Structure and cytotoxic activity of terpenoid-like chalcones

2.1 General procedure for synthesis of terpenoid-like chalcones 1–8

α-Ionone, β-ionone, and benzaldehyde derivatives were highest reagent grade commercial products. A water solution of sodium hydroxide (0.123 g) was added to mixture of ionone (1.5 mmol) and benzaldehyde derivative (2.5 mmol) in methanol (10 ml). The reaction mixture was stirred at room temperature for 12 h. At the end of reaction time, the mixture was diluted with CH₂Cl₂, treated with HCl solution (10%), and washed with water. The organic phase was dried over anhydrous Na₂SO₄ and filtered. The solvent was slowly evaporated and the crystals suitable for structure determination were isolated without further crystallization. Furthermore, the product of the reaction was chromatographed on a silica gel column, 120–230 mesh, and eluted with ethyl acetate: hexane (2:8).

2.2 Structure determination by single-crystal X-ray diffraction technique

Suitably shaped single crystals of 1, 3–8 were selected for the X-ray diffraction data collection on a Bruker-AXS Kappa Duo diffractometer with an APEX II CCD detector. Even though compound 5 has been determined previously (Zou et al., 2012), we performed the collection of its single-crystal X-ray data. Therefore, all geometrical parameters of 5 presented here are from our structure redetermination. We have deposited crystal data of 5 with Cambridge Crystallographic Data Centre (CCDC) under deposit code 1029769. Compound 2 was not redetermined in this study and its crystal data are from the literature (CCDC deposit code 818053) (Fernandes et al., 2013). Mo Kα radiation from an IμS microsource with multilayer optics was employed. Diffraction images were recorded by φ and ω scans set using APEX2 software (Bruker, 2012). This software was also employed to treat the raw dataset for indexing, integrating, reducing and scaling of the reflections. Next, the crystallographic softwares were used as follows: SIR2004 (structure solving) (Burla et al., 2005), SHELXL-97 (structure refinement) (Sheldrick, 2008), MERCURY and ORTEP-3 (structure analysis and representations) (Macrae et al., 2008; Farrugia, 1997). All structures were solved through identification of all non-hydrogen atoms in asymmetric units directly from the Fourier synthesis of the structure factors after retrieval of their phase using the direct methods. Each initial model was refined by full-matrix least squares method on F², adopting free anisotropic and constrained isotropic atomic displacement parameters for non-hydrogen and hydrogen atoms, respectively. In the case of hydrogens, their U_iso was set to either 1.2U_eq of the bonded carbon, except for methyl hydrogens in which their U_iso was set to 1.5U_eq of the corresponding carbon. Hydrogen coordinates were stereochemically defined and constrained in the refinements, oscillating as that of the bonded carbon to keep idealized bond angles and lengths fixed in 0.93 Å (C_sp²—H), 0.96 Å (C_sp³—H in CH₃ groups), 0.97 Å (C_sp³—H in CH₂ groups), and 0.98 Å (C_sp³—H in CH groups).

Compounds 5–8 were present with two twist conformations of the cyclohex-1-ene ring. This conformational variability has occurred due to static disorder into the six-membered ring. The static disorder was modeled by splitting the CH₂ carbons and hydrogens over two sites with occupancy factors constrained to the following values: 50% (both conformations of 6), 55% (major conformation of 5 and 7 whose carbon fractions were labeled as C14—C16), 65% (major conformation of 8 whose carbon fractions were labeled as C14—C16), 45% (minor conformation of 5 and 7 whose carbon fractions were labeled as C14′—C16′), or 35% (minor conformation of 8 whose carbon fractions were labeled as C14′—C16′). It is important to state that all site occupancy factors (SOFs) of these carbons were refined freely in trial refinements. The SOFs found from initial refinements were then constrained to the corresponding sites and in sequence hydrogens were stereochemically positioned using a riding model on carbon as described above. Such refinement procedure has resulted in better R-factors, the lowest residual electronic density in the unit cell and refinement convergence. It is striking to note that unusual values of R-factors were outputted from refinement of 1 and 3, which could be a consequence of a static disorder due to different cyclohex-2-ene ring puckering modes similar to those modeled reliably for terpenoid-like chalcones 5–8. In trial refinements constraining the primary site occupancy of cyclohex-2-ene carbon atoms to 50%, extra occupancy sites were observed but they were very close together with the primary ones. In attempt to refine them as carbon fractions, high correlation between all tested SOFs and atomic displacement parameters was observed applying the classical split-atom approach to the structural model of 1 and 3, as well as refinements were unstable without convergence. Therefore, we have used the classical harmonic model based upon a development of the atomic displacement parameters with just one site instead of the split-position model.

