J Mol Med DOI 10.1007/s00109-017-1589-2

ORIGINAL ARTICLE

A 15-amino acid C-terminal peptide of beta-defensin-3 inhibits bone resorption by inhibiting the osteoclast differentiation and disrupting podosome belt formation Ok-Jin Park 1 & Jiseon Kim 1 & Ki Bum Ahn 1,2 & Jue Yeon Lee 3 & Yoon-Jeong Park 4 & Kee-Yeon Kum 5 & Cheol-Heui Yun 6 & Seung Hyun Han 1

Received: 10 February 2017 / Revised: 22 August 2017 / Accepted: 29 August 2017 # Springer-Verlag GmbH Germany 2017

Abstract Human beta-defensin-3 (HBD3), which is secreted from cells in the skin, salivary gland, and bone marrow, exhibits antimicrobial and immunomodulatory activities. Its C-terminal end contains a 15-amino acid polypeptide (HBD3-C15) that is known to effectively elicit antimicrobial activity. Recently, certain antimicrobial peptides are known to inhibit osteoclast differentiation and, thus, we investigated whether HBD3-C15 hinders osteoclast

Ok-Jin Park and Jiseon Kim contributed equally to this work. Electronic supplementary material The online version of this article (https://doi.org/10.1007/s00109-017-1589-2) contains supplementary material, which is available to authorized users. * Seung Hyun Han [email protected] 1

Department of Oral Microbiology and Immunology, DRI, and BK21 Plus Program, School of Dentistry, Seoul National University, Building 86, 1 Gwanak-ro, Gwanak-gu, Seoul 08826, Republic of Korea

2

Research Division for Biotechnology, Korea Atomic Energy Research Institute, Jeongeup 56212, Republic of Korea

3

Central Research Institute, Nano Intelligent Biomedical Engineering Corporation (NIBEC), Seoul 03080, Republic of Korea

4

Department of Dental Regenerative Biotechnology, School of Dentistry, Seoul National University, Seoul 03080, Republic of Korea

5

Department of Conservative Dentistry, Seoul National University Dental Hospital, School of Dentistry, Seoul National University, Seoul 03080, Republic of Korea

6

Department of Agricultural Biotechnology and Research Institute for Agriculture and Life Sciences, Seoul National University, Seoul 08826, Republic of Korea

differentiation and bone destruction to assess its potential use as an anti-bone resorption agent. HBD3-C15 inhibited the receptor activator of nuclear factor κB ligand (RANKL)-induced osteoclast differentiation and formation of resorption pits. In addition, HBD3-C15 disrupted the formation of RANKL-induced podosome belt which is a feature typically found in mature osteoclasts with bone-resorbing capacity. HBD3-C15 downregulated cortactin, cofilin, and vinculin, which are involved in the podosome belt formation. Furthermore, bone loss induced by RANKL was significantly reduced in a mouse calvarial implantation model that was treated with HBD3-C15. Similar inhibitory effects were observed on the osteoclast differentiation and podosome belt formation induced by Aggregatibacter actinomycetemcomitans lipopolysaccharide (AaLPS). Concordantly, HBD3-C15 attenuated the resorption in the calvarial bone of AaLPS-implanted mouse. Collectively, these results suggest that HBD3-C15 has an anti-bone resorption effect in developing osteoclasts and that this occurs via its disruption of podosome belt formation. HBD3-C15 could be a potential therapeutic agent for the inhibition of bone destruction. Key messages & HBD3-C15 inhibits osteoclast differentiation and bone resorption capacity. & HBD3-C15 disrupts the podosome belt formation in osteoclasts. & HBD3-C15 alleviates the bone loss by RANKL or A. actinomycetemcomitans LPS in vivo.

