Biological evaluation of isoflavonoids from Genista halacsyi using estrogen-target cells: Activities of glucosides compared to aglycones

The purpose of this study was to evaluate the response of estrogen target cells to a series of isoflavone glucosides and aglycones from Genista halacsyi Heldr. The methanolic extract of aerial parts of this plant was processed using fast centrifugal partition chromatography, resulting in isolation of four archetypal isoflavones (genistein, daidzein, isoprunetin, 8-C-β-D-glucopyranosyl-genistein) and ten derivatives thereof. 7-O-β-D-glucopyranosyl-genistein and 7,4΄-di-O-β-D-glucopyranosyl-genistein were among the most abundant constituents of the isolate. All fourteen, except genistein, displayed low binding affinity for estrogen receptors (ER). Models of binding to ERα could account for the low binding affinity of monoglucosides. Genistein and its glucosides displayed full efficacy in inducing alkaline phosphatase (AlkP) in Ishikawa cells, proliferation of MCF-7 cells and ER-dependent gene expression in reporter cells at low concentrations (around 0.3 μM). ICI182,780 fully antagonized these effects. The AlkP-inducing efficacy of the fourteen isoflavonoids was more strongly correlated with their transcriptional efficacy through ERα. O-monoglucosides displayed higher area under the dose-response curve (AUC) of AlkP response relative to the AUC of ERα-transcriptional response compared to the respective aglycones. In addition, 7-O-β-D-glucopyranosyl-genistein and 7,4΄-di-O-β-D-glucopyranosyl-genistein displayed estradiol-like efficacy in promoting differentiation of MC3T3-E1 cells to osteoblasts, while genistein was not convincingly effective in this respect. Moreover, 7,4΄-di-O-β-D-glucopyranosyl-genistein suppressed lipopolysaccharide-induced tumor necrosis factor mRNA expression in RAW 264.7 cells, while 7-O-β-D-glucopyranosyl-genistein was not convincingly effective and genistein was ineffective. However, genistein and its O-glucosides were ineffective in inhibiting differentiation of RAW 264.7 cells to osteoclasts and in protecting glutamate-challenged HT22 hippocampal neurons from oxidative stress-induced cell death. These findings suggest that 7-O-β-D-glucopyranosyl-genistein and 7,4΄-di-O-β-D-glucopyranosyl-genistein display higher estrogen-like and/or anti-inflammatory activity compared to the aglycone. The possibility of using preparations rich in O-β-D-glucopyranosides of genistein to substitute for low-dose estrogen in formulations for menopausal symptoms is discussed.

of glucosides compared to aglycones using epithelial estrogen target cells yielded disparate results. For instance, Morito et al. [25] reported that genistein stimulated MCF-7 cell proliferation less than its 7-O-glucoside, although the latter was transcriptionally less active than the aglycone.
As part of our ongoing research for new phytoestrogens from Leguminous plants that form part of the Mediterranean flora [26,27], we used fast centrifugal partition chromatography (FCPC) [28], to isolate isoflavone aglycones and C-and O-glucosides from the methanolic extract of aerial parts of Genista halacsyi Heldr., an endemic plant of Greece. Using many different estrogen-responsive cell lines, we assessed transcriptional, estrogenic, osteoblastic, antiosteoclastic, anti-inflammatory and neuroprotective activities of the isolated isoflavonoids. Our findings suggest that low micromolar concentrations of genistein and its O-β-D-glucopyranosides display similar estrogen-like activities compared to low levels of the hormone, indicating that they might substitute for low-dose estrogen therapy of menopausal symptoms.

General
Chemicals and reagents were from Merck (Darmstadt, Germany). Evaporation of solvents was performed on a vacuum rotary evaporator (Rotavapor R-3000r, Buchi, Switzerland). FCPC was carried out on a Kromaton instrument equipped with a 1000-ml column, adjustable rotation of 200-2000 rpm and a preparative Laboratory Alliance pump with a pressure safety limit of 50 bar. NMR spectra in MeOD were recorded at 400 and 600 MHz (Bruker Advance III 600 MHz and DRX 400). 2D NMR experiments, including correlation spectroscopy (COSY), heteronuclear single-quantum correlation (HSQC) and heteronuclear multiple-bond correlation (HMBC) were performed using standard Bruker microprograms. Electrospray ionisation mass spectrometry (ESI-MS) experiments were performed on a LTQ-Orbitrap XL hybrid mass spectrometer (Thermo-Scientific, Bremen, Germany). Analytical TLC was performed on Merck Kieselgel 60 F 254 or RP-8 F 254 plates. Spots were visualized by UV light (254 and 365 nm) or by spraying with sulfuric vanillin. The plates were then heated for 2 min at 110˚C. Preparative TLC was conducted on PLC Silica gel 60 F254 plates (1 mm). The selected zones were scraped and extracted with ethyl acetate to separate the corresponding compounds. Column chromatography was performed on silica gel 70-230 mesh (63-200 μm). Size exclusion chromatography was performed on Sephadex LH-20.

