Cytotoxicity of some synthetic bis(arylidene) derivatives of cyclic ketones towards cisplatin-resistant human ovarian carcinoma cells

Symmetrical α,αʹ-bis(arylidene)ketones were prepared by acid-catalyzed aldol condensations between aliphatic ketones (e.g., cyclopentanone, 4-alkylcyclohexanones, tetrahydropyran-4-one, and tetrahydrothiopyran-4-one) and two equivalents of an aromatic hydroxyaldehyde (e.g., 4-hydroxybenzaldehyde, 3,4-dihydroxybenzaldehyde, vanillin, isovanillin, and 3-fluoro-4-hydroxybenzaldehyde). Most of the compounds were cytotoxic towards the cisplatin-resistant human ovarian cancer cell line A2780-CP70 as well as the non-resistant line A2780.


Introduction
Ovarian cancer is estimated to kill more than 180,000 women each year worldwide (World Health Organization 2019). A particular challenge is the frequent development of resistance towards cisplatin, one of the important chemotherapeutic agents (Galluzzi et al. 2014). Curcumin (1), which is a major constituent of the spice turmeric, inhibits the proliferation of cisplatin-resistant ovarian cancer cells by multiple mechanisms (Weir et al. 2007). Although curcumin has low toxicity, its low potency, rapid metabolism, and low oral availability limit its clinical use and have prompted the evaluation of analogues, particularly α,α′-bis (arylidene)ketones (Adams et al. 2004). Such compounds are easily prepared through chemical synthesis (Scheme 1) and considerable structural diversity can be obtained by combining sets of readily available starting materials. In our present study we have prepared a group of symmetrically substituted bis(arylidene)ketones containing new examples along with some previously known compounds. We have then examined the ability of these substances to inhibit the growth of both A2780 human ovarian carcinoma cells and the related cisplatin-resistant cells, A2780-CP70.

Material and methods
'Petrol' refers to the fraction of petroleum spirit with bp 40-60°C. Flash chromatography was performed on BDH silica gel (33-70 μm). All new compounds were >95% pure as assessed by TLC and high field NMR. Melting points were determined using a Reichert hot stage microscope and are uncorrected. NMR spectra were recorded on Jeol EX270 and Bruker AMX400 spectrometers. Mass spectra were provided by the EPSRC NMSF at Swansea University. Low resolution, EI and CI measurements were performed on a Quattro II triple quadrupole instrument. High resolution measurements were made using a Finnigan MAT 900 XCT high resolution spectrometer. DEPT and/or HSQC experiments were used to assist the assignment of all reported 13 C NMR spectra.

Method of cell culture and viability assay
Human ovarian carcinoma cells sensitive to cisplatin, A2780 and resistant to cisplatin, A2780-CP70 were grown and maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum at 37°C, 5% CO 2 . The cells were plated in 96-well culture plates at a density of 1 × 10 4 cells/ well and allowed to adhere at 37°C for 24 h. Then the next day the cells were treated with various doses of compounds ranging from 0.1 nM to 10 μM or vehicle. Cell viability assay, MTT, was performed following incubation time of 96 h. In such experiments the supernatant was removed and MTT (3-(4,5-dimethylthiazol-2-yl)-2,5diphenyltetrazolium bromide) was added for 4 h. The ability of cells to form formazan crystals by active mitochondrial respiration was determined by using a microplate reader after dissolving the crystals in DMSO. Cytotoxicity was expressed as a relative percentage of the absorbance measured at 540 nm in the control and drugtreated cells. MTT assay (Mosmann 1983) measures the conversion of the soluble tetrazolium salt (MTT) to an insoluble purple formazan by succinate dehydrogenase in the mitochondria of living cells. IC 50 values are reported as mean ± s.e. mean (n = 2-4 experiments). Carboplatin (20) was used as a reference chemotherapeutic drug.

