Differential effects of various soy isoflavone dietary supplements ( nutraceuticals ) on bacterial growth and human fibroblast viability

Flavonoids, polyphenolic compounds present in many food products, affect growth of different bacterial species when tested as purified or synthetic substances. They can also influence gene expression in human cells, like fibroblasts. Here, we asked if soy isoflavone extracts, commonly used in many products sold as anti-menopausal dietary supplements, influence bacterial growth similarly to a synthetic isoflavone, genistein. Four commercially available products were tested in amounts corresponding to genistein concentrations causing inhibition of growth of Vibrio harveyi (a model bacterium sensitive to this isoflavone) and Escherichia coli (a model bacterium resistant to genistein). Differential effects of various extracts on V. harveyi and E. coli growth, from stimulation, to no changes, to inhibition, were observed. Moreover, contrary to genistein, the tested extracts caused a decrease (to different extent) in viability of human dermal fibroblasts. These results indicate that effects of various soy isoflavone extracts on bacterial growth and viability of human cells are different, despite similar declared composition of the commercially available products.


INTRODUCTION
Flavonoids are polyphenolic compounds present in many food products, including vegetables, fruits, tea and others (Chen et al., 2018).They are known as molecules of various biological activities (Mierziak et al., 2014;Mouradov & Spangenberg, 2014).Among these activities, some have been recognized as beneficial for human health, and particularly the following features were reported: antioxidative properties, implication in coronary heart disease prevention, hepatoprotective, anti-inflammatory, and anticancer functions, as well as antiviral and antimicrobial actions (Kumar & Pandey, 2013).Experimental studies indicated that flavonoids, especially isoflavones, can be also considered as drugs for some genetic diseases, particularly cystic fibrosis, mucopolysaccharidosis (reviewed in Wegrzyn, 2012;Wegrzyn et al., 2010), and Huntington's disease (Pierzynowska et al., 2018).
It was previously demonstrated that certain flavonoids have antibacterial activities which may result from inhibition of synthesis of macromolecules (Ulanowska et al., 2006).Recent reports indicated that flavonoids can significantly influence human intestinal microbiota by inhibiting growth of certain bacterial species (Coppo & Marchese, 2014;Xie et al., 2015).This is also true for dietary flavonoids and those present in food supplements or nutraceuticals (Marín et al., 2015).
One of the most popular groups of dietary supplements and nutraceuticals are soy isoflavones (D 'Adamo & Sahin, 2014).Isoflavone-rich soy extracts are commonly used products recommended against menopausal symptoms (Franco et al., 2016).On the other hand, isoflavones are one of the strongest antimicrobial agents among flavonoids, with genistein giving the most pronounced results (Ulanowska et al., 2006;Ulanowska et al., 2007).Therefore, the aim of this work was to test effects of commercially available dietary supplements/nutraceuticals containing soy isoflavone extracts (rather than purified isoflavones or their mixtures with precisely defined content) on model bacterial species.We asked how these commonly available dietary supplements and nutraceuticals might influence growth of bacteria.As models, the most sensitive to isoflavones -Vibrio harveyi, and the most resistant to these compounds -Escherichia coli, bacterial species (as determined previously by Ulanowska et al., 2007), were chosen.Moreover, effects of these extracts on viability of human cells (dermal fibroblasts) were tested.This type of cells was chosen since it was demonstrated that isoflavones can cause changes in expression of many genes in human fibroblasts, including those involved in glycosaminoglycan metabolism and lysosome biogenesis (Moskot et al., 2015a), cell cycle and DNA replication (Moskot et al., 2015b), proliferation and migration (Kim et al., 2015;Moskot