Isovaleric acid ameliorates ovariectomy‐induced osteoporosis by inhibiting osteoclast differentiation

Abstract Osteoclasts (OCs) play important roles in bone remodelling and contribute to bone loss by increasing bone resorption activity. Excessively activated OCs cause diverse bone disorders including osteoporosis. Isovaleric acid (IVA), also known as 3‐methylbutanoic acid is a 5‐carbon branched‐chain fatty acid (BCFA), which can be generated by bacterial fermentation of a leucine‐rich diet. Here, we find that IVA suppresses differentiation of bone marrow‐derived macrophages into OCs by RANKL. IVA inhibited the expression of OC‐related genes. IVA‐induced inhibitory effects on OC generation were attenuated by pertussis toxin but not by H89, suggesting a Gi‐coupled receptor‐dependent but protein kinase A‐independent response. Moreover, IVA stimulates AMPK phosphorylation, and treatment with an AMPK inhibitor blocks IVA‐induced inhibition of OC generation. In an ovariectomized mouse model, addition of IVA to the drinking water resulted in significant decrease of body weight gain and inhibited the expression of not only OC‐related genes but also fusogenic genes in the bone tissue. IVA exposure also blocked bone destruction and OC generation in the bone tissue of ovariectomized mice. Collectively, the results demonstrate that IVA is a novel bioactive BCFA that inhibits OC differentiation, suggesting that IVA can be considered a useful material to control osteoclast‐associated bone disorders, including osteoporosis.

IVA inhibited the expression of OC-related genes. IVA-induced inhibitory effects on OC generation were attenuated by pertussis toxin but not by H89, suggesting a G icoupled receptor-dependent but protein kinase A-independent response. Moreover, IVA stimulates AMPK phosphorylation, and treatment with an AMPK inhibitor blocks IVA-induced inhibition of OC generation. In an ovariectomized mouse model, addition of IVA to the drinking water resulted in significant decrease of body weight gain and inhibited the expression of not only OC-related genes but also fusogenic genes in the bone tissue. IVA exposure also blocked bone destruction and OC generation in the bone tissue of ovariectomized mice. Collectively, the results demonstrate that IVA is a novel bioactive BCFA that inhibits OC differentiation, suggesting that IVA can be considered a useful material to control osteoclast-associated bone disorders, including osteoporosis.

K E Y W O R D S
branched-chain fatty acid, isovaleric acid, macrophage, osteoclast, osteoporosis neutrophil recruitment by regulating the expression of cytokineinduced neutrophil chemoattractant-2αβ and the adhesion molecule L-selectin. 8 SCFAs also regulate cellular differentiation. 9,10 Especially, butyric acid induces apoptosis to inhibit macrophage differentiation. 11 Dendritic cells treated with butyric acid promote regulatory T-cell differentiation from naïve T cells. 12 Regarding the target receptors for the SCFAs, several G-protein coupled receptors (GPCRs) including Olfr78 have been reported to mediate biological responses caused by the SCFAs. 7,13,14 Branched-chain fatty acids (BCFAs) are made from essential amino acids including leucine by microbial fermentation, to form isobutyric acid with four carbons, or 2-methylbutyric acid and isovaleric acid (IVA) with five carbons. 15,16 The levels of the leucine metabolite IVA are increased by the consumption of high protein foods and certain pathological conditions including colon cancer. 16,17 Previous reports showed that BCFAs affect energy metabolism and cause colon smooth muscle relaxation. 18,19 Another report demonstrated that enterochromaffin cells are activated by a microbial metabolite, IVA, via Olfr558. 20 However, much remains unknown regarding whether BCFAs regulate additional cellular activities and differentiation.
In this study, we examined whether IVA affects the activity of macrophages and their differentiation into OCs. The molecular mechanism involved in the modulation of OC differentiation by IVA was investigated as well. We also examined the in vivo effects of IVA on the regulation of osteoporosis, an OC-associated disorder.

| Generation of mouse bone marrow-derived macrophages (BMDMs)
The protocols for mouse experiments received the approval of the Institutional Review Committee for Animal Care and Use at Sungkyunkwan University (Suwon, Korea). Mouse BMDMs were generated as described before. 21 Briefly, total bone marrow cells were obtained from the femur and tibia of 5-week-old C57BL/6 mice (Orient Bio). After culturing the bone marrow cells in α-MEM (Gibco) with 10% FBS (Access Biological) for 1 day, non-adherent cells were further cultured for 3 days in α-MEM with 10% FBS and 30 ng/mL of M-CSF (Peprotech).

