IPP51, a chalcone acting as a microtubule inhibitor with in vivo antitumor activity against bladder carcinoma.

We previously identified 1-(2,4-dimethoxyphenyl)-3-(1-methylindolyl) propenone (IPP51), a new chalcone derivative that is capable of inducing prometaphase arrest and subsequent apoptosis of bladder cancer cells. Here, we demonstrate that IPP51 selectively inhibits proliferation of tumor-derived cells versus normal non-tumor cells. IPP51 interfered with spindle formation and mitotic chromosome alignment. Accumulation of cyclin B1 and mitotic checkpoint proteins Bub1 and BubR1 on chromosomes in IPP51 treated cells indicated the activation of spindle-assembly checkpoint, which is consistent with the mitotic arrest. The antimitotic actions of other chalcones are often associated with microtubule disruption. Indeed, IPP51 inhibited tubulin polymerization in an in vitro assay with purified tubulin. In cells, IPP51 induced an increase in soluble tubulin. Furthermore, IPP51 inhibited in vitro capillary-like tube formation by endothelial cells, indicating that it has anti-angiogenic activity. Molecular docking showed that the indol group of IPP51 can be accommodated in the colchicine binding site of tubulin. This characteristic was confirmed by an in vitro competition assay demonstrating that IPP51 can compete for colchicine binding to soluble tubulin. Finally, in a human bladder xenograft mouse model, IPP51 inhibited tumor growth without signs of toxicity. Altogether, these findings suggest that IPP51 is an attractive new microtubule-targeting agent with potential chemotherapeutic value.


Live cell imaging and video microscopy.
For time-lapse microscopy, HeLa cells expressing GFP-tubulin were plated in glass-bottom dishes then placed inside a video microscopy platform equipped with an incubator enabling the regulation of the temperature. Just before recording, the cell growth medium was replaced with the CO 2 independent growth medium DMEM/F12 with15 mM HEPES (Invitrogen Life Technologies) supplemented with 10% serum with or without 10 µM IPP51. Time-lapse Z series images (Z = 3) were collected with an inverted Olympus IX81 epifluorescence motorized microscope equipped with an environmental control chamber. The microscope was equipped with a motorized piezo stage (Ludl Electronic Products, Hawthorne, USA) and a Retiga-SRV CCD camera (QImaging, Surrey, Canada) driven by VOLOCITY software (Improvision, PerkinElmer, Waltham, USA) with a binning of 1, using a PlanApo 60xNA 1.42 objective (Olympus, Rungis, France). Images were acquired every 5 min and the acquisition time was 53 ms. For each Z series, the best focus images was chosen before the reconstitution of the movie.

Molecular modeling.
The crystallographic structure of tubulin from Bos Taurus (99% identity with the human tubulin) in complex with DAMA colchicine was retrieved from the Protein Data Bank (PDB ID: 1SA0) (1) . The receptor was prepared by using the Biopolymer tool of Sybyl-X2.0 (Tripos Inc., St. Louis, MO, USA): co-crystallized ligands and water molecules were extracted, missing side chain residues and hydrogen atoms were added, Tripos atom and bond types were assigned to both protein and extracted ligand (2) . The rigid binding site was defined by a 6 Å radius sphere around the co-crystallized ligand. 100 docking solutions were generated by using 100'000 GOLD Genetic Algorithm iterations (Preset option), ranked according to the GOLD score. Root-mean-square-deviation RMSD values between the docking solutions and the crystallized ligand of reference were calculated. Moreover, a RMSD clustering analysis was performed. By following the same approach, IPP51 was docked to the protein by applying the methodology described above. The best ranked pose in complex with tubulin was then used for carrying out 1ns Molecular Dynamics (MD) simulation with Amber 12 (3) by using Gaff and Amberff99SB force fields for the ligand and the protein, respectively. The partial charges of ligand were assigned with the AM1-BCC method, compatible with the Amber force field. The complex was then embedded in an octahedral box of 10 Å TIP3P water molecules neutralized by Na + ions and minimized along 1000 steps with restraints on the solute using the pmemd module of Amber 12. Complexes were then free to relax along further 2500 minimization steps. 100 ps of heating at constant volume (from 10 K to 300 K) with weak restraints on the solute (10.0 kcal/mol/Å) was applied, followed by 750 ps MD at constant pressure conditions (1 atm) reaching an equilibrium state. Trajectories were then collected each 2 ps along 1 ns of MD production phase. The Particle-Mesh-Ewald algorithm (PME) with a cut-off of 8 Å was used to treat the longrange electrostatic effects (4) . The SHAKE algorithm was also used to constrain the bonds connecting hydrogen atoms (5) . A MM-GBSA single trajectory approach was then performed on 1000 snapshots extracted every 1 ps for free energies of binding (ΔG MM-GBSA ) estimation (3) .

Generation of stable luciferase expressing RT112 cells.
A vector containing the luciferase reporter sequence was generated by subcloning an insert BamHI/HindIII excised from pGL3 basic vector (Promega France, Charbonnières, France) into the same restriction sites of pcDNA3+ (Life Technologies). Recombinant plasmid was transfected into RT112 cells using Trans-LT1 reagent (Mirus, Madison, USA). Cells were incubated overnight, then washed twice with PBS and grown for another 24 h in RPMI + 10% SVF. Cell clones stably expressing luciferase were selected with 200 µg/ml of G418. Resistant clones were amplified and we have selected the clone showing the highest luminescence signal when incubated with luciferase substrate.

Bioluminescence in vivo imaging.
All imaging was performed under inhalational anaesthesia (3% isoflurane) and administered to a free breathing mouse using a nose cone.
For bioluminescence imaging, mice received an intraperitoneal injection of D-luciferin potassium salt dissolved in sterile phosphate-buffered serum (150 mg/kg) 5 min before imaging (ORCAII-BT-512G, Hamamatsu Photonics, Massy, France), as described previously by Jin et Al. Semi-quantitative data were obtained from the bioluminescence images by drawing regions of interest on the area to be quantified. Images were acquired as 16-bit TIFF files, which can provide a dynamic of up to 65,535 greys levels. Measurement of the bioluminescence intensities, expressed as the number of relative light units (RLU) per pixel