ANTICANCER RESEARCH 37: 1091-1098 (2017) doi:10.21873/anticanres.11421 Quantitative Structure–Cytotoxicity Relationship of Chalcones HIROSHI SAKAGAMI1,2, YOSHIKO MASUDA2, MINEKO TOMOMURA2, SATOSHI YOKOSE3, YOSHIHIRO UESAWA4, NARUHIKO IKEZOE5, DAIKI ASAHARA5, KOICHI TAKAO5, TAISEI KANAMOTO6, SHIGEMI TERAKUBO6, HAJIME KAGAYA4, HIDEKI NAKASHIMA6 and YOSHIAKI SUGITA5 1Division of Pharmacology, 2Meikai Pharmaco-Medical Laboratory (MPL) and of Endodontics, Meikai University School of Dentistry, Sakado, Japan; 4Department of Clinical Pharmaceutics, Meiji Pharmaceutical University, Tokyo, Japan; 5Faculty of Pharmaceutical Sciences, Josai University, Sakado, Japan; 6Department of Microbiology, St. Marianna University School of Medicine, Kanagawa, Japan 3Division Abstract. Background: Fifteen chalcones were subjected to quantitative structure–activity relationship (QSAR) analysis based on their cytotoxicity and tumor specificity, in order to find their new biological activities. Materials and Methods: Cytotoxicity against four human oral squamous cell carcinoma cell lines and three oral mesenchymal cells was determined by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) method. Tumor specificity (TS) was evaluated by the ratio of the mean 50% cytotoxic concentration (CC50) against normal cells to that against tumor cell lines. Potency-selectivity expression (PSE) value was calculated by dividing TS by CC50 against tumor cells. Apoptosis markers were detected by western blot analysis. Physicochemical, structural and quantumchemical parameters were calculated based on the conformations optimized by force-field minimization. Results: Among 15 chalcone derivatives, (2E)-1-(2,4-dimethoxyphenyl)3-(4-methoxyphenyl)-2-propen-1-one had the highest TS and PSE values, comparable with those of doxorubicin and methotrexate, respectively. This compound also stimulated the cleavage of poly(ADP-ribose) polymerase and caspase-3. Chalone TS values were correlated with molecular shape and polarization rather than the types of substituted groups. None of the compounds had any anti-HIV activity. Conclusion: Chemical modification of the lead compound may be a potential choice for designing new types of anticancer drugs. This article is freely accessible online. Correspondence to: Hiroshi Sakagami, Division of Pharmacology, Department of Diagnostic and Therapeutic Sciences, Meikai University School of Dentistry, Sakado, Saitama 350-0283, Japan. Tel: +81 492792758, Fax: +81 492855171, e-mail: sakagami@dent.meikai.ac.jp Key Words: Chalcones, QSAR analysis, cytotoxicity, tumor selectivity, apoptosis induction, anti-HIV activity. Chalcone has a structure of 1,3-diaryl-2-propen-1-one in which the two aromatic rings are joined by a three-carbon α,β-unsaturated carbonyl system, representing a class of flavonoids that occur naturally in fruits and vegetables. Chalcones are also metabolic precursors of some flavonoids and isoflavonoids (1). Chalcones are promising lead antitumor/chemopreventive drugs due to three different activities: antioxidant, cytotoxic, and apoptosis induction (2). Several studies with murine xenograft models have shown that administration of chalcones significantly reduced the tumor volume by inducing apoptosis (3-10). The tumor specificity of chalcones has been reported in comparing sensitivity of hepatocarcinoma HepG2 cells to normal liver AML12 cells (11); osterosarcoma to bone marrow and smallintestinal epithelial cells (12); murine acute lymphoblastic leukemia cells L-1210 to normal human lymphocytes (13); and human prostate cancer cells PC3 and DU145 to normal human prostate epithelial cells (14). Although chalcones have been reported to induce apoptosis of human oral carcinoma cell line (HSC-3) (15) and cultured primary and metastatic oral cancer cell lines (16), the tumor specificity against these has not been investigated as far as we are aware of. In order to