Plant material extraction and isolation
The aerial parts of Genista halacsyi (Fabaceae) were collected from Mount Parnon in the Peloponnese, Greece. The plant material was identified by Dr. E. Kalpoutzakis. A Voucher specimen has been deposited in the herbarium of the Laboratory of Pharmacognosy and Natural Products Chemistry, Faculty of Pharmacy, University of Athens, Greece, under the number KL121.
Dried pulverized aerial parts of Genista halacsyi (1.5 kg) were extracted exhaustively by maceration using initially CH 2 Cl 2 (3 x 2L) and then MeOH (3 x 2L). The solvents were removed under reduced pressure to give 20.1 g of a crude CH 2 Cl 2 extract and 34.2 g of MeOH extract. The MeOH extract was submitted to fractionation using FCPC in a dual mode methodology. Fourteen solvent systems (Table A in S1 File) were selected and evaluated for their suitability for FCPC using a shaken tube test in combination with TLC. For the evaluation, a small amount of the sample was thoroughly mixed in a vial with equal volumes of the upper and lower phases of the solvent system to test and the solubility of the extract and the settling time of the biphasic system were recorded. The systems that were considered suitable were then evaluated for the distribution of the components of the extract in the two phases. Equal volumes of each phase were applied to a TLC plate and allowed to migrate in the presence of the two-phase solvent system. Optimal systems are expected to give equal distribution of the sample components between the two phases and Rf values of 0.2-0.5. This procedure showed that the biphasic system EtOAc:EtOH:H2O of 10:1:10 was the most appropriate for the fractionation of the MeOH extract of the aerial parts of Genista halacsyi. To proceed with FCPC, 6 L of this solvent system were prepared in a funnel and, after agitation, the phases were separated. Initially, the column (total volume of 1L) was filled with stationary (aqueous) phase using a flow rate of 20 ml/min and the revolution speed was set at 850 rpm. The extract (4.5 g) was dissolved in 30 ml of the two phases and the resulting mixture was introduced into the column, while the organic phase was passed (flow rate 10 mL/min) through the stationary phase in a tail to head or ascending mode. The effluent of the column was collected in 50 mL aliquots and after TLC analysis eight fractions (A-H) were collected. The aqueous layer was then used as mobile phase (flow rate 10 mL/min) in a head to tail or descending mode (reversed phase elution) and four aliquot-combining fractions (I-M) were collected. Fraction A (250.8 mg), containing several apolar flavonoids, as revealed by TLC analysis, was subjected to silica gel chromatography using CH 2 Cl 2 /MeOH of increasing polarity as mobile phase. Preparative TLC analysis of fractions A3 (20.1 mg) and A5 (24.8 mg), which were eluted with CH 2 Cl 2 / MeOH 98/2, afforded biochanin A (1, 4.7 mg) and 8-methoxy-formononetin (2, 8.1 mg), respectively. Furthermore, the A7 (6.9 mg) fraction, eluted with CH 2 Cl 2 /MeOH 95/5, was identified as pure genistein (3, 6.9 mg), while the A8 fraction (28.9 mg), eluted with CH 2 Cl 2 / MeOH 93/7, was chromatographed by preparative TLC and CH 2 Cl 2 /MeOH 90/10 as mobile phase, resulting in isolation of isoprunetin (4, 6.1 mg), daidzein (5, 2.7 mg), 3-methoxyisoprunetin (6, 2.5 mg) and 5-O-methylorobol (7, 3.9 mg). In addition, FCPC fractions D, F, H and K yielded directly in pure form 8-C-glucopyranosyl-genistein (8, 30.4 mg), 8-C-glucopyranosyl-orobol (9, 22.8 mg), 7-O-glucopyranosyl-isoprunetin (10, 24.1 mg) and 7,4-di-O-glucopyranosyl-genistein (11, 28.7 mg), respectively. Finally, 7-O-glucopyranosyl-genistein (12, 26.0 mg), 8-C-glucopyranosyl-3-O-methylorobol (13, 12.7 mg) and 8-C,4-O-diglucopyranosylgenistein (14, 7.9 mg) were isolated from fractions B (152.6 mg), E (124.5 mg) and L (110.2 mg), respectively, using Sephadex column chromatography. All compounds were identified by means of spectral data (HRMS, 1 H-NMR, 13 C-NMR, COSY, HSQC, HMBC) and direct comparison with the respective literature data [29][30][31][32][33].

Cell culture
The mouse preosteoblast cell line 'MC3T3-E1 subclone 4', which is capable of differentiating to mature mineralizing osteoblasts, was purchased from ATCC (ATCC CRL-2593). The cells were cultured in alpha-MEM medium (GIBCO) supplemented with 10% FBS (Biosera), 100 units/ml penicillin and 100 μg/ml streptomycin (Biochrom). Cells were subcultured before reaching confluence (approx. every 2 days). For differentiation to osteoblasts, MC3T3-E1 cells were cultured with DMEM-low glucose medium (Sigma-Aldrich) supplemented with 3% FBS, 10mM beta-glycerophosphate (Sigma-Aldrich) and 50μg/ml ascorbic acid (Sigma-Aldrich). RAW 264.7 mouse macrophages capable of differentiating to multinuclear osteoclasts were purchased from ATCC (ATCC TIB-71). The cells were maintained in alpha-MEM medium supplemented with 10% heat-inactivated ultra-low endotoxin FBS (Biosera), 100 units/ml penicillin and 100 μg/ml streptomycin. Cells were subcultured before reaching confluence (approx. every 2 days). For differentiation to osteoclasts, RAW 264.7 cells were cultured with alpha-MEM medium and 50 ng/ml Receptor Activator of Nuclear Factor kappaB ligand (RANKL, R&D Systems). ER-expressing HT22 neuronal cells were kindly provided by Dr David Schubert (The Salk Institute). The cells were maintained in Dulbecco's Modified Eagles Medium (DMEM) supplemented with 10% FBS at a confluence not greater than 50%. MCF-7 human breast adenocarcinoma cells (from ATCC) and Ishikawa human endometrial adenocarcinoma cells (from ECACC) were cultured as recommended by the suppliers. MDA-MB-231 cells (ATCC) were cultured as already described [34]. MCF-7:D5L cells, a clone of MCF-7 cells stably transfected with an Estrogen Response Element (ERE)-endowed reporter plasmid (pERE-Gl-Luciferase), were generated and cultured as previously described [26]. HEK:ERβ cells, a clone of HEK-293 human embryonic kidney cells stably transfected with an expression plasmid coding human ERβ and an ERE-endowed reporter plasmid (pERE-tk-Luciferase), were generated and cultured as previously described [35]. Unless stated otherwise, cell culture media were from Sigma-Aldrich and FBS from Invitrogen. Dextran-coated-charcoal-treated FBS (DCC-FBS), i.e. FBS treated with 10% DCC to remove endogenous steroids, was prepared as already described [36]. The effect of test compounds on the viability of plated cells was assessed using Trypan blue as already described [34].