2.3 Additional characterization

Other characterizations of synthesized compounds were performed as follows. Melting points were measured in a Micro Química MQAPF-301 apparatus. The infrared spectra were obtained in Bomem M102 Fourier Transform spectrometer. Samples were analyzed by the FTIR transmission technique as KBr pellets (mixtures comprising 200 mg of KBr and 1 mg of sample). NMR spectra were recorded on a Bruker Avance III, operating at 500.13 MHz for ¹H and 125.77 MHz for ¹³C, equipped with a 5 mm TBI probe (¹H, ¹³C and XBB).

2.4 Cytotoxicity assay

Terpenoid-like chalcones were tested for cytotoxicity against three tumor cell lines: SF-295 (central nervous system), HCT-116 (colon cancer) and OVCAR-8 (breast cancer). All cell lines were obtained from National Cancer Institute (NCI) (Skehan et al., 1990) and they are the first choice cancer cell lines used for the cytotoxicity screening in the Laboratório de Oncologia from Universidade Federal of Ceará, where the assays were performed. These cancer cell lines were cultivated in RPMI 1640 medium supplemented with fetal bovine serum (10%), glutamine (2 mm), penicillin (100 U/ml), and streptomycin (100 mg/ml), at 37 °C, with CO₂ (5%). The MTT assay was used for the cytotoxicity evaluation of terpenoid-like chalcones (Berridge et al., 1996; Mossman, 1983). Cells were plated at a concentration of 0.1 × 10⁶ cells/ml for strains SF-295 and OVCAR-8-3 and 0.7 × 10⁵ cells/ml for the HCT-116 strain.. Doxorubicin (5 μg/ml) and DMSO (0.05%) were used as positive and negative controls, respectively. After 24 h, the terpenoid-like chalcones were added to the wells to obtain the final concentration (25 μg/ml). After 72 h, plates were centrifuged (1500 rpm, 15 min), and the supernatants were removed. An aliquot (150 μl) from the MTT soln. (0.5 mg/ml) was added into each well, and the plates were incubated for 3 h under the same conditions described above. Then, DMSO (150 μl) was poured to dissolve the precipitate, and the absorbance was measured at 595 nm. This experiment was run in three replicates, and all absorbance values were converted into a cell growth inhibition percentage (GI-%) by the following equation: GI-% = 100 − [(T/C) × 100%]. C is the absorbance for the negative control, and T was the absorbance in the presence of the tested extract. Those compounds that presented more than 85% of activity were selected to be tested at concentrations varying from 0.078 to 5 μg/ml to determine IC₅₀ values by nonlinear regression using GraphPad Prism 4.0 software.

