Synthesis and SAR Analysis of Novel 4-Hydroxytamoxifen Analogues Based on Their Cytotoxic Activity and Electron-Donor Character

Utilizing McMurry reactions of 4,4′-dihydroxybenzophenone with appropriate carbonyl compounds, a series of 4-Hydroxytamoxifen analogues were synthesized. Their cytotoxic activity was evaluated in vitro on four human malignant cell lines (MCF-7, MDA-MB 231, A2058, HT-29). It was found that some of these novel Tamoxifen analogues show marked cytotoxicity in a dose-dependent manner. The relative ROS-generating capability of the synthetized analogues was evaluated by cyclic voltammetry (CV) and DFT modeling studies. The results of cell-viability assays, CV measurements and DFT calculations suggest that the cytotoxicity of the majority of the novel compounds is mainly elicited by their interactions with cellular targets including estrogen receptors rather than triggered by redox processes. However, three novel compounds could be involved in ROS-production and subsequent formation of quinone-methide preventing proliferation and disrupting the redox balance of the treated cells. Among the cell lines studied, HT-29 proved to be the most susceptible to the treatment with compounds having ROS-generating potency.


Introduction
Selective Estrogen Receptor Modulators (SERMs) represent one of the most important classes of drugs with a wide spectrum of possible applications, including chemotherapy for breast cancer and treatment of hormonal deficiencies related to Estrogen, and others [1]. Besides SERM activity, some members of this drug family feature other types of the mechanism of action capable of inducing a large variety of unexpected biological responses including effects not related to the original SERM mechanism. A well-known example is Tamoxifen, the most important representative of first-class SERMs. Tamoxifen's main application is the treatment and/or prevention of breast cancer most frequently occurring among post-menopausal women, but this drug proved to be applicable for the treatment of infertility both in women and men [2,3], or for the prevention of gynecomastia [4]. Aside from Tamoxifen's SERM activity explored in the indications listed above, its function in the generation of Reactive Oxygen Species (ROS) is also well-documented concerning adverse effects [3,5]. ROS including, e.g., superoxide anion and hydroxyl radicals, are involved in signal transmission pathways regulating cell growth, differentiation, survival, vention of gynecomastia [4]. Aside from Tamoxifen's SERM activity explored in the indications listed above, its function in the generation of Reactive Oxygen Species (ROS) is also well-documented concerning adverse effects [3,5]. ROS including, e.g., superoxide anion and hydroxyl radicals, are involved in signal transmission pathways regulating cell growth, differentiation, survival, inflammation and immune response. In an imbalanced cellular state, ROS with a concentration above the normal level induces oxidative stress, leading to cell death [6]. Upon the effect of ferrocene-containing analogues of Tamoxifen (e.g., Ferrocifen), the ROS level is significantly increased compared to that generated by Tamoxifen [7,8]. Consequently, ferrocene-containing Tamoxifen analogues display cytotoxicity not only on Estrogen Receptor positive (ER+) breast cancer cells, but also on ER-independent cell lines, due to the ROS-generating ability of the organometallic agents playing a predominant role in their activity [7]. The formation of ROS proceeds along a radical chain reaction pathway initiated by a radical cation-generating single electron transfer (SET); ionization is necessarily implicated in such processes [6]. The potential of a compound to undergo ionization can directly be assessed through its half-wave oxidation potential measured by Cyclic Voltammetry (CV) whose value is also affected by the experimental conditions. The intrinsic electron donor character of a molecule can also be estimated by DFT calculations providing HOMO energy and adiabatic ionization energy [9,10] considered as quantitative theoretical descriptors of the feasibility of the first SET to an oxidizing agent.
On the one hand, as mentioned above, certain side effects might be the consequence of ROS-generation, but on the other hand, ROS-dependent pathways can be explored in triggering cytotoxic effects in malignant cells [11][12][13][14][15]. In this context, the main purpose of this comparative study was to assess at least the approximate contribution of structurally fine-tunable ROS-generation to the cytotoxic activity of systematically designed 4-Hydroxytamoxifen analogues by comparing their cytotoxic activity to their redox properties determined by CV measurements and DFT calculations.