| OC differentiation and TRAP staining
Isolated mouse BMDMs (1 × 10 4 cells/well) were cultured with 30 ng/mL of M-CSF and 100 ng/mL of RANKL (Peprotech) in 96well plates for 3 days. 22 Differentiated immature OCs were further differentiated into mature OCs by adding α-MEM with 10% FBS, 30 ng/mL of M-CSF and 100 ng/mL of RANKL for 2 days. Before TRAP staining, mature OCs were fixed with paraformaldehyde (4%) for 20 minutes, as described previously. 23 Stained multinuclear cells (≥3nuclei) were considered to be OCs and were counted.

| Cell viability assay
To determine cell viability, BMDMs were treated by IVA (Sigma-Aldrich), and culture media were applied to the LDH assay kit (Promega) at 490 nm.

| Chemotaxis assay
Chemotactic migration of BMDMs was conducted using multi-well chambers (Neuroprobe Inc) as reported previously. 24 Briefly, 30 μL of chemoattractants such as IVA and WKYMVm (Anygen) were loaded on the lower well and then 25 μL of suspended BMDMs (1 × 10 6 cells/mL) in α-MEM were applied on the upper well, separated by a polyhydrocarbon filter. After 2 hours, cells that had migrated through the filter were stained with haematoxylin. Migrated cells in each well were counted under a microscope (Leica DM750).

| Western blot analysis
BMDMs were stimulated with 30 ng/mL of M-CSF and 100 ng/mL of RANKL in the absence or presence of 200 μM of IVA for the indicated lengths of time. Extracted proteins were separated by 10% SDS-PAGE and transferred to a 0.45 μm nitrocellulose membrane (Cytiva). After incubating the membranes with antibodies for target proteins, visualization with enhanced chemiluminescence followed.
Antibodies used for Western blot analyses were obtained from Cell Signalling Technology or Santa Cruz Biotechnology.

| Bone resorption assay
The pit formation assay was performed on Corning 96-well plates (Corning). BMDMs (1 × 10 4 cells/well) were cultured with 30 ng/mL of M-CSF and 100 ng/mL of RANKL in 96-well plates for 7 days. The media containing M-CSF and RANKL was changed every 2 or 3 days after seeding the cells. After 7 days of induction, the media was removed and cells were incubated in 100 μL of 10% bleach solution for 5 minutes at room temperature. Wells were washed with dH 2 O 2 times and dried at room temperature for 4 hours. The pit clusters generated by OCs were photographed by a microscope.

| RNA isolation
Total RNA from mouse BMDMs and immature OCs were isolated using TRIzol reagent (Invitrogen). cDNAs were synthesized using the RT Premix Kit (iNtRON). Femurs and tibias of ovariectomy (OVX) mouse models were rapidly frozen in liquid nitrogen and ground in a homogenizer to extract RNA using TRIzol reagent.

| Quantitative polymerase chain reaction (qPCR) analysis
qPCR analyses were conducted using the Rotor-gene Q (2plex on PC) instrument (QIAGEN) with SYBR Green qPCR Mix (Biofact). The primers used for qPCR analyses were presented in Table 1. Data were normalized against the expression of GAPDH as an internal control.

| OVX mouse model
OVX was performed with 8-week-old C57BL/6 female mice according to a previous report. 25 Mice were anaesthetized by inhalation anaesthesia, and ovaries were bilaterally removed via a dorsal approach as described previously. 26 To examine the effects of IVA on OVX mice, we added IVA to the drinking water (at a final concentration of 75 mM or 150 mM) from the day after surgery. The mice were sacrificed after 5 weeks, and the femurs were separated and used in subsequent experiments.

| Bone histology in OVX mice
All OVX model mice were sacrificed 5 weeks after surgery, and their bones were fixed in paraformaldehyde solution (4%) for 2 days.
Next, the bones were decalcified in 10% EDTA solution for 3 weeks and embedded with paraffin. Bones were sectioned into 4 μm by a microtome. After staining the samples with haematoxylin and eosin (H&E) solution, morphological analysis such as bone density was conducted. TRAP staining was conducted to determine OC formation.

| Statistical analysis
GraphPad Prism software was used to evaluate results. Statistical analysis was performed using Student's t-test or ANOVA (one-way or two-way). All results are expressed as the mean ± SEM. A Pvalue < .05 was considered statistically significant.

