Hippuric acid and 3‐(3‐hydroxyphenyl) propionic acid inhibit murine osteoclastogenesis through RANKL‐RANK independent pathway

Abstract Nutritional factors influence bone development. Previous studies demonstrated that bone mass significantly increased with suppressed bone resorption in early life of rats fed with AIN‐93G semi‐purified diets supplemented with 10% whole blueberry (BB) powder for 2 weeks. However, the effects of increased phenolic acids in animal serum due to this diet on bone and bone resorption were unclear. This in vitro and in ex vivo study examined the effects of phenolic hippuric acid (HA) and 3‐(3‐hydroxyphenyl) propionic acid (3‐3‐PPA) on osteoclastic cell differentiation and bone resorption. We cultured murine osteoclast (macrophage) cell line, RAW 264.7 cells, and hematopoietic osteoclast progenitor cells (isolated from 4‐week‐old C57BL6/J mice) with 50 ng/ml of receptor activator of nuclear factor κ‐Β ligand (RANKL). Morphologic studies showed decreased osteoclast number with treatment of 2.5% mouse serum from BB diet–fed animals compared with those treated with serum from standard casein diet–fed mice in both RAW 264.7 cell and primary cell cultures. HA and 3‐3‐PPA, but not 3–4‐PPA, had dose‐dependent suppressive effects on osteoclastogenesis and osteoclast resorptive activity in Corning osteo‐assay plates. Signaling pathway analysis showed that after pretreatment with HA or 3‐3‐PPA, RANKL‐stimulated increase of osteoclastogenic markers, such as nuclear factor of activated T‐cells, cytoplasmic 1 and matrix metallopeptidase 9 gene/protein expression were blunted. Inhibitory effects of HA and 3‐3‐PPA on osteoclastogenesis utilized RANKL/RANK independent mediators. The study revealed that HA and 3‐3‐PPA significantly inhibited osteoclastogenesis and bone osteoclastic resorptive activity.


| INTRODUCTION
Physiologic bone development, growth, and repair are coordinated by well-balanced bone formation and resorption, and this coordination is influenced by factors including nutrition, hormones, and weight bearing/physical activity (Office of the Surgeon General, US, 2004;Shams-White et al., 2018;Tan et al., 2014). For nutritional factors, consideration has been closely directed to micronutrients, such as calcium, vitamin D, and phosphate, and to macronutrients, such as proteins and fats. Moreover, functional dietary factors from dairy products, fruits and vegetables, and foods contributing to acidbase balance interact with local bone transcription factors and circulating endogenous hormones to facilitate bone development (Yan et al., 2016). It is of interest to identify specific bioactive compounds. Previous studies have shown that bone mass and size in early pubertal children are significantly dependent on intake of fruits and vegetables (Hardcastle, Aucott, Reid, & Macdonald, 2011;Lanham, 2006;Tylavsky et al., 2004). We have previously described the robust effect of a 10% blueberry (BB)-supplemented diet on the promotion of bone formation in young male and female rats (Chen et al., 2010). We hypothesized that the significant effects of BB diet on bone mass increase in rapidly growing rats may be associated with increased phenolic acid levels in blood to inhibit osteoclastic bone resorption and to stimulate osteoblast differentiation.
Phenolic acids (PAs), such as hippuric acid (HA) and 3-(3-hydroxyphenyl) propionic acid (3-3-PPA), are metabolites derived from BB pigment polyphenols that appear in the serum of BB-fed rats. These molecules have been recently characterized as bioactive in stimulating osteoblast activity and dose-dependently increasing bone mass in mice . HA and 3-3-PPA are produced by gut microflora through the breakdown of chlorogenic acid and are thereafter absorbed and oxidized in the liver before entering circulation (Marín, Miguélez, Villar & Lombó, 2015). Until recently, these small molecules were not known for functions on stimulating or inhibiting particular cell differentiation or activity; moreover, it has not been proposed that HA and 3-3-PPA suppress osteoclastic bone resorption.