Design and Synthesis of Novel Tamoxifen Analogues
Regarding the fact that 4-Hydroxytamoxifen (1) is an active metabolite of Tamoxifen with SERM activity, besides hydroxylation, the replacement of the solubility-enhancing N,N-dimethylaminoethoxy substituent for a hydroxyl group is expected to increase not only the target-binding affinity, but also the feasibility and rate of ROS production. Thus, we first selected 4-Hydroxytamoxifen-related diphenol 2 as a simplified reference model for our current comparative research directed to synthesis and biological evaluation of novel homo/heterobicyclic analogues of 2 (Scheme 1) complemented with electrochemical and DFT analysis of their redox properties. We considered the targeted 4-Hydroxytamoxifen analogues with general formula I, having ring size-controlled different steric arrangements and at least partly heteroatom-influenced tendency to undergo oxidation, as suitable models for the approximate assessment of the contribution of highly stereostructure-dependent receptor signaling pathways and ROS-related mechanisms to cytotoxic activity. (For the sake of simplicity, the term "4-Hydroxytamoxifen", describing the real molecular structures, was replaced by the term "Tamoxifen" in the residual part of the text.). It must be emphasized here that we have put particular focus on analogues of type I with embedded electron-donating aromatic groups capable of increasing their propensity to generate ROS. Accordingly, employing reported protocols, we prepared substituted 1,5,6,7-tetrahydro-4H-indol-4-one (3)(4)(5) [16] and 1,5,6,7-tetrahydro-4H-indazol-4-one derivatives (7,8) [17] (Scheme 2) planned to serve as coupling partners in such McMurry reactions that were expected to construct pyrrolo-and pyrazolo-condensed Tamoxifen mod-els. Attempting to achieve a selective synthesis of 7, we have also observed the formation of its regioisomer (8) in a ratio of 40% of the product mixture. As we have failed so far to separate the components, the mixture of 7 and 8 was used in the subsequent MCMurry coupling reaction.
For comparing the apoptotic efficiency of ROS-generation pathways to that of signal transductions initiated by target-binding, we also undertook the implication of ferrocene-containing Tamoxifen-analogues in our biological-, physicochemical-and DFT studies. Thus, besides the emblematic 4,4′-(2-ferrocenylbut-1-ene-1,1-diyl)diphenol 18 [18] and the racemic planar chiral ferrocene derivative (38), of which synthesis and in vitro cytotoxicity on pancreatic and breast cancer cell lines have also been reported by our team [19], we also envisaged the synthesis and comprehensive study of the racemic form of aza-analogues (35)(36)(37) expected to have extended basic target-binding regions along with significantly increased propensity to undergo ROS-generating oxidation by single electron-transfer (SET), which is facilitated by the presence of aminoferrocene moiety with enhanced HOMO energy and electron-density (Scheme 3). Secondary amine 12, the parent molecule of the crucial ferrocene-containing components of the McMurry coupling, was accessed by a two-step procedure involving iodoferrocene-mediated Goldberg-type N-ferrocenylation of azetidin-2-one (9) [20] followed by TFA-catalyzed ring expansion [21] of the resulting intermediate 11 (Scheme 3). The N-alkylation of highly air-sensitive 12 with benzyl bromide and 4-fluorobenzylbromide afforded less air-sensitive benzylated derivatives 13 and 14, respectively. Scheme 2. Synthesis of 1,5,6,7-tetrahydro-4H-indol-4-one and 1,5,6,7-tetrahydro-4H-indazol-4one derivatives.
For comparing the apoptotic efficiency of ROS-generation pathways to that of signal transductions initiated by target-binding, we also undertook the implication of ferrocenecontaining Tamoxifen-analogues in our biological-, physicochemical-and DFT studies. Thus, besides the emblematic 4,4 -(2-ferrocenylbut-1-ene-1,1-diyl)diphenol 18 [18] and the racemic planar chiral ferrocene derivative (38), of which synthesis and in vitro cytotoxicity on pancreatic and breast cancer cell lines have also been reported by our team [19], we also envisaged the synthesis and comprehensive study of the racemic form of aza-analogues (35)(36)(37) expected to have extended basic target-binding regions along with significantly increased propensity to undergo ROS-generating oxidation by single electron-transfer (SET), which is facilitated by the presence of aminoferrocene moiety with enhanced HOMO energy and electron-density (Scheme 3). Secondary amine 12, the parent molecule of the crucial ferrocene-containing components of the McMurry coupling, was accessed by a two-step procedure involving iodoferrocene-mediated Goldberg-type N-ferrocenylation of azetidin-2-one (9) [20] followed by TFA-catalyzed ring expansion [21] of the resulting intermediate 11 (Scheme 3). The N-alkylation of highly air-sensitive 12 with benzyl bromide and 4-fluorobenzylbromide afforded less air-sensitive benzylated derivatives 13 and 14, respectively. For the synthesis of the targeted Tamoxifen analogues, we have applied McMurry reactions in the course of which 4,4′-dihydroxybenzophenone (15) was coupled with a variety of acyclic-, cyclic/heterocyclic-and organometallic ketone components in the presence of zinc dust and TiCl4 (Scheme 4).
For the synthesis of the targeted Tamoxifen analogues, we have applied McMurry reactions in the course of which 4,4′-dihydroxybenzophenone (15) was coupled with a variety of acyclic-, cyclic/heterocyclic-and organometallic ketone components in the presence of zinc dust and TiCl4 (Scheme 4).  Most reactions performed by a well-established protocol [22] allowed the isolation of the targeted analogue in low-to-mediocre yield, while the reactions with coupling components 3, 12, 13 and 14 afforded the appropriate desired olefins (30, 35, 36 and 37) only in traces as detected by NMR analysis of the crude products which could not be isolated in pure form due their rapid uncontrolled decomposition.