Keywords Human beta-defensin-3 . Osteoclast . Podosome belt formation . Bone resorption

J Mol Med

Introduction Bone metabolism is a tightly regulated process, normally characterized by a balance between the development of boneresorbing osteoclasts and bone-forming osteoblasts [1]. Many bone diseases including osteoporosis and periodontal disease are caused by an imbalance between these two processes [2]. Osteoclasts originate from monocyte/macrophage lineage cells after having their differentiation stimulated by exposure to macrophage colony-stimulating factor (M-CSF) and receptor activator of nuclear factor κB ligand (RANKL) [3]. Of particular importance is the activation of RANKL/ RANK signaling, which is essential for normal osteoclast differentiation and function [4]. Indeed, RANKL induces the activation of c-Fos and NFATc1, both of which are important transcription factors that regulate osteoclast differentiation [4]. Consequently, RANKL upregulates osteoclast-specific genes, including tartrate-resistant acid phosphatase (TRAP), cathepsin K, and dendritic cell-specific transmembrane protein (DCSTAMP) [5, 6], and acts as a critical positive regulator of osteoclast development. Specific changes in the actin cytoskeletal organization of osteoclasts are critical for their adhesion to the bone matrix, which allows for subsequent bone resorption [7]. In particular, the formation of F-actin structures that contain the podosome is an essential step in the pathway eventually leading to bone resorption [8]. During osteoclast differentiation, podosomes exhibit three distinct structural patterns, commonly described as clusters, rings, and belts. Podosome clusters and rings are found in immature osteoclasts, while a large peripheral podosome belt is a typical feature of bone resorption which is seen in mature osteoclasts [9]. Additionally, the induction and activation of actin-binding proteins like vinculin, cofilin, and cortactin are closely related to the maturation of podosome [10]. Defensins are a family of mammalian antimicrobial peptides that normally contribute to innate host defense by causing antimicrobial activities that inhibit the growth, proliferation, and survival of bacteria, fungi, and viruses [11]. They are comprised of short cationic peptides that contain six disulfide-linked cysteine residues, and they are secreted by macrophages, neutrophils, and epithelial cells [12]. Defensins are classified into three subfamilies: α-, β-, and θ-defensin. Among them, α- and β-defensins are found in humans [11]. Human α-defensins are constitutively secreted, whereas β-defensins must be induced by various stimuli, including bacterial lipopeptide, lipoteichoic acid, lipopolysaccharide (LPS), and bacterial DNA [13]. Human βdefensin-3 (HBD3) has higher bactericidal activity than HBD1 or HBD2 [14]. It has been suggested that antimicrobial peptides including HBD3 have a weak stability [15], which could

be improved via a structure-based peptide design [16]. Additionally, a functional peptide analog with a short sequence would be economical compared to the fulllength protein [17]. A recent report states that a polypeptide consisting of 15 amino acids from the C-terminus of HBD3 (HBD3-C15) effectively inhibits bacterial and fungal biofilm formation [18, 19]. Additional studies have shown that many antimicrobial peptides, such as LL-37 and cathelin-related antimicrobial peptide (CRAMP), have an inhibitory effect on osteoclast differentiation [20, 21]. As such, we argue that this antimicrobial peptide holds a potential as a therapeutic against several bone diseases. To that end, in this study, we investigated the effect of HBD3-C15 on osteoclast differentiation and characterized its action mechanism.

Materials and methods Reagents and chemicals Recombinant mouse M-CSF and RANKL were purchased from PeproTech (Rocky Hill, NJ, USA). Recombinant HBD3 was purchased from R&D Systems (Minneapolis, MN, USA). The synthetic peptide HBD3-C15 (GKCSTRGRKCCRRKK) and control peptide (GACSTAGAACCAAAA) were manufactured via F-moc-based chemical solid-phase synthesis (NIBEC, Seoul, South Korea). Ascorbic acid, β-glycerophosphate, and 1α,25-dihydroxyvitamin D3 were obtained from Sigma-Aldrich (St. Louis, MO, USA). LPS was prepared from Aggregatibacter actinomycetemcomitans ATCC 43718 (American Type Culture Collection, Manassas, VA, USA) as described previously [22]. Fetal bovine serum (FBS) and alpha-minimum essential medium (α-MEM) were purchased from Gibco (Paisley, UK). Penicillin/streptomycin and trypsinEDTA were obtained from HyClone (Logan, UT, USA). Antibodies specific to cortactin, phospho-cortactin, cofilin, phospho-cofilin, vinculin, c-Fos, NFATc1, and β-actin were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). HRP-conjugated anti-rabbit IgG and anti-mouse IgG antibodies were purchased from Southern Biotech (Birmingham, AL, USA). Animals Animal experiments were approved by the Institutional Animal Care and Use Committee of Seoul National University. Five-week-old C57BL/6 mice were obtained from Orient Bio (Seongnam, South Korea). All mice were housed in a specific pathogen-free facility with a controlled temperature (22–24 °C) and 12-h day/night cycles and were allowed free access to water and food.

