Soy isoflavones and their metabolites modulate cytokine-induced natural killer cell function

Soybeans are a rich source of isoflavones that have been linked with anti-inflammatory processes and various health benefits. However, specific mechanisms whereby soy bioactives impact immune cell subsets are unclear. Isoflavones, such as genistein and daidzein, are metabolized by microbes to bioactive metabolites as O-desmethylangolensin (O-DMA) and equol, whose presence has been linked to health benefits. We examined how soy isoflavones and metabolites impact natural killer (NK) cell signaling and function. We observe no impact of isoflavones on viability of healthy donor peripheral blood mononuclear cells (PBMCs) or NK cells, even at high (25 µM) concentrations. However, pre-treatment of PBMCs with physiologically-relevant concentrations of genistein (p = 0.0023) and equol (p = 0.006) decreases interleukin (IL)-12/IL-18-induced interferon-gamma (IFN-γ) production versus controls. Detailed cellular analyses indicate genistein and equol decrease IL-12/IL-18-induced IFN-γ production by human NK cell subsets, but do not consistently alter cytotoxicity. At the level of signal transduction, genistein decreases IL-12/IL-18-induced total phosphorylated tyrosine, and phosphorylation MAPK pathway components. Further, genistein limits IL-12/IL-18-mediated upregulation of IL-18Rα expression on NK cells (p = 0.0109). Finally, in vivo studies revealed that C57BL/6 mice fed a soy-enriched diet produce less plasma IFN-γ following administration of IL-12/IL-18 versus control-fed animals (p < 0.0001). This study provides insight into how dietary soy modulates NK cell functions.

individuals, daidzein can be processed into its secondary metabolites, O-desmethylangolensin (O-DMA) and equol. It is proposed that this process is impacted by the presence of particular gut bacteria and their functional capabilities that vary for uncertain reasons among individuals. In various human populations it is estimated that approximately 30-50% of individuals have the ability to produce equol upon ingestion of soy, while 80-95% favor production of O-DMA 3 . We have previously observed that men with prostate cancer consuming a soy isoflavone-enriched bread experienced a change in circulating immune regulatory cytokine profiles consistent with a reduction in pro-inflammatory processes and immunosuppressive cell populations 2,4 . These data provide evidence for the immunomodulatory impact of soy isoflavones in a clinically-relevant setting.
Specific soy isoflavones appear to exhibit differential effects on inflammatory processes. For example, IFN-γ induced pSTAT1 was reduced in Caco-2 cells (human epithelial colorectal adenocarcinoma) upon treatment with genistein 5 . Similarly, other studies have shown that LPS-induced STAT1 activation is abrogated when microglial cells were cultured with genistein or equol 6 . These data are consistent with reports indicating that soy isoflavones can downregulate inflammatory cytokine production (IL-6, IL-8, TNF-α,  in several different immune cell subtypes 7 . Others have shown that phytoestrogens (genistein and daidzein) are weakly estrogenic and can activate natural killer (NK) cell activity at concentrations of 0.1 to 10 µM in vitro 8,9 . Thus, it is clear that soy isoflavones and their metabolites have modulatory effects on inflammatory signaling processes. However, the differential effects of individual soy isoflavones and their host or microbial metabolites on NK cell activity, one component of a complex immune response, is still poorly understood.
Understanding the role of isoflavones and their metabolites in regulating NK cell activity is critical due to the importance for NK cells as mediators of cytokine secretion and inflammatory responses including infection, tissue damage and immunosurveillance against tumors 10 . These actions are accomplished not only by the well-characterized cytotoxic effects of NK cells, but also by their ability to produce IFN-γ in response to inflammatory cytokine stimuli such as IL-12 and/or IL-18 11,12 . Together these cytokines shape the immune response in a manner that orchestrates inflammatory changes that regulate control of cancers and pathogens.
We hypothesized that soy isoflavones and their metabolites represent dietary components that impact NK cell mediated immune function. In this report, we describe how two soy isoflavones (genistein and daidzein) and two microbial metabolites (O-DMA and equol) impact NK cell activity. We found that genistein and equol were potent inhibitors of IL-12/IL-18 induced NK cell IFN-γ production, but had no effect on NK cell cytotoxicity against target cells. Additionally, treatment of NK cells with genistein reduced IL-12/IL-18 induced phosphorylation of ERK and total phosphorylated-tyrosine. Genistein also reduced the IL-12/IL-18 induced surface expression of IL-18 receptor-alpha (IL-18 Rα) by CD56 + NK cells, but had no effect on IL-12 receptor-beta-1 (IL-12Rβ1) expression. Further, murine experiments showed that mice fed a soy-enriched diet produced lower plasma IFN-γ levels following in vivo challenge with IL-12 and IL-18. Together our data provide novel evidence for soy phytochemicals as modulators of cytokine communication involving NK cells in vitro and in vivo. Importantly, these data suggest that exposure to soy isoflavone metabolites may impact the ultimate biologic response to a soy based dietary intervention. Demonstrating that dietary components impact NK cell mediated immune responses has implications for future clinical application.