It is also of note that the McMurry reaction of 4 afforded the targeted Tamoxifen analogue (31) and ketone 31a isolated in a non-negligible yield (18%). Its formation can be rationalized in terms of the pinacol rearrangement of primary McMurry intermediate 39 proceeding via a resonance-stabilized carbocation (39 ) that undergoes 1,2-aryl-migration accompanied by the regeneration of the carbonyl group (Scheme 5).
only in traces as detected by NMR analysis of the crude products which could not be isolated in pure form due their rapid uncontrolled decomposition.
It is also of note that the McMurry reaction of 4 afforded the targeted Tamoxifen analogue (31) and ketone 31a isolated in a non-negligible yield (18%). Its formation can be rationalized in terms of the pinacol rearrangement of primary McMurry intermediate 39 proceeding via a resonance-stabilized carbocation (39′) that undergoes 1,2-aryl-migration accompanied by the regeneration of the carbonyl group (Scheme 5).
Scheme 5. Proposed reaction pathway leading to 31 and 31a.
It has been demonstrated by Jaouen et al. that methoxy-substituents induce a smaller cytotoxic effect at a lower rate of ROS generation compared to those produced by the phenol counterparts [23]. Accordingly, by coupling 4,4′-dimethoxybenzophenone with three representative ketones (3, 16 and 41) under standard McMurry conditions, we prepared methyl ethers 42-44 as reference models (Scheme 6) of which comparative assays were expected to allow us to estimate the extra-contribution of the hydroxyl groups to ROS-related activity relative to that contributed by methoxy groups on the identical positions of the molecular skeleton. It has been demonstrated by Jaouen et al. that methoxy-substituents induce a smaller cytotoxic effect at a lower rate of ROS generation compared to those produced by the phenol counterparts [23]. Accordingly, by coupling 4,4 -dimethoxybenzophenone with three representative ketones (3, 16 and 41) under standard McMurry conditions, we prepared methyl ethers 42-44 as reference models (Scheme 6) of which comparative assays were expected to allow us to estimate the extra-contribution of the hydroxyl groups to ROSrelated activity relative to that contributed by methoxy groups on the identical positions of the molecular skeleton.
isolated in pure form due their rapid uncontrolled decomposition.
It is also of note that the McMurry reaction of 4 afforded the targeted Tamoxifen analogue (31) and ketone 31a isolated in a non-negligible yield (18%). Its formation can be rationalized in terms of the pinacol rearrangement of primary McMurry intermediate 39 proceeding via a resonance-stabilized carbocation (39′) that undergoes 1,2-aryl-migration accompanied by the regeneration of the carbonyl group (Scheme 5). It has been demonstrated by Jaouen et al. that methoxy-substituents induce a smaller cytotoxic effect at a lower rate of ROS generation compared to those produced by the phenol counterparts [23]. Accordingly, by coupling 4,4′-dimethoxybenzophenone with three representative ketones (3, 16 and 41) under standard McMurry conditions, we prepared methyl ethers 42-44 as reference models (Scheme 6) of which comparative assays were expected to allow us to estimate the extra-contribution of the hydroxyl groups to ROS-related activity relative to that contributed by methoxy groups on the identical positions of the molecular skeleton. resulted in a complex mixture of undefined materials. Under the same conditions, demethylation of 44 was accompanied by fission of the furane ring generating intermediate 45 of which Friedel-Crafts type recyclization on one of the pending 4-hydroxyphenyl rings afforded indene 46 (Scheme 7). In the subsequent experiments, we used 46 as trihydroxy analogue of 1 because we assumed that the electrochemical and ROS generating properties of 46 are similar to those of an acyclic counterpart with a cleaved five-membered ring. Scheme 7. Proposed reaction pathway leading to 46.
The assigned 1 H-and 13 C-NMR signals of the novel hybrid compounds, unambiguously identified by 2D-HMQC-and HMBC connectivities, are consistent with their structures. (The spectra are listed in the Supplementary Materials).

Evaluation of the Novel Tamoxifen Analogues for Their Antiproliferative Activity on Human Malignant Cell Lines
Applying the MTT assay, the in vitro cytostatic effect was studied on ER(+) MCF-7 [24] and ER(−) MDA-MB-231 [25] human breast adenocarcinoma cells, A2058 human melanoma [26] and HT-29 human colorectal carcinoma [27] cell lines. While breast cancer cell lines were obvious targets in our studies, the selection of A2058 cells can be justified by evidenced estrogen-mediated signaling disclosed in melanoma cells [28]. Finally, since Tamoxifen inhibits the growth of colorectal cancer cells [29], the HT-29 cell line was also included into these assays. To estimate the contribution of the ROS-induced apoptotic pathways to target-related antiproliferative effects, the flexible and conformationally rigid redox-active ferrocene derivatives 18 and 38, respectively, were used as reference compounds. On the other hand, isatin-derived analogue 28 served as a positive control with identified ER-mediated anticancer activity [22]. The MTT assays of the novel and reference compounds disclosed a dose-dependent cytostatic effect on the investigated cell lines. The majority of the tested compounds exhibited an antitumor effect within a concentration range of ca. 3−50 μM, in most cases in marked cell-line dependent manner (Table 1). It is of interest that, besides 28 with demonstrated ER transactivation-inhibitory activity and ferrocene-containing references (18 and 38), having mainly ROS-mediated effects, fused pyrrole 43 and indene 46 were identified as the most potent antiproliferative compounds in our cell-viability assays. The assigned 1 H-and 13 C-NMR signals of the novel hybrid compounds, unambiguously identified by 2D-HMQC-and HMBC connectivities, are consistent with their structures. (The spectra are listed in the Supplementary Materials).