J Mol Med

Calvarial bone resorption assay

Western blot analysis

Collagen sheets soaked with 20 μg of RANKL in the presence or absence of HBD3-C15 (200 μg) were subsequently implanted on mouse calvaria (n = 5 per group). After 7 days, the calvaria were harvested and scanned using X-ray microcomputed tomography (micro-CT) (Skyscan1172 scanner; SkyScan, Kontich, Belgium) at 40 kV, 250 mA, 10 W, 0.5 mm AI filter, and 13 μm per pixel scan resolution. The micro-CT images were reconstructed by the SkyScan CT analyzer software. The three-dimensional images were created by the SkyScan CT volume program. Bone-resorbed areas were measured with the ImageJ program.

Western blot analysis was performed as previously described [23]. Briefly, BMMs were incubated with 20 μg/ml HBD3C15 in the presence of 100 ng/ml RANKL and 20 ng/ml MCSF for 2 days. Cell lysate was prepared and subjected to 10% SDS-PAGE. Then, the gel was transferred onto a PVDF membrane (Millipore, Bedford, MA, USA). After blocking with 5% skim milk in Tris-buffered saline containing 0.05% Tween 20, the membrane was incubated with antibodies specific to c-Fos, NFATc1, cortactin, phospho-cortactin, cofilin, phospho-cofilin, vinculin, and β-actin, followed by incubation with HRP-conjugated secondary antibodies. The immuno-reactive band was detected by using the ChemiDoc MP system (Bio-Rad Laboratories, Inc., Hercules, CA, USA).

Osteoclast differentiation Reverse transcription polymerase chain reaction Bone marrow (BM) was isolated from mouse femurs and tibiae. BMs were incubated in α-MEM supplemented with 10% FBS, 100 U/ml penicillin, and 100 μg/ml streptomycin containing 5 ng/ml of M-CSF for 1 day. Non-adherent cells were induced to differentiate into bone marrow-derived macrophages (BMMs) by treating them with 25 ng/ml of M-CSF for 4 days. BMMs were plated onto 96-well culture plates at 2 × 104 cells/well and incubated with 100 ng/ml of RANKL and 20 ng/ml of M-CSF in the presence or absence of HBD3C15 for 3 days. The cells were fixed and then stained using the TRAP staining kit (Sigma-Aldrich) according to the manufacturer’s recommended protocol. The TRAP-positive multinucleated cells (MNCs) with three or more nuclei were enumerated as osteoclasts with an inverted phase-contrast microscope. In a separate experiment, BMMs were differentiated into committed osteoclasts with 100 ng/ml RANKL and 20 ng/ml of M-CSF.

In vitro bone resorption assay BMMs (3 × 105 cells/ml) were plated onto 96-well calcium phosphate plates (Corning, Tewksbury, MA, USA) or dentin discs (Immunodiagnostic Systems, Boldon, UK) and incubated with 100 ng/ml of RANKL and 20 ng/ml of M-CSF in the presence or absence of HBD3-C15 for 5 days. The cells were lysed with 5% sodium hypochlorite for 10 min and then allowed to air-dry. The resorption pits on calcium phosphate plates were photographed, and the resorbed areas were analyzed by using the ImageJ program (National Institutes of Health, Bethesda, MD, USA). Dentin disc resorption pits were scanned by using a confocal laser scanning microscope (LSM 5 Pascal; Carl Zeiss MicroImaging GmbH, Jena, Germany) and analyzed by using the LSM Image Browser software (Carl Zeiss).

