Structural determinants of resveratrol for cell proliferation inhibition potency: experimental and docking studies of new analogs

Read Also

We compared two series of molecules starting from (E)-stereoisomer and its (Z)-counterpart. It must be noted that to date the use of (Z)-resveratrol has not been reported. Some of the compounds tested demonstrate much more potencies than the natural parent molecule. Compared with (E)-resveratrol, which leads to a cell growth arrest in S phase, the methylated derivatives stop cell proliferation by inducing G₂/M failures and also a polyploidization of the SW480 cell line. (E)-resveratrol derivatives also induce cancer cell apoptosis since sub-G₁ peaks were found during the flow cytometry analysis.

Cell polyploidization, which can occur naturally to repair damaged DNA, is induced by all of the methylated compounds. (Z)-3,5,4′-trimethoxystilbene (4) has been described as an inhibitor of tubulin polymerization [1931]. Inhibition leads to a mitosis defect and cytokinesis impairment. The destiny of tetraploidy-induced cells is unknown; these cells could still proliferate and become more resistant [32] or die by mitotic catastrophe [33].

The study of synthetic (E)-resveratrol derivatives could offer a wide range of compounds that are potentially more active than (E)-resveratrol (about 66-fold for 4), but these molecules seem to have a different way of delaying cancer cell growth.

Our team previously tested similar resveratrol analogs on various tumor cell lines: DU-145 (androgen-nonresponsive human prostate cancer), LNCaP (androgen-responsive human prostate cancer tumor), M-14 (human melanoma), and KB (human mouth epidermoid carcinoma) [15]. In all cell lines, results indicate a stronger effect of (E)-3,5,4′-trimethoxystilbene (3) than (E)-resveratrol, especially toward DU-145 cells. The corresponding (Z)-trimethoxy analog 4 was very active toward the KB cell line but with a poor effect on M-14 cells.

Szekeres’s group [34] reported on the influence of several (E)-resveratrol analogs on HT 29 human colon cancer cell proliferation inhibition and apoptosis, and some results are similar to those obtained by our group, i.e., the poor effect of (E)-3,5,3′,4′-tetramethoxystilbene (11) and blockade on G₀-G₁. Other results are conflicting; a strong effect of (E)-3,5,4′-trimethoxystilbene (3) on HT 29 cells and a weaker effect on SW480 cells (our data).

More generally, methylated resveratrol analogs, although nonantioxidant molecules, have a stronger effect than the parent molecule. Indeed, they inhibit the human tumor necrosis factor alpha-induced activation of transcription factor nuclear factor kappa B [35].

In summary, while (E)-resveratrol is considered to be a promising molecule for fighting cancer [36], synthetic resveratrol analogues could offer a wide range of compounds that are potentially more active than (E)-resveratrol. These molecules seem to have a different way of delaying cancer cell growth. Resveratrol inhibits cells in S phase, while most of the other synthetic derivatives stop mitosis or block it in an unknown manner (7, 9, 11, 13). We can consider that these methylated derivatives, which are prevented from any hydroxyl group-conjugation dependency, would be less metabolized than resveratrol and potentially more bioavailable. Indeed, we have observed (unpublished results) limited metabolism of 4 after incubation with the SW480 cell line. Moreover, recent results from Lin and Ho [37, 38] are in agreement with this statement, since they reported that the pharmacokinetics of (E)-3,5,4′-trimethoxystilbene (3) and (E)-3,5,3’4′-tetramethoxystilbene (11)  in rat plasma are much slower than those of (E)-resveratrol, i.e., greater plasma exposure, longer elimination exposure, and lower clearance. In addition, the stronger effect of (Z)-methoxy derivatives with respect to their (E)-isomers is not related to the lack of an antioxidative effect (disappearance of hydroxyl groups) but is probably due to a steric-dependent mechanism leading to interference in different pathways as compared to the trans derivatives.

We learned of the recent work of Li et al., [39], while our work was just being completed and submitted. This work reports the effect of a pentamethoxy resveratrol derivative and the analysis of its apoptotic properties. Although from our respective analogs the (Z)-methoxy derivatives have both microtubule targets, our present paper does not have the same objective since our goal was the analysis of a structure–function relationship using a large series of analog (E)- and (Z)-derivatives, especially by combining experimental data and an original docking approach. Moreover, in our cell model and molecules we did not detect a strong apoptosis but a strong polyploidy (see Fig. 2).

With regard to the docking work, it is noteworthy that the adopted procedure, consisting in the extrusion of colchicin from the tubulin binding site, reconstruction of protein deficiencies (lacking in amino acid residues, hydrogen bondings, hydrogen orientations, etc.), restoring the appropriate ligand (combretastatin A-4), molecular mechanic minimization of the whole system, and, finally, submission to the AutoDock  docking program [40],  allowed us to obtain a model that was able to reproduce the IC₅₀ experimental values of other combretastatin analogs also reported in the literature with a high grade of precision that is, to our knowledge, the best obtained to date using docking methodology. Although the results obtained do not always coincide with the antitumoral activity, this is probably attributable to the ADME implications and not to the mechanism of actions. Then the described procedure results in a new valid approach for determining a potential tubulin ligand in the colchicin binding site.