Evaluation of the Novel Tamoxifen Analogues for Their Antiproliferative Activity on Human Malignant Cell Lines
Applying the MTT assay, the in vitro cytostatic effect was studied on ER(+) MCF-7 [24] and ER(−) MDA-MB-231 [25] human breast adenocarcinoma cells, A2058 human melanoma [26] and HT-29 human colorectal carcinoma [27] cell lines. While breast cancer cell lines were obvious targets in our studies, the selection of A2058 cells can be justified by evidenced estrogen-mediated signaling disclosed in melanoma cells [28]. Finally, since Tamoxifen inhibits the growth of colorectal cancer cells [29], the HT-29 cell line was also included into these assays. To estimate the contribution of the ROS-induced apoptotic pathways to target-related antiproliferative effects, the flexible and conformationally rigid redox-active ferrocene derivatives 18 and 38, respectively, were used as reference compounds. On the other hand, isatin-derived analogue 28 served as a positive control with identified ER-mediated anticancer activity [22]. The MTT assays of the novel and reference compounds disclosed a dose-dependent cytostatic effect on the investigated cell lines. The majority of the tested compounds exhibited an antitumor effect within a concentration range of ca. 3−50 µM, in most cases in marked cell-line dependent manner (Table 1). It is of interest that, besides 28 with demonstrated ER transactivation-inhibitory activity and ferrocene-containing references (18 and 38), having mainly ROS-mediated effects, fused pyrrole 43 and indene 46 were identified as the most potent antiproliferative compounds in our cell-viability assays.

Electrochemical Measurements
As we aimed at assessing the approximate contribution of ROS generation to the cytotoxic activity of the novel Tamoxifen analogues, we envisaged evaluating their ability to form a radical cation. Accordingly, all tested compounds including reference models 18, 28 and 38 were first subjected to cyclic voltammetric (CV) analysis providing quantitative information about their tendency to undergo one-electron oxidation as reflected from their cationic peak potential (E pc , Table 1). It must be noted here that the E pc values measured for 43 and 46, two of the promising novel organic Tamoxifen analogues, suggest that their marked antiproliferative potency can at least partly be attributed to ROS-initiated mechanism of action, as discussed in more detail in Section 3. data. As discussed in the experimental section, in the CV experiments, NaCl-saturated calomel electrode was used as a reference electrode and ferrocene was chosen as a standard reference material that shows a very stable one-electron redox process. The complete CV records and experimental parameters of the reference ferrocene and all the other Tamoxifen analogues are presented in Figures S1-S5 in the Supplementary Materials.

DFT Calculations
Since the Epc values are supposed to be markedly influenced by several factors not relevant to ROS-generation potency under biological conditions, including diffusion, ionic strength, interactions with the electrode surface and attenuated coordination of the

DFT Calculations
Since the E pc values are supposed to be markedly influenced by several factors not relevant to ROS-generation potency under biological conditions, including diffusion, ionic strength, interactions with the electrode surface and attenuated coordination of the solvent acetonitrile molecules to the cationic species [20,30,31], we also undertook complementary DFT studies envisaged to get intrinsic information encoded by molecular structure about the redox character of the tested Tamoxifen analogues. The computations were carried out by B3PW91 functional [32] using the DGTZVP basis set [33] for the neutral compounds and the corresponding radical cationic species with doublet electron configuration generated by removing one electron from the parent molecules. (Since the comparative modeling studies also implicated ferrocene-containing compounds, we opted for B3PW91 as functional, which has been found superior to B3LYP in providing a more reliable and realistic description of bonding parameters close to experimental data in metal-based fragments [34,35]). The adiabatic ionization energy [E i ] [9,10] was obtained as the difference in the total electron energy values calculated for the optimized structures of the analyzed compounds and their radical cationic counterparts. As solvation can highly modify the relative stability of the neutral parent compounds and the appropriate radical cations with significantly enhanced dipole moment, single-point energy calculations were performed on the optimized structures by the IEFPCM solvent model [36] using the dielectric constants of water and acetonitrile that represent the conditions of biological systems and CV experiments, respectively.
A comparison of E i data calculated for vacuum and the modeled solvents unambiguously indicates that ionization (Table 1) is highly facilitated by the polar media that render extra stability to the radical cations having substantially larger dipole moment than the neutral parent molecules. In this regard, acetonitrile and water seem to exert a similar impact on ionization energy. The relatively low E i values calculated for the ferrocene derivatives 18 and 38 are in agreement with their ROS-induced cytotoxicity. On the other hand, the exceptionally low cationic peak potentials (E pc ) measured for these organometallic models might refer to the cation-stabilizing coordination of the acetonitrile molecules to the iron center of the ferricenium cations [20].