Cell differentiation
Differentiation of MC3T3-E1 cells to osteoblasts has been extensively used to screen for compounds that may promote bone formation [37,38]. Differentiation of MC3T3-E1 cells was carried out in 96-well plates using 3,300 cells per well and was assessed via induction of, i) Alkaline Phosphatase (AlkP) activity after 6 days of treatment and, ii) mineralization of extracellular matrix after 21 days of treatment. Briefly, 24 h after plating, the cells were incubated with test compounds or vehicle i.e. the compound diluent (0.1% DMSO) and then exposed for 6 days to differentiation medium in presence or absence of differentiation factors (cf. Cell culture) with a change to fresh compounds and medium in 3 days. AlkP activity was assessed at 405 nm in a Safire II microplate reader using as substrate p-nitrophenyl-phosphate (pNPP, Sigma-Aldrich) as already described [39]. Mineralization of MC3T3 cells was assessed by staining with Alizarin red (Fluka). Cells were cultured and treated as described above for 21 days, with media and test compounds changed every 3 days, and calcium phosphate deposition was assayed as described by Gregory et al. [40]. Briefly, the cells were washed twice with PBS and fixed with 70% ethanol for 15 min on ice. The cells were stained with Alizarin red solution (40 mM, pH 4.2) for 30 min at room temperature, washed twice with distilled water and once with PBS. The dye from the stained mineral deposits was extracted with 33% acetic acid and the absorbance was measured at 405 nm using a Safire II microplate reader (Tecan).
Clones of RAW 264.7 cells competent to differentiate to multinuclear osteoclasts upon activation with RANKL provide valuable information on the regulation of osteoclast differentiation [41]. Differentiation-competent RAW cells (ATCC TIB-71) were seeded in 96-well plates at a density of 9,600 cells per well. The cells were plated in the presence of test compounds or compound diluent (0.1% DMSO) and, 4 h after plating, were exposed for 3 days to 50 ng/ml RANKL or to plain medium. Osteoclastic differentiation was assessed via induction of Tartrate-Resistant Acid Phosphatase (TRAP) activity. The cells were washed with PBS and incubated with 25μl lysis buffer (0.4 M NaCl, 25 mM Hepes pH 7.7, 1.5 mM MgCl 2 , 0.2 mM EDTA, 1% NP40) for 5 min on ice. Then, 25μl of assay solution (100 mM pNPP, 125 mM Sodium Acetate pH 5.2, 1mM L(+) Tartrate) were added followed by incubation at 37 o C for 10 min and the absorbance was measured at 405 nm using a Safire II microplate reader. efficacy of test compounds to prevent oxytosis of HT22 cells was assessed as already described [43]. Briefly, HT22 cells were plated in 96-well flat bottom plates at a density of 4,000 cells per well in 100 μl of DMEM (low glucose) containing 2% FBS. 24 h after plating, the cells were treated with test compounds or compound diluent (0.1% DMSO) and then challenged with 5 mM glutamate for 24 h. Relative numbers of viable cells were determined following conversion of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (MTT, Sigma-Aldrich) to coloured formazan, with the difference in optical density at 550 and 690 nm taken as a measure of viable cell number. Non-challenged cells served to assess test compound effects on cell viability, whereas challenged cells served to assess the neuroprotective activity of test compounds.

Cell proliferation
Test compound effects on the proliferation of MCF-7 and MDA-MB-231 cells were assessed as previously described [34]. Briefly, the cells were plated in 96-flat-bottom-well plates at a density of 8,000 cells per well in phenol-red-free MEM supplemented with 1 μg/mL insulin and 5% DCC-FBS. 24 h after plating, the cells were exposed for 72 h to 0.1 nM estradiol (Sigma-Aldrich), test compounds or vehicle (� 0.2% DMSO), in the absence or presence of 1 μM of the ER-degrader ICI182,780 (Tocris Bioscience), and viable cells were determined using MTT as described above. The difference in optical density at 550 and 690 nm was taken as a measure of the number of viable cells and used to derive the number of compound-treated cells relative to that of vehicle-treated cells. Relative numbers of cells exposed to estradiol or to ICI182,780 acted as positive and negative control, respectively.

Estrogen receptor activities
Transcriptional activity. Assessment of ER/ERE-dependent luciferase gene transcription in MCF-7:D5L (express ERα) and HEK:ERβ cells (express ERβ) and determination of ER transcriptional agonism was carried as already described [44]. Briefly, the cells were plated in 96-well plates at a density of 12,000 cells per well in phenol-red-free MEM (MCF-7:D5L cells) or phenol-red-free DMEM (HEK:ERβ cells) supplemented with 1 μg/mL insulin and 5% DCC-FBS. Three days after plating, the cells were exposed for 18 h to 0.1 nM estradiol, to the indicated concentrations of test compounds or to vehicle (�0.2% DMSO), in the absence or presence of 1 μM ICI182,780, and luciferase expression was assessed in a Safire II microplate reader (Tecan) using the Steady-Glo Luciferase Assay System (Promega). Cells exposed to vehicle, estradiol and/or ICI182,780 served as controls. Dose-response data of ERα/EREdependent luciferase expression were curve fitted using SigmaPlot10 (SPSS Inc) and the area under the dose-response curve (AUC) was calculated using >150 trapezoids and Microsoft Excel. The AUC of luciferase response of 3, 4, 5 and 8 (taken as the archetypal isoflavones) was calculated up to the concentration that provided a response most comparable to that of 0.1 nM estradiol (i.e. up to 0.3, 10, 1 and 3 μM for 3, 4, 5 and 8, respectively). The concentration of 3 (i.e. 0.3 μM) that mounted an estradiol-like luciferase response was used to calculate the AUC of 1, 11 and 12, while that of 4 (i.e. 10 μM) was used to calculate the AUC of 6, 7 and 10, that of 5 (i.e. 1 μM) was used for the AUC of 2, and that of 8 (i.e. 3 μM) was used for the AUC of 9, 13 and 14. The AUC of the isoflavonoids of each group were then expressed as % of the AUC of the respective archetypal isoflavone. Since AUC data combine the efficacy as well as the potency of cell response into a single parameter, the %AUC data were considered here as a more reliable measure of relative ER/ERE-dependent luciferase expression of the test compounds.
AlkP expression. Induction of AlkP expression of Ishikawa cells is considered as a reliable in vitro measure of estrogenic activity [45,46]. Assessment of AlkP expression in Ishikawa cells was carried as already described [26]. Briefly, the cells were plated in 96-well plates at a density of 12,000 cells per well in phenol-red-free MEM supplemented with 1 μg/mL insulin and 5% DCC-FBS. 24 h after plating, the cells were exposed for 72 h to 0.1 nM estradiol, to the indicated concentrations of test compounds or to vehicle (� 0.2% DMSO), in the absence or presence of 1 μM ICI182,780, and AlkP activity was assessed at 405 nm in a Safire II microplate reader using pNPP. Cells exposed to vehicle, estradiol and/or ICI182,780 served as controls. Data on the dependence of AlkP activity on isoflavonoid concentration were curve fitted using SigmaPlot10 (SPSS Inc) and the numerical data of curve fitting were used to determine AUC as described above, using for the calculation of AUC of a particular AlkP response the very same concentration used for the respective luciferase response. The AUC of AlkP responses of the isoflavonoids of a group were expressed as % of the AUC of the luciferase response of the respective archetypal isoflavone.
ER binding. Test compound affinity of binding to ERα and ERβ relative to that of estradiol (relative binding affinity, RBA) was assessed using competitor assay kits with full-length recombinant ERα and ERβ (Invitrogen) as already described [26]. Briefly, the concentration of 17β-estradiol or test compound that inhibited the binding of the fluorescent estrogen ES2 (Invitrogen) to ERα or ERβ by 50% (IC 50 ) was used to calculate RBAα, RBAβ and the ERβbinding selectivity as already described [26].

