Novel Set of Diarylmethanes to Target Colorectal Cancer: Synthesis, In Vitro and In Silico Studies

Distinctive structural, chemical, and physical properties make the diarylmethane scaffold an essential constituent of many active biomolecules nowadays used in pharmaceutical, agrochemical, and material sciences. In this work, 33 novel diarylmethane molecules aiming to target colorectal cancer were designed. Two series of functionalized olefinic and aryloxy diarylmethanes were synthesized and chemically characterized. The synthetic strategy of olefinic diarylmethanes involved a McMurry cross-coupling reaction as key step and the synthesis of aryloxy diarylmethanes included an O-arylation step. A preliminarily screening in human colorectal cancer cells (HT-29 and HCT116) and murine primary fibroblasts (L929) allowed the selection, for more detailed analyses, of the three best candidates (10a, 10b and 12a) based on their high inhibition of cancer cell proliferation and non-toxic effects on murine fibroblasts (<100 µM). The anticancer potential of these diarylmethane compounds was then assessed using apoptotic (phospho-p38) and anti-apoptotic (phospho-ERK, phospho-Akt) cell survival signaling pathways, by analyzing the DNA fragmentation capacity, and through the caspase-3 and PARP cleavage pro-apoptotic markers. Compound 12a (2-(1-(4-methoxyphenyl)-2-(4-(trifluoromethyl)phenyl) vinyl) pyridine, Z isomer) was found to be the most active molecule. The binding mode to five biological targets (i.e., AKT, ERK-1 and ERK-2, PARP, and caspase-3) was explored using molecular modeling, and AKT was identified as the most interesting target. Finally, compounds 10a, 10b and 12a were predicted to have appropriate drug-likeness and good Absorption, Distribution, Metabolism and Excretion (ADME) profiles.

Colorectal cancer (CRC) is the third most common cancer-related death after prostate and lung cancer in western societies, with more than half a million annual deaths worldwide. The incidence is high and is steadily increasing. In France, CRC is the second leading cause of death (17,117 deaths/year) [26][27][28]. The current therapies used in the clinic of CRC still have several side effects and resistance issues. Therefore, there is an urgent need to find novel therapeutical alternatives to reduce these drawbacks.
We described previously the synthesis of ferrocenyl DAM using McMurry cross coupling and the antibacterial activity evaluation [2,29]. Inspired by this study, and by the evidence that endows that DAM are well described for their anti-cancer activities, we were interested in the design and synthesis of new pyridyl diarylmethanes-type to develop novel therapeutic agents to target CRC.
Hence, the pyridyl DAM olefinic and their oxygenated analogues were synthesized, and their antiproliferative activity was assayed in vitro on human CRC cell lines. The potential cytotoxicity of the synthesized compounds was evaluated on normal murine fibroblast. Several target receptors involved in the inhibition of cancer cell proliferation were analyzed in vitro to suggest the likely mechanism of action of the DAM drug candidates. Moreover, we complemented our endeavors with a docking study performed on some biological targets that may potentially be involved in the biological activity. Finally, an in silico prediction of biopharmaceutical properties allowed us to investigate the Absorption, Distribution, Metabolism, and Excretion (ADME) profile and druglikeness of the most promising anticancer DAM.

Chemistry
All reagents were obtained from commercial sources unless otherwise noted and used as received. Heated experiments were conducted using thermostatically controlled heating mantles and were performed under an atmosphere oxygen-free in oven-dried glassware when necessary. The reactions were monitored by analytical Thin Layer Chromatography (TLC). TLC was performed on aluminum sheets precoated silica gel plates (60 F254, Merck). TLC plates were visualized using irradiation with light at 254 nm. Flash column chromatography (FCC) was carried out when necessary, using silica gel 60 (particle size 0.040-0.063 mm, Merck). A mixture of cyclohexane (CyHex) and ethyl acetate (EtOAc) was used as mobile phase.

Physical Measurements
Melting points were recorded on a Kofler hot block Heizbank type 7841 and were uncorrected. The structures of the products were checked by comparison of Nuclear Magnetic Resonance (NMR), Infrared (IR) and Mass spectrometry (MS) data and by the TLC Colorectal cancer (CRC) is the third most common cancer-related death after prostate and lung cancer in western societies, with more than half a million annual deaths worldwide. The incidence is high and is steadily increasing. In France, CRC is the second leading cause of death (17,117 deaths/year) [26,27]. The current therapies used in the clinic of CRC still have several side effects and resistance issues. Therefore, there is an urgent need to find novel therapeutical alternatives to reduce these drawbacks.
We described previously the synthesis of ferrocenyl DAM using McMurry cross coupling and the antibacterial activity evaluation [2,28]. Inspired by this study, and by the evidence that endows that DAM are well described for their anti-cancer activities, we were interested in the design and synthesis of new pyridyl diarylmethanes-type to develop novel therapeutic agents to target CRC.
Hence, the pyridyl DAM olefinic and their oxygenated analogues were synthesized, and their antiproliferative activity was assayed in vitro on human CRC cell lines. The potential cytotoxicity of the synthesized compounds was evaluated on normal murine fibroblast. Several target receptors involved in the inhibition of cancer cell proliferation were analyzed in vitro to suggest the likely mechanism of action of the DAM drug candidates. Moreover, we complemented our endeavors with a docking study performed on some biological targets that may potentially be involved in the biological activity. Finally, an in silico prediction of biopharmaceutical properties allowed us to investigate the Absorption, Distribution, Metabolism, and Excretion (ADME) profile and druglikeness of the most promising anticancer DAM.

Chemistry
All reagents were obtained from commercial sources unless otherwise noted and used as received. Heated experiments were conducted using thermostatically controlled heating mantles and were performed under an atmosphere oxygen-free in oven-dried glassware when necessary. The reactions were monitored by analytical Thin Layer Chromatography (TLC). TLC was performed on aluminum sheets precoated silica gel plates (60 F254, Merck). TLC plates were visualized using irradiation with light at 254 nm. Flash column chromatography (FCC) was carried out when necessary, using silica gel 60 (particle size 0.040-0.063 mm, Merck). A mixture of cyclohexane (CyHex) and ethyl acetate (EtOAc) was used as mobile phase.