Besides ionization energy, the HOMO level of the parent neutral molecules can also be regarded as the measure of their propensity to undergo ionization by one-electron oxidation; thus, E HOMO values of the tested models calculated in vacuum and the polar solvents are also listed in Table 1. Since solvation also decreases the total energy of the polar, yet neutral molecules as the consequence of lowering the energy level of their bonding MO's, the marked polarity-induced drop in the E HOMO data of the tested Tamoxifen analogues would suggest that polar media suppress ionization of the tested products. However, the polarity-induced stabilization of radical cations having enhanced dipole moments is substantially larger than that computed for the appropriate neutral Tamoxifen analogues with a smaller dipole moment. Consequently, the relative stability of a radical cation and its neutral parent molecule, which is equivalent to the adiabatic ionization energy of the latter one, is significantly lowered when calculated in a polar medium relative to that calculated in vacuum. Thus, adiabatic ionization energy (E i ) can be regarded as a more reliable theoretical descriptor for the redox character of a neutral molecule than its HOMO energy level.

Discussion
Except for the ferrocene-based reference compounds 18 and 38 comparison of the measured IC 50 values with E pc , E HOMO and E i values do not show unambiguous correlation between the cytotoxicity of the novel Tamoxifen analogues and their ability to form radical cations suggesting that the receptor-binding mediated mechanism of actions rather than electron-transfer-related processes are dominant in their antiproliferative effect. However, it is of note that, although reference model 28 has been identified as an antiproliferative agent with confirmed SERM [22] activity, we found this compound efficient on HT-29 and triple-negative MDA-MB-231 cell lines pointing to its action(s) by other mechanism(s) which might be target-mediated and/or partly associated with ROS-production. The view about the feasibility of ROS-mediated mechanism in the action of 28 is apparently in line with the low cationic peak potential (E pc = 0.437 V) measured in acetonitrile for this model. Due to its extensively delocalized "push-pull" electron system comprising two donor 4-hydroxylphenyl moieties and acceptor carbonyl group at the terminal positions, 28 is assumed to undergo O-deprotonation more readily compared to other investigated Tamoxifen analogues without any electron-acceptor terminal. Thus, the low E pc value might be associated with the lowered redox potential of a significant amount of anionic species with enhanced overall electron-density formed in equilibrium deprotonation under the experimental conditions. Thus, the relatively high adiabatic ionization energy (E i,w = 5.430 eV) calculated in water for the neutral molecule does not seem a relevant descriptor of the redox properties of this isatin-derived model. We suggest that under physiological conditions in a tumor cell ROS-mediated dehydrogenation of 28 taking place with sequential deprotonation-promoted SET steps constructs quinone-methide 47 also comprising an N-acylimine-activated cyclic alkene as an additional electrophilic residue (Scheme 8a). According to our hypothesis in the subsequent step 47 alkylates, two equivalents of cellular nucleophiles including, e.g., glutathione, thioredoxine-and ribonucleotid reductases (47→48), finally disrupted redox balance and proliferation in cancer cells. In this regard, it is of pronounced importance that thioredoxine reductases are often overexpressed in cancer cells [37][38][39][40][41][42][43][44]. Supporting this view about the mechanism of action, it has been disclosed that electrophilic quinone-methides (QMs), produced in ROS-mediated sequential SET steps promoted by deprotonation, are active species that exert a cytotoxic effect through reacting with cellular nucleophilic targets critical to maintaining homeostasis of the tumor cells [45][46][47][48][49][50][51][52].
The relatively low cationic peak potential and ionization energy determined for 43 (E pc = 0.742 V, E i,w =4.78 eV) also emerged as one of the most potent compounds presented in this work. This could be connected to its ROS-mediated mechanism of action suggested for 28, implicating the formation of a reactive bis-Michael acceptor 49 (Scheme 8b). In the subsequent step, this alkylating agent also reacts with two equivalents of sulfur and/or selenium-donor nucleophiles vital to proliferation and/or redox balance (49→50), triggering such cellular processes that eventually lead to cell death. On the other hand, based on its highly electron-donor character N-methyl derivative 31 (E pc = 0.637 V, E i,w = 4.76 eV), the closely related structural analogue of 43 would also be expected to exert an antitumor effect via the formation and subsequent transformations of a reactive intermediate type 49 (Scheme 8b). However, this electron donor purely organic compound displayed convincing activity exclusively on HT-29 cells (Table 1). This suggests that, besides target-binding initiated signal-transductions, redox-based mechanisms might prominently contribute to an antiproliferative effect on this cell line. This assumption could be supported by recent studies demonstrating that ROS is extensively implicated in drug-induced antiproliferative effects on HT-29 and other colon cancer cells [43,44,[53][54][55].