Results and discussion
The aldol condensation reaction between aromatic aldehydes and aliphatic ketones in appropriate stoichiometric proportions gives α,α′-bis(arylidene)ketones (Scheme 1). Use of hydroxybenzaldehyde starting materials readily allows the construction of symmetrical products that resemble curcumin in the sense that two phenol or catechol rings are connected by a linking group containing an unsaturated ketone. Some substances of this general type are known to have activity against various cancer cell lines. These include HeLa human cervix cancer (Costi et al. 2004), leukemia , and HEKT293T human embryonic kidney cancer (Leow et al. 2014).
There is literature precedent for conducting aldol condensation reactions under either acidic or basic conditions. However, base would deprotonate phenolic hydroxyl groups that are present in the aldehyde substrates, probably necessitating the protection of the hydroxyl groups (Artico et al. 1998). We therefore decided to perform the reactions directly on the hydroxybenzaldehydes in the presence of hydrochloric acid (general approach of Sardjiman et al. 1997). Examples were selected to investigate the effect of including methoxyl or fluorine atom substituents in addition to the phenolic hydroxyl groups.
Synthetic ketones 2-19 were readily isolated as crystalline solids that had NMR, IR, and high resolution mass spectra consistent with the expected structures. The cyclopentanone derivative 2 did not show significant inhibition of the growth of either ovarian cell line at concentrations up to 10 μM (Table 1); cyclopentanone 4 was also ineffective against cisplatin-resistant cells. However, all of the cyclohexanones 5-15 and heterocyclic analogs 16-19, showed inhibition with IC 50 < 10 μM, including against cisplatin-resistant cells. This contrasts with the results obtained using carboplatin 20, which is used to treat ovarian cancer clinically, but is comparatively ineffective against cisplatin-resistant cells. It is notable that the cyclohexanone and the 4ethylcyclohexanone derivatives 5 and 11 are much more active than cyclopentanone 2, which has identically substituted aromatic rings. Alkyl substitution of the 4position of the cyclohexanone ring is well tolerated, even with the bulky tert-butyl group in analogs 14 and 15. It is also evident that the pyranones 16-18 and the thiopyranone 19 have high levels of activity against ovarian cancer cells. These results should be compared with previous studies in which thiopyranone derivatives were found to be highly active against both NB4 acute promyelocytic leukemic cells and NB4-R1 retinoic acid resistant cells , and a piperidin-4-one derivative was active against cisplatin-resistant ovarian cancer cells .
With regard to the substituents on the aryl groups, we observe that replacing R 1 = OMe in 8 by R 1 = F in 9 did not improve activity, neither did replacement of R 1 = OH in 16 to give 18, whereas substituting R 1 = H in compound 10 by R 1 = F to give 13 was of limited benefit.
It is known that the enones such as these are able to react as electrophiles, undergoing addition of soft nucleophiles such as thiols to the β-carbon of the C=C bond. It has previously been shown that such reactions are responsible for the induction of phase 2 enzymes such as quinone reductase (Dinkova-Kostova et al. 2001). To compare the electrophilicity of representative compounds from the present study, we ran a series of 1 H NMR spectra on samples in which individual samples of cyclopentanone 2, cyclohexanone 5, and tetrahydropyranone 17 were mixed with a model thiol (3mercaptopropionic acid) in a 1:20 molar ratio. After a few hours it was observed that the signal intensities due to the C=CH protons diminished (Fig. 1). These signals are singlets which occur downfield of the aromatic signals, with chemical shifts between 7.3 and 7.6 ppm. The appearance of new signals between 6.6-7.0 and 3.7-4.0 ppm was consistent with Michael addition having occurred in each case, with the cyclopentanone 2 clearly being less reactive than the six-membered ring analogs. Analysis of data obtained at different times indicated that the pseudo first order rate constant for consumption of cyclohexanone 5 was 5.1 times greater than that of the analogous cyclopentanone 2, whereas that for the tetrahydropyranone 17 was 8.0 times greater than that of 2. Thus, for these three compounds, the trend in electrophilicity correlates with cytotoxicity. The fact that cisplatin-resistant cells were sensitive to the cytotoxicity induced by these compounds also indicates that the mechanism(s) of action of these compounds are likely to be independent of nucleus events associated with the DNA repair cascade as platinum compounds primarily affect the DNA within the nucleus. Organic compounds that target DNA, such as nitrogen mustards and anthraquinones, show strong cross-resistance with cisplatin (Hamaguchi et al. 1993). However, other non DNA-related damage can also lead to apoptosis or cell death. For example, it has been shown that changes in the mitochondrial function following the release of cytochrome c into the cytosol resulted in caspase activation, which in turn leads to apoptosis (Liu et al. 1996). This indicated that apoptosis can also be initiated in the cell cytosol and not just the nucleus (Cory and Adams 2002). Considering that many patients do develop resistance to platinum drugs would highlight the importance of developing analogues that at least initially activate the apoptotic pathway in the cytoplasm of the cells, which bypasses the resistance pathway. Further experiments are required to substantiate the mechanism of action of these compounds.

Conclusions
Most of the synthetic curcumin analogues that were studied showed significant inhibition of ovarian cancer cell growth, including inhibition of cisplatin-resistant cells. The ability of the enone C=C groups in these analogues react electrophilically with thiol groups of biological molecules is likely to be relevant to their activity. The derivatives of the heterocyclic ketones stand out as being particularly active and worthy of further investigation. Fig. 1 400 MHz 1 H NMR spectra of mixtures of individual ketones (1.00 × 10 -5 mol) with 3-mercaptopropanoic acid (2.00 × 10 -4 mol) in DMSO-d 6 (0.60 mL) at 21°C. Spectra shown were recorded 0.5 h (left) and 4.0 h (right) after sample preparation. Ketones used were cyclopentanone 2 (top spectra), cyclohexanone 5 (middle spectra), and tetrahydropyranone 17 (bottom spectra)