Soy isoflavones and their metabolites do not affect immune cell viability. Daidzein and genistein
represent two prominent isoflavones present in soy that are metabolized by the intestinal microbiota into unique secondary metabolites such as equol and O-DMA ( Fig. 1) [13][14][15] . In vitro studies first examined whether these particular compounds impact immune cell viability. At physiological and pharmacologic concentrations (25 µM Genistein and Equol inhibit IL-12/IL-18-induced IFN-γ production. We next examined whether soy isoflavones or metabolites could alter the in vitro response of immune cells to canonical inflammatory stimuli. For these studies, the response to IL-12/IL-18 was first examined, as these cytokines are critical orchestrators of immune function and represent a potent inflammatory stimulus. PBMCs were cultured with varying concentrations of soy isoflavones and metabolites for 4 hours and then stimulated for 72 hours with IL-12/IL-18. PBMCs remained incubated with soy isoflavones/metabolites for the duration of IL-12/IL-18 stimulation. A significant decrease in IFN-γ production in cell culture supernatants from PBMCs treated with genistein ( Fig. 3a; p = 0.0023) or equol ( Fig. 3d; p = 0.006) was observed. In contrast, culture of PBMCs with daidzein ( Fig. 3b) or O-DMA ( Fig. 3c; p = 0.1020) did not significantly reduce IFN-γ production in vitro.
Genistein and Equol decrease IL-12/IL-18-induced IFN-γ production by NK cells. NK cells are a predominant cell type activated in response to stimulation with IL-12 and IL-18, although differentiated T lymphocytes and NKT cells can also be impacted to a lesser degree 16,17 . To determine the impact of soy phytochemicals on IFN-γ production from each of these immune cell subsets, bulk PBMCs were cultured with soy isoflavones or their metabolites for 4 hours prior to a 24 hour stimulation with IL-12/IL-18. Cells were then stained for intracellular IFN-γ via flow cytometry. We found that CD3 − CD56 Bright NK cells were the primary producer of IL-12/ IL-18 induced IFN-γ (Fig. 4), which significantly decreased when incubated with genistein ( Fig. 4b; p = 0.0147) or equol ( Fig. 4c; p < 0.0001). We also observed similar decreases of intracellular IFN-γ by CD3 − CD56 dim NK cells when treated with genistein (p = 0.0153) and equol, (p = 0.0003) however overall percentages of positive cells were lower compared to CD3 − CD56 Bright NK cells (Fig. 4b). Similar to data obtained with cell culture supernatants, incubation of CD3 − CD56 dim NK cells with ODMA also reduced intracellular IFN-γ compared to DMSO control ( Fig. 4b; p = 0.0078). Percentages of IFN-γ + CD3 − CD56 Bright or CD3 − CD56 dim were not affected when www.nature.com/scientificreports www.nature.com/scientificreports/ incubated with daidzein (Fig. 4c). Intracellular staining also revealed a decrease in IFN-γ from CD3 + CD56 + NKT cells when cultured with genistein and equol (Supplemental Fig. 2). The overall frequency of IFN-γ producing NKT cells was much lower than NK cells, however, this trend in response to genistein and equol was consistent across multiple donors. Further, the overall IFN-γ production by CD3 + T cells and CD14 + monocytes was considerably lower than NK cells, however not appreciably impacted by the soy isoflavones or metabolites.