| IVA stimulates BMDM migration in a pertussis toxin (PTX)-sensitive manner
Various extracellular stimuli elicit the activation of crucial kinases including Akt and MAPKs during cellular activation. 27,28 In this study, we investigated whether IVA stimulates BMDMs by measuring the phosphorylation of Akt and ERK. Stimulation of BMDMs with IVA caused apparent phosphorylation of Akt and ERK at 30 minutes ( Figure 1A). These results suggested that IVA may activate the BMDMs. Chemotactic migration is one of the important functional activities of BMDM, which is needed for subsequent cellular responses. 29,30 Next, we investigated the effects of IVA on macrophage chemotaxis at several different concentrations. Addition of IVA significantly elicited BMDM chemotactic migration, in an IVA concentration-dependent manner ( Figure 1B), maximal migrationinducing activity was observed at 500 μM IVA, which is comparable

TA B L E 1 Primers used for qPCR analysis
to that of a well-known macrophage chemoattractant, WKYMVm ( Figure 1B). Previously, macrophage chemotactic migration was reported to be regulated by PTX-sensitive GPCRs. 31 We also examined the possible role of PTX-sensitive GPCRs in IVA-induced BMDM chemotactic migration. Addition of PTX to BMDMs prior to the chemotaxis assay with IVA resulted in significant inhibition of BMDM chemotactic migration ( Figure 1C). These results suggest that IVA stimulates BMDMs leading to chemotactic migration in a PTX-sensitive manner.

| IVA inhibits RANKL-induced OC differentiation
Since IVA stimulates the activation of macrophages, which can differentiate into OCs, the effects of IVA on OC generation induced by RANKL was examined. Addition of M-CSF and RANKL to BMDMs generated multinucleated OCs, which can be observed by TRAP staining (Figure 2A)  Figure 2D). In order to check whether IVA inhibits OC generation by inducing cell death, we conducted a cell viability analysis and found that IVA did not affect cell viability during OC generation at 100 ~ 500 μM ( Figure 2E). Collectively, IVA strongly inhibits RANKL-induced OC differentiation without affecting cell viability.
Mature OCs show bone resorbing activity. 32,33 Since we found that IVA suppressed OC formation, we next tested the effects of IVA on OC function by using dentine slices. We found that OC generated by M-CSF + RANKL showed strong bone resorbing activity, while addition of IVA in the presence of M-CSF + RANKL blocked this activity ( Figure 2F). Taking our results together, IVA appears to block OC differentiation, leading to inhibition of bone resorbing activity.

| IVA inhibits OC-related gene expression
Our finding that IVA suppresses the differentiation of BMDMs into these results indicate that IVA inhibits OC generation by suppressing the expression of not only OC-related genes but also fusogenic genes.

| IVA-induced inhibitory effects on OC generation are partly mediated by PTX-sensitive signalling and AMPK activity
Since IVA has been reported to be recognized by Olfr558, 20  The results suggest that IVA may inhibit OC generation in a PTXsensitive manner.
Diverse metabolites absorbed in the gut are associated with host energy metabolism. 18,39 AMPK plays an essential role in energy homeostasis. 40,41 AMPK also has crucial roles in regulating osteoclastogenesis. Previous reports demonstrate that activated AMPK suppresses osteoclastogenesis. [42][43][44] Here, we found that IVA elicits AMPK phosphorylation ( Figure 4D). The role of AMPK in the IVA-elicited inhibitory effects on OC generation was also investigated using an AMPK inhibitor, compound C. Addition of compound C markedly reversed the inhibitory effects of IVA on OC generation ( Figure 4E,F).

