Microwave-assisted preparation, structural characterization, lipophilicity, and anti-cancer assay of some hydroxycoumarin derivatives

Abstract A new series of hydroxycoumarin derivatives has been synthesized using conventional synthesis. The syntheses were accelerated by microwave assistance. Yields in both cases were comparable (59–69 %). The structures were established by 1H and 13C NMR spectroscopy and high-resolution mass spectrometry. Five compounds (5-hydroxy-4,7-dimethylcoumarin, 6-acetyl-5-hydroxy-4,7-dimethylcoumarin, 4-(cyanomethoxy)chromen-2-one, 5-(cyanomethoxy)-4,7-dimethylchromen-2-one, and 6-acetyl-5-(cyanomethoxy)-4,7-dimethylchromen-2-one) were assayed for anti-cancer activity. For all presented coumarin derivatives, lipophilicity was measured using reversed-phase TLC in different eluent systems with standardization. In addition, the crystal structure of 6-acetyl-5-hydroxy-4,7-dimethylcoumarin has been solved by X-ray structure analysis of single crystals. Graphical abstract


Introduction
The cytotoxic activity of plant-derived hydroxycoumarins and their derivatives, synthetic analogues, has been reviewed in a number of studies [1]. The anti-tumor activity of 7-hydroxycoumarin (umbelliferone) and 4-hydroxycoumarin against human tumor cell lines, including prostate cancer, malignant melanoma, and metastatic renal cell carcinoma have been reported [2][3][4]. Geiparvarin, a naturally occurring compound bearing an umbelliferone residue has been shown to possess a significant inhibitory activity against a variety of cell lines including sarcoma 180, Lewis lung carcinoma, P-388 lymphotic leukemia, and Walker 256 carcinosarcoma [5,6].
Physicochemical properties have been widely applied to guide absorption, distribution, metabolism, and elimination (ADME) properties and pharmacological activities of discovery molecules, from small synthetic [9] to large natural or semisynthetic derivatives [10]. Lipophilicity has been one of the most used physicochemical properties useful in drug design, since it considerably influences bioavailability of compounds. Lipophilicity, as expressed by the logarithm of octanol/water partition coefficient logP (or distribution coefficient logD for ionizable compounds), plays an important role in ADME properties, as well as in the pharmacodynamic and toxicological profile of drugs [11,12]. For lipophilicity assessment, partition chromatographic techniques and, in particular, reversedphase HPLC offer several practical advantages compared to the traditional shake-flask method. These include speed, reproducibility, broader dynamic range, online detection, insensitivity to impurities or degradation products, and reduced sample handling and sample sizes [13,14].
7-Hydroxycoumarins and 4-hydroxycoumarin derivatives have been the targets of our research [15][16][17][18]. A similar system, 5-hydroxycoumarin which is a promising target as a scaffold for new therapeutic agents was studied much less intensively, and the library of derivatives of 5-hydroxycoumarin is definitely less prominent. To fill the gap in this study, we present a new series of 4-and 5-cyanomethoxy derivatives of coumarin ( Fig. 1) with an expected anti-cancer activity which have been synthesized using microwave irradiation and conventional synthesis and characterized by various methods.
The coumarins 2, 3, and 9-11 were resynthesized according to previously published papers [19][20][21][22]. The microwave-assisted synthesis has been used because it has many advantages over traditional methods, which are operational simplicity, good yields, short reaction times, and easy workup procedures [23]. For all presented compounds, lipophilic properties were determined experimentally and by calculating the partition coefficient by reversed-phase TLC technique in different solvent systems with measurement standardization. To complete the structural characterization, we also report the results of the X-ray crystallographic studies for 6-acetyl-5-hydroxy-4,7-dimethylcoumarin (3).
The O-alkylation reaction of starting coumarins 1-3 with the alkylating agent chloroacetonitrile under microwave irradiation afforded the products within few minutes. The syntheses were carried out under reflux in acetone or at the temperature 130-140°C in 1-methyl-2-pyrrolidone, using anhydrous potassium carbonate as a base. The synthesis under microwave irradiation gave the same products 4-6. The yields were a little higher or comparable with the yields of syntheses carried out in the conventional way. The major achievement of this procedure was the considerable reduction of reaction times: from 6 h to 12 min for compound 4, from 10 h to 12 min for compound 5, and from 12 h to 15 min for compound 6 ( Table 1). Spectroscopic data ( 1 H, 13 C NMR, and mass) confirmed the structures of all products.