Soy isoflavones and NK cytotoxicity.
Another critical function of NK cells is their ability to elicit cytotoxic activity against target cells 10,18,19 . This cytotoxic function of NK cells can be enhanced in the presence of cytokines such as IL-12 20 . We tested whether soy isoflavones or their metabolites alter NK cytotoxicity. For these experiments, CD56 + NK cells negatively selected from healthy normal donors were pre-incubated for 4 hours with 25 µM of the soy compounds, and then stimulated with and without IL-12 overnight (Fig. 5a,b). Following stimulation, NK cells were co-cultured with Cr 51 labeled K562 target cells for 4 hours. As expected, IL-12 enhanced NK cytotoxicity against K562 target cells. However no consistent difference in cell killing was observed across all effector:target ratios when cells were pre-incubated with soy isoflavones or metabolites. For instance, in unstimulated NK cells, only three individual comparisons (equol vs. DMSO at a ratio of 5:1, genistein vs. DMSO at ratios of 20:1 and 2.5:1) were below the 0.0625 significance level, where only one would be expected by chance. For NK cells stimulated with IL-12, only one comparison (equol vs. DMSO at a ratio of 10:1) was below the threshold, the same as would be expected by chance.

Selective modulation of cytokine-induced signal transduction events by genistein. The impact
of genistein on canonical cytokine-induced phosphorylation of STAT1 (pSTAT1), ERK (pERK), and total tyrosine (pTyr) was next examined in cells pre-treated with genistein and exposed to IL-12/IL-18 stimulation. These signaling proteins are of interest given the abundance of IFN-γ produced in response to IL-12/IL-18, and the known role of ERK in signaling downstream of the IL-18R 21,22 . The phosphorylation of ERK ( Fig. 6a,b, Supplemental Fig. 3) and total tyrosine (Fig. 6e, Supplemental Fig. 3) was abrogated by a 4 hour genistein pre-treatment in both untreated PBMCs and those exposed to IL-12/IL-18 stimulation as compared to controls. Similar trends were observed for STAT1 phosphorylation although these data did not reach statistical significance (Fig. 6c,d).

Genistein decreases expression of IL-18Rα on NK cells.
Given the ability of genistein to limit cytokine-induced signaling via ERK, we were interested in examining the expression of IL-12 and IL-18 receptors on NK cells. Healthy donor PBMCs were pre-incubated with genistein, stimulated with IL-12/18, and analyzed for expression of IL-12 and IL-18 receptor expression on CD56 + NK cells (Fig. 7a,b). Genistein pre-treatment did not impact expression of IL-12Rβ1 on CD56 + NK cells (Fig. 7c). However, genistein reduced the expression of IL-18Rα on CD56 + NK cells stimulated with IL-12/18 compared to unstimulated cells ( Fig. 7d; p < 0.01).

Soy-enriched diet reduces IL-12/18 induced IFN-γ responses in vivo.
To better approximate the impact of physiologically relevant soy phytochemical concentrations consumed in the diet, we modeled the potent inflammatory stimulus of IL-12 and IL-18 in vivo. Previously published work from our group reported that C57BL/6 mice injected with the combination of IL-12 and IL-18 resulted in IFN-γ from primarily only the www.nature.com/scientificreports www.nature.com/scientificreports/ NK cell population 21 . For these studies, C57BL/6 mice were fed a control or soy-isoflavone enriched diet for one week and injected intraperitoneally with IL-12/IL-18 (Fig. 8a). Plasma levels of IFN-γ were significantly reduced in mice fed a soy-enriched diet compared to the control diet ( Fig. 8b; p < 0.0001).