| IVA attenuates OVX-induced osteoporosis
Since IVA inhibits OC generation induced by RANKL in vitro (Figure 2), the effects of IVA in an experimental osteoporosis model using OVX mice were examined. OVX surgery mimics postmenopausal osteoporosis showing body weight gain as well as bone density decrease and increased OC generation in bone tissue. 45,46 We also found that OVX mice gained body weight compared to control mice. Addition of IVA at the final concentrations of 75 mM and 150 mM to drinking water significantly decreased the body weight gain in OVX mice ( Figure 5A).
At day 32, OVX mice gain about 25% body weight; however, sham mice gain about 10% body weight. OVX mice exposed to 75 mM and 150 mM IVA gain about 15% and 10.25% body weight, respectively ( Figure 5B). Body weight increase in OVX mice is accompanied with decreased PGC-1α expression, which is reversed by IVA ( Figure 5C). In addition, OVX surgery significantly augmented the expression of OC-related genes including NFATc1, Ctsk and TRAP, which was almost completely reversed by addition of IVA to the drinking water ( Figure 5D).
Fusogenic genes such as Atp6v0d2, DC-STAMP and OC-STAMP were also strongly increased by OVX surgery, and IVA exposure significantly decreased the levels of these genes ( Figure 5D). However, the expression of MafB, a negative regulator of OC generation, was decreased by OVX and recovered by IVA exposure in OVX mice ( Figure 5D). We also performed qPCR analysis of bone turnover markers such as ALP, Osx and Runx2. We found that addition of IVA in the OVX mice model upregulates the expression of several bone formation-associated genes including Osx and Runx2 ( Figure 5D). Histological analysis with H&E staining of the bone tissue of OVX mice showed that OVX surgery markedly increased the porous area, which was strongly decreased by IVA exposure (Figure 5E 46 Here, we found that a BCFA, IVA, suppresses RANKL-induced OC generation without affecting cellular viability (Figure 2). The IVA-induced inhibitory effects on OC generation were also strongly supported by the sup-  Figures 1C and 4B,C). The results suggest that IVA may act on PTX-sensitive G i -protein-coupled receptors in BMDMs ( Figure 6). In a separate experiment, we found that BMDMs express Olfr558 (data not shown), which has been reported to mediate IVA-induced cellular responses in enterochromaffin cells. 20 Although the functional role of Olfr558 in IVA-induced suppression of OC formation is yet to be clarified, IVA may act on G icoupled GPCRs in BMDMs to block osteoclastogenesis ( Figure 6).

| D ISCUSS I ON
In a previous report, serum amyloid A inhibits osteoclastogenesis by shedding c-fms. 23 Unlike serum amyloid A, IVA did not induce shedding of c-fms (data not shown), suggesting a different mode In this study, we found that IVA not only stimulates chemotactic migration of BMDMs but also suppresses OC formation ( Figures   1 and 2). Regarding the functional relationship between the migration of OCs (or osteoclast precursors) and regulation of osteoclast differentiation, a previous report demonstrated that two chemokines, CCL19 and CCL21, stimulate osteoclast migration and bone resorption, mediating bone destruction. 54 In this study, however, we found that IVA stimulates migration of osteoclast precursor cells, but suppresses osteoclast differentiation. At this point, it is not easy to connect the functional relationship of these two different findings.
However, according to previous reports that demonstrate the functional relationship of the gut-bone axis, 6,55 it might be possible that IVA produced by the gut as a commensal bacterial metabolite could regulate bone metabolism leading to inhibition of osteoclast differentiation. Further studies are necessary to clarify the relationship between macrophage migration and regulation of OC differentiation.
Previously, IVA was shown to be produced from a leucine-rich diet based on animal foods such as egg and cheese through fermentation by the human gut microbiome. 17 Here, we observed that  Figure 5C). Since a recent study showed that PGC-1α regulates osteoclastogenesis through metabolic regulation, 56 it would be necessary to examine the effects of IVA on the regulation of metabolic activity in OC precursors. Our results suggest that IVA, a BCFA, can be used to control osteoporosis. In a previous paper, beta-hydroxy-beta-methylbutyrate, another leucine metabolite, was reported to be associated with bone growth in newborn pigs. 57 Thus, we suggest that metabolites derived from leucine-rich diets may have potential beneficial effects against bone metabolic disorders including osteoporosis. In addition, isovaleric acidemia is an inborn error of leucine metabolism caused by a congenital deficiency of isovaleryl CoA dehydrogenase with elevated urinary IVA metabolites. Although the clinical syndrome of isovaleric acidemia is dominated by the neurologic findings, haematologic abnormalities are frequent. 58,59 Also, the unpleasant odour of IVA in urine is typical. Our study may provide some molecular mechanisms to explain the clinical phenotype of isovaleric acidemia.
In conclusion, we found that IVA, a BCFA metabolite of a leucinerich diet, can regulate macrophage activity and suppress osteoclastogenesis through G i -coupled GPCR and AMPK-dependent, but PKA-independent pathways. The therapeutic effects of IVA against an experimental osteoporosis OVX model suggest that IVA and a leucine-rich diet, which can be used to produce IVA, can be considered as a potentially new approach against OC-related bone diseases.

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.