TNFα mRNA expression
To assess TNFα mRNA expression, RAW 264.7 cells were plated in 6-well plates in alpha-MEM medium supplemented with 10% DCC-FBS, at a density of 300,000 cells per well, maintained in culture for 72 h and then exposed to vehicle (0.3% DMSO) or 3 μΜ test compound for 20 min or 24 h prior to stimulation with 100 ng/ml lipopolysaccharide (Sigma L4391) for 1 h. Extraction of total RNA, reverse transcription (RT) and quantitative PCR (qPCR) analysis of mRNA expression levels were performed as previously described [47]. Relative gene expression levels were calculated by the comparative Ct method using the formula 2 (-ΔCt) . Tumor necrosis factor alpha (TNFα) mRNA levels were normalized to the respective levels of glycerol-3-phosphate dehydrogenase (GAPDH). Test compound effects on the relative number of viable cells were assessed using crystal violet and a Safire II microplate reader (Tecan) as previously described [39]. The difference in optical density at 550 and 690 nm was taken to measure the actual number of viable cells. The following primers were used: Mouse Receptor and ligand preparation. PDB entries 1X7R [48] and 2P15 [49] were used as starting structures for the ligand-binding domain of ERα having H12 in agonist position, in complex with genistein and ortho-trifluoromethyl-phenylvinyl estradiol (EZT) respectively. The protein preparation wizard was utilized as implemented on Maestro Software. All crystallographic water molecules were deleted except the one between ARG 394 and GLU 353. HIS 524 was protonated on NE2. All ligands were designed using Maestro 10.  [50]. The IFD protocol models induced fit docking of ligands using the following steps: 1. An optional constrained minimization of the receptor (protein preparation, refinement only) with an RMSD cutoff of 0.18 Å. Normally, this is done when preparing the protein with the Protein Preparation Wizard.
2. Initial Glide docking of each ligand using a softened potential (van der Waals radii scaling), and optional removal of side chains and application of constraints. By default, a maximum 20 poses per ligand are retained.
3. Prime side-chain prediction for each protein-ligand complex, on residues within a given distance of any ligand pose (default 5 Å), with optional inclusion or exclusion of other residues, and an optional implicit membrane model. Prime minimization of the same set of residues and the ligand for each protein-ligand complex pose. The receptor structure in each pose reflects an induced fit to the ligand structure and conformation.
4. Glide redocking of each protein-ligand complex structure within a specified energy of the lowest-energy structure (default 30 kcal/mol). The ligand is rigorously docked, using default Glide settings, into the induced-fit receptor structure.
6. Estimation of the binding energy (IFDScore) for each output pose.

Statistical analysis
Data were expressed as mean ± standard error of the mean (SEM) of at least three independent experiments carried out in triplicate. Statistically significant differences were determined using one-way ANOVA and SPSS 13.0 software unless stated that t-test was used. Differences were considered significant for values of p � 0.05.

Isolation of isoflavones and isoflavone glucosides
In continuation of a screening programme investigating plants of the Mediterranean basin for compounds with estrogen-like activity, we observed that the methanolic extract of the aerial parts of Genista halacsyi displayed considerable AlkP-inducing activity in Ishikawa cells. In order to fractionate the extract rapidly and effectively, we used FCPC. Initially, fourteen solvents systems ( Table A in S1 File) were evaluated for the solubility of the extract, the settling time of the biphasic system (should be shorter than 30 seconds) and the distribution of the components in the two phases by TLC. The biphasic systems 9, 11, 13 and 14 were rejected due to the formation of emulsion. On the other hand, TLC revealed an inappropriate partition of the ingredients in two phases in the case of systems 1-7, 10 and 12. Finally, the biphasic system EtOAc : EtOH : H 2 O (10 : 1 : 10) was selected as the most suitable. Fractionation of 4.5 g of the extract using 2 L of organic solvents yielded in 4 h several fractions of 4.4 g total weight amounting to 98.2% recovery. The FCPC separation process yielded directly in pure form four isoflavone glucosides (8)(9)(10)(11). In addition, this process, in combination with Sephadex column chromatography, provided three more isoflavone glucosides (12)(13)(14) and, in combination with preparative TLC, seven aglycones (1-7). The structure of the isolated components (Fig 1) was elucidated by High Resolution Mass Spectrometry and 1D and 2D NMR spectroscopy, and by comparison with literature data (Tables B-O in S1 File). The quantities of pure isoflavonoids isolated from 4.5 g of extract ranged from 2.5 (6) to 30.4 mg (8). The total amount of genistein (6.1 mg) and its 7-O-and 7,4di-O-glucosides (28.7 and 26.0 mg, respectively) amounted to 33% of the total amount of isolated isoflavonoids. For the ensuing biological evaluation, the isoflavonoids were taken to belong to 4 groups, each comprising an archetypal isoflavone (underlined) and members deriving from minimal addition(s) to the core structure of the archetypal entity (underlined): the genistein group (1, 3, 11, 12), the isoprunetin (5-methyl-genistein) group (4, 6, 7, 10), the daidzein group (2, 5) and the 8-C-glucopyranosyl-genistein group (8,9,13,14).
We calculated the mean value (and the range) of previously reported RBAα and RBAβ values to be, respectively, 1.  [4,25,52], in good agreement with the data of Table 1.