Periods of rapid osteoblastic bone formation are essential early in life as well as in adulthood to maintain skeletal health whereas osteoclastic bone resorption is essential to shape and keep an appropriate amount of bone (Clarke, 2008). While osteoclastic bone resorption is an important physiological cellular function in skeletal development, suppression of osteoclastogenesis is effective as a therapeutic approach to bone-destructive diseases such as osteoporosis and rheumatoid arthritis. Osteoblasts and osteoclasts originate from different lineage of cells. Osteoblasts are derived from mesenchymal stem cells; osteoclasts are derived from hematopoietic lineage (Grigoriadis et al., 2010). Osteoclast differentiation and bone resorptive activities of mature osteoclasts are regulated by series of cellular molecules, such as receptor activator of nuclear factor κ-Β ligand (RANKL) and macrophage colony-stimulating factor (Khosla, 2001). RANKL, a secreted protein, is and thought to be produced chiefly by osteoblasts and also by other cell types such as mesenchymal stromal cells, osteocytes, preadipocytes, and chondrocytes (O'Brien, 2010;Wang et al., 2014). Secreted RANKL stimulates receptor activator of nuclear factor kappa-B (RANK) signaling (Khosla, 2001).
The osteoclastogenic signaling starts from binding of RANKL to its receptor RANK on the surface of osteoclast precursors; this binding triggers the recruitment of tumor necrosis factor (TNF) receptor-associated factors (Park, Lee, & Lee, 2017). The RANK-TNF receptor-associated factor complex activates downstream signal pathways, including the nuclear factor kappa-light-chain-enhancer of activated B cells (NFκB) and mitogen-activated protein kinase pathways, which lead to the induction and activation of transcription factors such as Fos proto-oncogene (cFos) and nuclear factor of activated T-cells, cytoplasmic 1 (NFATc1; Boyce, 2013), resulting in enhanced expression of osteoclast-specific genes (Kim et al., 2018;Moreaux et al., 2011). In the current report, we hypothesize that natural plant-derived PAs, HA, and 3-3-PPA, directly and transcriptionally inhibit cFos and NFATc1 to decrease osteoclastogenesis via a RANKL-RANK independent mechanism.
Bone marrow cells were flushed from femurs and then cultured in T175 flasks (Corning® Cell Culture Flasks; Sigma-Aldrich) for 2 days to let stromal cells attach. Two days later, nonadherent cells were collected and cultured in appropriate plates with appropriate cell densities. These cells are considered as hematopoietic osteoclast progenitor cells (Chen et al., 2005). Conditional serum for treatment of cells was taken from female rats either fed 10% BB supplemental diet or standard rodent casein diet for 4 weeks as described in our previous study (Chen et al., 2010). Cell cultures were performed in α-Minimum Essential Medium (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (FBS; Hyclone, Logan, UT), penicillin (100 units/ml), streptomycin (100 µg/ml), and glutamine (4 mM). Cells were seeded in 96-, 12-, or 6-well cell culture plates at appropriate density of cells per well for morphology, RNA, and protein expression experiments. At 85% confluence, cells in 96-well plates were treated with 2.5% rat serum (7.5% FBS) in the presence of 50 ng/ml of soluble RANKL for osteoclastic cell morphologic study.

| Osteoclast resorption activity and proliferation assay
For osteoclast resorption activity assay, RAW264.7 cells or nonadherent bone marrow cells were seeded in six-well collagen-coated plates (BD Biosciences) at a density of 1 × 10 5 cells/well, and cells were treated with 50 ng/ml RANKL for 2-3 days. When osteoclasts begin to differentiate to mature cells on day 3, the cells were dissociated and the same number of osteoclastic cells were cultured onto hydroxyapatite-coated plates (CLS3989; Corning). The cells were treated with different concentrations of HA, 3-3-PPA, and 3-4-PPA, or 2.5% conditional serum for another 48 hr in the presence or absence of RANKL. Cells in the culture plates were fixed using 2.5% glutaraldehyde with or without Von Kossa staining.
The areas of hydroxyapatite resorption were observed by light microscopy and analyzed using the Image J software (imagej.nih.gov/ij/). Cell viability was measured by 3-(4,5-dimethylthiazol-2-yl)−2,5-diphenyltetrazolium bromide (MTT) assay. Briefly, fixed cells in the culture plates were washed with phosphate-buffered saline and incubated with MTT (0.5 mg/ml) at 37°C for 3 hr. After rinsing out MTT, 100 µl of dimethylformamide was used to dissolve the reduced formazan crystals, and microplate reader was used to determine the absorbance of each well at 540 nm.