Physical Measurements
Melting points were recorded on a Kofler hot block Heizbank type 7841 and were uncorrected. The structures of the products were checked by comparison of Nuclear Magnetic Resonance (NMR), Infrared (IR) and Mass spectrometry (MS) data and by the TLC behavior. 1 H-and 13 C-NMR spectra were recorded on a Bruker BioSpin GmbH spectrometer 400 MHz, at room temperature. Chemical shifts are reported in δ units, parts per million (ppm). Coupling constants (J) are measured in hertz (Hz). Splitting patterns are designed as follows: s, singlet; d, doublet; dd, doublet of doublets; m, multiplet; t, triplet; td, triplet of doublet; ddd: doublet of doublet of doublet. Distortionless enhancement with polarization transfer (DEPT) experiments and various 2D techniques such as COr-relation SpectroscopY (COSY), Heteronuclear Single Quantum Coherence (HSQC) and Heteronuclear Multiple Bond Correlation (HMBC) were used to establish the structures and to assign the signals. Conventional adopted to assign signal of 1 H-and 13 C-NMR spectra are described in Figure 2. Gas chromatography-mass spectrometry (GC-MS) analysis was performed with an Agilent 689 0N instrument equipped with a dimethyl polysiloxane capillary column (12 m × 0.20 mm) and an Agilent 5973N MS detector-column temperature gradient 80-300 • C (method 80): 80 • C (1 min); 80 • C to 300 • C (12.05 • C/min); 300 • C (2 min). Electrospray ionization (ESI)-Low resolution mass spectra (LRMS) were performed from ionization by electrospray on a Waters Micromass ZQ2000. Infrared spectra were recorded over the 400-4000 cm −1 range with an Agilent Technologies Cary 630 Fourier-transform infrared spectroscopy (FTIR)/Attenuated Total Reflectance (ATR)/ ZnSe spectrometer. High-resolution mass spectra (HRMS) analyses were acquired on a Thermo Scientific LTQ Orbitrap mass spectrometer.
behavior. 1 H-and 13 C-NMR spectra were recorded on a Bruker BioSpin GmbH spectrometer 400 MHz, at room temperature. Chemical shifts are reported in δ units, parts per million (ppm). Coupling constants (J) are measured in hertz (Hz). Splitting patterns are designed as follows: s, singlet; d, doublet; dd, doublet of doublets; m, multiplet; t, triplet; td, triplet of doublet; ddd: doublet of doublet of doublet. Distortionless enhancement with polarization transfer (DEPT) experiments and various 2D techniques such as COrrelation SpectroscopY (COSY), Heteronuclear Single Quantum Coherence (HSQC) and Heteronuclear Multiple Bond Correlation (HMBC) were used to establish the structures and to assign the signals. Conventional adopted to assign signal of 1 H-and 13 C-NMR spectra are described in Figure 2. Gas chromatography-mass spectrometry (GC-MS) analysis was performed with an Agilent 689 0N instrument equipped with a dimethyl polysiloxane capillary column (12 m × 0.20 mm) and an Agilent 5973N MS detector-column temperature gradient 80-300 °C (method 80): 80 °C (1 min); 80 °C to 300 °C (12.05 °C/min); 300 °C (2 min). Electrospray ionization (ESI)-Low resolution mass spectra (LRMS) were performed from ionization by electrospray on a Waters Micromass ZQ2000. Infrared spectra were recorded over the 400-4000 cm −1 range with an Agilent Technologies Cary 630 Fouriertransform infrared spectroscopy (FTIR)/Attenuated Total Reflectance (ATR)/ZnSe spectrometer. High-resolution mass spectra (HRMS) analyses were acquired on a Thermo Scientific LTQ Orbitrap mass spectrometer. Figure 2. Convention adopted to assign signals of 1 H-and 13 C-NMR spectra. Only the chemical characterization of final DAM obtained are described herein. Intermediates are detailed as SI-1.

General Procedure for the Synthesis of Olefinic Diarylmethanes
To a suspension of zinc (2.86 mmol, 182 mg, 6 eq) in anhydrous tetrahydrofuran (THF) (2.6 mL) under argon was added dropwise TiCl4 (1.872 mmol, 0.2 mL, 4 eq). The reaction mixture was stirred for 2 h at 85 °C. A solution of and (4-methoxyphenyl) (pyridin-2-yl)methanone 5 (0.468 mmol, 100 mg, 1 eq) and the corresponding aromatic aldehyde 2 (0.486 mmol, 1.04 eq) in THF (1 mL) was then added dropwise via a syringe. After reaction completion, the mixture was cooled at room temperature and then poured into the water and extracted with dichloromethane (DCM). The combined organic extracts were dried over anhydrous MgSO4, filtered and concentrated. The crude was purified by FCC on silica gel.
Z  . Convention adopted to assign signals of 1 H-and 13 C-NMR spectra. Only the chemical characterization of final DAM obtained are described herein. Intermediates are detailed as SI-1.

General Procedure for the Synthesis of Olefinic Diarylmethanes
To a suspension of zinc (2.86 mmol, 182 mg, 6 eq) in anhydrous tetrahydrofuran (THF) (2.6 mL) under argon was added dropwise TiCl 4 (1.872 mmol, 0.2 mL, 4 eq). The reaction mixture was stirred for 2 h at 85 • C. A solution of and (4-methoxyphenyl) (pyridin-2-yl)methanone 5 (0.468 mmol, 100 mg, 1 eq) and the corresponding aromatic aldehyde 2 (0.486 mmol, 1.04 eq) in THF (1 mL) was then added dropwise via a syringe. After reaction completion, the mixture was cooled at room temperature and then poured into the water and extracted with dichloromethane (DCM). The combined organic extracts were dried over anhydrous MgSO 4 , filtered and concentrated. The crude was purified by FCC on silica gel.

General Procedure for the Synthesis of aryloxyDAM
To a suspension of molecular sieves (480 mg) in anhydrous dichloromethane (4.8 mL) under argon was added carbinol (0.46 mmol, 100 mg), arylboronic acid (1.38 mmol, 3 eq), Cu(OAc) 2 (0.46 mmol, 84.5 mg, 1 eq) and anhydrous pyridine (0.92 mmol, 0.074 mL, 2 eq). The reaction mixture was refluxed during 24 h at 40 • C under argon. After reaction completion, the mixture was cooled at room temperature, filtered under celite. To recover the product, the celite was washed using dichloromethane and ethyl acetate. The filtrate  Human CRC HCT116 and HT-29 adherent cell lines were purchased from the American Type Culture Collection (ATCC-LGC Standards, Molsheim, France). We chose these human CRC cell lines because they are of different stages in order to evaluate possible resistance to our treatments: the HCT116 CRC line was isolated from a stage I colorectal carcinoma of an adult male. The HT-29 CRC line was derived from a stage II colorectal adenocarcinoma from a 44-year-old woman.
Cells were grown in DMEM medium for HT-29 cells and RPMI 1640 medium for HCT116 cells, supplemented with 10% FBS, 1% L-glutamine and 100 U/mL penicillin and 100 µg/mL streptomycin. Cultures were maintained in a humidified atmosphere containing 5% CO 2 at 37 • C. Stock solutions of each compound were used at 10 −2 M in DMSO and then diluted in culture medium to obtain the appropriate final concentrations. The same amount of vehicle (percentage of DMSO did not exceed 0.5%) was added to control cells. L929 cell line is a non-cancer cell line derived from L-strain (L cells) and has been grown in the Peptinov laboratory for years. L929 cells are murine adherent fibroblasts from subcutaneous connective tissue (areolar and adipose tissues). Cells were grown in DMEM medium supplemented with 10% FBS, 1% L-glutamine and 100 U/mL penicillin and 100 µg/mL streptomycin and maintained in a humidified incubator at 37 • C, 5% CO 2 . When confluence is at 80%, cells are trypsinyzed for 3 min and diluted in fresh medium. Each batch of cells is kept for 15 passages before being discarded and a new batch thawed.