7-Acetoxy-4-methylcoumarin (10) was synthesized by acetylation of compound 8 with acetic anhydride in pyridine in the presence of catalytic amount of (dimethylamino)pyridine (DMAP) [21]. 6-Acetyl-7hydroxy-4-methylcoumarin (11) was obtained from 7-acetoxy-4-methylcoumarin (10) by heating with anhydrous aluminum trichloride [22]. The structures of coumarins 1-3 and 7-11 were established on the basis of their spectral data ( 1 H, 13 C NMR, and mass) and comparison of their melting points and spectral data with those reported in the literature. The lipophilicity descriptors on analyzed compounds on which the entire classification has been performed are listed in Table 2. The lipophilicity descriptors on analyzed  compounds on which the entire classification has been performed are listed in Table 3.
The logP values of the tested compounds were determined by the series of standards with known logP value (Table 4) [24] using procedures described previously [18]. The regression correlation coefficients corresponding to Eq. 2 (see ''Experimental'', ''Lipophilicity'') are having a good linearity with R M values through the molar fraction of organic modifier in mobile phase. The correlation coefficients (R 2 ) were between 0.9384 and 0.9997. The best results with the highest linearity were obtained for isopropanol/water mobile phase, which suggest that this system is appropriate for the further analysis of this group of compounds (Tables 3, 4).
The linear relationship between known logP values and experimental R M0 parameter for standards had been used for calculating experimental logP values for tested compounds. All results are shown in Tables 5 and 6.
Analysis of the data obtained revealed that even small structural changes can produce substantial differences in lipophilicity of derivatives investigated, which is in accordance with previous studies on this class of compounds [25]. The lowest lipophilicity was found for compounds 7 (umbelliferone), 8 (7-hydroxy-4-methylcoumarin), 10 (7-acetoxy-4-methylcoumarin), and 4 (4-(cyanomethoxy)chromen-2-one). Such effect may be connected to the presence of the free hydroxyl groups (compounds 7 and 8), acetoxy group (compound 11), or cyanomethoxy group (compound 4). Presence of both methyl and acetyl moieties (compounds 9 and 10; 2 and 3) increases lipophilicity. It is interesting that for 8-acetyl-7hydroxycoumarin (9) and 6-acetyl-7-hydroxycoumarin (10) the lipophilicity differs by only ca. 0.2. It might suggest that substituents in position 6 would slightly increase lipophilicity when substituents in position 8 would lower it. In the group of 5-hydroxycoumarin derivatives, the presence of cyanomethoxy moiety lowers the lipophilicity in respect of compounds with free hydroxyl group (derivatives 2 and 5; 3 and 6).
5-Hydroxy-4,7-dimethylcoumarin (2), 6-acetyl-5hydroxy-4,7-dimethylcoumarin (3), 4-(cyanomethoxy)chromen-2-one (4), 5-(cyanomethoxy)-4,7-dimethylchromen-2-one (5), and 6-acetyl-5-(cyanomethoxy)-4,7dimethylchromen-2-one (6) were accepted for cytotoxicity testing. Initially they had been evaluated in the two-cell line panel consisting of the B16-F10 (melanoma) and DU145 (prostate). Compounds 2 and 3 were slightly active, while compounds 4-6 were considered inactive in the primary screen. Considering the cytotoxicity of the 4-,  5-, 7-cyanomethoxy derivatives of coumarins, only the introduction of cyanomethoxy group in 7-position seems to have pronounced effect [8]. Considering the lipophilicity in correlation with cytotoxic effect suggest that slight increase of lipophilicity (from 5-cyanomethoxy derivatives to 5-hydroxy derivatives lipophilicity differs by ca. 0.59), decreased the cytotoxicity of these compounds. The molecular structure of 6-acetyl-5-hydroxy-4,7dimethylcoumarin (3) in solid state was analyzed by single-crystal X-ray diffraction technique. Only for this compound we have obtained suitable crystals. Compound 3 crystallizes in the P2 1 /c space group. Crystal data and structure refinement parameters for 3 are collected in Table 7. Thermal ellipsoid plot and packing diagrams are presented in Figs. 2 and 3, respectively. The independent part of the crystal lattice of presented compound consists of two chemical moieties numbered A and B. While the B molecule is almost flat in the A one chromene skeleton is slightly twisted along the longer axis of the molecule. This is visible in Fig. 4 presenting overlay of molecules A and B calculated in Mercury program [26].
In both moieties methyl group (C9) is visibly deflected from hydroxyl group. The angle C4A-C4-C9 is comparable in both cases and yields 123.3(2)°and 122.1(2)°in A and B molecules, respectively. There are no strong intermolecular interactions in the crystal structure of 3. This is due to the fact that OH group (O10) is engaged in intramolecular hydrogen bond with carbonyl oxygen atom (O13). Distance O…O is equal to 2.435(2) Å and 2.430(2) Å in molecules A and B, respectively. In the crystal lattice molecules are forming stacks along [100] direction. In each stack molecules A and B are located alternately and average intermolecular distance is equal ca. 3.4 Å (Fig. 5).
The mixture was refluxed (acetone) or heated at the temperature 130-140°C (1-methyl-2-pyrrolidone) and monitored by TLC on silica-gel plates (eluent CHCl 3 -MeOH 10:0.25). After completion of the reactions as indicated by TLC, the mixture was poured into the flask with 50 cm 3 water and ice and stirred for 30 min. The precipitate was filtered out and dried. The analytical samples were crystallized from ethanol.

