Inhibition of Cancer Derived Cell Lines Proliferation by Synthesized Hydroxylated Stilbenes and New Ferrocenyl-Stilbene Analogs. Comparison with Resveratrol

Further advances in understanding the mechanism of action of resveratrol and its application require new analogs to identify the structural determinants for the cell proliferation inhibition potency. Therefore, we synthesized new trans-resveratrol derivatives by using the Wittig and Heck methods, thus modifying the hydroxylation and methoxylation patterns of the parent molecule. Moreover, we also synthesized new ferrocenylstilbene analogs by using an original protective group in the Wittig procedure. By performing cell proliferation assays we observed that the resveratrol derivatives show inhibition on the human colorectal tumor SW480 cell line. On the other hand, cell viability/cytotoxicity assays showed a weaker effects on the human hepatoblastoma HepG2 cell line. Importantly, the lack of effect on non-tumor cells (IEC18 intestinal epithelium cells) demonstrates the selectivity of these molecules for cancer cells. Here, we show that the numbers and positions of hydroxy and methoxy groups are crucial for the inhibition efficacy. In addition, the presence of at least one phenolic group is essential for the antitumoral activity. Moreover, in the series of ferrocenylstilbene analogs, the presence of a hidden phenolic function allows for a better solubilization in the cellular environment and significantly increases the antitumoral activity.

Zhang et al. have confirmed that trans-resveratrol was known to be active only in its E configuration while some methoxylated derivatives proved to be active in the Z configuration [41]. In order to deepen our understanding of the mechanism of action and to highlight compounds with enhanced effects on colorectal tumor SW480 and hepatoblastoma HepG2 cell lines, we synthesized a series of E-stilbenes, including three new original ferrocenylstilbene analogs, by improved Wittig and Heck methods [46]. Each compound was submitted to evaluation for biological properties (antiproliferative activity and cell cycle disturbance of SW480 colon cancer and hepatic HepG2 cancer cells). To obtain an inhibitory effect, the chemical parameters studied are the following: (a) the presence of a hydroxy group in position 4; (b) the increased effect due to the presence of a methoxy group (a decrease of the polar character leading to an increase in lipophilic property); (c) the lack (or masked form) of other hydroxy groups. In the series of ferrocenylstilbene analogs, the presence of a phenolic function as an ester greatly increases the antitumoral activity. Most of synthetic compounds are more efficient towards colorectal SW480 cells than liver-derived HepG2 cells. Furthermore, the lack of effects on non-tumor cells (IEC18 intestinal epithelium cells) demonstrates the selectivity of these molecules for cancer cells, which is an important aspect for possible therapeutic applications.

Synthesis of E-4-Hydroxystilbenes
Given the importance of the free phenolic function in position 4 [30,31], we focused on the preparation of derivatives bearing a free phenolic group in position 4 and substituents on the ring B of the stilbenes (compounds 1-6; Figure 2a) or on the A and B rings of the stilbenes (compounds 7-9; Figure 2b). The methoxy group was often chosen as a substituent to improve the membrane permeability of the stilbenes. To highlight the importance of the presence and the position of the phenolic function in the activity of the stilbenes towards tumor cell lines, one derivative with OH group in position 3 was prepared (compound 10; Figure 2c) and four resveratrol analogs without a free phenolic function were synthesized (compounds 11-14; Figure 2d). Compound 10 was already studied by Zhang et al. for its effects on NQO1 induction in hepatoma cells, but its synthesis was not described [41].
On the contrary, compounds 1-4, 6, 7, 12 and 13 were already synthesized by different method, including Horner-Emmons-Haworth [35,47,48], Perkin [49][50][51] and Mizoroki-Heck reactions [52]. Previously, our group has reported the synthesis of compounds 1-14 by two standard methods [46]. Stilbenes 4, 7-13 were prepared by palladium-catalyzed Heck coupling using ferrocenylphosphane ligands. In our protocol, the hydroxylated stilbenes were obtained without the need of protection/deprotection steps on the phenolic functions. Stilbenes 1-3, 5, 6 and 14 were prepared by Wittig reactions; the protection on the hydroxy groups of aromatic aldehydes was achieved using the labile trimethylsilyl group, rarely used in this case. This protective group was easily cleaved during the aqueous work-up following the Wittig reaction.