Cell Metabolic Activity
All compounds were tested on the metabolic activity of the cells using the MTT colorimetric assay [29]. Briefly, cells were seeded in 96-well microplates at 8 × 10 3 cells/well for human CRC HT-29 cells and 5 × 10 3 cells/well for human CRC HCT116 cells and grown for 24 h in appropriate culture medium prior to exposure or not to compounds (6-27) with concentration ranges from 1 to 50 µM. After 48 h of treatment, MTT (5 g/L in Phosphate-buffered saline (PBS)) was added and incubated for another 3 h. The MTT was then removed from the wells and 100 µL/well of DMSO were added to dissolve formazan. The optical density was detected with a microplate reader (Thermoscientific, Multiskan FC) at 550 nm and cell viability was expressed as a percentage of each treatment condition compared to control cells. IC 50 values were calculated for all compounds from the dose-response curve.
Cell viability of L929 cell line was evaluated in presence of synthesized compounds on. Briefly, cells were trypsinyzed and seeded in 96-well microplates at 4 × 10 4 cells/well for 24 h in a humidified incubator at 37 • C, 5% CO 2 . The following day, 100 µL of a mix containing serial diluted compounds, in constant 0.5% DMSO, were added to the cells and plates were incubated for 24 h in a humidified incubator at 37 • C, 5% CO 2 . 100 µL of a mix containing 0.5% of DMSO was added to untreated cells (control cells). After removing the mix, 100 µL of 0.5 mg/mL of MTT were added to the cells and plates were incubated for 2 h in an humidified incubator at 37 • C, 5% CO 2 . MTT was then removed and 200 µL of DMSO were added to the wells to dissolve formazan crystals. Optical density was then read with a spectrophotometer (Multiskan, Fisher Scientific, Illkirch, France) at 560 nm. Cell viability was expressed as percentages compared to untreated cells.

Protein Extraction and Western Blot Analysis
Human CRC HT-29 and HCT116 cells were treated or not with the determined IC 50 values of compounds (12a, 10a, 10b) for indicated times (6, 12, 24 and 48h) and then harvested with trypsin. For total protein extraction, collected samples of each condition were washed with PBS. Then, the total cell pool was centrifuged at 200× g for 5 min at 4 • C and homogenized in RIPA lysis buffer (50 mM HEPES, pH 7.5, 150 mM NaCl, 1% sodium deoxycholate, 1% NP-40, 0.1% Sodium Dodecyl Sulfate (SDS), 20 mg/mL of aprotinin) containing protease inhibitors according to the manufacturer's instructions as previously described [30]. Proteins (60 µg) were separated on 12.5% SDS-PAGE gels and transferred to polyvinylidene difluoride (PVDF) membranes. Membranes were probed with respective human antibodies against caspase-3, cleaved caspase-3, PARP and Akt, ERK, p38 MAP Kinases and its phosphorylated forms according to the manufacturer's instructions. After incubation with appropriate secondary antibodies, blots were developed using the «Immobilon Western» substrate following the manufacturer's protocol and G:BOX system (Syngene, Ozyme). Membranes were then reblotted with human anti-β-actin used as a loading control.

Apoptosis Quantification by DNA Fragmentation Analysis
Human CRC HT-29 and HCT116 cells were treated or not with the determined IC 50 values of compounds (12a, 10a, 10b) for 24 and 48 h and then harvested with trypsin. Histone release from the nucleus during apoptosis was analyzed using the Cell Death Detection ELISA PLUS as previously described [6]. Cytosol extracts from 10 5 cells of each condition were obtained and DNA fragmentation was measured according to the manufacturer's protocol. Results were reported as n-fold compared to control cells.

Statistical Analysis
Data are expressed as the arithmetic means ± standard error of the mean (SEM) of at least three separate experiments. Statistical significance was evaluated by the two-tailed unpaired Student's t-test and expressed as: * p < 0.05; ** p < 0.01 and *** p < 0.001.

Protein and Compounds Preparation
The protein structures of the five potential targets were extracted from the Protein Data Bank (PDB): [31] AKT (PDB ID: 6S9W), ERK-1 (PDB ID: 4QTB), ERK-2 (PDB ID: 6SLG), PARP (PDB ID: 4ZZZ) and caspase-3 (PDB ID: 6CKZ). The structure of ERK-2 was superimposed on the structure of ERK-1 to allow comparison of the binding modes obtained on these two isoforms. All the structures were prepared using the DockPrep tool from UCSF Chimera [32] and MGL tools [33].
Three dimensional structures of compounds 12a and 10a were generated using iCon [34], the LigandScout v.4.3. conformer generator (defaults settings of the BEST option were used, except for the maximum number of conformations that was setted to 2000). Compounds were protonated at pH 7.4 using the cxcalc plugin of the Chemaxon suite [35] and converted in .pdbqt format.

Docking Study and Protein-Ligand Interactions Analysis
For all the five target, the docking study was conducted using smina and the vinardo scoring function [36]. For each target, a search space was defined with a size of 20 Å × 20 Å × 20 Å and the following x, y, z grid center coordinates: −12.729, −15.248, 13.193  The predicted binding mode of each ligand in each binding site was analyzed using the Protein-Ligand Interaction (PLIP) webserver.

ADME Profile and Drug-Likeness
Predictions of ADME properties and drug-likeness for the most promising compounds were conducted in SCHRÖDINGER Maestro v11.9 software, using the Molecular Properties Panel from QikProp v5.9 platform [37]. Structure Minimization was performed using Force field OPLS3e, [38] the Polak-Ribier method of Conjugate Gradient (PRCG), with a convergence threshold of 0.05 and a maximum of 2500 iterations.

Chemistry
The general synthetic pathway followed to synthesize two series of functionalized DAM: olefinic and aryloxy is shown in Scheme 1. Details about the synthetic protocol and chemical characterization of all intermediates are given in the Supplementary Information (SI-1).