Discussion
Individual soy isoflavones and their microbial metabolites may have distinct effects on immune function. The present study focuses on how these compounds impact the diverse biologic functions of NK cells. In particular, we demonstrate that the soy isoflavone, genistein and the daidzein metabolite, equol are particularly adept at abrogating IL-12/IL-18 induced production of IFN-γ by NK cells. A previous report by Nishio et al. indicate higher concentrations of genistein (10 µg/mL) can decrease cytotoxicity of IL-2 stimulated NK cells 23 . In contrast, these soy isoflavones or their metabolites did not alter cell viability or the cytotoxic function of NK cells either with or without IL-12 stimulation. Concentrations used in our study were lower (25 µM), and this prior study was limited to only genistein in the context of IL-2 stimulation, potentially accounting for the different observations related to cytotoxic activity. Importantly, the ability of soy components to reduce IL-12/IL-18 induced IFN-γ was recapitulated in mice receiving a soy-enriched diet, highlighting the physiological relevance of these observations. Together these results indicate the need for further mechanistic study on how individual soy components or metabolites derived from the host or microbiota can modulate the function of specific cellular components of the immune system. This data has important implications for understanding how NK cells may regulate chronic and acute inflammatory responses 10,18,19 . www.nature.com/scientificreports www.nature.com/scientificreports/ Our motivation to examine how individual soy components or metabolites impact immune function stems in part from results of a prior human clinical study 4 . We observed that men with prostate cancer received a dietary soy intervention that resulted in down-regulation of several pro-inflammatory cytokines in the plasma 4 . A particularly informative companion study using blood and urine from these same men was also conducted before and after intervention 2 . This pharmacokinetic study demonstrated that men clustered into four distinct soy metabolite signatures based on their ability to metabolize daidzein. We postulated that this observation may open further lines of investigation as to how the proportions of isoflavones or their metabolites can alter inflammation and the immune response.
There is currently a gap in our understanding as to how soy components impact specific facets of the immune response, and few studies have systematically explored the complexity of isoflavone metabolites as they relate to immune function. For example, in vitro studies investigating soy isoflavones and immune function have focused on the parent isoflavones, genistein and daidzein. Past studies have investigated the effects of genistein on IFN-γ responses, however these studies have utilized stimuli such as PMA/Ionomycicn, ConA, or PHA to induce a response [24][25][26] . These stimuli are of more limited physiologic relevance and these studies were focused primarily on genistein rather than other soy phytochemicals. Indeed, our prior clinical studies with dietary soy intervention 2,4 and others have more recently appreciated the complexity of soy isoflavone metabolism with the need to expand our investigation of isoflavone metabolite byproducts, O-DMA and equol. In clinical trials of patients with prostate cancer receiving a soy bread intervention, detectable genistein, daidzein, O-DMA and equol were noted in both plasma and urine from patients 2 . Although beyond the scope of the current study, it is important to note that understanding the host and microbial metabolism of soy isoflavones into multiple metabolites with local or systemic bioactivity is only beginning to emerge. In support of a role for microbial regulation of isoflavone metabolism are studies completed in rats lacking soy metabolizing gut microbiota. These animals were unable to metabolize isoflavones unless inoculated with bacterial strains known to metabolize isoflavones 27 . Taken together, these data suggest the microbiome is a relevant mediator of isoflavone metabolism. Thus, there is an emerging opportunity to delineate the multi-directional communication between diet, microbiota and immune response and its impact upon inflammation or disease 2,28,29 .
In theory, the ability of different populations and patients to metabolize soy isoflavones could also influence the potential outcomes and effect on immune interventions targeting inflammatory conditions. For example, a percentage of individuals effectively metabolize the soy isoflavone daidzein into equol more efficiently 3 , which www.nature.com/scientificreports www.nature.com/scientificreports/ could lead to a more drastic reduction in NK cell IFN-γ production compared to non-metabolizers. Thus, the individual ability to metabolize isoflavones could potentially lead to differential effects on immune responses. This emphasizes the importance and novelty of this report as we observed not only the effect of soy isoflavones on NK cell function but also their metabolite byproduct equol. The monitoring of isoflavone metabolism in clinical settings will be informative when administering dietary soy with the intent of health benefit.
Several questions remain when considering the activity of soy on immune response, and in particular, its actions on NK cell biology as demonstrated in this study. It will be important in future studies to determine whether these actions of soy on NK cells may be beneficial or detrimental to immune responses in the settings of  10 . Thus, depending on the context, soy effects on NK cell activity might have unpredictable effects on these disease processes or the response to vaccines or other immune modulatory therapies. Additionally, we reported that genistein can reduce the expression of IL-18Rα on CD56 + NK cells and phosphorylated tyrosine protein levels. The ability of soy isoflavones to reduce IL-18 receptor signaling could be a possible mechanism for dampened IFN-γ production by NK cells, although the robust changes in multiple cellular signaling pathways imply there are likely pleiotropic effects mediated by soy that culminate in global reductions in cytokine production. Together these results allow for a greater understanding of the complexity of isoflavone metabolites and their impact upon NK cell biology.

Methods
Cell culture and Reagents. All cells were cultured in RPMI-1640 (Gibco, Grand Island, NY) with 5% FBS, 1% L-Glutamine and antibiotic/antimycotic (Gibco). Soy isoflavones, genistein and daidzein, were purchased from LC Laboratories with greater than 99% purity (Woburn, MA). Equol was purchased from Toronto Research Chemicals (Toronto, ON). O-demtheylangolensin (O-DMA) was purchased from Planteck UK (Berkshire, UK). The control diet was the AIN-93 G (Table 1). A soy-enriched irradiated diet was prepared by Envigo (Madison, WI) consisting of the AIN-93G diet with 7.246 g of soy extract per 1 kg of diet substituting 7% corn oil, for soybean oil, in the AIN-93G diet formulation. The AIN-93G diet with 7% corn oil (in lieu of soybean oil) was used as a base for both the soy-enriched and vehicle control diets. The soy extract (Solgen 40 from Solbar, Ashdod, Israel) is a 40.25% soy isoflavone (w/w) soy bean extract, and when incorporated results in approximately 2.92 g/ isoflavones per 1 kg of diet. Recombinant human and murine IL-12 was purchased from Peprotech, Inc. (Rocky Hill, NJ). Recombinant human and murine IL-18 was purchased from R&D Systems Inc.