Models of binding to ERα
To better understand the binding affinity of the 14 isoflavonoids, we run docking calculations and we analyzed the physicochemical properties as described by their logP values. For docking calculations two different structures of ERα were utilized, having helix-12 (H12) of ERα in the agonist position, one in complex with genistein (PDB entry 1X7R) [48] and the other in complex with ortho-trifluoromethyl-phenylvinyl estradiol (EZT, PDB entry 2P15) [49]. The latter is a very good example where the plasticity of the receptor allows fitting of a relatively voluminous estradiol analog (Fig 2A). The binding pocket of ERα has a hydrophobic binding cavity, with only two hydrophilic spots, R394/E353 and H524 in a distance of 12Å. Glucosides 8-14 Estrogen-like activities of isoflavone glucosides compared to aglycones Estrogen-like activities of isoflavone glucosides compared to aglycones displayed very low logP values (Table P in S1 File), reflecting their hydrophilic profile which perturbs binding to ERα and lowers binding affinity for the receptor. Upon Induced Fit Docking, all compounds except 11 and 14 fitted inside the binding pocket of 2P15 ligand binding domain. However, only the aglycons (1-7) were able to bind on 1X7R, while the remainder isoflavonoids were rejected because of their volume. Compounds 11 and 14 contain two glucose residues, of which one is at position 4', and are the most hydrophilic of the fourteen isoflavonoids (Table P in S1 File). Compound 8 bound to ERα very similarly to EZT and genistein (Fig 2B and 2D). However, although 12 was able to bind to ERα with similar orientation (Fig 2C), the inability to form a hydrogen bond with H524 due to the absence of a OH group at position 7 increased the theoretical free energy of binding by 1.2 kcal/mol compared to compound 8.

Agonism of gene transcription, AlkP expression and cell proliferation
We then examined whether the isoflavonoids induce, i) ERE-dependent luciferase gene transcription in MCF-7:D5L cells (known to express ERα), ii) expression of AlkP in Ishikawa endometrial adenocarcinoma cells (known to express ERα and ERβ), iii) stimulate proliferation of MCF-7 breast adenocarcinoma cells (known to express ERα but not ERβ) and, iv) affect proliferation of MDA-MB-231 breast adenocarcinoma cells (known to express neither ERα nor ERβ). We also examined whether ICI182,780, a full antagonist of ER, suppresses modulation of these activities by the 14 isoflavonoids and whether the isoflavonoids suppress induction of these activities by 0.1 nM estradiol (postmenopausal level of the hormone). The isoflavonoids were tested at 1 μM, since higher concentrations are hardly attainable in vivo and hence of little practical interest. Fig 3A shows that luciferase gene transcription was induced 3.36-fold in the presence of 0.1 nM estradiol but not in the presence of 1 μM ICI182,780; that 3, 11 and 12 induced gene transcription to levels very similar to estradiol; that isoflavonoid-induced gene transcription was fully suppressed by ICI182,780; and that none of the isoflavonoids was able to suppress estradiol-induced gene transcription. Fig 3B shows that AlkP expression was induced 3.68-fold in the presence of 0.1 nM estradiol but not in the presence of 1 μM ICI182,780; that 3, 11 and 12 induced AlkP expression to levels very similar to estradiol; that isoflavonoid-induced AlkP expression was fully suppressed by ICI182,780; and that none of the isoflavonoids was able to suppress estradiol-induced AlkP expression. Some isoflavonoids (e.g. 12) displayed somewhat lower AlkP expression compared to vehicle in the presence of ICI182,780; however, these effects were not statistically significant. Fig 3C shows that proliferation of MCF-7 cells was induced 1.80-fold in the presence of 0.1 nM estradiol but not in the presence of 1 μM ICI182,780; that several isoflavonoids induced cell proliferation to levels similar or somewhat lower to estradiol; that all these cell proliferation effects were fully suppressed by ICI182,780; and that none of the isoflavonoids was able to suppress estradiol-induced cell proliferation. Finally, Fig 3D shows that the proliferation of MDA-MB-231 cells was not affected by estradiol, ICI182,780 or any of the isoflavonoids, with the exception of few very minor stimulatory effects which are considered haphazard.
The data of Fig 3 were used to deduce estrogen agonist effects of isoflavonoids at 1 μΜ as compared to the respective agonist effect of 0.1 nM estradiol. Regarding induction of AlkP, only four isoflavonoids displayed statistically significant agonist effects that were similar to (3,11,12) or lower than (1) the effect of estradiol ( Table 2, column 2).
With reference to induction of MCF-7 cell proliferation, eight isoflavonoids displayed statistically significant agonist effects that were similar to (1, 3, 5, 8, 11, 12) or lower than (4, 6) the effect of estradiol (Table 2, column 3). With respect to ERα-dependent induction of gene transcription, seven isoflavonoids displayed statistically significant agonist effects that were higher than (3,12), similar to (1, 11) or lower than (5, 6, 8) the effect of estradiol ( Table 2, column 4). Finally, in relation to ERβ-dependent induction of gene transcription, nine isoflavonoids displayed statistically significant agonist effects that were similar to (3, 6, 11, 12) or lower than (1,4, 7, 8, 14) the effect of estradiol (Table 2, column 5). Agonist effects higher than or similar to the respective effect of estradiol were mostly classified as full, effects lower than the effect of estradiol were classified as partial or weak, while effects that were not statistically significant were classified as marginal (Table 2). Interestingly, the data of Table 2 revealed that AlkP-inducing agonisms were more strongly correlated with transcriptional agonisms through ERα (Pearson's r = 0.945; 2-tailed p = 0.000; n = 14) than through ERβ (Pearson's r = 0.734; 2-tailed p = 0.004; n = 13). Estrogen-like activities of isoflavone glucosides compared to aglycones

Dose-response variations of AlkP-inducing and ERα-transcriptional activities
Having shown that the agonist effects of the O-glucosides of genistein rivaled those of the aglycone when tested at 1 μM, we set out to investigate in more detail how glucose moieties and other substituents in the structure of genistein affect the AlkP-inducing response compared to the ERα-transcriptional response in a dose-dependent manner. Figs 4A and 5A show that luciferase and AlkP responses of 4 and 8 were considerably weaker compared to 3, while those of 11 and 12 were fairly comparable to 3. Similarly, Figs 4B and 5B show that luciferase and AlkP responses of 9, 13 and 14 were weaker compared to 8 and Figs 4C and 5C show that luciferase and AlkP responses of 4, 7 and 10 were considerably weaker compared to 6. Table 3 shows IC 25 (columns 2 and 5) and IC 50 data (columns 3 and 6) from several experiments such as those shown in Figs 4 and 5 as well as AUC data (columns 4 and 7), as normalized on a group basis using the AUC of ERα/ERE-dependent luciferase response of 3 (genistein group), 4 (isoprunetin group), 5 (daidzein group) and 8 (8-C-glucopyranosyl-genistein group) up to 0.3, 10, 1.0 and 3.0 μM, respectively. The AUC data were moderately correlated inversely with IC 50 data (Pearson's r = -0.609; 2-tailed p = 0.027; N = 13) but not with the Table 2. Agonism of AlkP expression, cell proliferation and gene transcription.  N = 13). The ratios of normalized AUC responses of 1 and 11 were lower than the ratio of 3 while the ratio of 12 was higher (Table 3,  column 9). Likewise, the ratio of normalized AUC response of 6 was lower than that of 4 while the ratio of 10 was higher. As well, the ratio of normalized AUC response of 9 was lower than that of 8 while the ratio of 14 was higher. Evidently, monoglucosides 10, 12 and 14 displayed higher AUC ratios compared to the respective archetypal isoflavone, while their 4-methoxy-(1), 3-methoxy-(9) and 3-hydroxy-derivatives (6) and the di-glucoside 11 displayed lower ratios. While seven isoflavonoids displayed an AUC ratio significantly different from the ratio of the respective archetypal isoflavone, only four isoflavonoids displayed an IC 50 ratio significantly different from the IC 50 of the respective archetypal isoflavone, possibly indicative of lower discriminative power of IC 50 -based comparisons relative to AUC-based ones.