2.4 | RNA isolation, real-time reverse transcription-polymerase chain reaction RAW264.7 cells were cultured in 12-well plates (1.2 × 10 5 cells/well) with or without RANKL (50 ng/ml) in the presence or absence of 1× HA for 1, 2, 3, or 4 days. RNA from cultured cells were extracted using TRI Reagent (MRC Inc., Cincinnati, OH) according to the manufacturerʼs recommendation followed by DNase digestion and column cleanup using QIAGEN mini columns (Chen et al., 2016). Reverse transcription was carried out using an iScript cDNA Synthesis Kit from Bio-Rad (Hercules, CA). Real-time reverse transcription-polymerase chain reaction (RT-PCR) was carried out using SYBR Green and an ABI 7000 Sequence Detection System (Applied Biosystems, Foster City, CA); gene expression data were normalized by housekeeping gene glyceraldehyde 3-phosphate dehydrogenase (Chen et al., 2016). All primers for RT-PCR analysis used were designed using Primer Express Software 2.0.0 (Applied Biosystems) and are listed in Table 1.

| Statistical analyses
All analyses used Stata 12.0 (Stata Corporation, College Station, TX) and Prism 5 (GraphPad, San Diego, CA). Data are expressed as the mean ± standard error; n equals to the number of samples/group.
Differences within groups were evaluated using t test or one-way analysis of variance followed by Tukeyʼs post hoc test with p < .05 considered significant. Representative images from three separate cell culture experiments are displayed. Significant dose-or time-dependent effects of tested compounds were assessed using Cruickʼs nonparametric test for trend (Cuzick, 1985).

| RESULTS
3.1 | HA and 3-3-PPA dose-dependently inhibit osteoclastogenesis in RAW264.7 cells and nonadherent mouse bone marrow cells HA and 3-3-PPA are PAs. BB diet is associated with high concentrations of these PAs in animal blood; their structures have been characterized previously (Chen et al., 2010). RAW 264.7 cells were treated with 2.5% serum either from standard casein diet-fed mice (Cas serum) or BB diet-fed mice (BB serum) in the presence of 50 ng/ml RANKL for 5 days.
Osteoclast morphology after TRAPase staining (Figure 1a Independent experiments were repeated more than three times, obtaining similar results each time, *p < .05, significant differences versus control by t test, dose response was assessed using Cruzickʼs nonparametric test for trend. FBS, fetal bovine serum; RANKL, receptor activator of nuclear factor κ-Β ligand; TRAPase, tartrate-resistant acid phosphatase regulate both osteoblast and osteoclast activity, however, β-catenin is one of these signaling molecules. We therefore next checked if HA and 3-3-PPA regulate β-catenin signaling in pre-osteoclasts. We

| DISCUSSION
Our novel data suggest direct effects of PAs, HA, and 3-3-PPA on inhibiting osteoclastogenesis through a RANKL-RANK independent mechanism. It is known that RANKL is a type II transmembrane protein produced mainly by osteoblasts and other cell types including mesenchymal stromal cells, osteocytes, preadipocytes, and chondrocytes (O'Brien, 2010;Wang et al., 2014). It is cleaved by proteases to yield soluble form RANKL. Secreted RANKL in turn stimulates preosteoclastic RANK signaling (Feng & Teitelbaum, 2013) and is able to directly induce the differentiation of precursor cells, such as bone marrow-derived macrophages, into mature and active osteoclasts (Feng & Teitelbaum, 2013). RANKL has its specific decoy receptor, molecules that can regulate osteoclastogenesis or inhibit osteoclast activity has been an important clinical goal. It is also important to identify food-based approaches to improve bone health and optimize bone growth. We have previously reported that BB diet suppressed bone resorption by downregulating RANKL expression in stromal cells and/or peroxisome proliferator-activated receptor γ in preadipocytes. We suggested that bone marrow stromal cells are targets for BB diet to regulate RANKL expression and osteoclast differentiation in young rapidly growing rats. In addition, we have also recently reported on the effect of HA on osteoblastic bone formation, and together to our current findings, it appears that HA and 3-3-PPA have effects on anti-osteoclastic cell resorption. Clearly, more in vivo studies are needed, in additional models such as ovariectomized mouse models or collagen-induced rheumatoid arthritis animal models to comprehensively test the hypothesis that HA and 3-3-PPA have value to inhibit bone resorption.
We have previously described that BB diet contains abundant bioactive compounds derived from polyphenols, which mainly include proanthocyanidins, anthocyanidins, flavones, phenolic acid, and stibenes.