Chemistry
The general synthetic pathway followed to synthesize two series of functionalized DAM: olefinic and aryloxy is shown in Scheme 1. Details about the synthetic protocol and chemical characterization of all intermediates are given in the Supplementary Information (SI-1). First, the synthesis of carbinol 3 was performed from bromopyridine 1 and anisaldehyde 2 by a bromine-magnesium exchange using isopropylmagnesium chloride in tetrahydrofuran at room temperature [40]. The reaction was also performed by a brominelithium exchange following the procedure of Seto et al. [41]. Nevertheless, this latter protocol led to only 22% of the corresponding carbinol and to a by-product 4 not previously described in the literature. The corresponding arylketone 5 was synthesized from carbinol 3 in excellent yield (98%) via a base-promoted aerobic oxidation using air as a free and clean oxidant [42].
The key step to obtain the desired olefinic diarylmethanes involved a McMurry crosscoupling reaction between the aryllketone 5 and the corresponding para-substituted aromatic aldehyde in presence of TiCl4/Zn in THF. After the in situ formation of the catalytic First, the synthesis of carbinol 3 was performed from bromopyridine 1 and anisaldehyde 2 by a bromine-magnesium exchange using isopropylmagnesium chloride in tetrahydrofuran at room temperature [39]. The reaction was also performed by a bromine-lithium exchange following the procedure of Seto et al. [40]. Nevertheless, this latter protocol led to only 22% of the corresponding carbinol and to a by-product 4 not previously described in the literature. The corresponding arylketone 5 was synthesized from carbinol 3 in excellent yield (98%) via a base-promoted aerobic oxidation using air as a free and clean oxidant [41].
The key step to obtain the desired olefinic diarylmethanes involved a McMurry crosscoupling reaction between the aryllketone 5 and the corresponding para-substituted aromatic aldehyde in presence of TiCl 4/ Zn in THF. After the in situ formation of the catalytic entity (TiCl 2 ) by the reduction of TiCl 4 using Zn in THF at reflux for 2 h, an equimolar mixture of the arylketone 5 and the benzaldehyde 2 in THF was added dropwise to the reaction medium. Depending on the aromatic aldehyde used, containing electron-withdrawing, electron-donating and halogen groups, the reaction time varies between 10 and 30 min after the addition of reagents. The expected DAM were obtained in two separable E and Z isomers except for compound 11 (Scheme 1). No selectivity was observed in the formation of major isomers. The Z and E isomers were easily distinguished by NMR analysis. Thus, it is noted that the chemical shift of the olefinic proton in NMR for E isomer is higher than for Z isomer (7.7 ppm for E isomer versus 7.0 ppm for Z isomer).
The introduction of the N-oxide moiety has been highly considered in medicinal chemistry programs. Indeed, N-oxides provide interesting physicochemical properties such as improved solubility and the capability to increase affinity with receptor sites [28]. Considering these arguments and based on biological results (Section 2.2), the two Z isomers of the olefinic DAM 10a and 12a as well as the two aryloxyDAM 17 and 22, were N-oxidized using m-chloroperbenzoic acid (m-CPBA) in dichloromethane at room temperature.
For the olefinic compounds 10a and 12a, the oxidation was performed using only 1 equivalent of m-CPBA in order to avoid undesirable oxidation of the double bond. However, for the aryloxyDAM 17 and 22, as these derivatives do not present any other oxidizable site, 5 equivalents of m-CPBA were used. The olefinic oxidized compounds were obtained in yields of 45% and 50%, respectively. Furthermore, the corresponding pyridine N-oxides of aryloxyDAM are obtained in 99% and 88% yields (Scheme 2).
ysis. Thus, it is noted that the chemical shift of the olefinic proton in NMR for E is higher than for Z isomer (7.7 ppm for E isomer versus 7.0 ppm for Z isomer).
The introduction of the N-oxide moiety has been highly considered in m chemistry programs. Indeed, N-oxides provide interesting physicochemical pr such as improved solubility and the capability to increase affinity with receptor si Considering these arguments and based on biological results (Section 2.2), the tw mers of the olefinic DAM 10a and 12a as well as the two aryloxyDAM 17 and 22, oxidized using m-chloroperbenzoic acid (m-CPBA) in dichloromethane at room te ture.
For the olefinic compounds 10a and 12a, the oxidation was performed using equivalent of m-CPBA in order to avoid undesirable oxidation of the double bond ever, for the aryloxyDAM 17 and 22, as these derivatives do not present any other able site, 5 equivalents of m-CPBA were used. The olefinic oxidized compounds w tained in yields of 45% and 50%, respectively. Furthermore, the corresponding p N-oxides of aryloxyDAM are obtained in 99% and 88% yields (Scheme 2).

Cell Viability, Cell Proliferation Inhibition, and IC50 Determination
First, cell viability was evaluated using MTT assay. Eighteen olefinic DA eleven aryloxyDAM previously synthesized) were evaluated on human CRC c HT-29 and HCT116 at 50 µ M for 48 h. In this screening, olefinic DAM series disp

Cell Viability, Cell Proliferation Inhibition, and IC 50 Determination
First, cell viability was evaluated using MTT assay. Eighteen olefinic DAM and eleven aryloxyDAM previously synthesized) were evaluated on human CRC cell lines HT-29 and HCT116 at 50 µM for 48 h. In this screening, olefinic DAM series displayed a higher cell viability proliferation inhibition than their aryloxyDAM analogues. In addition, for two series HT-29 cells seems to be more sensitive to the compounds than HCT116 cells. This observation is more pronounced for olefinic DAM (Figures 3 and 4).
cells. This observation is more pronounced for olefinic DAM (Figures 3 and 4).
The determination of a median inhibitory concentration (IC 50 ) at 1 to 50 µM (1, 10, 20, 30, 40 and 50 µM) was performed at 48 h on the two human CRC cell lines used for compounds with cell viability <50% (Table 1).  Figure 4. Screening of aryloxyDAM series on human HT-29 and HCT116 CRC cell lines viability. Cell viability (%) after exposure to the DAM was measured by MTT assay. Compounds 17-27 were assayed at 50 μM for 48 h. Each chart represents the mean percentage ± SEM from at least three independent experiments (n = 3). Cell viability lower than 50% was considered cytotoxic. * p < 0.05; ** p < 0.01 and *** p < 0.001 significantly different from the control.
The determination of a median inhibitory concentration (IC50) at 1 to 50 µ M (1, 10, 20, 30, 40 and 50 µ M) was performed at 48 h on the two human CRC cell lines used for compounds with cell viability <50% (Table 1).
The obtained results highlight that only five compounds displayed an IC50 lower than 30 µ M (10a-b, 12a, 13a, 14a) for HT-29 and three compounds (10a, 12a, 13b) displayed an IC50 below 35 µ M for HCT116. These results allowed us to establish the first elements of the structure-activity relationships (see Section 3.2.2).

Structure Activity Relationship Considerations
The cell proliferation inhibition induced by the DAM allowed a comprehensive structure-activity relationship (SAR) analysis. We considered the spatial configuration of either E or Z, the impact of olefinic carbon or the oxygen atom, the nature of the functionalization and the influence of N-oxide moiety.  The obtained results highlight that only five compounds displayed an IC 50 lower than 30 µM (10a-b, 12a, 13a, 14a) for HT-29 and three compounds (10a, 12a, 13b) displayed an IC 50 below 35 µM for HCT116. These results allowed us to establish the first elements of the structure-activity relationships (see Section 3.2.2).