Synthesis of Stilbenes Bearing Ferrocenylstilbene Analogs
In addition to these stilbenes bearing classical substituents, we developed original ferrocenyl-analogs of stilbenes 15-17 ( Figure 3). Indeed, since the discovery of the antitumoral properties of cisplatin [53], the therapeutic interests in metallic complexes and organometallic compounds has increased steadily [54], especially for ferrocenyl derivatives [55]. Several organometallic compounds bearing a ferrocenyl group display better biological properties than their organic counterparts, such as chloroquine and ferroquine used in the treatment of malaria [56]. A key example of an anticancer ferrocene derivative is the anti-breast cancer ferrocifen series. Jaouen's group has synthesized different derivatives of the ferrocen complexes of tamoxifen and has shown complementary activities of these compounds [57,58]. Therefore, in the aim to improve the antitumoral activities of the polyphenols, we have targeted the synthesis of an original stilbene molecular structure wherein a ferrocenyl ring replaced a benzenic ring; the position 4 of the remaining benzenic ring was substituted by a free phenolic function. The proposed strategy to access this series of ferrocenylstilbene analogs is to react under Wittig reaction conditions  ferrocenecarbaldehyde (18) or ferrocene-1,1'-dicarbaldehyde (19) [59] with a benzylphosphonium bromide bearing a protected phenolic function 20 ( Figure 4). The precursor of 20 is 4-hydroxybenzylic alcohol (21), the corresponding bromide 22 is not commercially available and cannot be prepared by bromination of 21 because of its instability [60] (Scheme 1). Thus, the protection of the phenolic function has to be carried out before the bromination of the benzylic alcohol and in addition, the protective group should be stable to the bromination reagent. These conditions preclude the use of the trimethylsilyl group [46]. Therefore, the phenolic function has been protected as an ester function by reacting 21 with para-toluoyl chloride in the presence of K 2 CO 3 and acetone as a solvent [61]. The benzylphosphonium bromide 20 was obtained by reacting benzylic alcohol 23 successively with N-bromosuccinimide in CH 2 Cl 2 [62] and triphenylphosphine in toluene (Scheme 1).

Scheme 1. Synthesis of benzylphosphonium bromide 20.
Finally, the benzylphosphonium bromide 20 was reacted with ferrocenecarbadehyde (18) in the presence of butyl lithium in THF. The cleavage of phenolic esters was carried out by KOH in methanol [63] and the ferrocenylstilbene analog 15 was recovered in 52% yield. In the same manner, the ferrocenyl derivative was obtained from 20 and ferrocene-1,1'-dicarbaldehyde (19) in 47% yield (Scheme 2).

Biological Effects
We compared the potency of the new resveratrol synthetic analogs towards the human colorectal tumor cell line SW480, the human hepatoblastoma HepG2 cell line and the rat normal intestine epithelium IEC18 cell, comparing their effect with the natural reference molecule, i.e., trans-resveratrol. Firstly, we have determined the sensitivity of human tumoral colorectal cell line SW480 towards the newly synthesized stilbene derivatives and compared them to resveratrol, the parent molecule. Figure 5 shows, as expected and in agreement with the literature [64], that resveratrol at 30 µM decreases drastically cell viability which is of 40% compared to the control ( Figure 5). Interestingly, compounds 1-5 exhibit higher cytotoxicity than resveratrol. These derivatives bear, like resveratrol, at least one phenol group in the para position of the stilbene ring. The only structural differences between these molecules are the positions and numbers of methoxy groups. The efficiency of compound 1 indicates that its activity is due to the phenolic group, despite the absence of methoxy groups on its skeleton. Compound 14, a tetramethoxylated derivative, shows similar activity as resveratrol, suggesting that these substituents are not essential for the activity. However, the fact that compounds 9, 11-13 have only weak effects seems to indicate that a free phenolic group in the para position of the aromatic ring is needed for toxicity. To further explore the mechanisms by which the most efficient compounds exert their antiproliferative potencies, we studied their effects on the cell cycle distribution of SW480 cells ( Figure 6). The treatment of cells with compound 2, which bears a hydroxy group in position 4 and a methoxy group in position 4', induces an accumulation of SW480 cells in S phase in the same manner as resveratrol ( Figure 6). Interestingly, compound 4, bearing hydroxy groups at positions 4 and 4' and a methoxy group at position 3, leads to an increase of S phase which is better than that of resveratrol and compound 2. In contrast, pterostilbene (3) does not show any effect on the cell cycle, while it inhibits cell proliferation. This derivative has been reported to induce a blockade of HL60 intestine cancer cells in the G 1 phase, and to induce apoptosis [65]. The distribution of cells in the different cell cycle phases is reported in Figure 1 of the supplementary material.
One of the mechanisms by which resveratrol modulates carcinogenesis is the blockage of cells in S phase [66]. However, these effects at the cell cycle are complex and depend on the cell type, the resveratrol concentration and the duration of the treatment. Indeed, a low concentration of resveratrol induces accumulation of cells in S phase while at higher concentrations it leads to cell accumulation in G 1 or G 2 /M phases [67]. Moreover, many cytotoxic agents also induce cell death by apoptosis. We have previously shown in SW480 and in HepG2 cell lines that resveratrol induces accumulation of cells in early S phase by action on the p21 protein and on the cyclin/cdk complexes formation and activity [68]. In the structural core of resveratrol, the phenol group in position 4 would be responsible for the antiproliferative effect by its action on DNA polymerases alpha and gamma [69,70]. Indeed, the increase of number of hydroxy groups on the stilbene moiety of resveratrol derivatives led to an increase of inhibition of tumor cell proliferation [71]. On the other hand, She et al. [72] have shown that trans-3,3',4',5-tetrahydroxystilbene and trans-3,3',4',5,5'-pentahydroxystilbene exhibit a higher apoptotic effect than resveratrol on the epidermal JB6 cell line. Figure 6. Influence of stilbene derivatives on the cell cycle phases of the SW480 cells line. Cells were grown for 48 h in the presence of 30 µM resveratrol (or no RSV in a control experiment) or 30 µM stilbene derivatives (numbered on the x-axis). After treatment, nuclear DNA was labeled with propidium iodide. The cell cycle effect of the tested compounds was done analysing cell distribution in the different phases of the cell cycle (mean ± standard deviation of two independent experiments).