PAs are efficiently absorbed and oxidized in the liver and are one of the final metabolites derived from the polyphenols in vivo. HA and 3-3-PPA were identified two phenolic acid metabolites derived from BB pigment polyphenols that appear in the serum of BB diet-fed rats with the highest concentrations (Chen et al., 2010). Using HPLC/MS, we found that their concentrations were increased with increased percentage of BB consumption in the diet (Zhang et al., 2013). HA and 3-3-PPA inhibited RANKL-induced osteoclast differentiation in a significant dose-dependent manner. The inhibitory effects of HA and 3-3-PPA were not due to any obvious toxic effects on RAW264.7 cells. Although their synergistic effect will be tested in future studies, HA and 3-3-PPA each significantly suppressed the expression of osteoclastogenesis-related marker genes and proteins. HA and 3-3-PPA significantly inhibited NFATc1 expression, with a subsequent reduction in expression of downstream osteoclastogenic marker genes. As a required protein for RANKL-induced osteoclast formation, a highly phosphorylated form of NFATc1 is in the cytoplasm (Sheridan, Heist, Beals, Crabtree, & Gardner, 2002), and its expression is regulated by cFos (Mohamed et al., 2007). Following interactions between RANKL and RANK, NFATc1will be activated and enters into the nucleus to stimulate osteoclast-specific gene expression, including integrin β3, Acp5 (TRAcP), and Ctsk (Crotti et al., 2008, Matsumoto et al., 2004. It has been reported that severe bone sclerosis observed in cFos mutant mice is due to the blockage of osteoclast formation (Matsuo et al., 2000). In addition, it is well recognized that RANKL-RANK interaction results in TNF receptor-associated factor 6 (TRAF6) recruitment and nuclear factor kappa-B, c-Jun amino-terminal kinase, extracellular signalregulated kinase, and p38 signaling pathways activation (Jules, Ashley, & Feng, 2010). However, further investigations are required to determine if HA and 3-3-PPA inhibit osteoclast differentiation through these subsignaling pathways.
To provide evidence of HA and 3-3-PPA inhibiting osteoclastogenesis through mediators different than RANK in cell membrane of pre-osteoclasts, we examined GPR109A expression during osteoclast differentiation. We previously determined that HA and 3-3-PPA bind to GPR109A, a G-protein coupled receptor (Ren et al., 2009;Tunaru et al., 2003). Herein, we were surprised to find that both HA and 3-3-PPA significantly inhibited GPR109A gene expression in pre-osteoclasts in the absence or presence of RANKL. Evidence suggested that activation of GPR109A results in reduced cAMP levels which may affect activity of cAMP-dependent protein kinase A and phosphorylation of target proteins, leading to neutrophil apoptosis (Insel, Zhang, Murray, Yokouchi, & Zambon, 2012). We found that GPR109A gene expression is inhibited by HA and 3-3-PPA resulting in increased levels of intracellular cAMP. It is unknown if the decreased expression of osteoclastic gene markers is the result of a cell type-specific phenomenon or increased levels of intracellular cAMP in pre-osteoclasts by HA and 3-3-PPA.
Intracellular cAMP is known as an important second messenger that mediates a diverse set of extracellular signals, and GPRs evoke cAMP-mediated signaling via Gs (a guanine nucleotidebinding protein related to adenylyl cyclase activation). It was reported that stimulation of cAMP/PKA signaling suppressed osteoclast differentiation (Weivoda et al., 2016).
In summary, we demonstrated for the first time that HA and 3-3-PPA, PAs found in the highest concentrations in blood from mice fed a BB diet, dose-dependently inhibited osteoclastogenesis through a RANKL/RANK independent mechanism to protect against increased osteoclastic bone resorption. Our data suggested that the effect of HA and 3-3-PPA on osteoclastogenesis is through modulating cell membrane GPR109A to control cAMP levels in cytoplasm and hence osteoclastogenic gene expression. Our current research provides first-ever insights into mechanisms driving the effects of food-derived PAs HA and 3-3-PPA to regulate osteoclastogenesis and bone resorption. F I G U R E 6 HA and 3-3-PPA inhibit GPR109A expression in RAW264.7 cells. Real-time PCR shows both HA (a) and 3-3-PPA (b) dosedependently suppressed GPR109A gene expression in RAW264.7 cells. *p < .05 versus control by t test, dose response were assessed using Cruzick's nonparametric test for trend. (c) Western blot analysis shows RANKL activated GPR109A protein expression after RAW264.7 cells were treated for three days. Both 1× HA and 3-3-PPA inhibited RANKL-induced GPR109A protein expression. β-actin is a protein loading control. Real-time PCR shows both HA (d) and 3-3-PPA (e) do not change RANK gene expression in RAW264.7 cells. HA, hippuric acid; RANK, receptor activator of nuclear factor kappa-B; RANKL, receptor activator of nuclear factor κ-Β ligand