Structure Activity Relationship Considerations
The cell proliferation inhibition induced by the DAM allowed a comprehensive structure-activity relationship (SAR) analysis. We considered the spatial configuration of either E or Z, the impact of olefinic carbon or the oxygen atom, the nature of the functionalization and the influence of N-oxide moiety.
In the olefinic DAM series, compounds with Z configuration displayed better antiproliferative activity compared to their corresponding E analogues except for compounds functionalized with bulky alkyl groups (isopropyl 9b vs. 9a and tert-butyl 10b vs. 10a) on human CRC HT-29 cells. The Z isomers containing Cl, Br, and CF 3 groups (14a, 13a and 12a) as well as the two Z and E isomers bearing the tert-butyl group (10a-b) showed a better antiproliferative activity on the human CRC HT-29 cell line with IC 50 values of 28.48, 24.03, 23.02, 25.70 and 25.15 µM, respectively. The brominated E isomer (13b), the Z isomer bearing a CF 3 moiety (12a) as well as a Z isomer bearing a tert-butyl group (10a) displayed the best antiproliferative activity on the human CRC HCT116 cell line with IC 50 of 26.13, 31.44 and 33.61 µM, respectively (Table 1).
These results suggested that in a general, Z isomers have a more interesting antiproliferative activity than the E isomers. These results are in agreement with those described in the literature for ferrocenyl derivatives of tamoxifen with IC 50 = 11 µM for the Z isomer and IC 50 = 60 µM for the E isomer [43].
Concerning the aryloxy DAM series, most of the compounds were found to be inactive towards the cell proliferation inhibition at concentrations lower than or equal to 50 µM. Only the brominated derivative 22 showed a potential antiproliferative activity on the two human CRC cell lines evaluated. These results suggested that the introduction of an oxygen atom decreases activity. Thus, this structural modification did not appear to be crucial for the antiproliferative activity on the human CRC cell lines evaluated (Table 1).
To study the influence of the introduction of an N-oxide pyridine motif, compounds 10a and 12a from the olefinic series and compounds 17 and 22 from the aryloxyDAM series were N-oxidized. The molecules were selected based on their antiproliferative activity and an inactive compound 17 was also selected in order to compare. The synthesized N-oxides were also tested at a concentration between 1 to 50 µM for 48 h.
For two series, the results displayed that pyridine N-oxides induce a loss of antiproliferative activity on both human CRC cell lines at a concentration of up to 50 µM. These results could suggest that non-substituted nitrogen atom in the pyridine ring is required for the biological activity.

Mechanism of Action Investigation
The compounds 10a, 10b and 12a, that showed the best biological activity were selected for further investigation of the mechanism of action. To elucidate the potential target of the antiproliferative activity on the human CRC cell lines, the study of some anti-apoptotic cell survival signaling pathways (phospho-ERK, phospho-Akt) and apoptotic signaling pathways (phospho-p38) were performed. In addition, the evaluation of pro-apoptotic markers (caspase-3 and PARP cleavage, DNA fragmentation) was also carried out.
Evaluation of pro-apoptotic markers of the survival and apoptosis signaling pathways was performed to complete the study of the mechanism of action. P-Akt, Akt, P-ERK, ERK, P-p38, p-38 sourced from human CRC HCT116 cells line were investigated. These cells were treated or not with IC 50 values of compounds 12a and 10a for 6 and 12 h. Total lysates were collected and expression of Akt, ERK and p38 MAPK proteins and their phosphorylated forms were determined by Western blot analysis.
Protein kinase B (Akt), is a protein involved in the cell death and survival process, playing a pivotal role in several interconnected cell signaling mechanisms ultimately engaged in cell metabolism, growth and division, apoptosis suppression and angiogenesis. Once phosphorylated, this protein generates consequently the P-Akt (phosphorylated Akt) which ultimately participates in the process of oxidative stress and plays a prognostic role in cancer. The inhibition of the Akt as well as the signaling pathway for its phosphory-lation prevents cell regeneration and thereby causes cell death. Analyzing the effects of compounds 12a and 10a on this Akt and the P-Akt, we observed that P-Akt is downregulated after 6 h of treatment with 12a, and this inhibition effect is enhanced at 12 h. This inhibition is also observed with 10a but this effect is not time-dependent as the level of P-Akt expression remains the same between 6 and 12 h ( Figure 5).
Once phosphorylated, this protein generates consequently the P-Akt (pho Akt) which ultimately participates in the process of oxidative stress and plays a role in cancer. The inhibition of the Akt as well as the signaling pathway for it ylation prevents cell regeneration and thereby causes cell death. Analyzing t compounds 12a and 10a on this Akt and the P-Akt, we observed that P-Akt is lated after 6 h of treatment with 12a, and this inhibition effect is enhanced a inhibition is also observed with 10a but this effect is not time-dependent as th Akt expression remains the same between 6 and 12 h ( Figure 5). Extracellular signal-regulated kinases (ERKs), are member of the mitoge protein kinase (MAPK) involved in a series of physiological processes, such lation of cell survival and proliferation as well as cell differentiation [45][46][47]. C Extracellular signal-regulated kinases (ERKs), are member of the mitogen-activated protein kinase (MAPK) involved in a series of physiological processes, such as the regulation of cell survival and proliferation as well as cell differentiation [44][45][46]. Colon tumor epithelial cells are dependent on mitogen activated protein kinase (MAPK) p38 for proliferation and survival [47,48]. We investigated the role of ERK and p38 in the observed antiproliferative effect. The results showed that ERK appeared unaltered compared to the control. On the other hand, compounds 12a and 10a drastically decreased ERK phosphorylation as early as 6 h of treatment, with the latter remaining at the same level of expression after 12 h of treatment with compound 10a. In addition, no p38 activation was observed, suggesting that compounds 12a and 10a do not influence this pro-apoptotic signaling pathway ( Figure 5).
Evaluation of pro-apoptotic markers.
Once the DNA in cancer cells is fragmented, enzymes such as poly(ADP-ribose) polymerase (PARP), involved in DNA reparation and consequently regeneration of the cancer cells, are mutated or inactivated by a pro-apoptotic process. Likewise, caspase-3 is an apoptosis related protein that is involved and activated during apoptosis [49]. In order to determine whether the inhibition of human CRC HCT116 and HT-29 cells line indeed affected the DNA reparation, the potential of the compounds 10a, 10b and 12a to inhibit these apoptosis related proteins PARP and caspase-3 were analyzed [50,51]. Cells were treated or not for 24 and 48 h, with IC 50 values of compounds 10a, 10b and 12a, for human CRC HT-29 cells and compounds 10a and 12a for human CRC HCT116 cells. Total lysates were collected, and expression of apoptosis-related proteins was determined by Western blot analysis.
Only compound 12a showed inhibition of both pro-apoptotic proteins. Treatment of human CRC HT-29 cells with the compound 12a showed a high expression of the cleaved form of caspase-3 within 48 h of treatment ( Figure 6A). Similarly, the treatment of human CRC HCT116 cells, provides the cleavage of caspase-3 over 24 h of treatment and its expression is enhanced after 48 h ( Figure 6B).  In human HT-29 cells, the cleaved PARP expression is slightly enhanced for the compound 12a within 24 h and turns overexpressed at 48 h. However, for compounds 10a and 10b native PARP remains constant and cleaved PARP has never been expressed either at 24 h or 48 h ( Figure 6A). In human CRC HCT116 cells, compound 12a induces a significant increase of cleaved PARP at 24 and 48 h ( Figure 6B).  On human CRC HT-29 cells, compounds 10a and 10b showed a slight DNA fragmentation compared to the control. However, significant fragmentation was observed with compound 12a at 24 and 48 h, 2.90-and 3.25-fold, respectively compared to the control ( Figure 7A). On human CRC HCT116 cells, no significant DNA fragmentation was observed after 24 h showing only 1.46-fold with compound 12a compared to the control. However, very high fragmentation of the DNA was observed at 48 h showing 6.97-fold compared to the control as illustrated in Figure 7B. It can be noticed that compound 12a has an immediate effect on human CRC HT-29 cells while on HCT116 cells, the effect is more delayed and is observed only at 48 h.