find new types of anticancer drugs active against human oral cancer, we first investigated the tumor specificity of 15 chalcone derivatives (Figure 1), using four human oral squamous cell carcinoma (OSCC) cell lines (Ca9-22, HSC-2, HSC-3, HSC-4) and three human normal oral cells (gingival fibroblast, HGF; periodontal ligament fibroblast, HPLF; pulp cells, HPC) as target cells, and then the apoptosis-inducing activity of the most active compound. The cytotoxicity data were used to perform the quantitative structure–activity relationship (QSAR) analysis. Since very few articles have been published on the antiviral activity of chalcones, we also investigated whether any of these compounds has any anti-human immunodeficiency virus (HIV) activity. 1091 ANTICANCER RESEARCH 37: 1091-1098 (2017) phenyl)-2-propen-1-one (5), (2E)-3-(4-chlorophenyl)-1-(2-hydroxyphenyl)-2-propen-1-one (6), (2E)-3-(4-bromophenyl)-1-(2-hydroxyphenyl)-2-propen-1-one (7), (2E)-1-(2-hydroxy-4-methoxyphenyl)3-(4-hydroxyphenyl)-2-propen-1-one (8), (2E)-1-(2-hydroxy-4methoxyphenyl)-3-phenyl-2-propen-1-one (9), (2E)-1-(2-hydroxy-4methoxyphenyl)-3-(4-methoxyphenyl)-2-propen-1-one (10), (2E)-3(3,4-dimethoxyphenyl)-1-(2-hydroxy-4-methoxyphenyl)-2-propen-1one (11), (2E)-3-(4-bromophenyl)-1-(2-hydroxy-4-methoxyphenyl)2-propen-1-one (12), (2E)-1-(4-methoxyphenyl)-3-(4-methoxyphenyl)-2-propen-1-one (13), (2E)-3-(2,4-dimethoxyphenyl)-1-(4methoxyphenyl)-2-propen-1-one (14), (2E)-1-(2,4-dimethoxyphenyl)-3-(4-methoxyphenyl)-2-propen-1-one (15) (structures shown in Figure 1) were synthesized by base-catalyzed condensation of the appropriate acetophenone with selected benzaldehyde derivatives according to previous methods (17). All compounds were dissolved in DMSO at 40 mM and stored at –20˚C before use. Cell culture. Human normal oral mesenchymal cells (HGF, HPLF, HPC), established from the first premolar tooth extracted from the lower jaw of a 12-year-old girl (18), and human OSCC cell lines [Ca9-22 (derived from gingival tissue); HSC-2, HSC-3, HSC-4 (derived from tongue)], purchased from Riken Cell Bank (Tsukuba, Japan), were cultured at 37˚C in DMEM supplemented with 10% heat-inactivated FBS, 100 units/ml, penicillin G and 100 μg/ml streptomycin sulfate under a humidified 5% CO2 atmosphere. HGF, HPLF and HPC cells at 10-18 population doubling levels were used in the present study. Figure 1. Structure of fifteen chalcones. Materials and Methods Materials. The following chemicals and reagents were obtained from the indicated companies: Dulbecco’s modified Eagle’s medium (DMEM), from GIBCO BRL, Grand Island, NY, USA; fetal bovine serum (FBS), 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), doxorubicin, azidothymidine, 2’,3’-dideoxycytidine from Sigma-Aldrich Inc., St. Louis, MO, USA; dimethyl sulfoxide (DMSO), dextran sulfate (molecular mass, 5 kDa) from Wako Pure Chem. Ind., Osaka, Japan; methotrexate from Nacalai Tesque, Inc., Kyoto, Japan; curdlan sulfate (molecular mass: 79 kDa) from Ajinomoto Co. Ltd., Tokyo, Japan. Culture plastic dishes and plates (96-well) were purchased from Becton Dickinson (Franklin Lakes, NJ, USA). Synthesis of test compounds. (2E)-1-(2-Hydroxyphenyl)-3-phenyl-2propen-1-one (1), (2E)-1-(2-hydroxyphenyl)-3-(4-hydroxyphenyl)-2propen-1-one (2), (2E)-1-(2-hydroxyphenyl)-3-(4-methoxyphenyl)2-propen-1-one (3), (2E)-3-(3,4-dimethoxyphenyl)-1-(2-hydroxyphenyl)-2-propen-1-one (4), (2E)-3-(4-fluorophenyl)-1-(2-hydroxy- 1092 Assay for cytotoxic activity. Cells were inoculated at 2.5×103 cells/0.1 ml in a 96-microwell plate (Becton Dickinson Labware, Franklin Lakes, NJ, USA).After 48 h, the medium was replaced with 0.1 ml of fresh medium containing different concentrations of single test compounds. Cells were incubated further for 48 h and the relative viable cell number was then determined by the MTT method (19). The relative viable cell number was determined by the absorbance of the