Effects on osteoblastic and osteoclastic differentiation and on neuronal oxytosis
Since micromolar concentrations of isoflavonoids 3, 11 and 12 displayed AlkP-inducing activity and ERα-dependent transcriptional activity comparable to physiological concentrations of estradiol but much higher compared to micromolar concentrations of the other isoflavonoids, we also compared their effects on the differentiation of MC3T3-E1 to osteoblasts using induction of AlkP and Alizarin red staining as markers of osteoblastic differentiation and mineralization, respectively. Incubation of MC3T3-E1 cells for 6 days with differentiation factors (DF: 10 mM beta-glucerophosphate, 50 μg/ml ascorbic acid plus test compound vehicle) or DF-free medium (plus test compound vehicle) resulted in 3.7-fold induction of AlkP expression in the presence of DF ( HT22 cells are known to undergo oxidative stress-induced cell death (oxytosis) within 18-24 h following exposure to 5 mM glutamate. We therefore sought to determine whether estradiol, 3, 11 or 12 could affect the viability of HT22 hippocampal neurons under normal growth conditions or following exposure to glutamate. Fig 6 shows that under normal growth Estrogen-like activities of isoflavone glucosides compared to aglycones conditions, the viability of HT22 cells growing in the presence of 1 nM estradiol or 1 μΜ 3, 11 or 12 was not different compared to that of vehicle-treated HT22 cells (set equal to 100%; p>0.05, t-test). Treatment with glutamate reduced cell viability to �40% of that of glutamatefree cells, whether the treatment was carried out in the presence of vehicle, 1 nM estradiol or 1 μM 3, 11 or 12.

Effects on TNFα mRNA expression of RAW 264.7 cells
The mouse macrophage cell line RAW 264.7 is known to express ERα and ERβ. Since the ERdependent activities of 3, 11 and 12 were much higher compared to the activities of the other  Estrogen-like activities of isoflavone glucosides compared to aglycones isoflavonoids, we tested the ability of these three isoflavonoids to modulate transcription of the gene of pro-inflammatory cytokine TNFα in estrogen-free RAW 264.7 cells following 1-h stimulation of the cells with lipopolysaccharide (LPS). Pre-treatment of the cells with 3, 11, 12 or vehicle for 20 min prior to stimulation with LPS resulted in~30-fold induction of TNFα mRNA expression in the presence of vehicle and in 30% reduction of this induction in the presence of 11 (Fig 7A). Pre-treatment of the cells with the test compounds for 24 h prior to the stimulation resulted in~50-fold induction of TNFα mRNA expression in the presence of vehicle and in 15, 29 or 20% reduction of this induction in the presence of 3, 11 or 12, respectively, with the effect of 11 being statistically significant (Fig 7B). None of the isoflavonoids could affect TNFα mRNA expression in the absence of LPS. In addition, none of the isoflavonoids affected cell viability whether in the absence or presence of LPS (Fig 7C).  Estrogen-like activities of isoflavone glucosides compared to aglycones