Finally, triphenol 46 with a modified Hydroxytamoxifen skeleton of enhanced rigidity also emerged as a potent antiproliferative structure among the novel Tamoxifen analogues. We assume that its cytostatic effect can be mainly due to signal-inducing binding interactions with targets, including estrogen receptors in MCF-7 cells, attenuated by a mechanism of action involving ROS-mediated stepwise generation of quinone-methides 51, 52 and 56 (Scheme 8c), which knockdown homeostasis-and proliferation-maintaining sulfur and/or selenium-donor cellular nucleophiles by alkylation (51 → 54, 52 → 55 and 56 → 57). Although the experimental and theoretical findings on the electron-donor character of 46 (E pc = 0.752 V, E i,w = 5.09 eV) would not justify the operation of a redox-initiated mechanism, it is reasonable to assume that equilibrium deprotonation on the trihydroxylated skeleton takes place readily under physiological conditions to generate electron-rich phenolates enhancing the feasibility of ROS-mediated SET processes. mediate type 49 (Scheme 8b). However, this electron donor purely organic compound displayed convincing activity exclusively on HT-29 cells (Table 1). This suggests that, besides target-binding initiated signal-transductions, redox-based mechanisms might prominently contribute to an antiproliferative effect on this cell line. This assumption could be supported by recent studies demonstrating that ROS is extensively implicated in drug-induced antiproliferative effects on HT-29 and other colon cancer cells [43,44,[53][54][55].

Materials and Methods
All fine chemicals were obtained from commercially available sources ((Merck (Budapest, Hungary), Fluorochem (Hadfield, UK), Molar Chemicals (Halásztelek, Hungary), VWR (Debrecen, Hungary)) and used without further purification. DMF and TEA were distilled from calcium hydride, and THF was distilled from Na-benzophenone. Merck Kieselgel (230-400 mesh, 60 Å) was used for flash column chromatography. The 1 H-and 13 C-NMR spectra of all compounds were recorded in CDCl 3 or DMSO-d 6 solution in 5 mm tubes at RT, on a Bruker DRX-500 spectrometer (Bruker Nano GmbH, Karlsruhe, Germany) at 500. 13 ( 1 H) and 125.76 ( 13 C) MHz and Avance NEO 400 spectrometer (Bruker Nano GmbH, Karlsruhe, Germany) at 400. 16 ( 1 H) and 100.62 ( 13 C) MHz, with the deuterium signal of the solvent as the lock and TMS as the internal standard. The HSQC, HMBC, COSY and NOESY spectra, which support the exact assignments of 1 H-and 13 C-NMR signals, were obtained by using the standard Bruker pulse programs. For each compound characterized in this session, the numbering of atoms used for the assignment of 1 H-and 13 C-NMR signals do not correspond to IUPAC rules reflected in the given systematic names. The identification of the positions of the hydroxyphenyl rings on the molecular skeleton was based on the cross-peaks gathered from NOESY experiments. Elemental analysis was performed on a Vario EL III CHN analyzer (Elementar Analysensysteme GmbH, Langenselbold Germany). All DFT calculations determining optimized structures and energy of the neutral compounds, and their radical cations were carried out using Gaussian 09 software (Gaussian Incorporation, Pittsburgh, PA, USA) package [56]. The optimized structures are available from the authors. A Metrohm Autolab PGSTAT 302N (SelectScience, Corston, UK) electrochemical workstation controlled by the Autolab Nova software was used in all electrochemical experiments. Azetidin-2-one (0.36 g, 5.0 mmol, 1.0 eq.) was dissolved in DMF-TEA 1:1 (10 mL), then K 2 CO 3 (2.05 g, 15.0 mmol, 3.0 eq.), CuI (0.95 g, 5.0 mmol, 1.0 eq.), PdCl 2 (PPh 3 ) 2 (87.5 mg, 0.125 mmol, 2.5%) and 1:1 mixture of iodoferrocene and ferrocene (3.40 g, 5.5 mmol, 1.1 eq.) were added. The reaction mixture was stirred under argon, at 110 • C for overnight, cooled down, filtrated and brine (30 mL) was added to the solution. The precipitated solid was filtered off, washed with brine and water then purified by column chromatography on silica, using DCM as eluent. The product was crystallized EtOH. 1-Ferrocenylazetidin-2-one 11 (2.60 g, 10.2 mmol, 1.0 eq.) was dissolved in 1,2-dichloroethane (40 mL), and the solution was cooled to 0 °C. To the reaction mixture, trifluoromethanesulfonic acid (1.80 mL, 3.10 g, 20.6 mmol, 2.0 eq.) was added