Normal Cell Line Viability
Cell viability of L929 cell line (murine fibroblasts) was evaluated in presence of synthesized compounds using the MTT colorimetric assay. At the concentrations evaluated (between 0.78 and 100 µ M), all compounds have no effect on the cell viability of these normal cells. The figures and graphs concerning these results are reported in the supplementary data section (SI-2). On human CRC HT-29 cells, compounds 10a and 10b showed a slight DNA fragmentation compared to the control. However, significant fragmentation was observed with compound 12a at 24 and 48 h, 2.90-and 3.25-fold, respectively compared to the control ( Figure 7A). On human CRC HCT116 cells, no significant DNA fragmentation was observed after 24 h showing only 1.46-fold with compound 12a compared to the control. However, very high fragmentation of the DNA was observed at 48 h showing 6.97-fold compared to the control as illustrated in Figure 7B. It can be noticed that compound 12a has an immediate effect on human CRC HT-29 cells while on HCT116 cells, the effect is more delayed and is observed only at 48 h.

Normal Cell Line Viability
Cell viability of L929 cell line (murine fibroblasts) was evaluated in presence of synthesized compounds using the MTT colorimetric assay. At the concentrations evaluated (between 0.78 and 100 µM), all compounds have no effect on the cell viability of these normal cells. The figures and graphs concerning these results are reported in the supplementary data section (SI-2).

Molecular Modeling
In order to study the potential binding mode of compounds 10a and 12a in the biological targets previously identified, a docking approach was used.
For AKT (PDB ID: 6S9W), the predicted binding modes obtained for compounds 10a and 12a were quite identical and also very similar to a part of the co-crystallized ligand ( Figure 8A). Indeed, our compounds were predicted to share similar hydrophobic interactions with the residues of the AKT binding site (with W80, L264, V270 and Y272) and πstacking (W80), than the co-crystallized ligand ( Figure 8B). Additionally, to these shared interactions, our compounds 10a and 12a are predicted to establish a hydrogen bond (HB) with K268.

Molecular Modeling
In order to study the potential binding mode of compounds 10a and 12a in the biological targets previously identified, a docking approach was used.
For AKT (PDB ID: 6S9W), the predicted binding modes obtained for compounds 10a and 12a were quite identical and also very similar to a part of the co-crystallized ligand ( Figure 8A). Indeed, our compounds were predicted to share similar hydrophobic interactions with the residues of the AKT binding site (with W80, L264, V270 and Y272) and πstacking (W80), than the co-crystallized ligand ( Figure 8B). Additionally, to these shared interactions, our compounds 10a and 12a are predicted to establish a hydrogen bond (HB) with K268. We also evaluated the potential binding mode of our compounds 10a and 12a in the two isoforms of the ERK proteins, namely ERK-1 and ERK-2. The binding site of these two proteins are highly similar, however the predicted binding mode in these proteins were dissimilar ( Figure 9). In the ERK-1 binding site (PDB ID: 4QTB), our compounds are predicted to present different binding modes ( Figure 10A) but with some similarities. Indeed, compounds 10a and 12a are both predicted to be able to establish hydrophobic interactions with ERK-1 residues, among which Y53, K71 and L173, are also involved in hydrophobic interactions with the 4QTB co-crystallized ligand. Additionally, compound 12a predicted binding mode includes a halogen bond with the ERK-1 binding site residue E88, which is engaged in a salt bridge with the 4QTB co-crystallized ligand ( Figure 10B). For compound 10a, We also evaluated the potential binding mode of our compounds 10a and 12a in the two isoforms of the ERK proteins, namely ERK-1 and ERK-2. The binding site of these two proteins are highly similar, however the predicted binding mode in these proteins were dissimilar ( Figure 9).