cell lysate at 562 nm, using a microplate reader (Sunrise Rainbow RC-R; TECAN, Männedorf, Switzerland). Control cells were treated with the same amounts of DMSO and the cell damage induced by DMSO was subtracted from that induced by test agents. The concentration of compound that reduced the viable cell number by 50% (CC50) was determined from the dose– response curve and the mean value of CC50 for each cell type was calculated from duplicate assays. Calculation of tumor-selectivity index (TS). TS was calculated using the following equation: TS=mean CC50 against normal cells/mean CC50 against tumor cells [(D/B) in Table I]. Since both Ca9-22 and HGF cells were derived from the gingival tissue (20), the relative sensitivity of these cells was also compared [(C/A) in Table I]. We have confirmed that the TS value thus determined reflects the antitumor potential of test samples, although normal and tumor cells are derived from different tissues (mesenchymal or epithelial tissues, respectively) (21). We did not use human normal oral keratinocytes as controls, since doxorubin and 5-fluorouracil showed potent cytotoxicity against these epithelial cells by an as yet unidentified mechanism (19, 22, 23). Calculation of potency-selectivity expression (PSE). PSE was calculated using the following equation: PSE=TS/CC50 against tumor cells ×100 (24) [that is, (D/B2) ×100 (HGF, HPLF, HPC vs. Sakagami et al: QSAR of Chalcones Table I. Cytotoxic activity of 15 chalcones against human oral malignant and non-malignant cells. Each value represents the mean of duplicate determinations. CC50 (μM) Human oral squamous cell carcinoma cell lines (A) Compound Ca9-22 Human normal oral cells TS PSE (D/B2) (C/A2) ×100 ×100 HSC-2 HSC-3 HSC-4 (B) Mean SD (C) HGF HPLF HPC (D) Mean SD (D/B) (C/A) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 18.2 31.2 21.7 24.2 15.1 8.7 9.5 29.6 9.8 21.5 14.6 10.3 15.8 15.8 4.2 28.6 53.0 36.2 45.7 22.2 13.2 10.0 32.7 11.2 23.1 20.9 11.0 21.8 30.0 6.6 19.2 29.4 21.0 26.0 13.4 10.1 9.9 27.8 10.7 21.6 15.4 10.1 18.9 17.9 3.9 33.2 51.0 42.0 41.1 26.6 16.4 14.3 44.6 20.5 40.2 31.8 19.4 28.7 19.8 <3.1 24.8 41.1 30.2 34.2 19.3 12.1 10.9 33.6 13.0 26.6 20.7 12.7 21.3 20.8 <4.4 7.3 12.6 10.5 10.7 6.2 3.4 2.2 7.6 5.0 9.1 7.9 4.5 5.5 6.3 1.5 41.5 115.5 56.5 83.0 34.7 21.1 22.8 73.7 31.6 47.4 64.4 38.1 32.5 30.7 23.1 72.0 281.5 145.0 153.5 69.3 35.0 47.0 151.0 51.8 234.0 97.3 98.5 52.0 68.8 50.0 71.8 167.0 69.5 112.1 46.2 30.2 36.6 153.0 55.3 197.5 73.4 81.3 70.4 70.1 40.8 61.8 188.0 90.3 116.2 50.1 28.8 35.5 125.9 46.2 159.6 78.4 72.6 51.6 56.5 37.9 17.5 85.0 47.8 35.4 17.6 7.1 12.1 45.2 12.8 98.9 17.0 31.1 19.0 22.4 13.6 2.5 4.6 3.0 3.4 2.6 2.4 3.3 3.7 3.5 6.0 3.8 5.7 2.4 2.7 >8.6 2.3 3.7 2.6 3.4 2.3 2.4 2.4 2.5 3.2 2.2 4.4 3.7 2.1 1.9 5.6 10 11 10 10 13 20 30 11 27 23 18 45 11 13 >194 13 12 12 14 15 28 26 8 33 10 30 36 13 12 134 DXR MTX 0.089 10.7 <0.078 10.9 <0.078 <7.8 <0.078 <7.8 <0.08 <9.3 0.006 1.8 0.17 203.0 0.64 962.5 0.54 1000.0 0.45 721.8 0.25 449.7 >5.5 >77.8 1.9 18.9 >6865 >838 2102 177 HGF, Human gingival fibroblast; HPC, human pulp cells; HPLF, human periodontal ligament fibroblast; Ca9-22 (derived from gingival tissue), HSC-2, HSC-3 and HSC-4 (derived from tongue), oral squamous cell carcinoma cell lines; CC50, 50% cytotoxic concentration; DXR, doxorubicin; MTX, methotrexate. TS, tumor-selectivity; PSE, potency-selectivity expression. Ca9-22, HSC-2, HSC-3, HSC-4) and (C/A2) ×100 (HGF vs. Ca922 in Table II). Table II. Anti-HIV activity of chalcones and chemotherapeutic agents. Each value represents the mean of triplicate determinations. Estimation of CC50 values. Since the CC50 values had a distribution pattern close to a logarithmic normal distribution, we used the pCC50 (i.e., the −log CC50) for the comparison of the cytotoxicity between the compounds. The mean pCC50 values for normal cells and tumor cell lines were defined as N and T, respectively (19). Compound Calculation of chemical descriptors. The 3D-structure of each chemical structure (drawn by Marvin) was optimized by CORINA Classic (Molecular Networks GmbH, Germany) and force-field calculations (amber-10: EHT) in Molecular Operating Environment (MOE) version 2014.09 (Chemical Computing Group Inc., Quebec, Canada). The number of structural descriptors calculated from MOE and Dragon 7.0 (Kode srl., Pisa, Italy) after the elimination of overlapped descriptors were 295 and 2797, respectively. The following 12 Dragon descriptors and 4 MOE descriptors were significantly correlated with T, N and T-N. Dragon descriptors (25): (a) B10[O-O]: Presence/absence of O O at topological distance 10; (b) CATS3D_10_DA: CATS3D Donor-acceptor BIN 10 (10.000-11.000 Å); (c) F10[O-O]: Frequency of O-O at topological distance 10; (d) VE2_H2: average coefficient of the last eigenvector (absolute values) from reciprocal squared distance matrix (2D matrix-based descriptors); (e) L3m: 3rd component size directional WHIM index/weighted by mass (WHIM descriptors); (f) L3s: 3rd component size directional WHIM 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Positive controls Dextran sulfate (μg/ml) Curdlan sulfate (μg/ml) Azidothymidine 2’,3’-Dideoxycytidine CC50 (μM) EC50 (μM) SI 323.627 178.21 >400 387.87 80.73 200.04 34.44 38.78 205.35 191.73 369.64 32.54 235.92 142.54 >400 >400 >400 >400 >400 >400 >400 >400 >400 >400 >400 >400 >400 >400 >400 >400 <1 <1 ><1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 ><1 232.68 >1000 53.004 1858.629 0.777 0.172 0.026 1.113 300 >5805 2017 1670 CC50, 50% Cytotoxic concentration; EC50, 50% effective concentration; SI: selectivity index (=CC50/EC50). 1093 ANTICANCER RESEARCH 37: 1091-1098 (2017) index/weighted by atomic ionization state (WHIM descriptors); (g) HATS6p: leverage-weighted autocorrelation of lag 6/weighted by polarizability (GETAWAY descriptors); (h) R5v+: R maximal autocorrelation of lag 5/weighted by van der Waals volume (GETAWAY descriptors); (i) R6p R autocorrelation of lag 6/weighted by polarizability (GETAWAY descriptors); (j) R6v: R autocorrelation of lag 6/weighted by van der Waal’s volume (GETAWAY descriptors); (k) RDF010s: Radial distribution function - 010/weighted by atomic ionization state (RDF descriptors); (l) RDF035u: Radial distribution function - 035/unweighted (RDF descriptors); MOE descriptors: (m) vsurf_IW6: Hydrophilic integy moment 6 in the vsurf_ descriptors which are similar to the VolSurf descriptors (26); (n) h_logS: Log of the aqueous solubility (mol/L) using a 7 parameter model based on Hueckel theory (27); (o) PEOE_VSA-6: Sum of vi where qi is less than –0.30 in the partial equalization of orbital electronegativities (PEOE) method of calculating atomic partial charges (27); (p) Q_VSA_PNEG: Total negative polar van der Waals surface area (28). Western blot analysis. The cells were washed with PBS and processed for western blot analysis, as described previously (29). Antibodies against cleaved caspase-3 (Cell Signaling Technology Inc., Beverly, MP, USA), poly(ADP-ribose) polymerase (PARP) (Cell Signaling Technology Inc.) and glyceraldehyde 3-phosphate dehydrogenase (GAPDH; Trevigen, Gaithersburg, MD, USA) were used as primary antibodies. As secondary antibodies, we used αrabbit IgG (DAKO Japan) antibodies which were conjugated with horseradish peroxidase. Assay for anti-HIV activity. HTLV-I-carrying human T-cell line MT4 cells, highly sensitive to human immunodeficiency virus-1 (HIV1), were infected with HIV-1IIIB at a multiplicity of infection of 0.01. HIV- and mock-infected (control) MT-4 cells were incubated for 5 days with different concentrations of samples and the relative viable cell number was determined by the MTT assay. The CC50 and 50% effective concentration (EC50) were determined from the dose–response curve for mock-infected and HIV-infected cells, respectively (30). All