4-(Cyanomethoxy
Chromatographic conditions TLC analysis was performed on HPTLC silica gel, 10 9 10 cm, RP-18 WF 254s glass plates (Merck, Germany). Samples and standards solutions were applied as a 1-mm spot onto plate using 0.5-mm 3 thin glass capillary tube (Camag, Switzerland). The distance between each application was 10 and 10 mm distance from low edge of the plate. TLC-image analysis method Air-dried in room temperature HPTLC plate was visualized under UV-light at 254 and 366 nm (TLC-Visualiser, Camag) and saved as lossless JPEG file. The image was opened with winCATS software (Camag, Switzerland) and the R F parameter was calculated.
Chromatographic parameters R M values for tested compounds ( Table 2) and standards (Table 3) were calculated from the experimental R F by the use of the Eq. 1.
The calculated R M values were extrapolated to 0 % organic modifier concentration (R M0 ) by use of Eq. 2: where u (molar fraction) describes concentration of organic modifier in mobile phase, S is a slope of the trend line and R M0 is extrapolated value of the lipophilicity for 100 % of water.

Cytotoxicity against cancer cell lines
Reagents Synthesized substances were dissolved in DMSO (Sigma-Aldrich) to obtain 20 mM stock solutions and kept in 4°C prior to use.

Cell culture
The human prostate cancer cells DU145 and mouse melanoma cells B16F10 were maintained in humidified incubator containing 5 % CO 2 at 37°C. DU145 cells were cultured in RPMI medium and B16F10 were cultured in DMEM medium. The media were supplemented with 10 % of fetal calf serum (FCS) and 1 % of antibiotic antimycotic (Sigma). Cells were passaged every 2-3 days. All cell lines were obtained from ATCC (The Global Bioresource Center).
Cytotoxicity assays PrestoBlue (test based on resazurin) was used for analysis cytotoxicity/cytostatic effects. DU145 and B16F10 cells were seeded on 96-well plate (10 9 10 3 cells per well). Examined substances were added to culture after overnight cell incubation. Two concentrations of substances (10 and 100 lM) were used. Control cells and control cells with solvent (DMSO) were applied. After 48 h PrestoBlue test was performed.

Crystallography
The X-ray measurement of 3 was performed at 100(2) K on a Bruker D8 Venture Photon100 diffractometer equipped with a TRIUMPH monochromator and a MoKa fine focus sealed tube (k = 0.71073 Å ). A total of 660 frames were collected with Bruker APEX2 program [27]. The frames were integrated with the Bruker SAINT software package [28] using a narrow-frame algorithm. The integration of the data using a monoclinic unit cell yielded a total of 25,586 reflections to a maximum h angle of 26.50°(0.80 Å resolution), of which 4,330 were independent (average redundancy 5.909, completeness = 100.0 %, R int = 4.35 %, R sig = 2.61 %) and 3,378 (78.01 %) were greater than 2r(F 2 ). The final cell constants of a = 14.0681(8) Å , b = 7.5749(4) Å , c = 19.6802(11) Å , b = 95.465(2)°, volume = 2,087.7(2) Å 3 , are based upon the refinement of the XYZ-centroids of 8,583 reflections above 20 r(I) with 4.842°\ 2h \ 54.16°. Data were corrected for absorption effects using the multi-scan method (SADABS) [29]. The ratio of minimum to maximum apparent transmission was 0.837. The calculated minimum and maximum transmission coefficients (based on crystal size) are 0.9730 and 0.9913.
The structure was solved and refined using SHELXTL software package [30, 31] using the space group P 1 21/c 1, with Z = 8 for the formula unit C 13 H 12 O 4 . The final anisotropic full-matrix least-squares refinement on F 2 with 315 variables converged at R1 = 5.12 %, for the observed data and wR2 = 13.50 % for all data. The goodness-of-fit was 1.121. The largest peak in the final difference electron density synthesis was 0.374 e/Å 3 and the largest hole was -0.206 e/Å 3 with an RMS deviation of 0.060 e/Å 3 . On the basis of the final model, the calculated density was 1.478 g/ cm 3 and F(000) 976 e.
The non-hydrogen atoms were refined anisotropically. All hydrogen atoms were placed in calculated positions and refined within the riding model. In addition CH 3 and OH groups were free to rotate along C-C and C-O bonds, respectively. The temperature factors of hydrogen atoms were not refined and were set to be equal to either 1.2 or 1.5 times larger than U eq of the corresponding heavy atom. The atomic scattering factors were taken from the International Tables [32]. Molecular graphics were prepared using program Diamond 2.1 [33]. Thermal ellipsoids parameters are presented at 50 % probability level.