Discussion
In an ongoing search for phytoestrogens from Mediterranean flora we found that the methanolic extract of the aerial parts of Genista halacsyi induced AlkP in Ishikawa cells. Induction of AlkP activity in Ishikawa cells is considered as a reliable measure of estrogenic activity [45]. This was not unexpected given that isoflavonoids with estrogen-like activity reportedly feature among the secondary metabolites of Genista species [1]. However, the phytochemical composition of Genista halacsyi had not yet been studied, and conventional chromatographic techniques had mainly been used for the isolation of compounds from plants belonging to this genus. We therefore exploited the advantages of FCPC (low risk of sample degradation, reversible adsorption of ingredients and low organic solvent consumption) in order to fractionate the extract. Several two-phase solvent systems have been reported for the chromatographic isolation of isoflavonoids from various natural sources [28]. Using a biphasic system consisting of EtOAc, EtOH and H 2 O we succeeded in fractionating the methanolic extract of G. halacsyi rapidly as well as effectively, and in isolating 14 isoflavonoids in pure form and in substantial amounts. Assessment of RBAα and RBAβ of the isolated isoflavonoids revealed that 3 displayed ERbinding affinities and ERβ selectivity considerably higher than those of the other isoflavonoids. It has been reported that ERα and ERβ bind compounds with OH groups at a distance of 9.7-12.3 Å, through hydrogen bond formation with Glu353/Arg394 and His524 of ERα and Glu305/Arg346 and His475 of ERβ; that the ERβ selectivity of 3 likely depends predominantly on van der Waals contacts between its A and C rings and Ile373 and Met336 of ERβ, respectively; that the 5-OH and the keto group of the flavone part of 3 likely participate in intramolecular hydrogen-bond formation that gives rise to a planar 3-ring flavone system; and that the expanded planar structure of 3 in conjunction with the smaller ligand-binding pocket of ERβ compared to ERα likely provide for tighter packing of ERβ binding pocket residues around the flavone ring system of 3, and hence for pronounced ERβ affinity and selectivity [53][54][55]. It is therefore likely that the smaller planar structure of 5 compared to 3, due to the absence of a 5-OH in the former, reduces the number of ER pocket residues that could pack around the ligand and hence the ER-binding affinity and selectivity of 5 compared to 3. It is also likely that introduction of bulkiness by methylation of 5-OH or by addition of a C-8 glucose moiety to generate 4 or 8, respectively, prevents proper alignment for hydrogen bond formation through 4-OH and 7-OH, thus reducing their affinity and selectivity compared to 3. Notably, 6 and 13 displayed higher RBAα compared to 4 and 8, respectively, likely reflecting a comparatively higher hydrogen bonding ability of the 4-OH of 6 and 13, due to the high electron donor potential of their 3-methoxy group. Similarly, the electron donor potential of 3-OH may account for the higher RBAα of 7 and 9 compared to 4 and 8, respectively. Finally, as expected, 1, 10, 11, 12 and 14, which carry modified 4-OH and/or 7-OH groups, displayed considerably weaker RBA compared to 3, 4 and 8. Although 11 had RBA<0.01, 12 and 14 had measurable RBA for both ER subtypes. On 11, the 7-OH and 4'-OH are substituted, while on 12 and 14, 4'-OH and 7-OH, respectively, are free for hydrogen bond interaction. Although 12 was able to bind to ERα with very low affinity, it displayed estrogen-like activities much higher than excepted based on its RBAα, suggesting that these activities were due to hydrolytic release of the aglycone. Induced fit docking failed to predict binding mode for compound 14. Probably, the binding cavity of 2P15 is not large enough to accommodate compound 14 and more residues should be shifted for binding. It is possible that 14 may have suffered limited hydrolysis at position 4' during its isolation.
It is widely assumed that, although isoflavones display much lower ER-binding affinity compared to estradiol, their estrogen-like activity in vivo can match that of the hormone, given that their concentration in the circulation can reach levels up to 1 μΜ (i.e.~10,000-fold higher than postmenopausal levels of the hormone) [13][14][15]56]. We showed that 0.1 nM estradiol displayed full AlkP-inducing, proliferative and ERα-transcriptional agonisms at concentrations � 0.1 nM and this was the case also with 3, 11 and 12 at concentrations � 0.3 μM. We also showed that all isoflavonoid agonisms were inhibited by ICI182,780, suggestive of ERdependent responses. In addition, we showed that AlkP-inducing agonisms were more strongly correlated with transcriptional agonisms through ERα than through ERβ. This correlation bias was clearly reflected in that e.g. 4 and 6 at 1 μΜ displayed partial and full, respectively, ERβ-dependent transcriptional agonisms, but marginal and weak, respectively, AlkPinducing agonisms and ERα-dependent transcriptional agonisms. Thus, it appears that, although both ER subtypes are involved in regulating AlkP expression [51], the regulation is predominantly driven by ERα, presumably due to the dominant role that the latter reportedly plays in heterodimers with ERβ [9]. Moreover, we showed that none of the isoflavonoids displayed ER antagonist activity at 1 μΜ, suggesting that the substituents introduced in the archetypal structures failed to give rise to entities with antiestrogen-like and/or antiproliferative activity. However, this was reportedly not the case with some alkyl substitutions of the 7-OH of daidzein [57].
While the archetypal isoflavones 3, 5 and 8 displayed higher activities compared to the other isoflavonoids of the respective group, 6 displayed higher activities compared to 4, in accordance with their RBAα. However, the RBAα of 7, 9 and 13 could not account for their lower activities compared to 4 and 8, possibly reflecting variations in the ability of isoflavone-bound ERα to recruit transcriptional coactivators and/or to variations in cell uptake, metabolism and efflux of the aglycones. It has been reported that differential cofactor recruitment by phytoestrogen-bound ER may cause induction potencies to deviate considerably from predictions based solely on RBA [6]. In addition, it has been reported that the cellular availability of o-diphenolic compounds such as 7 and 9 could be drastically reduced by oxidation to ortho-quinones and by phase II metabolism to rapidly excreted sulfonate and glucuronide esters [13,14,58,59].
Unlike aglycones, which are more or less readily transported across plasma membranes, O-glucosides need processing by β-glucosidases before transport [11,12]. Deglycosylation by intestinal β-glucosidases is critically involved in the uptake of O-glucosides in humans and likely accounts for the significant variations observed in aglycone bioavailability [10,11,12]. O-glucosides 11 and 12, in particular, displayed fairly similar AlkP-inducing and ERα-transcriptional activities compared to 3. The presumptive beta-glucosidase activity that released 3 from 11 and 12, is unlikely to have originated from the culture medium or from dead cells, given that cell viability was >95%, and serum was heat-inactivated for 1 h [60]. Most likely, this activity was secreted to the medium by living cells and/or was located in their plasma membrane [61]. In contrast to 11 and 12, O-monoglucoside 14 displayed lower AlkP-inducing and ERα-transcriptional activities compared to 8, implying that release of the 4-O-linked glucose moiety of 14 could be impeded by its 8-C-linked glucose moiety.