3.1 Conformational analysis from crystal structures

The terpenoid-like chalcones 1–8 were synthetized by Claisen–Schmidt condensation reaction between α- and β-ionone with substituted benzaldehydes dissolved in methanol (MeOH). The reactions were catalyzed by sodium hydroxide (NaOH) in aqueous solution at room temperature for 12 h. The structure of all compounds is presented in Scheme 1. All products were characterized by melting point, infrared and NMR spectroscopy, and single-crystal X-ray diffraction technique. Except for compounds 2 and 5, their crystal structure is reported for the first time here. In addition, we have also redetermined the structure of 5 in this study. Such structure redetermination allowed us to know correctly the population of each cyclohexene twist conformation. Zou et al. (2012) have wrongly modeled the disordered methylene groups of cyclohexene over two sets of sites with occupancy ratios of 0.50:0.50 and 0.60:0.40. This does result obligatorily in an unexpected planar cyclohexene conformation of 10% occupancy in addition to the two expected twist conformations. Furthermore, an extremely short C_sp³—C_sp³ bond length is observed in this planar conformation (according to Zou et al. (2012) atom labeling scheme, C15—C16′ measures 1.21 Å). Therefore, our redetermination does correct the previous model to only two twist conformations of the cyclohexene ring with either 55% or 45% occupancy. In our model, all bond lengths into cyclohexene ring are also in agreement with the expected values.

Here, structures of six terpenoid-like chalcones (1, 3, 4, 6–8) have been elucidated by single-crystal X-ray diffraction technique for the first time, while analogs 2 and 5 were already determined crystallographically (Zou et al., 2012; Fernandes et al., 2013) but their crystal data will be also presented here for comparison purposes. Terpenoid-like chalcones 1–5 and 8 crystallized in centrosymmetric space groups as expected from either the absence of chiral centers in their molecular backbones (compounds 2, 4, 5 and 8) or achiral synthesis (compounds 1 and 3) (Table 1). Even without molecular chirality, compounds 6 and 7 were solved in a noncentrosymmetric space group as a consequence of frozen-in chirality upon crystallization, i.e., just one conformation is assumed in the crystal lattice.

The asymmetric unit of all terpenoid-like chalcones is composed by just one molecule, even though cyclohexene ring can adopt two puckering modes in a same structure which resulted in disordered occupancy sites (Fig. 1).

Close

Figure 1 Asymmetric units of terpenoid-like chalcones crystallographically elucidated in this study. Non-hydrogen atoms are represented as 30% probability ellipsoids, while hydrogens are shown as arbitrary radius spheres. Ellipsoids of the carbon fractions in disordered positions of minor occupancy (or equal occupancy in 6) are drawn as boundaries, and open lines draw bonds between these atom fractions.

Export to PPT

As can be viewed in Fig. 1, a 30% probability ellipsoids representation of the asymmetric units determined in this study, the stereochemistry around both C7⚌C8 and C10⚌C11 double bonds is unambiguously assigned as an (E)-configuration. In compounds 2, 4–8, all derived from β-ionone, another double bond was clearly identified between C12 and C17 carbons, while in compounds 1 and 3, both derived from α-ionone, C12 is sp³-hybridized and the cyclohexene double bond does occur between C16 and C17 carbons.

Cyclohexene ring adopts half-chair and twist conformations in chalcones. Half-chair puckering is found in 1 and 3, with C13 deviated by 0.598(7) Å and 0.486(4) Å from the least-squares plane calculated through the other five cyclic coplanar atoms (r.m.s.d. of the C12—C17—C16—C15—C14 fitted atoms is 0.0617 Å in 1 and 0.0889 Å in 3). However, C13 is on the same side of C11 if the cyclohexene mean plane of 1 is taken as reference, while these atoms are on opposite sides relative to the corresponding plane of 3. This conformational feature can be related to the equatorial position of C11 in 1 and the axial one in 3 (see Fig. 1).

Just one twist conformation is observed in 4, with C14 and C15 carbons deviating from the mean plane formed through the C13—C12—C17—C16 atoms by −0.256(3) Å and 0.455(3) Å, respectively (r.m.s.d. of the fitted atoms is 0.0293 Å). This twist conformation was also reported for the conformation with minor occupancy of 2 (C14 and C15 deviations of 0.305(18) Å and −0.170(12) Å, r.m.s.d. of the fitted coplanar atoms is 0.0668 Å). The other conformation was present with a half-chair puckering with C15 at the flap (C15 deviation of −0.467(13) Å, r.m.s.d. of the fitted coplanar atoms is 0.0384 Å).