BMMs were incubated with 20 μg/ml HBD3-C15 in the presence of 100 ng/ml RANKL and 20 ng/ml M-CSF for 1 day. Reverse transcription polymerase chain reaction (RT-PCR) was performed as previously described [24]. The primers used were as follows: c-Fos: 5′-GCCGACTACGAGGCGTCATC C-3′ and 5′-TCCGGCACTTGGCTGCAGC-3′; NFATc1: 5′TCGGATCGAGGTGCAGCCCA-3′ and 5′-GGGC TGACCGCTGGGAACAC-3′; osteoclast-associated receptor (OSCAR): 5′-GTCCCTCCCCTGGCCTGCAT-3′ and 5′AGACCAGACACCTGGCCCCC-3′; DC-STAMP: 5′ACATGTGGGTGCTGTTTGCCG-3′ and 5′-CGGT TTCCCGTCAGCCTCTCTC-3′; cathepsin K: 5′-TCAA GGTTCTGCTGCTA-3′ and 5′-GAGCCAAGAGAGCA TAT-3′; TRAP: 5′-AACCGTGCAGACGATGGGCG-3′ and 5′-GCCAGGACAGCTGAGTGCGG-3′; calcitonin receptor (CTR): 5′-GGCAAACCCACGAGGCCGAG-3′ and 5′GGAGGCCATCTGGCTCAGCG-3′; and β-actin: 5′-GTGG GGCGCCCCAGGCACCA-3′ and 5′-CTCCTTAATGTCAC GCACGATTTC-3′. Immunofluorescence staining Cells were fixed with 4% paraformaldehyde and then permeabilized in 0.2% Triton X-100 for 15 min. After blocking with 1% bovine serum albumin in phosphate-buffered saline (PBS) for 1 h, cells were incubated with Alexa Fluor 488-conjugated phalloidin (Invitrogen, Grand Island, NY, USA) for 1 h. After washing with PBS, cells were incubated with Hoechst 33258 (Sigma-Aldrich). Finally, cells were washed with PBS and then mounted. Images were obtained using a confocal fluorescence microscope (Zeiss LSM 700, Carl Zeiss). In a separate experiment, double-stained cells with anti-vinculin antibody and Alexa Fluor 488-conjugated phalloidin were observed by a confocal fluorescence microscope (Zeiss LSM 800 with Airyscan, Carl Zeiss).

J Mol Med

Statistical analysis Statistical significance was assessed using Student’s t test. An asterisk (*) indicates a significant (P < 0.05) difference from the control group. All mice used in this study were randomly assigned to experimental or control groups. No animals were excluded from analysis.

Results HBD3-C15 inhibits RANKL-induced osteoclast differentiation First, we examined the effect of HBD3 on RANKL-induced osteoclast differentiation in vitro. As expected, RANKL increased the number of TRAP-positive MNCs. However, HBD3 inhibited the RANKL-increased TRAP-positive MNC formation, suggesting that HBD3 directly attenuates RANKL-induced osteoclast differentiation (Fig. 1a, b). Previous studies have shown that peptides comprised of the C-terminus of HBD3 effectively elicit antimicrobial and antifungal activities [19, 25]. Thus, we next investigated whether HBD3-C15 effectively controls osteoclast differentiation. Similar to HBD3, HBD3-C15 inhibited RANKL-induced osteoclast differentiation in a dose-dependent manner (Fig. 1c– e). Notably, HBD3 and HBD3-C15 did not affect cell viability (Supplemental Figs. 1a, b). However, unlike HBD-C15, the control peptide did not show such inhibitory effect (Supplemental Fig. 2a). Furthermore, the number of TRAPpositive MNCs containing more than ten nuclei was strikingly reduced after HBD3-C15 treatment, suggesting that HBD3C15 inhibits the late stages of osteoclast differentiation in particular (Fig. 1d). Since c-Fos and NFATc1 are essential transcription factors that positively regulate osteoclast differentiation [3], we examined the effect of HBD3-C15 on c-Fos and NFATc1 expression. We found that HBD3-C15 effectively inhibited RANKL-induced mRNA and protein expression of both c-Fos and NFATc1 in a dose-dependent manner (Fig. 1f, g, Supplemental Fig. 3a). In addition, HBD3-C15 also inhibited RANKL-induced expression of osteoclastogenesis-related genes including OSCAR, DCSTAMP, cathepsin K, TRAP, and CTR (Supplemental Fig. 3b). To determine the stage of osteoclast differentiation influenced by HBD3-C15, BMMs were treated with HBD3C15 at various time points during RANKL-induced osteoclast differentiation (Fig. 1h). When BMMs were treated with HBD3-C15 at day 0, the number of TRAP-positive MNCs with more than three nuclei was attenuated by HBD3-C15 (Fig. 1i). However, BMMs treated with HBD3-C15 at days 1 and 2 did not show the inhibitory effect of osteoclast differentiation. Interestingly, TRAP-positive multinucleated cells with more than ten nuclei were inhibited by HBD3-C15 in