Molecular Modeling
In order to study the potential binding mode of compounds 10a and 12a in the biological targets previously identified, a docking approach was used.
For AKT (PDB ID: 6S9W), the predicted binding modes obtained for compounds 10a and 12a were quite identical and also very similar to a part of the co-crystallized ligand ( Figure 8A). Indeed, our compounds were predicted to share similar hydrophobic interactions with the residues of the AKT binding site (with W80, L264, V270 and Y272) and πstacking (W80), than the co-crystallized ligand ( Figure 8B). Additionally, to these shared interactions, our compounds 10a and 12a are predicted to establish a hydrogen bond (HB) with K268. We also evaluated the potential binding mode of our compounds 10a and 12a in the two isoforms of the ERK proteins, namely ERK-1 and ERK-2. The binding site of these two proteins are highly similar, however the predicted binding mode in these proteins were dissimilar (Figure 9). In the ERK-1 binding site (PDB ID: 4QTB), our compounds are predicted to present different binding modes ( Figure 10A) but with some similarities. Indeed, compounds 10a and 12a are both predicted to be able to establish hydrophobic interactions with ERK-1 residues, among which Y53, K71 and L173, are also involved in hydrophobic interactions with the 4QTB co-crystallized ligand. Additionally, compound 12a predicted binding mode includes a halogen bond with the ERK-1 binding site residue E88, which is engaged in a salt bridge with the 4QTB co-crystallized ligand ( Figure 10B). For compound 10a, In the ERK-1 binding site (PDB ID: 4QTB), our compounds are predicted to present different binding modes ( Figure 10A) but with some similarities. Indeed, compounds 10a and 12a are both predicted to be able to establish hydrophobic interactions with ERK-1 residues, among which Y53, K71 and L173, are also involved in hydrophobic interactions with the 4QTB co-crystallized ligand. Additionally, compound 12a predicted binding mode includes a halogen bond with the ERK-1 binding site residue E88, which is engaged in a salt bridge with the 4QTB co-crystallized ligand ( Figure 10B). For compound 10a, additional interactions are π-stacking with Y53 and π-cation interaction with K71 ( Figure 10C). additional interactions are π-stacking with Y53 and π-cation interaction with K71 ( Figure  10C). In the ERK-2 binding site (PDB ID: 6SLG), our compounds are predicted to interact in a similar way, but differently from both the co-crystallized ligand ( Figure 11A) and as previously mentioned, the predicted ERK-1 binding modes. Our compounds are predicted to interact with the ERK-2 binding sites ( Figure 11B) through numerous hydrophobic interactions (with residues A18, Y19, V22, K37, I39 and D150) and hydrogen bonds (with residues Y19, K37 and R50). The predicted binding modes of compounds 12a and 10a are very similar in the PARP binding site (4DZZ) but they are not superimposed with the co-crystallized ligand NMS-P118 due to a lack of structural similarity ( Figure 12A). However, our compounds shared hydrophobic interactions (with residues Y896 and Y907) and π-stacking (with Y907) with the reference NMS-P118/PARP complex ( Figure 12B). Moreover, our compounds are predicted to establish additional hydrophobic interactions (with Y889), π-stacking (with 896Y and 889Y) and HB (with M890 vs. E863 in the reference NMS-P118/PARP complex). In the ERK-2 binding site (PDB ID: 6SLG), our compounds are predicted to interact in a similar way, but differently from both the co-crystallized ligand ( Figure 11A) and as previously mentioned, the predicted ERK-1 binding modes. Our compounds are predicted to interact with the ERK-2 binding sites ( Figure 11B) through numerous hydrophobic interactions (with residues A18, Y19, V22, K37, I39 and D150) and hydrogen bonds (with residues Y19, K37 and R50).
additional interactions are π-stacking with Y53 and π-cation interaction with K71 ( Figure  10C). In the ERK-2 binding site (PDB ID: 6SLG), our compounds are predicted to interact in a similar way, but differently from both the co-crystallized ligand ( Figure 11A) and as previously mentioned, the predicted ERK-1 binding modes. Our compounds are predicted to interact with the ERK-2 binding sites ( Figure 11B) through numerous hydrophobic interactions (with residues A18, Y19, V22, K37, I39 and D150) and hydrogen bonds (with residues Y19, K37 and R50). The predicted binding modes of compounds 12a and 10a are very similar in the PARP binding site (4DZZ) but they are not superimposed with the co-crystallized ligand NMS-P118 due to a lack of structural similarity ( Figure 12A). However, our compounds shared hydrophobic interactions (with residues Y896 and Y907) and π-stacking (with Y907) with the reference NMS-P118/PARP complex ( Figure 12B). Moreover, our compounds are predicted to establish additional hydrophobic interactions (with Y889), π-stacking (with 896Y and 889Y) and HB (with M890 vs. E863 in the reference NMS-P118/PARP complex). The predicted binding modes of compounds 12a and 10a are very similar in the PARP binding site (4DZZ) but they are not superimposed with the co-crystallized ligand NMS-P118 due to a lack of structural similarity ( Figure 12A). However, our compounds shared hydrophobic interactions (with residues Y896 and Y907) and π-stacking (with Y907) with the reference NMS-P118/PARP complex ( Figure 12B). Moreover, our compounds are predicted to establish additional hydrophobic interactions (with Y889), π-stacking (with 896Y and 889Y) and HB (with M890 vs. E863 in the reference NMS-P118/PARP complex). The caspase-3 binding site is formed by different sub-pockets. Our compounds 12a and 10a are less flexible than the co-crystallized ligands of caspase-3 and are thus predicted to establish interactions with different subpockets than other co-crystallized ligands. Nevertheless, the pyridine moiety of our compounds is predicted to be located in a similar hydrophobic pocket than part of a few co-crystallized ligands of caspase-3 ( Figure  13A). Compounds 10a and 12a are predicted to establish both π-stacking and hydrophilic interactions with residues Y204, W206 and F256 of the caspase-3 binding site ( Figure 13B). The predicted binding modes of our compounds in the AKT, ERK-1 and ERK-2, PARP and caspase-3 binding sites complement and support the biological results obtained for these targets. According to the analysis of these predicted binding modes (and also the docking scores), the most promising biological target for compounds 10a and 12a seems to be the AKT protein. The caspase-3 binding site is formed by different sub-pockets. Our compounds 12a and 10a are less flexible than the co-crystallized ligands of caspase-3 and are thus predicted to establish interactions with different subpockets than other co-crystallized ligands. Nevertheless, the pyridine moiety of our compounds is predicted to be located in a similar hydrophobic pocket than part of a few co-crystallized ligands of caspase-3 ( Figure 13A). Compounds 10a and 12a are predicted to establish both π-stacking and hydrophilic interactions with residues Y204, W206 and F256 of the caspase-3 binding site ( Figure 13B). The caspase-3 binding site is formed by different sub-pockets. Our compounds 12a and 10a are less flexible than the co-crystallized ligands of caspase-3 and are thus predicted to establish interactions with different subpockets than other co-crystallized ligands. Nevertheless, the pyridine moiety of our compounds is predicted to be located in a similar hydrophobic pocket than part of a few co-crystallized ligands of caspase-3 ( Figure  13A). Compounds 10a and 12a are predicted to establish both π-stacking and hydrophilic interactions with residues Y204, W206 and F256 of the caspase-3 binding site ( Figure 13B). The predicted binding modes of our compounds in the AKT, ERK-1 and ERK-2, PARP and caspase-3 binding sites complement and support the biological results obtained for these targets. According to the analysis of these predicted binding modes (and also the docking scores), the most promising biological target for compounds 10a and 12a seems to be the AKT protein.  Figure 13. Predicted binding modes of compounds 10a (in salmon) and 12a (in orange) in the caspase-3 binding site (PDB ID: 6CKZ, 1GFW) (A) in comparison with the 6CKZ co-crystallized ligand (in blue) and the 1GFW co-crystallized ligand (in cyan) and (B) with detailed information about the interactions established between our compounds and the caspase-3 binding site obtained using PLIP [52] (green dotted lines: πstacking; grey dotted lines: hydrophobic).

Prediction of ADME Properties and Druglikeness
The predicted binding modes of our compounds in the AKT, ERK-1 and ERK-2, PARP and caspase-3 binding sites complement and support the biological results obtained for these targets. According to the analysis of these predicted binding modes (and also the docking scores), the most promising biological target for compounds 10a and 12a seems to be the AKT protein.