Isolation and differentiation of human blood cells. Peripheral blood mononuclear cells (PBMCs) were
isolated from source leukocytes of healthy adult donors (American Red Cross, Columbus, OH) with Ficoll-Paque (Amersham, Uppsala, Sweden) as previously described 30,31 . For NK cells, CD56 + cells were isolated from human donor blood by negative selection RosetteSep antibody cocktail (STEMCELL Technologies, Vancouver, BC).

Analysis of IFN-γ by ELISA.
PBMCs from multiple healthy human donors were pre-incubated with DMSO vehicle control or isoflavones/metabolites for 4 hours, and were subsequently stimulated with 20 ng/ml of recombinant human IL-12 and 50 ng/ml of IL-18 for 72 hours. Cell culture supernatants were analyzed for the presence of IFN-γ. PBMCs remained incubated with soy isoflavones/metabolites for the duration of IL-12/IL-18 NK cytotoxicity assay. CD56 + NK cells isolated from normal human donors by negative selection (RosetteSep antibody cocktail (STEMCELL Technologies) were plated in 96-well V-bottomed plates treated with or without IL-12 (10 ng/ml) overnight in RPMI-1640 supplemented with 10% human AB serum media at 37 °C. Eighteen hours later, 51 Cr-labeled K562 tumor cells were incubated with NK cells at various effector:target (E:T) ratios. Following a 4-hour incubation, supernatants were harvested and chromium release was assayed and percent lysis was calculated as previously described 32 . Briefly, spontaneous release represents 51 Cr release from target cells in medium alone (minimum value), and maximum release is 51 Cr release from target cells in medium + 10% Figure 6. Genistein modulates cellular signaling events. PBMC from healthy donors were cultured with 25 µM of genistein or DMSO (vehicle control) for 4 hours then stimulated with 20 ng/ml IL-12 and 50 ng/ml IL-18 for either 24 hours of additional incubation (long-term), or 15 minutes of additional incubation (short-term) to analyze canonical signal transduction events. (a) Immunoblot analysis of pERK was conducted in short-term lysates to focus on signaling events proximal to initial IL-18 receptor engagement. (b) Densitometry analysis of western blot pERK levels across conditions normalized to total ERK and β-actin levels. (c) Immunoblot analysis of pSTAT1 was conducted in long-term lysates where IFN-γ, an upstream STAT1 activating stimulus was at high levels (d) Densitometry analysis of western blot phosphor-STAT1 levels across conditions normalized to total STAT1 and β-actin levels. (e) Immunoblot analysis of pTyr was conducted in long-term lysates where IFN-γ and pan-pTyr activation was at high levels. Data are presented from three separate donors with similar results. In two of three experiments, total STAT1 or ERK, along with β-actin are included as controls and were run using identical lysates on separate gels, given similar molecular weight of some proteins. In one of three experiments, each phospho-specific Ab blot was stripped and re-probed for detection of loading controls. (Veh = vehicle; 12/18 = IL = 12 + IL-18 stimulation) Immunoblots were cropped for presentation purposes and original blots are provided as Supplemental Fig. 4. www.nature.com/scientificreports www.nature.com/scientificreports/ SDS. The NK cells were plated at a concentration of 1 × 10 6 /ml to obtain an initial ratio of 20:1, followed by a serial dilution for the other E:T ratios. The calculation for the percent lysis is an average of samples plated in triplicate,  www.nature.com/scientificreports www.nature.com/scientificreports/ control diet or an AIN-93G-based 0.3% soy-enriched diet (Envigo) for 7 days. Mice were then injected with 0.1 µg of recombinant murine IL-12 (R&D Systems) and 1 µg IL-18 (R&D Systems) or PBS per mouse on days 7 and 8 33,34 . Mice were euthanized after 48 hours (Day 9) of exogenous cytokine injections to test for plasma IFN-γ concentrations.