all experimental groups (Fig. 1j). These results suggest that treatment with HBD3-C15 at the early stage effectively inhibits the osteoclast differentiation, in particular, to large osteoclasts with more than ten nuclei rather than small osteoclasts with less than three nuclei. HBD3-C15 inhibits RANKL-induced bone resorption capacity In order to further investigate the effect of HBD3-C15 on osteoclast, we performed an in vitro resorption assay using calcium phosphate-coated plates and dentin discs. RANKLinduced resorption on calcium phosphate-coated plates containing osteoclasts is inhibited by HBD3-C15 (Fig. 2a, b), but not by the control peptide (Supplemental Fig. 2b), suggesting that HBD3-C15 effectively inhibits not just osteoclast differentiation but also its function of mature osteoclasts. Furthermore, both the depth and area of resorption on the dentin discs were significantly reduced in the HBD3-C15treated group compared to the RANKL-treated group (Fig. 2c–e). This indicates that HBD3-C15 effectively inhibits RANKL-induced osteoclast activation. HBD3-C15 attenuates RANKL-induced podosome belt formation in osteoclasts Dysregulation of podosome organization can lead to impaired bone resorption [26] that is positively correlated with osteoclast podosome belt formation [27]. As such, we hypothesized that HBD3-C15 reduces the capacity of RANKL-induced bone resorption at least in part by disrupting the formation of the podosome belt. We found that RANKL-induced mature osteoclasts formed a podosome belt at the cell periphery (Fig. 3a). Treatment with HBD3-C15, however, dramatically decreased the degree of podosome belt formation in osteoclasts. The number of osteoclasts with an evident podosome belt was significantly reduced by treatment with HBD3-C15 (Fig. 3b). Podosome belt thickness was also significantly lower in HBD3-C15-treated osteoclasts compared to the control group (Fig. 3c, d). Especially, RANKL effectively induced a process of podosome belt formation including cluster, ring, and belt during osteoclast differentiation. However, HBD3C15 treatment prevented the transition between cluster/ring and belt, suggesting a decreased number of mature osteoclasts (Fig. 3e). Proper expression and activation of cytoskeletal molecules such as cortactin, cofilin, and vinculin are important for podosome belt formation in osteoclasts. Thus, we next examined the expression of each of those genes in the BMMs in the presence and absence of HBD3-C15. We found that HBD3-C15 inhibited RANKL-induced expression of cofilin and vinculin and that it led to decreased RANKL-activated phosphorylation of cortactin (Fig. 3f, Supplemental Figs. 4ae). Colocalization of actin with vinculin was observed in