data represent the mean values of triplicate measurements. The anti-HIV activity was evaluated by selectivity index (SI) (=CC50/EC50). Statistical treatment. The relation among cytotoxicity, tumor specificity index, anti-UV activity and chemical descriptors was investigated using simple regression analyses by JMP Pro version 12.2.0 (SAS Institute Inc., Cary, NC, USA). The significance level was set at p<0.05. Results Cytotoxicity. A total of 15 chalcone derivatives (Figure 1) generally showed higher cytotoxicity against human OSCC lines (Ca9-22, HSC-2, HSC-3, HSC-4) (mean CC50=4.441.1 μM, mean 21.7 μM) (B) than against human mesenchymal normal oral cells (HGF, HPLF and HPC) (CC50=28.8-188.0 μM, mean 80.0 μM), yielding an averaged TS of 3.7 (Table I). Among them, compounds 10, 12 and 15 had higher TS (5.7-8.6) than other compounds, comparable to that of anticancer drugs, doxorubicin (5.5). When tumor selectivity was calculated using cells both 1094 Figure 2. Apoptosis induction in HSC-2 human oral squamous cell carcinoma cell line by compound 15. derived from gingival tissue (Ca9-22 vs. HGF), 14 compounds 1-13 and 15 had much higher TS (2.1-5.6), exceeding that of doxorubicin (1.9). Compound 15 had the highest TS values in both cases. In order to identify compounds which have both good potency and are selectively toxic to neoplasms, the PSE values for the compounds were calculated. When all three normal cells (HGF, HPLF and HPC) and all four OSCC cell lines (Ca9-22, HSC-2, HSC-3 and HSC-4) were used, doxorubicin had the highest PSE value (>6,865), followed by methotrexate > compound 15 > compounds 1-14 (Table I). When HGF and Ca9-22 cells (both derived from gingival tissues) were used, the same pattern was found. Compound 15 had the highest PSE value among the 15 chalcones, approaching that of methotrexate. Western blot analysis demonstrated that compound 15 stimulated the cleavage of PARP and caspase-3, suggesting the induction of apoptosis (Figure 2). Anti-HIV activity of chalcones. In contrast to popular antiHIV agents (dextran sulfate, curdlan sulfate, azidothymidine, 2’,3’-dideoxycytidine) (SI=300-5,805), none of the chalcones protected cells from the cytopathic effect of HIV infection (SI<1) (Table II). Based on these data, the subsequent QSAR analysis was focused on the cytotoxicity of chalcones. Computational analysis. We next performed the QSAR analysis of chalcone derivatives in regards to their cytotoxicity against tumor cells and normal cells. Among a total of 3,092 descriptors (295 MOE and 2797 Dragon descriptors), 16 descriptors described below correlated well with cytotoxicity and tumor specificity. Cytotoxicity of chalcones against human OSCC cell lines was correlated with HATS6p (polarizability) (r2=0.541, p=0.0018), vsurf_IW6 (hydrophilic interaction energy moment 6) (r2=0.537, p=0.0019), R6v (van der Waal’s volume) (r2=0.524, p=0.0023), R6p (polarizability) (r2=0.491, p=0.0036), h_logS Sakagami et al: QSAR of Chalcones Figure 3. Determination of correlation coefficient between chemical descriptors and cytotoxicity of chalcones against tumor cells. The mean values of pCC50 (i.e., the −log of the concentration causing 50% cytotoxicity) for tumor cell lines were defined as T. (aqueous solubility) (r2=0.465, p=0.0051) and RDF035u (spherically averaged information on the atomic correlation, unweighted) (r2=0.455, p=0.0058) (Figure 3). Cytotoxicity of chalcones against human normal oral mesenchymal cells was correlated with CATS3D_10_DA (donor-acceptor BIN) (r2=0.732, p<0.0001), RDF010s (atomic ionization state) (r2=0.696, p=0.0001), Q_VSA_PNEG (total negative polar van der Waals surface area) (r2=0.633, p=0.0004), PEOE_VSA-6 (atomic partial charges) (r2=0.633, p=0.0004), B10[O-O] (presence/absence of O-O at topological distance 10) (r2=0.620, p=0.0005) and F10[O-O] (frequency of O-O at topological distance 10) (r2=0.620, p=0.0005) (Figure 