Similar to 4, cyclohexene ring of compounds 5–8 does assume twist conformations with C14 and C15 carbons away from the other four coplanar atoms. However, there are two twist conformations featured by flipping C14 and C15 carbons in the last four terpenoid-like chalcones. In 6, C14 and C15 carbons are deviated by 0.276(13) Å and −0.501(12) Å from the mean plane formed through the C13—C12—C17—C16 atoms (r.m.s.d. of the fitted atoms is 0.0116 Å) in one twist conformation, and by −0.321(12) Å and 0.486(10) Å in another one (r.m.s.d. of the C13—C12—C17—C16’ fitted atoms is 0.0020 Å). Both twist conformations are equally distributed over the crystal of 6, i.e., site occupancy factors of 50% resulted in the best refinement statistics. On contrary, the major populations of 5 and 7 (both with 55% occupancy) and 8 (65%) are present with a twist conformation featured by C14 and C15 deviating 0.412(7) Å and −0.436(9) Å in 5, 0.35(2) Å and −0.558(11) Å in 7, and 0.325(11) Å and −0.443(11) Å in 8 from the least-square mean plane passing through the four coplanar atoms (r.m.s.d. of the fitted atoms is 0.0209 Å in 5, 0.0177 Å in 7 and 0.0513 Å in 8). In the twist conformation with minor occupancy of either 45% in 5 and 7 or 35% in 8, those respective carbons do deviate −0.447(9) Å and 0.412(11) Å in 5, −0.247(17) Å and 0.425(10) Å in 7, and −0.39(2) Å and 0.45(2) Å in 8 from the least-square mean plane (r.m.s.d. of the fitted atoms is 0.0208 Å in 5, 0.0204 Å in 7 and 0.0179 Å in 8).

Endocyclic torsion angles also describe the cyclohexene conformations observed in terpenoid-like chalcones (Table 2). The lowest dihedral angles within the cyclohexene are those on the C15—C16 and C16—C17 bonds in the C13-flap half-chair conformation of 1 and 3 (torsions C14—C15—C16—C17 and C12—C17—C16—C15). For the C15-flap half-chair conformation of 2, the lowest dihedral angles into cyclohexene are those on the C12—C13 and C12—C17 bonds (torsions C17—C12—C13—C14 and C16—C17—C12—C13). Likewise, the lowest endocyclic cyclohexene torsion angle is on the C12—C17 bond in the twist conformation observed in 2, 4–8 (torsion C13—C12—C17—C16).

Another difference between the molecular conformations of terpenoid-like chalcones resides in the C11—C12 bond axis rotation, resulting in different orientations for cyclohexene rings relative to core carbonyl and olefin moieties. In chalcones 1 and 3 came from α-ionone, the cyclohexene mean plane, calculated through coplanar carbon atoms, does form an angle of ca. 90° [79.7(2)° in 1 and 84.14(14)° in 3] with the least-squares mean plane encompassing carbonyl atoms and neighboring α-carbons (r.m.s.d. of the fitted C8, C9, C10 and O1 atoms is 0.0120 Å in 1 and 0.0030 Å in 3). Their C10—C11—C12—C13 and C10—C11—C12—C17 torsion angles are likewise similar (Table 2). In the chalcones derived from β-ionone, the molecular skeleton is not so bent as in those made up of α-ionone. These corresponding mean planes are bent by 70(2)° and 44(3)° in the major and minor populations of 2 (r.m.s.d. of the fitted C8, C9, C10 and O1 atoms is 0.0045), by 52.93(8)° in 4 (r.m.s.d. of the fitted C8, C9, C10 and O1 atoms is 0.0144), by 5.39(15) in both major and minor populations of 5 (r.m.s.d. of the fitted C8, C9, C10 and O1 atoms is 0.0014), by 36.50(18)° and 36.52(19)° in the equally distributed populations of 6 (r.m.s.d. of the fitted C8, C9, C10 and O1 atoms is 0.0151), by 37.9(2) in both major and minor populations of 7 (r.m.s.d. of the fitted C8, C9, C10 and O1 atoms is 0.0258), and by 58.6(2)° in both major and minor populations of 8 (r.m.s.d. of the fitted C8, C9, C10 and O1 atoms is 0.0126).