et al., 2016).These studies were also substantiated by the fact that previous reports indicated that various commercially available soy extracts differ considerably in the actual amounts of isoflavones relative to those declared by manufacturers (Setchell et al., 2001;Chua et al., 2004;Piotrowska et al., 2010).Therefore, we assumed that testing biological activities of these products is of particular interest.Our results indicated different effects of various soy isoflavone and diluted to five different concentrations for obtaining calibration plots.Two tablets from one blister of each product were pulverized to powder (after removal of the coating, if necessary) and suspended in DMSO.The suspensions were placed in ultrasonic bath for 30 min, then each of them was filtered through a 0.45 µm membrane and analyzed by UPLC after dilution to suitable concentration.The chromatography was performed using Shimadzu NexeraX2 system.Compounds were separated on a 100×2.1 mm, 2.6 µm, C8 Kinetex column using 0.1% aqueous TFA (solvent A) and 80% acetonitrile + 0.1% TFA (solvent B) as a mobile phase with a linear gradient 0-100% B for 15 min.Shimadzu ELSD-LT II detector was connected to this LC system.Purified nitrogen was used as a nebulizing gas.The drift-tube temperature and gain of the detector were 40°C and 9, respectively.
To prepare calibration plots, five standard working solutions containing different concentrations of genistein were injected in triplicate.Integrated peak area (Y) was plotted against the amount of the substance in each standard sample (X, µg).Calibration plots were constructed directly on the basis of linear regression analysis.As the intensity of the scattered light is a function of the mass of the scattering particles, the approximate amount of isoflavones (daidzein, genistein) and their glycosides (daidzin, genistin), in each dietary supplement was calculated from the calibration plot obtained for genistein.The obtained data are presented in Table 2.The identity of each compound was confirmed by ESI-MS using Shimadzu LC MS-IT-TOF apparatus (exemplary data for Extract 4 are presented in Table 3).
Monitoring of bacterial growth.Bacteria were cultured overnight under conditions described above.Following dilution of the culture with fresh medium, at the v/v ratio 1:100, bacteria were cultured under the same conditions, and A 595 was measured every 15 min.The extracts were purchased in the form of tablets, which were crushed and dissolved in DMSO.Such extracts were added to the bacterial cultures at the early exponential phase of growth (A 595 =0.1), to final concentrations corresponding to 30, 60 and 100 µM isoflavones (included in each extract).In control experiments, either DMSO (final concentration 0.1%; negative control) or genistein (final concentration 30, 60 or 100 µM; positive control) were added.Growth curves were analyzed, and bacterial generation times were calculated on the basis of A 595 values at the exponential phase of bacterial culture growth.
MTT cell viability assay.The assay was performed according to Wasilewski and coworkers (Wasilewski et al., 2017), with some modifications.Briefly, 4 × 10 3 HDFa cells were passaged in each well of a 96-well plate, and allowed to attach overnight.Cells were then treated with DMSO (final concentration 0.1%; negative control), or 30, 60 or 100 µM genistein (positive control), or extracts (at final concentrations corresponding to 30, 60 or 100 µM isoflavones, included in each extract) at 37°C.After 24 or 48 h incubation, 25 µl of MTT solution (4 mg/ml) was added to each well.Following 3 h incubation at 37°C, formazan crystals, formed in living cells, were dissolved in 100 µl of DMSO.Absorbance was measured at 570 nm and 620 nm (reference wavelength) in a Victor 3 microplate reader.
Statistical analyses.One-way ANOVA test was used for testing statistical significance of obtained results versus control (no added compound).Two-way ANOVA test was employed to assess correlation between the tested extract and isoflavone concentration.In both tests, differences were considered statistically significant when p < 0.05.