dropwise over 5 min, and the mixture was stirred at 0 °C for an additional 30 min; during this period, the reaction was followed by TLC. When the reaction was completed, the mixture was poured into water and extracted with DCM six times. The combined organic phases were extracted with brine, dried on Na2SO4, and the solvent was evaporated. The residue was purified by filtration through silica gel. 1-Ferrocenylazetidin-2-one 11 (2.60 g, 10.2 mmol, 1.0 eq.) was dissolved in 1,2dichloroethane (40 mL), and the solution was cooled to 0 • C. To the reaction mixture, trifluoromethanesulfonic acid (1.80 mL, 3.10 g, 20.6 mmol, 2.0 eq.) was added dropwise over 5 min, and the mixture was stirred at 0 • C for an additional 30 min; during this period, the reaction was followed by TLC. When the reaction was completed, the mixture was poured into water and extracted with DCM six times. The combined organic phases were extracted with brine, dried on Na 2 SO 4 , and the solvent was evaporated. The residue was purified by filtration through silica gel. 1-Ferrocenylazetidin-2-one 11 (2.60 g, 10.2 mmol, 1.0 eq.) was dissolved in 1,2-dichloroethane (40 mL), and the solution was cooled to 0 °C. To the reaction mixture, trifluoromethanesulfonic acid (1.80 mL, 3.10 g, 20.6 mmol, 2.0 eq.) was added dropwise over 5 min, and the mixture was stirred at 0 °C for an additional 30 min; during this period, the reaction was followed by TLC. When the reaction was completed, the mixture was poured into water and extracted with DCM six times. The combined organic phases were extracted with brine, dried on Na2SO4, and the solvent was evaporated. The residue was purified by filtration through silica gel. 1,2-Dihydroferroceno[b]pyridin-4(3H)-one (12) (0.16 g, 0.63 mmol, 1.0 eq.) benzyl bromide (0.11 mL, 0.95 mmol, 1.5 eq.) and Cs 2 CO 3 (0.62 g, 1.90 mmol, 3.0 eq.) were dissolved in DMF (8 mL). The reaction mixture was stirred under argon, at room temperature for overnight. After 12 h of stirring, the reaction mixture was poured into water, extracted with EtOAc three times. The combined organic phases were washed with brine and LiCl to remove DMF completely, dried on Na 2 SO 4 , and the solvent was evaporated. The residue was purified by column chromatography on silica gel, using DCM:MeOH (80:1) as eluent. The product was crystallized from water. 1-Benzyl-1,2-dihydroferroceno[b]pyridin-4(3H)-one (13) (0.20 g, 0.80 mmol, 1.0 eq.), 4-flourobenzyl bromide (0.15 mL, 1.20 mmol, 1.5 eq.) and Cs2CO3 (0.78 g, 2.40 mmol, 3.0 eq.) were dissolved in 10 mL DMF. The reaction mixture was stirred under argon, at room temperature for overnight. After 12 h of stirring, the reaction mixture was poured into water, extracted with EtOAc three times. The combined organic phases were washed with brine and LiCl to remove DMF completely, dried on Na2SO4, and the solvent was evaporated. The residue was purified by column chromatography on silica gel, using DCM:MeOH (80:1) as eluent. The product was crystallized from water. 1-Benzyl-1,2-dihydroferroceno[b]pyridin-4(3H)-one (13) (0.20 g, 0.80 mmol, 1.0 eq.), 4-flourobenzyl bromide (0.15 mL, 1.20 mmol, 1.5 eq.) and Cs 2 CO 3 (0.78 g, 2.40 mmol, 3.0 eq.) were dissolved in 10 mL DMF. The reaction mixture was stirred under argon, at room temperature for overnight. After 12 h of stirring, the reaction mixture was poured into water, extracted with EtOAc three times. The combined organic phases were washed with brine and LiCl to remove DMF completely, dried on Na 2 SO 4 , and the solvent was evaporated. The residue was purified by column chromatography on silica gel, using DCM:MeOH (80:1) as eluent. The product was crystallized from water. 1-Benzyl-1,2-dihydroferroceno[b]pyridin-4(3H)-one (13) (0.20 g, 0.80 mmol, 1.0 eq.), 4-flourobenzyl bromide (0.15 mL, 1.20 mmol, 1.5 eq.) and Cs2CO3 (0.78 g, 2.40 mmol, 3.0 eq.) were dissolved in 10 mL DMF. The reaction mixture was stirred under argon, at room temperature for overnight. After 12 h of stirring, the reaction mixture was poured into water, extracted with EtOAc three times. The combined organic phases were washed with brine and LiCl to remove DMF completely, dried on Na2SO4, and the solvent was evaporated. The residue was purified by column chromatography on silica gel, using DCM:MeOH (80:1) as eluent. The product was crystallized from water.