As can be viewed in Table 2, the C10—C11—C12—C13 torsion angle of the β-ionone derived chalcones (except 5) is larger than the corresponding ones of 1 and 3 whereas in turn their C10—C11—C12—C17 torsion is lower than those of 1 and 3. This reflects the fact that cyclohexene ring is less bent in β-ionone-type compounds than in α-ionone derived chalcones present with cyclohexene ring almost perpendicular to the core molecular plane. This high planarity of β-ionone chalcones is highlighted in compound 5. Besides its lowest angle value between the carbonyl/α-carbons and cyclohexene mean planes, its C10—C11—C12—C13 and C10—C11—C12—C17 torsion angles are close to 0° and 180°, respectively (Table 2).

The central mean plane formed with carbonyl atoms and α-carbons and the substituted aromatic ring are not much bent in all terpenoid-like chalcones, forming angles between them of 37.53(14)°, 13.49(13)°, 2.59(12)°, 31.40(7)°, 22.46(9)°, 26.64(9)°, 25.33(11)°, and 21.0(11)° in 1–8, respectively. The torsions around the C1—C7 bond axis are also near to either 0° (torsion C2—C1—C7—C8) or 180° (torsion C6—C1—C7—C8) in terpenoid-like chalcones 1, 2, 4–8 and therefore reinforce the coplanarity feature between aromatic ring and the α,β-unsaturated carbonyl moiety. In terpenoid-like chalcone 3, the torsion C6—C1—C7—C8 is close to 0° while the torsion C2—C1—C7—C8 is near to 180°. This indicates that such coplanarity does also occur in this compound but it is present with a rotation of ca. 180° about the C8—C9 bond axis relative to all other terpenoid-like chalcones studied here (Fig. 2). The torsions around this bond measure near to 180° (O1—C9—C8—C7) or 0° (C7—C8—C9—C10) in 3, while there is inversion of their values in all other terpenoid-like chalcones, i.e., O1—C9—C8—C7 and C7—C8—C9—C10 are close to 0° and 180°, respectively (Table 2). Likewise, the torsions on the C9—C10 bond are similar in 2, 3 and 5, but their values are changed by ca. 180° in the other terpenoid-like chalcones (Table 2). Therefore, a twofold rotation on the C9—C10 bond axis can be also described for these compounds when compared to terpenoid-like chalcones 1, 4, 6–8 (Fig. 2). It is also noticeable the high coplanarity between aromatic ring and the α,β-unsaturated carbonyl moiety in compound 3 due to the negligible value of the angle between their least-square planes.

Close

Figure 2 Molecular overlay through carbonyl atoms and neighboring α-carbons of (A) compounds 1 (green) and 3 (blue); and (B) compounds 2 (yellow) and 3 (blue). The 180° rotations on the C8—C9 and C9—C10 bond axes converting one molecule conformation into another one are highlighted (circular arrows). (C) Selected torsions describing the conformations around C8—C9 and C9—C10 bonds found in the chalcone derivatives studied here.

Export to PPT

Nitro group is noticeably twisted only in compound 5, with its plane forming an angle of 38.41(7)° with the phenyl mean plane. In all other nitro-substituted terpenoid-like chalcones, these two moieties are considerably coplanar with angles between them of 12.9(4)°, 7.3(3)°, 7.8(3)°, and 4.82(17)° in 1–4, respectively. In addition, inspection of bond lengths indicates that electron delocalization does not encompass the molecule backbone of all chalcones (Table 3). Bond lengths of the α,β-unsaturated carbonyl moiety and of the bridges between this moiety and six-membered rings are in good agreement with expected values for pure single and double bonds.