Evaluation of Toxicity Level of Stilbene Derivatives Towards Non-Cancerous Intestinal Epithelial Cells
With the aim of possible therapeutic applications using resveratrol derivatives in mind it was important to evaluate the specificity of cytoxicity towards normal cells. Hence, we evaluated the effect of potent derivatives on the proliferation of intestine epithelium IEC18 cells. The results shown in Figure 7 indicate no significant toxic effect of compounds 2-4 at 30 µM, except for compound 5 (presence of vinyl group in position 4). At higher concentration (100 µM) all compounds, including resveratrol, slightly inhibit cell proliferation, but much less than with the tumor SW480 cell line.

Comparison of Resveratrol Analogs on Cytotoxicity of Colorectal Tumor Cells and on Hepatoblastoma Cells
To have an overall view of the mechanisms involved in the inhibitory effect of the compounds, we performed a concentration-dependent analysis of the cytotoxicity evaluated by the crystal violet method. The crystal violet assay was chosen for the screening of the dose-effect of numerous molecules despite its lower sensitivity compared to some other cytotoxicity methods [73]. The results are presented as IC 50 values. These IC 50 values have been determined both on human tumor colorectal SW480 cell line and on human hepatoblastoma HepG2 cell line (Table 1). All tested molecules have lower IC 50 than resveratrol towards SW480 cell line. Compounds 2 and 4 show a similar activity, indicating that the additional hydroxy group does not increase the activity of the stilbene. Comparison of the IC 50 values between compounds 2 and 10 confirm the importance of the position 4 of the phenolic group [30,31]. In the series of ferrocenylstilbene analogs, compound 17 without a free phenolic function is the most active. This may be explained by a better lipophilicity due to the ester group while the antitumor activity can be attributed to the ferrocenyl moiety. Five of the most active derivatives (compounds 1, 2, 5, 6 and 8) have been subsequently tested on the HepG2 cell line (Table 1).
Compounds 1, 2, 5 and 6 exhibit a lower potency on HepG2 than on SW480 cell line. Compounds 7 and 10 are the least active towards SW480 cells. Interestingly, compounds 5 (vinyl group in position 4') and 8 (carbinol group in position 3 and methoxy in position 4') exhibit a higher activity towards SW480 cell lines than HepG2 cell lines, while the bromine in position 4' (compound 6) has an opposite effect. In the case of compound 8, its metabolism by HepG2 cells may explain its weaker activity towards these cells. The difference between the resveratrol IC 50 cytotoxicity value (68.1 µM), (Table 1) and its inhibitory efficiency (30 µM) on cell proliferation ( Figure 5) towards SW480 cell line would be attributed to the difference in the experimental approaches.  Table 1. Compound 17 shows the highest inhibitory activity in both cell lines with a very low IC 50 value (5.9 µM), more than 10-fold higher compared to the resveratrol activity. Ferrocene used as a control does not induce any cytotoxic effect against SW480 cell line. Compound 16 (a deprotected version of compound 17) shows a higher IC 50 value (IC 50 > 100 µM) than compound 17. This data can be explained by the low solubility of 16 in DMSO in the cell medium. E-(4-vinylphenol)ferrocene (15), the closest isostere of resveratrol presented in this study shows a similar antiproliferative activity to resveratrol despite a lower solubility in the medium.