J Mol Med

Fig. 1 HBD3-C15 inhibits RANKL-mediated osteoclast differentiation by suppressing c-Fos and NFATc1 expression. BMMs were differentiated into osteoclasts by adding 100 ng/ml RANKL and 20 ng/ml M-CSF in the presence or absence of a vehicle (20 μM HCl) or b 50 μg/ml of HBD3 or c–e 0, 0.2, 2, or 20 μg/ml of HBD3-C15 for 3 days. Cells were subjected to TRAP staining to assess osteoclast differentiation. Cells were then photographed, and TRAP-positive MNCs were enumerated through microscopic analysis. Scale bars = 100 μm. N.D. not detected. f, g BMMs were stimulated with 100 ng/ml RANKL and 20 ng/ml M-CSF in the

presence or absence of HBD3-C15 for 1 or 2 days. The expression levels of c-Fos, NFATc1, and β-actin were determined by RT-PCR (f) and Western blotting (g). The ratio of each gene to β-actin was obtained by using a densitometer. h–j BMMs were differentiated into osteoclast by 100 ng/ml RANKL and 20 ng/ml M-CSF and treated with 20 μg/ml HBD3-C15 at day 0, 1, or 2. At day 3, the cells were subjected to TRAP staining. TRAP-positive MNCs with more than three nuclei or with more than ten nuclei were enumerated through microscopic analysis. *P < 0.05. Data are representative of three independent experiments

RANKL-treated osteoclasts, but not HBD3-C15-treated osteoclasts (Fig. 3g, h). These results suggest that HBD3-C15 effectively inhibits osteoclast differentiation via the disruption of podosome belt formation.

treatment with HBD3-C15 (Fig. 4a, b). Normally, RANKL implantation increases the number of TRAP-positive osteoclasts (Fig. 4c, d). In contrast, the number of osteoclasts after RANKL-implantation was significantly reduced by additional implantation with HBD3-C15 (Fig. 4c, d). These results suggest that HBD3-C15 inhibits RANKL-induced bone resorption by downregulating osteoclast activation.

HBD3-C15 attenuates RANKL-induced bone resorption in vivo Next, we examined the effect of HBD3-C15 on the regulation of bone resorption using a mouse calvarial implantation model. After implanting RANKL-soaked collagen sheets, we observed excessive resorption in the calvarial bone. RANKL-induced bone resorption was significantly reduced, however, by the

HBD3-C15 alleviates A. actinomycetemcomitans LPS-induced bone resorption in vivo A. actinomycetemcomitans is a Gram-negative bacterium closely associated with localized aggressive periodontitis

J Mol Med

Fig. 2 HBD3-C15 inhibits the capacity of RANKL-induced bone resorption. a, b BMMs on a calcium phosphate-coated plate were differentiated into osteoclasts by adding 100 ng/ml RANKL and 20 ng/ml of M-CSF in the presence or absence of 20 μg/ml HBD3-C15 for 5 days. Cells were then lysed with 5% sodium hypochlorite for 10 min and subsequently dried. The resorbed areas were photographed and analyzed by using the ImageJ program. Scale bars = 100 μm. *P < 0.05. c–e BMMs on dentin

discs were differentiated into osteoclasts by adding 100 ng/ml of RANKL and 20 ng/ml of M-CSF in the presence or absence of 20 μg/ml HBD3C15 for 5 days. c The resorbed areas and depth on the dentin discs were photographed. Scale bars = 20 μm. The resorbed depth (d) and resorbed areas (e) were measured with the LSM Image Browser software. *P < 0.05. Data are representative of three independent experiments

accompanying rapid bone destruction [28]. LPS, which is a major virulence factor of Gram-negative bacteria, induces bone resorption by promoting the generation of bone-resorbing osteoclasts [29]. We next sought to examine the anti-resorption effect of HBD3-C15 using an A. actinomycetemcomitans LPS (AaLPS)-induced bone loss model. When we implanted collagen sheets soaked with AaLPS, we observed a distinct calvarial bone resorption (Fig. 5a). The resorbed area was significantly reduced in calvaria implanted with AaLPS plus HBD3-C15 compared to the AaLPS-only implanted calvaria (Fig. 5b). Furthermore, HBD3-C15 decreased the amount of AaLPSinduced TRAP (Fig. 5c, d), demonstrating that HBD3-C15 inhibits the formation of osteoclasts in vivo. These results indicate that HBD3-C15 alleviates AaLPS-induced bone resorption through its inhibition of osteoclast activation.

of the osteoclast podosome belt. However, AaLPS-induced podosome belt formation was disrupted by HBD3-C15 (Fig. 6e, f). Concordantly, HBD3-C15 inhibited AaLPSinduced phosphorylation of cortactin and expression of cofilin and vinculin in a dose-dependent manner (Fig. 6g, Supplemental Figs. 4f-j and 5b). Collectively, HBD3-C15 inhibits AaLPS-induced bone destruction by downregulating osteoclast differentiation and inducing the formation of the osteoclast podosome belt.