Although potency is widely used to report data on in vitro biological evaluation of test compounds, an AUC metric fitted over the entire range of comparatively relevant test compound  (3 μΜ) or vehicle (0.3% DMSO) prior to stimulation of the cells with LPS (100ng/ml) for 1h. Values are expressed as % of TNFα mRNA expression following treatment of vehicle pre-treated cells with LPS. (C) Relative numbers of RAW 264.7 cells pre-treated for 24 h with test compounds or vehicle and exposed or not to LPS, as assessed using crystal violet. Absorbance at 550 nm of vehicle pre-treated cells not exposed to LPS was set equal to 100. Values are mean ±SEM of at least three independent experiments carried out in triplicate. � , p<0.05 vs TNFα mRNA levels in vehicle pre-treated LPS-treated cells (ANOVA). https://doi.org/10.1371/journal.pone.0210247.g007 Estrogen-like activities of isoflavone glucosides compared to aglycones concentrations combines both efficacy and potency and is considered as a more reliable measure of bioactivity [62,63]. The AUC of AlkP and ERα-transcriptional responses of isoflavonoids of a group were normalized using the AUC of the transcriptional response of the respective archetypal isoflavone. The relative AUC ratios revealed that the 3-OH and 3-methoxy derivatives 6, 7, 9 and 13 displayed AlkP responses~30% lower than ERα-transcriptional responses, possibly reflecting a lower availability of these derivatives compared to the archetypal isoflavones that predominantly affected the AlkP response. Similarly, the lower relative AUC ratio of 11 compared to 3 may indicate a lower availability of di-glucoside-derived aglycone compared to the free one that disfavored the AlkP response. In contrast, monoglucosides 10, 12 and 14 displayed AlkP responses~40% higher than ERα-transcriptional responses, possibly indicating a relative availability of monoglucoside-derived aglycones compared to the free ones that favored the AlkP response. Indeed, comparison of the rank order of AlkP-induction efficacies of 3, 11 and 12 at 0.3 μM (3 = 12>11; Fig  5A) to the rank order of the respective transcriptional efficacies (3>12 = 11; Fig 4A) could partly account for the finding that 11 and 12 displayed lower and higher, respectively, AUC ratio of AlkP response to ERα-transcriptional response compared to the aglycone.
Differentiation of MC3T3-E1 cells to mineralizing osteoblasts reportedly recapitulates in vitro many estrogen and ERα-dependent effects that promote osteoblastic differentiation in vivo and therefore has been extensively used to screen for phytoestrogens with osteoprotective activity [37,38]. We showed that estradiol, 11 and 12 promoted differentiation of MC3T3-E1 cells to osteoblasts, while 3 was not convincingly effective in this respect. In addition, estradiol and 12 promoted mineralization of differentiated MC3T3-E1 cells, while 11 was not convincingly effective in this respect. Part of these findings is in accordance with previous reports on effects of estradiol and genistein on the osteoblastic differentiation and mineralization of MC3T3-E1 cells [64]. In addition, they show that the O-glucosides of genistein stimulate osteoclastic differentiation of MC3T3-E1 cells as effectively as estradiol.
We also looked for effects of estradiol, 3, 11 and 12 on osteoclast differentiation of RAW 264.7 cells following treatment with RANKL (50 ng/ml) using induction of TRAP expression as differentiation marker. We observed no change of the basal or the RANKL-induced TRAP activity in the presence of 1 nΜ estradiol or 1 μM of 3, 11 or 12. These findings are in accordance with previous reports that estradiol and genistein can inhibit osteoclastic differentiation of RAW 264.7 cells at concentrations �10 μM [65,66]. Similarly, it has been reported that the inhibition of differentiation of RAW 264.7 cells to osteoclasts by resveratrol and other flavonoids depends predominantly on their antioxidant activity and that the inhibition is appreciable provided these compounds are tested at high concentrations [65,67]. Thus, it appears that estradiol, 3, 11 and 12 at physiologically and/or pharmacologically relevant concentrations are unable to inhibit osteoclastic differentiation of RAW 264.7 cells. In line with this notion, it has been reported that stimulation of osteoblastic differentiation of MC3T3-E1 cells by soybean extract or silibinin (a flavonolignan) endows the culture medium with factors capable of inhibiting osteoclastic differentiation of RAW 264.7 [68,69], implying that isoflavones inhibit differentiation of monocytes to osteoclasts indirectly.
Using HT22 cells, a ER-expressing hippocampal cell line used for screening chemical libraries for neuroprotective compounds [42,43,70,71], we showed that estradiol at 1 nM and 3, 11 or 12 at 1 μM failed to display neuroprotective activity. Estradiol is reportedly able to protect glutamate-challenged HT22 cells from oxytosis in an ER-dependent manner, although at supraphysiological concentrations (IC50>800 nM) and in an ICI182,780-independent manner at high physiological concentrations (50 nM) [71], implying that the involvement of ER in estrogen-dependent oxytosis prevention in HT22 cells is minor, if not obsolete. Others have also reported that even 1 μM estradiol or 2.5 μM genistein are unable to prevent oxytosis of glutamate-challenged HT22 cells [72,73]. However, given that the half maximal concentration for HT22 oxytosis inhibition by renowned antioxidants such as resveratrol and fisetin is �3 μM [43,70], one cannot exclude that 3, 11 and 12 may inhibit oxytosis at higher (yet not pharmacologically relevant) concentrations, as already shown for estradiol [72].
LPS is predominantly involved in stimulating TNFα mRNA expression and production of this cytokine following infection of monocytes. We have shown that compared to 3 which was ineffective and to 12 which was marginally effective, 11 displayed a~30% suppression of LPSstimulated TNFα mRNA expression in estrogen-free RAW 264.7 cells pre-incubated with 3 μM isoflavonoid prior to the stimulation. The higher efficacy of the 7,4-di-O-glucoside of genistein to prevent LPS-induced TNFα transcription compared to the respective 7-O-monoglucoside and the aglycone may reflect a higher availability of di-glucoside-derived aglycone compared to the mono-glucoside-derived one and to the free aglycone. However, others have reported that pre-incubation with low micromolar concentrations of genistein is capable of causing 30-40% inhibition of TNFα gene transcription in RAW 264.7 cells [74]. These findings indicate that the extent to which genistein can moderate LPS-induced TNFα mRNA expression of RAW 264.7 cells may depend on cell culture conditions and/or on using specific clones of RAW264.7 cells [41]. Although the mouse macrophage cell line RAW 264.7 is known to express ERα and ERβ, estradiol was reportedly unable to suppress TNFα mRNA expression and release of this cytokine following stimulation of estrogen-free RAW 264.7 cells with LPS [75]. Others have reported, however, that pretreatment of estrogen-free RAW 264.7 cells with 1-10 μM equol reportedly resulted in dose-dependent suppression of LPS-induced TNFα mRNA expression and cytokine release in an ER-independent manner [76].
The genitourinary syndrome of menopause (GSM) is known to affect more than 50% of menopausal women, especially breast cancer survivors, to whom even low-dose vaginal estrogen may be unacceptable due to concerns over possible breast or endometrial cancer risk [77,78]. While orally administered isoflavones are extensively metabolized and rapidly conjugated and excreted, topically or parenterally administered isoflavones are found in the circulation in unconjugated form in appreciable amounts [12]. Accordingly, topically administered isoflavones may reportedly relieve GSM as effectively as vaginal estrogen [79]. The present data on the estrogen-like activity of isoflavonoids may be taken to suggest that formulations rich in O-glucosides of genistein could be useful for topical treatment of GSM.

Conclusion
We have determined the estrogen-like activities of 14 isoflavonoids isolated from the aerial parts of Genista halacsyi using FCPC. We have shown that the estrogen-like activities of O-glucosides were similar to those of the respective aglycones, while the activities of C-glucosides were lower. We have also shown that low micromolar concentrations of genistein and its 7-Omono-and 7,4-di-O-glucosides were at least as active as physiological concentrations of estradiol. Our findings could be taken to indicate that topical use of formulations rich in O-glucosides of genistein could substitute for low-dose vaginal estrogen for the treatment of GMS. However, further studies using the ovariectomized mouse model of menopause are required in order to determine whether topical application of genistein and/or its glucosides may have GSM chemopreventive and/or therapeutic potential.
Supporting information S1 File. Biphasic solvent systems tested for the FCPC analysis; Spectroscopic data for compounds 1-14; LogP values of compounds 1-14 as predicted from QikProt software. (PDF)