General Experimental Procedures
Wittig reactions were performed under an inert atmosphere of argon using conventional vacuum-line and glasswork techniques. THF was degassed and distilled by refluxing over sodium and benzophenone under argon. The organic reagents were received from commercial sources and used without further purification. Separations by flash chromatography were performed on silica gel (230-400 mesh). 1 H-NMR, 13 C-NMR and 31 P-NMR spectra (δ, ppm) were recorded in CDCl 3 solutions on a Bruker 300 MHZ spectrometer, HRMS on MicroTOF Q-Bruker (ESI ionization). Spectroscopic analyses were performed at the Pôle de Chimie Moléculaire de l'Université de Bourgogne (23): To a mixture of 4-hydroxybenzylic alcohol (21, 100 g, 80.65 mmol) and potassium carbonate (13.4 g, 96.6 mmol) in acetone (300 mL) was added over 30 min at 0 °C a solution of para-toluoyl chloride (16 mL, 121 mmol) in acetone (100 mL). Then, the mixture was refluxed for 6 h. After cooling, the inorganic salts were filtrated and washed with acetone. The solvent was removed under vacuum and the crude product was purified by chromatography (EtOAc/heptane: 1/4) to give pure 4-toluoyloxybenzylic alcohol (23) in 47% yield. 1 (24): To a mixture of 23 (9 g, 37.70 mmol) and triphenylphosphine (14.9 g, 56.53 mmol) in CH 2 Cl 2 (150 mL) was added a solution of N-bromosuccinimide (10 g, 56.53 mmol) in CH 2 Cl 2 (100 mL). After stirring for one hour, the mixture was poured into a separatory funnel and was washed with water. The organic phase was dried over MgSO 4 . After removal of the solvent, the crude product was crystallized from ethanol (64%). 1 (20): A mixture of 24 (18.7 g, 33 mmol) and triphenylphosphine (9.7 g, 36.3 mmol) in toluene (50 mL) was refluxed for five hours. The reaction mixture was cooled down to room temperature and a first crop of product was collected by filtration. The filtrate was then refluxed for five additional hours and a second crop of product precipitated. Two other crops were then collected and the combined fractions were crystallized from ethanol (86%).  (25): Under argon atmosphere, butyllithium (1.6 M, 2.8 mL, 4.48 mmol) was slowly added to a solution of 4-toluoyloxybenzyltriphenylphosphonium bromide (20, 2.5 g, 4.41 mmol) in THF (40 mL) at −78 °C. The resulting solution was allowed to warm at room temperature. A solution of ferrocenecarbaldehyde [59] (18, 0.95 g, 4.41 mmol) in THF (15 mL) was added dropwise and the reaction mixture was then stirred overnight. Ice-cold water (500 mL) was added and the mixture stirred for an additional hour. The aqueous layer was extracted with ethyl acetate; the combined organic layers were washed with water and dried over MgSO 4 . After evaporating the solvent, 52% of a crude mixture of isomers Z and E was isolated. The E isomer was isolated by chromatography (heptane/EtOAc: 9/1), yield 34%. 1 (17): Under an argon atmosphere, butyl lithium (1.6 M, 5.6 mL, 8.96 mmol) was slowly added to a solution of 4-toluoyloxybenzyltriphenylphosphonium bromide (20, 5 g, 8.82 mmol) in THF (80 mL) at −78 °C. The resulting solution was allowed to warm at room temperature. A solution of ferrocene-1,1'-dicarbaldehyde [59] (19, 0.95 g, 4.41 mmol) in THF (15 mL) was added dropwise and the reaction mixture was stirred overnight. Ice-cold water (500 mL) was added and the mixture was stirred for an additional hour. The aqueous layer was extracted with ethyl acetate; the combined organic layers were washed with water and dried over MgSO 4 . After evaporating the solvent, 47% of a crude mixture of EE/EZ/ZZ isomers was obtained. The EE isomer was isolated by chromatography (heptane/EtOAc: 9/1), yield 25%. 1

Cell Culture
The human colon carcinoma cell line SW480 obtained from ATCC (American Type Culture Collection, Manassas,VA, USA) was cultured in RPMI-Medium with 10% fetal bovine serum (FBS) and 1% antibiotics. Human derived hepatoblastoma cell line HepG2 was obtained from the ECACC (European collection of cell culture, Salisbury, UK) and non-cancerous IEC18 cells from ileum epithelium of Rattus norvegicus (ATCC) were grown in monolayer culture system and maintained in phenol-red Dulbecco's Modified Eagle's Medium (DMEM) supplemented with 2 mM L-glutamine, 1% non-essential amino-acids, and 10% FBS (v/v) in a humidified atmosphere of 5% CO 2 at 37 °C.