HBD3-C15 inhibits osteoclast differentiation via disruption of podosome belt formation in osteoclasts Because HBD3-C15 inhibited AaLPS-induced bone destruction, we hypothesized that it could also suppress osteoclastogenesis and podosome belt formation. AaLPS-induced osteoclast differentiation was decreased by HBD3-C15 treatment in a dose-dependent manner (Fig. 6a, b) through downregulation of c-Fos and NFATc1 expression (Fig. 6c, d, Supplemental Fig. 5a). Like RANKL, AaLPS also induced the formation

Discussion In this study, we demonstrated that HBD3-C15, a polypeptide consisting of 15 amino acids from the C-terminus of HBD3, reverses RANKL-induced bone resorption by suppressing osteoclast differentiation. Furthermore, HBD3-C15-attenuated bone resorption may be at least partially the result of its disruption of podosome belt formation in osteoclasts. HBD3C15 also alleviated AaLPS-induced bone resorption. These results suggest that HBD3-C15 acts as an antiresorptive agent to be an effective treatment for patients with bone diseases such as osteoporosis and localized aggressive periodontitis. Our results show that HBD3-C15 downregulated the expression of NFATc1 and c-Fos transcription factors, which are

J Mol Med

Fig. 3 HBD3-C15 disrupts RANKL-induced podosome belt formation in osteoclasts by downregulating podosome-related protein expression. BMMs on coverslips were differentiated into osteoclasts by adding 100 ng/ml RANKL and 20 ng/ml of M-CSF in the presence or absence of 20 μg/ml HBD3-C15 for 5 days (a–d, g, h) or indicated time points (e). Cells were fixed, permeabilized, and stained with Alexa Fluor 488conjugated phalloidin (green) and Hoechst 33258 (blue). a, b Podosome belt formation in osteoclasts was visualized by using a digital inverted fluorescence microscope. White arrows indicate the podosome cluster. Scale bars = 100 μm. c, d Podosome belt thickness was analyzed by a confocal fluorescence microscope. The distance between two white arrows represents podosome belt thickness. Scale bars = 10 μm. e The

percentage of osteoclasts showing podosome belt formation was enumerated by using microscopic analysis. f BMMs were stimulated with 100 ng/ml of RANKL and 20 ng/ml of M-CSF in the presence or absence of 20 μg/ml HBD3-C15 for 2 days. The expression levels of phosphocortactin, cortactin, cofilin, phospho-cofilin, vinculin, or β-actin were determined by Western blotting. g, h Cells were fixed, permeabilized, and stained for vinculin (red), actin (green), and nuclei (blue). Colocalization of vinculin and actin was analyzed by a confocal fluorescence microscope. Scale bars = 20 μm (left) or 2 μm (right). The vinculin/ actin ratio was measured by using the ImageJ program. N.D. not detected. *P < 0.05. Data are representative of three independent experiments

key positive regulators of osteoclast differentiation. The inhibitory effect on osteoclast differentiation by other antimicrobial peptides has been also reported. For instance, LL37, an antimicrobial peptide in the cathelicidin family, has been shown to inhibit osteoclast differentiation [21]. Additionally, CRAMP inhibited osteoclast differentiation in human peripheral blood mononuclear cell culture [20]. Together with our findings, these results suggest that anti-osteoclastogenic activity is likely to be a common feature of antimicrobial peptides. The specific inhibitory mechanism for each peptide, however, appears to be distinct. For example, LL37 suppresses osteoclast differentiation by inhibiting the nuclear translocation of NFATc1 as a result of downregulated calcineurin activity [21], while CRAMP reduced the degree of LPS- and flagellin-induced osteoclast differentiation by downregulating RANKL in the osteoblast/osteoclast co-culture system [20]. In