3.2 Cytotoxic activity

All synthesized terpenoid-like-chalcones were evaluated for their in vitro cytotoxic activity against three human cancer cell lines (SF-295, HCT-116 and OVCAR-8). The antiproliferative activity assay results are shown in Tables 4 and 5.

As can be viewed in Table 4, terpenoid-like chalcones have almost entirely inhibited all cancer cell lines tested, except for compounds 4 and 6 inhibiting ca. 60% the SF-295 cell line and compound 5 inhibiting also ca. 60% the HCT-116 cell line. Therefore, a very broad spectrum of antiproliferative effects was observed for all compounds. In concentration of 50 μg/ml, the compounds displayed significant antiproliferative effects against SF-295, HCT-116 and OVCAR-8 cell lines. The relative potency of cell growth inhibition is shown in Table 5 in which as low the IC₅₀ value of a terpenoid-like chalcone as high its cytotoxic activity. It is noteworthy that compounds 1 and 5 possess antiproliferative effect against HCT-116 cells. In fact, compounds 1 and 2, both of them with para-nitro moiety and derived from α- and β-ionone, respectively, and compound 3, with meta-nitro group and derived from α-ionone, and compound 5, with ortho-nitro group and derived from β-ionone, were notably the most active terpenoid-like chalcones among all those tested. Their IC₅₀ values range from 2.88 to 10.84 μmol/L (compound 1), from 3.99 to 10.26 μmol/L (compound 2), from 4.33 to 6.20 μmol/L (compound 3), and from 2.70 to 8.85 μmol/L (compound 5). Compound 4 with meta-nitro group and derived from β-ionone was slightly less active than all other nitro-substituted analogs (IC₅₀ ranging from 5.22 to 20.28 μmol/L). This reveals the presence of one nitro moiety at any position of aromatic ring is decisive for antiproliferative activity of the terpenoid-like chalcones, whereas cyclohexene type does not affect much cytotoxic activity. Against other cancer lines, such as prostate (PC-3), breast (MCF), central nervous system (IMR32), and cervix (HeLa), the presence of one nitro group at para position of phenyl increases much the cytotoxicity of the β-ionone derived chalcone 2a (Sharma et al., 2013), which did also occur in another chalcone with two phenyl moieties (bearing three methoxy groups in the B-ring and one para-nitro group in the A-ring) (Bandgar et al., 2010). The presence of nitro moiety in the 2-position of phenyl rings has been likewise responsible to improve the anticancer activity of curcumin related compounds similar to chalcones (Bukhari et al., 2014a).