Cell Viability Assays
Proliferation inhibition assays were performed in 24-well plates in triplicate, and each experiment was conducted two to three times. 30,000 cells were seeded per well, after 24 h cells were incubated in medium containing either 0.1% dimethylsulfoxide-solubilized trans-resveratrol, resveratrol derivatives, or 0.1% dimethylsulfoxide (DMSO) only as control. After 48 h, cells were harvested and the number of live cells was quantified using the trypan blue exclusion test which is based on the ability of a viable cell with an intact membrane to exclude trypan blue dye using a haemocytometer in microscopic counting. Results were expressed as percentage of control values.

Cell Proliferation Assays
After 48 h of incubation at 37 °C, medium was carefully removed from wells and the plates were washed gently with PBS 1X warmed at room temperature. Then the crystal violet solution was added and incubated for 10 min. Thereafter, plates were washed several times with tap water. The nucleus-incorporated crystal violet was dissolved using a sodium citrate solution and plates were agitated on orbital shaker until the color became uniform with no areas of dense coloration at the bottom of wells. The absorbance was read on each plate at 540 nm with a spectrophotometer (Dynex MRX-TC Revelation, Manassas, VA, USA). The absorbance is proportional to the relative density of cells adhering to multi-well dishes in regard to the absorbance of control well-plate (5% DMSO). After 48 h, IC 50 values were determined by performing 0.75 to 100 µM treatments and the IC 50 values were obtained after parametric regressions on the percentages of viable cells versus the control.

Cell Cycle Analysis
Cell cycle analysis was performed as described previously [67,74,75]. Briefly, cells were seeded 24 h before treatment into 25 cm 2 flasks. After treatment, the detached and adherent cells were pooled, fixed with ethanol, and stained with propidium iodide (PI) for subsequent analyses with a CyFlow Green flow cytometer and the fluorescence of PI was detected above 630 nm. For each sample 20,000 cells were acquired. Furthermore, data were analyzed with the MultiCycle software (Phoenix Flow Systems, San Diego, CA, USA); the x-axis corresponds to the DNA content and the y-axis to the number of cycling cells. The maximum value on the y-axis is inversely proportional to the altered cells level (non-cycling cells) which is excluded by gating.

Conclusions
While trans-resveratrol is considered a promising molecule for fighting cancer [76], a wide range of synthetic resveratrol analogs are potentially more active than trans-resveratrol. Some of these new synthetic molecules have interesting effects. Compounds 2 and 17 are the most active, while compounds 10 and 16 show the lowest activity. The comparison between compounds 16 and 17 indicates that the presence of a protecting group lead to a better efficacy which could be due to a better solubilisation in DMSO. It appears that the lack of substituents at position 3 and 5 (compound 1) leads to a better inhibitory effect. Moreover, a limited number of methoxy groups (compounds 2, 3 and 4) provides better lipophilic properties. In most cases, the efficacy of the synthetic compounds is lower towards liver derived HepG2 cells than towards colorectal SW480 cells, except for compound 6 and mostly 17, which is the most powerful derivative. These differences can be explained by the high xenobiotic metabolizing activities of HepG2 cells. Furthermore, the lack of effect on non-tumor cells (IEC18 intestinal epithelium cells) demonstrates the selectivity of these molecules for cancer cells, which is an important aspect for potential therapeutic applications. Concerning the possible targets of resveratrol analogs, an inhibition of the TNF alpha-induced activation NFkB by polyhydroxylated resveratrol derivatives i.e., the hexahydroxystilbene in leukemia HL60 cells has been reported [70]. In terms of the structure-activity relationship, it appears that in order to obtain an inhibitory effect, the chemical parameters are the following: (a) the presence of a hydroxy group in position 4; (b) an increased inhibitory effect by the presence of a methoxy group (a decrease of the polar character leading to an increase in lipophilicity); (c) the lack (or masked form) of other hydroxy groups. In addition, (E,E)-1,1'-bis[(1-para-toluoyloxy-4-vinyl)benzene]ferrocene (17) a new compound, shows the highest efficacy.