General Method for McMurry Reactions
Zinc dust (1.57 g, 24.0 mmol, 6.0 eq.) was suspended in THF (50 mL) and 2.19 mL (3.79 g, 20 mmol, 5.0 eq.) TiCl 4 was added dropwise to the suspension which was then held at reflux for 2 h and cooled down to 0 • C. To the resulting mixture, the appropriate ketone (4.0 mmol (1.0 eq.) and 4,4 -dihydroxybenzophenone 15 (0.86 g, 4.0 mmol, 1.0 eq.) or bis(4-methoxyphenyl)methanone 40 (0.97 g, 4.0 mmol, 1.0 eq.) dissolved in 15 mL of THF was added in one portion. The reaction mixture was stirred at reflux temperature for 5 h, cooled down to room temperature, and poured into a solution of 15 g K 2 CO 3 dissolved in 200 mL water. The formed precipitate was filtered off, and the filtrate was extracted with DCM three times. The combined organic phase was washed with brine, dried on Na 2 SO 4 , and the solvent was evaporated. The residue was purified by column chromatography on silica gel, using DCM:n-hexane (1:1-10:1) or DCM or DCM:MeOH (60:1-10:1) as eluent. The products characterized below were crystallized by different mixtures of MeOH and water. 4,4 -(2-Phenylbut-1-ene-1,1-diyl)bis(methoxybenzene) 42 has been characterized [57] THF was added in one portion. The reaction mixture was stirred at reflux temperature for 5 h, cooled down to room temperature, and poured into a solution of 15 g K2CO3 dissolved in 200 mL water. The formed precipitate was filtered off, and the filtrate was extracted with DCM three times. The combined organic phase was washed with brine, dried on Na2SO4, and the solvent was evaporated. The residue was purified by column chromatography on silica gel, using DCM:n-hexane (1:1-10:1) or DCM or DCM:MeOH THF was added in one portion. The reaction mixture was stirred at reflux temperature for 5 h, cooled down to room temperature, and poured into a solution of 15 g K2CO3 dissolved in 200 mL water. The formed precipitate was filtered off, and the filtrate was extracted with DCM three times. The combined organic phase was washed with brine, dried on Na2SO4, and the solvent was evaporated. The residue was purified by column chromatography on silica gel, using DCM:n-hexane (1:1-10:1) or DCM or DCM:MeOH

Cell Culturing and Cytostasis Assay
MCF-7 and MDA-MB-231 cells were cultured in DMEM medium supplemented with 10% FBS, 2 mM L-glutamine, penicillin-streptomycin antibiotics mixture (50 IU/mL and 50 µg/mL, respectively), 1 mM sodium pyruvate and 1% non-essential amino acid mixture. A2058 and HT-29 cells were cultured in RPMI-1640 medium supplemented with 10% FBS, 2 mM L-glutamine, penicillin-streptomycin antibiotics mixture (50 IU/mL and 50 µg/mL, respectively). The cultures were maintained at 37 • C in a humidified atmosphere with 5% CO 2 . The cells were grown to confluency and 24 h before the treatment; they were divided into 96-well tissue culture plates with the initial cell number of 5.0 × 10 3 cells/well. The cells were treated with the compounds in 200 µL final volume containing 1.0 v/v% DMSO at 6.4 × 10 −4 -50 µM concentration range overnight at 37 • C, whereas control cells were treated with serum-free medium only or with DMSO (c = 1.0 v/v%) at the same conditions. After incubation, cells were washed twice with serum-free medium. Following that, cells were cultured for further 72 h in 10% serum-containing medium at 37 • C; then, MTTsolution (at c = 0.37 mg/mL final concentration) was added to each well. The respiratory chain [62] and other electron transport systems [63] reduce MTT and thereby form nonwater-soluble violet formazan crystals within the cell [64]. The amount of these crystals can be determined by spectrophotometry and serves as an estimate for the number of mitochondria and hence the number of living cells in the well [65]. After 3 h of incubation with MTT, the cells were centrifuged for 5 min at 2000 rpm and then the supernatant was removed. The obtained formazan crystals were dissolved in DMSO (100 µL) and optical density (OD) of the samples was measured at λ = 540 nm and 620 nm, respectively, using ELISA Reader (iEMS Reader, Labsystems, Finland). OD 620 values were subtracted from OD 540 values. The percent of cytostasis was calculated with the following equation: Cytostatic effect (%) = [1 − (OD treated /OD control )] × 100 where values OD treated and OD control correspond to the optical densities of the treated and the control wells, respectively. In each case, two independent experiments were carried out with 4 parallel measurements. Cytostasis was plotted as a function of concentration, and the half maximal inhibitory concentration was calculated based on a sigmoid curve fitted on the data points using Microcal™ Origin2018 software. IC 50 represents the concentration of a compound that is required for 50% inhibition expressed in micromolar units.

Conclusions
A series of novel 4-Hydroxytamoxifen analogues was synthetized and evaluated for their in vitro cytostatic and redox properties. The results of cell-viability assays, CV measurements and DFT calculations suggest that the antiproliferative activity of the potent members of the novel Tamoxifen analogues is mainly elicited by their interactions with targets including estrogen receptors. However, it can be assumed that-prominently in HT-29 cells-the target-related effect of model compounds comprising fused pyrrole ring or three hydroxyphenyl groups is probably attenuated by ROS-mediated action. This oxidative mechanism is supposed to implicate the formation of highly reactive bis-Michael acceptor quinone methide-based intermediates capable of alkylating sulfur-and selenium donor cellular nucleophiles such as glutathione, thioredoxine-and ribonucleotid reductases vital to maintaining proliferation and redox balance of the cancer cells. The structureactivity relationships (SAR) and cell-selectivity disclosed in this contribution might be explored in the design of further small potent molecules having electronically tunable Tamoxifen-related scaffolds of which enhanced anticancer potency is associated with synergistic cooperation of ROS-mediated effects and non-oxidative signal transductions initiated by binding to different targets.

Data Availability Statement:
The data generated and analyzed during our research are not available in any public database or repository but will be shared by the corresponding author upon reasonable request.