly contributes to TRAIL resistance in tumors ( 22–24 ), as exemplified by our observation in three different cancer cell lines (MCF-7, A-549 and SW-620) in our study. Robust activation of p53 and its subsequent recruitment to the promoters of death receptors followed by their functional activation by CPI-7c make it an ideal candidate for including in combination with TRAIL for cancer therapy. Death receptor-mediated CPI-7c’s mode of action suggests that even low amount of TRAIL can pose cytotoxic effects in the presence of augmented death receptors on cell surface. Collectively, this study demonstrated that targeting the MDM2–p53 interactions as well as inducing MDM2 degradation with a single small molecule is feasible. To the best of our knowledge, this is the first report regarding the dual targeting of MDM2 using small-molecule inhibitor. Further studies on the optimization of CPI-7c or its pharmacophore are underway.

Supplementary material

Supplementary Figures 1 and 2 can be found Supplementary Data

Funding

CSIR BSC0106 (to D.D., A.S.); CSIR fellowship grants (to A.K.S., S.S.C., S.M.); UGC fellowship grants (to R.K.A.).

Author contributions

A.K.S., A.S., R.K.A. and S.M. performed most of the experiments. S.S.C. and P.M.C. synthesized the molecule. V.V.V. carried out bioinformatics. S.K.S., A.K.S., M.S.A. and J.S. involved in binding studies. V.M.R. provided intellectual input in study designing and edited the manuscript. A.K.S. and D.D. wrote the manuscript. D.D. involved in study design, data interpretation, writing of the manuscript and overall supervision. All authors read and approved the final manuscript.

Conflict of Interest Statement : None declared.

Glossary

Abbreviations

ChIP

chromatin immunoprecipitation

HRP

horseradish peroxidase

MDM2

mouse double minute 2

SRB

sulforhodamine-B

TRAIL

TNF-related apoptosis-inducing ligand

WT

wild type