Prediction of ADME Properties and Druglikeness
The prediction of several properties and molecular descriptors allowed us to suggest the ADME profile and drug-likeness of the three promising drug candidates 10a, 10b, and 12a ( Figure 14). The prediction of several properties and molecular descriptors allowed us to suggest the ADME profile and drug-likeness of the three promising drug candidates 10a, 10b, and 12a ( Figure 14). . SASA: Total solvent accessible surface area in square angstroms using a probe with a 1.4 Å radius (range: 300-1000). donorHB (Hydrogen bond donors): number of hydrogen bonds that would be given by the solute to water molecules in an aqueous solution (recommended value: 0.0-6.0). accptHB (Hydrogen bond acceptor): number of hydron bonds that would be accepted by the solute from the water molecules in an aqueous solution (recommended value: 2.0-20.0). PSA: Van der Waals surface area of polar nitrogen and oxygen atoms and carbonyl carbon atoms (range: 7-200). Ppolrz: Predicted polarizability in cubic angstroms (range: 13-70). PlogPo/w: Predicted octanol/water partition coefficient (recommended value: −2.0 to 6.5). PlogS: Predicted aqueous solubility, log S. S in mol dm −3 is the concentration of the solute in a saturated solution that is in equilibrium with the crystalline solid (recommended value: −6.5 to 0.5). PlogBB: Predicted brain/blood partition coefficient for orally delivered drugs (recommended value: −3.0 to 1.2). PPMDCK: Predicted permeability of MDCK cells (mimic of the blood-brain barrier) in nm/sec (<25 is poor and >500 is great). PlogKp: Predicted skin permeability (recommended value: −8.0 to −1). PlogKhsa: Prediction of binding to human serum albumin (recommended value: −1.5 to 1.5). PPCaco: Predicted permeability of Caco-2 cells (model of the intestinal barrier) in nm/s (<25 is low and >500 is great). HOA%: Predicted human oral absorption on 0 to 100% scale. (<25% is poor and >80 is high). Metab: Number of likely metabolic reactions (recommended <8). Stars: Property or descriptor values that fall outside the 95% range of similar values for known drugs (recommended <5). RuleOf3: Number of violations of Jorgensen's rule of three. The three rules are: PlogS > −5.7, PPCaco > 22 nm/s, #Primary Metabolites < 7. Compounds with fewer (and preferably no) violations of these rules are more likely to be orally available. RuleOf5: Number of violations of Lipinski's rule of five. The rules are MW < 500, PlogPo/w < 5, donorHB ≤ 5, accptHB ≤ 10. Compounds that satisfy these rules are considered drug-like (recommended < 4). . SASA: Total solvent accessible surface area in square angstroms using a probe with a 1.4 Å radius (range: 300-1000). donorHB (Hydrogen bond donors): number of hydrogen bonds that would be given by the solute to water molecules in an aqueous solution (recommended value: 0.0-6.0). accptHB (Hydrogen bond acceptor): number of hydron bonds that would be accepted by the solute from the water molecules in an aqueous solution (recommended value: 2.0-20.0). PSA: Van der Waals surface area of polar nitrogen and oxygen atoms and carbonyl carbon atoms (range: 7-200). Ppolrz: Predicted polarizability in cubic angstroms (range: 13-70). PlogPo/w: Predicted octanol/water partition coefficient (recommended value: −2.0 to 6.5). PlogS: Predicted aqueous solubility, log S. S in mol dm −3 is the concentration of the solute in a saturated solution that is in equilibrium with the crystalline solid (recommended value: −6.5 to 0.5). PlogBB: Predicted brain/blood partition coefficient for orally delivered drugs (recommended value: −3.0 to 1.2). PPMDCK: Predicted permeability of MDCK cells (mimic of the blood-brain barrier) in nm/sec (<25 is poor and >500 is great). PlogKp: Predicted skin permeability (recommended value: −8.0 to −1). PlogKhsa: Prediction of binding to human serum albumin (recommended value: −1.5 to 1.5). PPCaco: Predicted permeability of Caco-2 cells (model of the intestinal barrier) in nm/s (<25 is low and >500 is great). HOA%: Predicted human oral absorption on 0 to 100% scale. (<25% is poor and >80 is high). Metab: Number of likely metabolic reactions (recommended <8). Stars: Property or descriptor values that fall outside the 95% range of similar values for known drugs (recommended <5). RuleOf3: Number of violations of Jorgensen's rule of three. The three rules are: PlogS > −5.7, PPCaco > 22 nm/s, #Primary Metabolites < 7. Compounds with fewer (and preferably no) violations of these rules are more likely to be orally available. RuleOf5: Number of violations of Lipinski's rule of five. The rules are MW < 500, PlogPo/w < 5, donorHB ≤ 5, accptHB ≤ 10. Compounds that satisfy these rules are considered drug-like (recommended < 4).
Most of the physicochemical properties of the three selected candidates were within the recommended ranges. However, low aqueous solubility (PlogS) and octanol/water partition coefficient (PlogPo/w) values outside the recommended limit for the compounds were predicted. The ADME properties that refer to the capacity to cross the blood-brain barrier (PlogBB and PPMDCK), the intestinal barrier (PPCaco), or to bind to human serum albumin PlogKhsa), as well as the predicted human oral absorption (HOA%) were all predicted as good for the three compounds. Only the skin permeability (PlogKp) was not predicted within the values considered recommended. Finally, several criteria (i.e., stars, RuleOf3, and RuleOf5) demonstrated the drug-likeness compliance of the studied molecules.

Conclusions
In summary, two original series of DAM compounds, olefinic DAM and aryloxyDAM, were straightforwardly synthetized using efficient synthetic strategies, characterized, and biologically evaluated. The effects of the novel 33 DAM derivatives (18 olefinic DAM and 11 aryloxyDAM and 4 N-oxides derivatives) were evaluated in vitro on human CRC cell lines HT-29 and HT116. The cell-based bioassays revealed that compounds 10a, 10b and 12a of the olefin series decreased the viability of cancer cells. The Z isomers seem to be more active than their E analogues and the pyridine cycle is important for the activity. The mechanism of action of compounds 10a, 10b, and 12a was investigated. We can conclude that compound 12a has a very interesting anti-cancer potential. Thus, the effects of 12a in inducing caspase-3 cleavage, and its inhibitory effect on PARP activity is correlated with the increase of DNA fragmentation in cancer cells. Moreover, in silico molecular docking studies were conducted to predict binding modes of compounds 10a and 12a in the AKT, ERK-1 and ERK-2, PARP and caspase-3 binding sites confirming the biological results. According to the analysis of the predicted binding modes and docking scores, the AKT protein seems to be the most interesting biological target to study the antiproliferative activity of the DAM. The docking study herein conducted can be used to guide further optimization of compounds 10a and 12a.
The structural adequacy, absence of cytotoxicity as well as druglikeness and favorable ADME profile, allow to suggest 10a, 10b, and 12a are a new leads compounds in the study anticancer drugs. In general, these results could provide valuable insights to design new potent anticancer drugs based on the DAM scaffold.