4). Tumor specificity of chalcones was correlated with R6p (polarizability) (r2=0.601, p=0.0007), R5v+(van der Waal’s volume) (r2=0.598, p=0.0007), L3m (mass) (r2=0.581, p=0.0009), VE2_H2 (average coefficient of the last eigenvector from reciprocal squared distance matrix) (r2=0.575, p=0.0010), L3s (atomic ionization state) (r2=0.565, p=0.0012) and HATS6p (polarizability) (r2=0.563, p=0.0013) (Figure 5). Discussion The present study demonstrated that 15 chalcones showed relatively higher cytotoxicity against four OSCC cell lines compared to that against human normal oral mesenchymal oral cells; among them, compound 15 had the highest TS and PSE values, although this is not a new compound. It should be noted that the TS value of 15 was comparable with that of doxorubicin, and the PSE value of 15 was comparable with that of methotrexate (Table I). It is ideal to use human epithelial cells as control normal cells in comparison with OSCC cell lines. However, we recently found that doxorubicin induced apoptosis in human oral keratinocytes (i.e. loss of cell surface microvilli, chromatin condensation, nuclear fragmentation, caspase-3 activation) at the concentration that affected the viability of OSCC cell lines (31). Until the mechanism of keratinocyte toxicity is clarified and a preventive method is explored, the use of human oral mesenchymal cells rather than normal epithelial cells may be the only choice for us to use in comparison with tumor cells. We calculated the possible contribution of substituted groups to the expression of cytotoxicity against OSCC cell lines and normal oral mesenchymal cells and tumor-specificity (Table III). Most of the substituents listed did not affect these activities (p=0.1067-0.9465) except for hydroxyl group (p=0.0167) or oxygen (p=0.0248) at R5 in determining cytotoxicity against normal cells. These data, suggest that tumor specificity of chalcones was rather correlated with molecular structure and polarization (Figure 5). We also found that correlated parameters differed between tumor cells and normal cells. For example, R6P (which represents polarizability) is correlated with cytotoxicity against tumor cells (Figure 3) and with tumor selectivity (Figure 5), but not 1095 ANTICANCER RESEARCH 37: 1091-1098 (2017) Figure 4. Determination of correlation coefficient between chemical descriptors and cytotoxicity of chalcones against normal cells. The mean values of pCC50 (i.e., the −log of the concentration causing 50% cytotoxicity) for normal cells were defined as N. Figure 5. Determination of correlation coefficient between chemical descriptors and tumor specificity of chalcones (defined as the difference between the −log of the concentration causing 50% cytotoxicity in tumor cells and that for normal cells (T-N). 1096 Sakagami et al: QSAR of Chalcones Table III. Substituted groups that affect the cytotoxicity against OSCC cell lines (T) and normal oral mesenchymal cells (N) and tumor specificity (T-N). T T T T T T T T N N N N N N N N T-N T-N T-N T-N T-N T-N T-N T-N Factor p-Value R1 R2Sbst R2OH R4 R5Sbst R5OH R5O R5X R1 R2Sbst R2OH R4 R5Sbst R5OH R5O R5X R1 R2Sbst R2OH R4 R5Sbst R5OH R5O R5X 0.3433 0.7795 0.146 0.4231 0.9008 0.0851 0.2186 0.2054 0.9465 0.5165 0.2111 0.3247 0.5449 0.0167 0.0248 0.0538 0.1067 0.1729 0.6177 0.9208 0.5252 0.6722 0.3378 0.5744 so with cytotoxicity against normal cells (Figure 4). These data indicate that an increase of polarizability of chalcones may increase their antitumor potential. The present study demonstrated that 15 chalcones did not have any anti-HIV activity. This finding is not contradictory with recent reports that chalcones exert anti-HIV activity partially or in a very narrow range of concentrations (32, 33). In conclusion, compound 15 is a potential lead compound for synthesizing more potent compounds targeted to OSCC cells. Conflicts of Interest We wish to confirm that there are no known