RESULTS AND DISCUSSION
It was reported previously that determination of growth rate, particularly generation time, in a liquid culture can be considered as an optimal method for estimation of effects of flavonoids on bacteria (Ulanowska et al., 2007).This is because methods based on diffusion are problematic due to low rate of diffusion of flavonoids, and other factors (like inoculum size, disk size, incubation period) may influence the results (Cushnie & Lamb, 2005).Therefore, to determine if the tested soy isoflavone extracts influence growth of two bacterial species, V. harveyi (previously demonstrated to be sensitive to genistein by Ulanowska et al., 2006;Ulanowska et al., 2007), and E. coli (previously demonstrated to be resistant to genistein by Ulanowska et al., 2006;Ulanowska et al., 2007), we monitored growth of bacterial liquid cultures and calculated generation times in the presence and absence of isoflavone(s).
Representative growth curves are presented in Fig. 1 (for V. harveyi) and Fig. 2 (for E. coli), and calculated generation times are shown in Table 4 (for V. harveyi) and Table 5 (for E. coli).Generation times were calculated for two time periods, first, 0-30 min, where 0 is  the time of the tested compund or extract addition), and second, between 60 and 120 min after tested compound or extract addition (i.e.60-120 min).As expected, genistein impaired growth of V. harveyi in both time periods, while it had no significant effect on E. coli.However, differential effects were observed for various soy isoflavone extracts.Extract 1 stimulated V. harveyi growth shortly after its addition to the culture (0-30 min), while strongly inhibiting the growth later on.Similar, though less pronounced effects were observed for extract 3. On the other hand, extract 2 caused only some stimulation of V. harveyi growth shortly after its addition to the culture,  and had no significant effect in the 60-120 min time interval.Extract 4 revealed the weakest effect on V. harveyi growth, as the only considerable change in growth rate was noted at the highest tested concentration between 60 and 120 min after culture supplementation.Although genistein did not influence growth of E. coli, extracts 1 and 2 caused a significant decrease in the generation time immediately after their addition to the culture.Moreover, addition of extract 3 at the highest concentration resulted in a longer generation time of the E. coli culture in the 60-120 min time interval.
Statistical analyses performed with the use of two-way ANOVA indicated that there is a significant correlation between the kind of extract and isoflavone concentration when measuring effects on bacterial growth, at both tested time periods (p < 0.05 in each case).These analyses confirmed the conclusions presented in preceding paragraphs.
The results described above indicated that despite using the same concentrations of soy isoflavones present in various commercially available products, their effects on bacterial growth were different.These differences might arise from activities of other compounds (either identified molecules or unidentified substances from extracts or both) present in the products (see Table 1).Different extracts, investigated in this work, contain various additional compounds, including well-characterized molecules, like calcium, vit.D3, vit.E, thiamine, riboflavin, vit.B6, vit.B12, biotin, folic acid, cellulose, titanium dioxide, magnesium stearate, talc, and glycerol.In fact, some of them are known as stimulators of bacterial growth (e.g.vitamins, glycerol) while others might be suspected to inhibit it, particularly those included in poorly characterized extracts from cone hops and flax seeds, as well as bee wax and Carnauba wax.On the other hand, discrepancies between the declared (by manufacturers) and actual compositions of extracts were Generation times were calculated for two periods: between 0 and 30 min after addition of genistein or extract (or 0.1% DMSO in control experiments), and between 60 and 120 min after addition of genistein or extract (or 0.1% DMSO in control experiments).Results are presented as mean values from at least 3 independent experiments ± S.D. p values were calculated using one-way ANOVA, statistically significant differences were assumed when p < 0.05 (values marked with asterisks).N/A, not applicable.
reported previously for other products (Piotrowska et al., 2010;Andres et al., 2015).Irrespective of the cause of this phenomenon, differential effects of tested extracts on bacterial growth are evident.This might be especially important in the light of recent studies which demonstrated that flavonoids can significantly influence the human gut microbiome Coppo & Marchese, 2014;Marín et al., 2015).
Other differential biological effects of various soy isoflavone extracts were reported previously, when synthesis of glycosaminoglycans was tested in human fibroblasts (Piotrowska et al., 2010).Therefore, we have also tested the effects of investigated extracts on viability of human cells (the HDFa cell line).By using the MTT test, we estimated the fraction of viable cells after different times of treatment with the tested extracts at different concentrations.