The inspection of IC₅₀ values does also show another interesting structure–activity relationship. There is a trend of decrease in cytotoxic activity with weak electron-withdrawing groups at phenyl ring. This has been reported previously for anticancer (Singh et al., 2014; Bandgar et al., 2010) and anti-inflammatory properties of chalcones (Bukhari et al., 2012). More precisely, the presence and strength of electron withdrawing groups in the phenyl ring of β-ionone derived chalcones have been found to be decisive for their cytotoxicity (Sharma et al., 2013) as well as in methoxychalcones (Singh et al., 2014; Bandgar et al., 2010). In agreement with these previous findings, the presence of either bromine or chlorine, which are less electron-withdrawing than nitro moiety, at para-position does decreases cytotoxic activity of terpenoid-like chalcones 6 and 7 (IC₅₀ ranging from 12.55 to 19.28 μmol/L for compound 6 and from 9.27 to 22.61 μmol/L for compound 7) relative to nitro derivatives 1 and 2. Similar IC₅₀ values have been found for related halogenated chalcones bearing two phenyl moieties bonded at 1,3-positions of prop-2-en-1-one (Dias et al., 2013). Chalcones with ortho-hydroxyphenyl group bonded to carbonyl carbon and one either meta or para halogenated (F, Cl or Br) phenyl ring bonded to β-carbon have shown IC₅₀ values in the 10.9–20.9 μmol/L range against colorectal carcinoma cell line HCT-116 (Dias et al., 2013). Also, ortho-hydroxychalcones with halogen substitution at para position were less active than those halogenated at meta position. Against the same cell line HCT-116, para-halogenated ortho-hydroxychalcones with Br and Cl had similar IC₅₀ values of 19.1 and 20.9 μmol/L (Dias et al., 2013), whereas these average values for our related compounds 6 and 7 were 19.23 and 9.27 μmol/L. This indicates that β-ionone and a further α,β-unsaturated set of carbons can increase anticancer activity from Br to Cl. Similar trend of increasing cytotoxicity as halogen substitution in the ortho-position of the phenyl ring bonded to β-carbon goes from Br to F was also observed in chalcones with 1,3-diphenyl moieties against a colon cell line (COLO-205) (Jain et al., 2014). Such an increasing was followed by the increase in the lipophilicity of the substituent in the para-position of phenyl ring bonded to carbonyl carbon from methoxy to phenoxy (Jain et al., 2014). When compared with the cytotoxic profile of compounds 6 and 7 against a colorectal cell line, it appears that this biological activity of chalcones increases as the lipophilicity of the moiety neighbor to carbonyl and as electron-withdrawing strength of one halogen in the phenyl ring attached at the other tail of prop-2-en-1-one, regardless of its position.

In our compounds, antiproliferative effect reduction is still highlighted in compound 8 present with para-methyl moiety (IC₅₀ ranging from 15.86 to 35.39 μmol/L). Indeed, this last compound was the less active terpenoid-like chalcone against all cancer cell lines.

Several articles show the anticancer potential of chalcones and their derivatives with selectivity for cancer cells, demonstrating the therapeutic potential of this chemical class (Jandial et al., 2014). Champelovier et al. (2013) showed selectivity of two chalcones (A and B) for the LN229 lineage (glioblastoma tumor cells) relative to the mononuclear cells from peripheral blood and skin fibroblasts. Both chalcones were able to cause intracellular oxidative stress, accumulation of cells in G2/M, besides chalcone A activating the caspase apoptotic pathway and chalcone B inducing cell death by autophagy without affecting proliferation of normal cells.

Tang et al. (2010) also showed that the compound flavokawain B, the chalcone kava, was effective in reducing the viability of cell lines DU145, PC3, LAPC4 and LNCaP, all strains of prostate cancer, with minimal cytotoxicity against normal cells from prostate epithelium and stroma. In cancer cells, this chalcone induces apoptosis by increasing the expression of pro-apoptotic proteins such as death receptor-5, Bim and Puma and by decreasing the expression of the apoptosis inhibitors such as XIAP and survivin.

In another work, the series of boronic-chalcone derivatives were synthesized and shown to be effective in inhibiting the proliferation of breast cancer cells at micromolar concentrations (MDA-MB-435, MDA-MB-231, Wt, MCF7), without affecting the viability and proliferation of normal mammary epithelial cells, MCF-10A and MCF-12A (Kumar et al., 2003). Furthermore, it was shown that indanocine, a chalcone-like compound, is an antimitotic and cytotoxic compound that blocks the polymerization of tubulin, inducing cell death by apoptosis in cancer cells resistant to multiple drugs at concentrations that are non-cytotoxic to non-carcinogenic cells (Leoni et al., 2000).

These literature data therefore reveal that besides synthetic chalcones having anticancer potential, many compounds of this class exhibit selectivity for tumor cells, which allow us to believe that the compounds reported in this study can also show selectivity for the cancer cells such as SF-295, HCT-116 and OVCAR-8 assessed here. So we can infer that the compounds derived from chalcones, as terpenoid-like chalcones, are promising for the development of new drugs to fight cancer.