conflicts of interest associated with this publication and there has been no significant financial support for this work that could have influenced its outcome. Acknowledgements This work was partially supported by KAKENHI from the Japan Society for the Promotion of Science (JSPS) (15K08111, 16K11519). The annual license of the statistical software, JMP Pro, was supported by the grant-in-aid of the oncology specialists promotion program by the Ministry of Education, Culture, Sports, Science and Technology, Japan. References 1 Das M and Manna K: Chalcone scaffold in anticancer armamentarium: a molecular insight. J Toxicol 2016: 7651047, 2016. 2 León-González AJ, Acero N, Muñoz-Mingarro D, Navarro I and Martín-Cordero C: Review. Chalcones as promising lead compounds on cancer therapy. Curr Med Chem 22(30): 34073425, 2015. 3 Wang LH, Li HH, Li M, Wang S, Jiang XR, Li Y, Ping GF, Cao Q, Liu X, Fang WH, Chen GL, Yang JY and Wu CF: SL4, a chalcone-based compound, induces apoptosis in human cancer cells by activation of the ROS/MAPK signalling pathway. Cell Prolif 48(6): 718-728, 2015. 4 Jiang C, Wang Q, Xu Z, Li WS, Chen C, Yao XQ and Liu FK: Cyclooxygenase-2 knockdown using retinoic acid chalcone (RAC), a promising therapeutic strategy for colon cancer. Am J Cancer Res 5(6): 2012-2021, 2015. 5 Loch-Neckel G, Bicca MA, Leal PC, Mascarello A, Siqueira JM and Calixto JB: In vitro and in vivo anti-glioma activity of a chalcone-quinoxaline hybrid. Eur J Med Chem 90: 93-100, 2015. 6 Singh N, Sarkar J, Sashidhara KV, Ali S and Sinha S: Antitumour activity of a novel coumarin-chalcone hybrid is mediated through intrinsic apoptotic pathway by inducing PUMA and altering BAX/BCL-2 ratio. Apoptosis 19(6): 1017-1028, 2014. 7 Zhang Y, Srinivasan B, Xing C and Lü J: A new chalcone derivative (E)-3-(4-methoxyphenyl)-2-methyl-1-(3,4,5-trimethoxyphenyl)prop-2-en-1-one suppresses prostate cancer involving p53-mediated cell cycle arrests and apoptosis. Anticancer Res 32(9): 3689-3698, 2012. 8 Lin E, Lin WH, Wang SY, Chen CS, Liao JW, Chang HW, Chen SC, Lin KY, Wang L, Yang HL and Hseu YC: Flavokawain B inhibits growth of human squamous carcinoma cells: Involvement of apoptosis and cell cycle dysregulation in vitro and in vivo. J Nutr Biochem 23(4): 368-378, 2012. 9 Zi X and Simoneau AR: Flavokawain A, a novel chalcone from kava extract, induces apoptosis in bladder cancer cells by involvement of BAX protein-dependent and mitochondriadependent apoptotic pathway and suppresses tumor growth in mice. Cancer Res 65(8): 3479-3486, 2005. 10 Hayashi A, Gillen AC and Lott JR: Effects of daily oral administration of quercetin chalcone and modified citrus pectin on implanted colon-25 tumor growth in Balb-c mice. Altern Med Rev 5(6): 546-552, 2000. 11 Kuete V, Nkuete AH, Mbaveng AT, Wiench B, Wabo HK, Tane P and Efferth T: Cytotoxicity and modes of action of 4’hydroxy-2’,6’-dimethoxychalcone and other flavonoids toward drug-sensitive and multidrug-resistant cancer cell lines. Phytomedicine 21(12): 1651-1657, 2014. 12 Ji T, Lin C, Krill LS, Eskander R, Guo Y, Zi X and Hoang BH: Flavokawain B, a kava chalcone, inhibits growth of human osteosarcoma cells through G2/M cell cycle arrest and apoptosis. Mol Cancer 12: 55, doi:10.1186/1476-4598-12-55, 2013. 13 Pedrini FS, Chiaradia LD, Licínio MA, de Moraes AC, Curta JC, Costa A, Mascarello A, Creczinsky-Pasa TB, Nunes RJ, Yunes RA and Santos-Silva MC: Induction of apoptosis and cell cycle arrest in L-1210 murine lymphoblastic leukaemia cells by (2E)3-(2-naphthyl)-1-(3’-methoxy-4’-hydroxy-phenyl)-2-propen-1one. J Pharm Pharmacol 62(9): 1128-1136, 2010. 1097 ANTICANCER RESEARCH 37: 1091-1098 (2017) 14 Yun JM, Kweon MH, Kwon H, Hwang JK and Mukhtar H: Induction of apoptosis and cell cycle arrest by a chalcone panduratin A isolated from Kaempferia pandurata in androgenindependent human prostate cancer cells PC3 and DU145. Carcinogenesis 27(7): 1454-1464, 2006. 