In control experiments, according to previously published results (Kloska et al., 2012), we found no signifi-cant cytotoxic effects of synthetic genistein (Figs. 3 and  4).Although genistein is a known inhibitor of tyrosine kinases (Peterson, 1995), it can also stimulate activities of various transcription factors, like TFEB (Moskot et al., 2014) affecting expression of many genes, and it was demonstrated that isoflavones may promote human dermal fibroblast proliferation and migration via TGF-β signaling (Kim et al., 2015).Thus, inhibition of tyrosine kinases' activities might be compensated, at least at certain genistein concentrations.However, 24 h incubation with the tested extracts was enough to decrease viability of human fibroblasts by extract 1 (at all tested concentrations) and extracts 2 and 3 at certain concentrations (Fig. 3).Even more pronounced cytotoxic effects were observed after 48 h incubation, when viability of human cells was significantly decreased, particularly in the presence of extracts 1 and 2, and to some extent in the presence of extracts 3 and 4 (Fig. 4).Generation times were calculated for two periods: between 0 and 30 min after addition of genistein or extract (or 0.1% DMSO in control experiments), and between 60 and 120 min after addition of genistein or extract (or 0.1% DMSO in control experiments).Results are presented as mean values from at least 3 independent experiments ± S.D. p values were calculated using one-way ANOVA, statistically significant differences were assumed when p < 0.05 (values marked with asterisks).Soy extracts and growth of bacterial and human cells These results indicated again that various soy isoflavone extracts exhibit different biological activities despite the same content of isoflavones declared by their manufacturers.This is true for effects on both, bacterial and human cells.Although synthetic genistein demonstrated a high level of safety both, in vitro (Kloska et al., 2012) and in vivo during clinical studies (Kim et al., 2013), as well as beneficial profile of antimicrobial effects (Ulanowska et al., 2007), confirmed in this study, some extracts revealed significant effects on E. coli growth, and their cytotoxicity to human fibroblasts was also detected.Since it was demonstrated that different flavonoids can shape the unique profile of gut microbiota (Huang et al., 2016), one might assume a potential influence of isoflavone-rich diet supplements or nutraceuticals on human intestinal flora.It is worth to mention that studies from different countries indicated huge variability in compositions of different soy isoflavone extracts.Setchell and coworkers (2001) analyzed 33 phytoestrogen supplements and extracts, derived from different countries (USA, the Netherlands, and UK) and found considerable differences in the isoflavone content from the one claimed by the manufacturers; in some products the differences were 5 to 10-fold, and one product contained no detectable amount of isoflavones.The level of impurities was high in most of the products.Chua et al. (2004) found that among 13 commercially available extracts (purchased in the Washington State, USA), only 4 contained at least 90% of the isoflavone content claimed on the label.Moreover, in 2 products, impurities exceeded 40% of the total mass of the extract.
Piotrowska and coworkers (2010) analyzed 7 commercially available soy isoflavone extracts, and determined that amounts of isoflavones were similar to those declared by manufacturers in only 3 of them.Other products contained significantly less isoflavones than claimed on the label, while one contained a 200-fold lower level.
Considerable contaminations were detected in isoflavone-poor extracts.
Determination of isoflavone amounts in extracts investigated in this work, performed by using UPLC-ELSD and MS, indicated that the actual contents of these compounds also differ considerably from those declared by manufactures.According to commercial description of the products, the amounts of isoflavones per tablet were 50, 50, 35, and 26 mg in extracts 1, 2, 3, and 4, respectively, while the values estimated experimentally by us were approximately 130, 27, 30, and 65 mg, respectively (when considering glycosides and aglycans together).Moreover, glycosides were highly predominant over aglycans which might influence biological activities of the extracts, beside actions of other compounds present in the tested products.
In conclusion, results presented in this report and the facts discussed above should be taken into account when the use of soy isoflavone extracts is considered, especially as local concentrations of the compounds included in a given tablet can reach high levels in the human intestine after oral administration of these nutraceuticals or dietary supplements.

Figure 1 .
Figure 1.Effects of genistein and different extracts on V. harveyi BB7 growth.Bacteria were cultured in the BOSS medium at 30oC, and different compounds were added at time 0 to indicated concentrations.Representative growth curves are shown.

Figure 2 .
Figure 2. Effects of genistein and different extracts on E. coli MG1655 growth.Bacteria were cultured in the LB medium at 37oC, and different compounds were added at time 0 to indicated concentrations.Representative growth curves are shown.

Figure 3 .
Figure 3. Effects of genistein and different extracts on human cells (the HDFa line) after 24 h incubation.Either genistein or different extracts were added to the culture medium to indicated final concentrations.DMSO was added in control experiments to 0.1% final concentration.The MTT test was performed.Mean values from 3 independent experiments are shown with error bars indicating S.D. Statistically significant differences (p < 0.05 in one-way ANOVA) are marked with asterisks either above particular columns (differences vs. DMSO control) or above brackets (differences vs. genistein).

Table 2 . The approximate amount (in mg) of daidzin, genistin, daidzein and genistein in one tablet of each dietary supplement ex- amined.
The values were calculated from results of 3-6 assays and are presented as mean value ± S.D.