15 Hseu YC, Lee MS, Wu CR, Cho HJ, Lin KY, Lai GH, Wang SY, Kuo YH, Kumar KJ and Yang HL: The chalcone flavokawain B induces G2/M cell-cycle arrest and apoptosis in human oral carcinoma HSC-3 cells through the intracellular ROS generation and downregulation of the AKT/p38 MAPK signaling pathway. J Agric Food Chem 60(9): 2385-2397, 2012. 16 Lee YM, Jeong GS, Lim HD, An RB, Kim YC and Kim EC: Isoliquiritigenin 2’-methyl ether induces growth inhibition and apoptosis in oral cancer cells via heme oxygenase-1. Toxicol In Vitro 24(3): 776-782, 2010. 17 Devakaram R, Black DS, Andrews KT, Fisher GM, Davis RA and Kumar N: Synthesis and antimalarial evaluation of novel benzopyrano[4,3-b]benzopyran derivatives. Bioorg Med Chem 19: 5199-5206, 2011. 18 Kantoh K, Ono M, Nakamura Y, Nakamura Y, Hashimoto K, Sakagami H and Wakabayashi H: Hormetic and anti-radiation effects of tropolone-related compounds. In Vivo 24: 843-852, 2010. 19 Sakagami H, Uesawa Y, Ishihara M, Kagaya H, Kanamoto T, Terakubo S, Nakashima H, Takao K and Sugita Y: Quantitative structure–cytotoxicity relationship of oleoylamides. Anticancer Res 35: 5341-5355, 2015. 20 Horikoshi M, Kimura Y, Nagura H, Ono T and Ito H: A new human cell line derived from human carcinoma of the gingiva. I. Its establishment and morphological studies. Jpn J Oral Maxillofac Surg 20: 100-106, 1974 (in Japanese). 21 Suzuki R, Matsuo S, Sakagami H, Okada Y and Shirataki Y: Search of new cytotoxic crude materials against human oral squamous cell carcinoma using NMR metabolomics. Anticancer Res 34(8): 4117-4120, 2014. 22 Sakagami H, Shimada C, Kanda Y, Amano O, Sugimoto M, Ota S, Soga T, Tomita M, Sato A, Tanuma S, Takao K and Sugita Y: Effects of 3-styrylchromones on metabolic profiles and cell death in oral squamous cell carcinoma cells. Toxocol Re 2: 1281-1290, 2015. 23 Uesawa Y, Sakagami H, Kagaya H, Yamashita M, Takao K and Sugita Y: Quantitative structure-cytotoxicity relationship of 3benzylidenechromanones. Anticancer Res 36(11): 5803-5812, 2016. 1098 24 Das S, Das U, Sakagami H, Umemura N, Iwamoto S, Matsuta T, Kawase M, Molnar J, Serly J, Gorecki DKJ and Dimmock JR: Dimeric 3,5-bis(benzylidene)-4-piperidones : A novel cluster of tumour-selective cytotoxins possessing multidrug-resistant properties. Eur J Med Chem 51: 193-199, 2012. 25 https://chm.kode-solutions.net/products_dragon_descriptors.php 26 Cruciani G, Crivori P, Carrupt P-A and Testa B: Molecular Fields In Quantitative Structure-Permeation Relationships: The VolSurf Approach. J Mol Struct (Theochem) 503: 17-30, 2000. 27 Labute P: MOE h_mr, h_logP and h_logS Models unpublished. Source code in $MOE/lib/svl/quasar.svl/q_eht.svl (2015). 28 Gasteiger J and Marsili M: Iterative partial equalization of orbital electronegativity – a rapid access to atomic charges. Tetrahedron 36: 3219, 1980. 29 Suzuki R, Matsushima Y, Okudaira N, Sakagami H and Shirataki Y: Cytotoxic components against human oral squamous cell carcinoma isolated from Andrographs paniculata. Anticancer Res 36(11): 5931-5935, 2016. 30 Nakashima H, Murakami T, Yamamoto N, Sakagami H, Tanuma S, Hatano T, Yoshida T and Okuda T: Inhibition of human immunodeficiency viral replication by tannins and related compounds. Antiviral Res 18(1): 91-103, 1992. 31 Sakagami H, Okudaira N, Masuda Y, Amano O, Yokose S, Kanda Y, Suguro M, Natori T, Oizumi H and Oizumi T: Induction of apoptosis in human oral keratinicyte by doxorubicin. Anticancer Res 37(3): in press, 2017. 32 Pan W, Liu K, Guan Y, Tan GT, Hung NV, Cuong NM, Soejarto DD, Pezzuto JM, Fong HH and Zhang H. Bioactive compounds from Vitex leptobotrys. J Nat Prod 77(3): 663-667, 2014. 33 Cole AL, Hossain S, Cole AM and Phanstiel O 4th: Synthesis and bioevaluation of substituted chalcones, coumaranones and other flavonoids as anti-HIV agents. Bioorg Med Chem 24(12): 2768-2776, 2016. Received December 21, 2016 Revised January 31, 2017 Accepted February 1, 2017