yet another distinct mechanism, we found that HBD3-C15 inhibited osteoclast differentiation by suppressing podosome belt formation while also downregulating the NFATc1 and cFos transcription factors. The work presented here demonstrates that HBD3-C15 substantially interferes with RANKL- or AaLPS-induced podosome belt formation in the osteoclasts. Podosomes are integrin-based and actin-rich circular structures that arise from the plasma membrane [30]. They are typically involved in the adhesion, spreading, migration, and bone resorption of osteoclasts [8]. Interference with podosome belt formation has been found to cause serious side effects, reducing the degree of osteoclast bone resorption activity [31]. It has recently been suggested that the bone resorption inhibitory effect of C21, which is a chemical inhibitor of Dock5-mediated Rac1 activation, is mediated via disruption of podosome organization

J Mol Med Fig. 4 HBD3-C15 reduces RANKL-induced bone loss in vivo. Collagen sheets soaked with RANKL in the presence or absence of HBD3-C15 were implanted on mouse calvaria. After 7 days, the calvaria were obtained and scanned by micro-CT. a Three-dimensional images of the calvaria were obtained from one of three similar results. b The bone-resorbed areas were measured by using the ImageJ program. ROI = 4.25 × 4.25 mm2; *P < 0.05. c The calvaria were subjected to TRAP staining and photographed. d TRAP-positive areas were measured by using the ImageJ program. ROI = 4.25 × 4.25 mm2; *P < 0.05. Data are representative of three independent experiments

[32]. FAK-depleted osteoclasts similarly experience bone resorption impairment as a result of disrupted podosome structures [33]. Additionally, protein tyrosine kinase Pyk2 deficiency is known to induce an osteopetrotic phenotype due to the Fig. 5 HBD3-C15 reverses AaLPS-induced bone loss in vivo. Collagen sheets soaked with AaLPS in the presence or absence of HBD3-C15 were implanted on mouse calvaria. After 7 days, the calvaria were obtained and scanned by micro-CT. a Threedimensional images of the calvaria were obtained from one of three similar results. b The bone-resorbed areas were measured by using the ImageJ program. ROI = 4.25 × 4.25 mm2; *P < 0.05. c Calvaria were subjected to TRAP staining and photographed. d TRAP-positive areas were measured by using the ImageJ program. ROI = 4.25 × 4.25 mm2. *P < 0.05. Data are representative of three independent experiments

dysregulation of podosome organization [26]. Therefore, the ability of HBD3-C15 to target podosome belt formation reveals a unique and important mechanism for suppressing osteoclast differentiation and bone resorption function.

J Mol Med

Fig. 6 HBD3-C15 inhibits AaLPS-induced osteoclastogenesis by disrupting osteoclast podosome belt formation. To generate committed osteoclast precursors, BMMs were treated with 100 ng/ml RANKL and 20 ng/ml M-CSF for 2 days. a, b Committed osteoclast precursors were differentiated into osteoclasts by adding 5 μg/ml AaLPS and 20 ng/ml MCSF in the presence or absence of HBD3-C15 for 3 days. Cells were photographed with magnification at × 40 (a), and TRAP-positive MNCs were enumerated through microscopic analysis (b). c, d Committed osteoclast precursors were stimulated with AaLPS and MCSF in the presence or absence of HBD3-C15 for 24 h. The expression levels of c-Fos, NFATc1, and β-actin were determined by using Western blotting. d The ratio of c-Fos or NFATc1 to β-actin was obtained by using a densitometer. e Committed osteoclast precursors on coverslips were

differentiated into osteoclasts by adding AaLPS and M-CSF in the presence or absence of HBD3-C15 for 5 days. Cells were fixed, permeabilized, and stained with Alexa Fluor 488-conjugated phalloidin (green) and Hoechst 33258 (blue). Osteoclast podosome belt formation was visualized with a digital inverted fluorescence microscope. Scale bars = 100 μm (upper) and 50 μm (lower). f The number of osteoclasts with podosome belt was analyzed by a confocal fluorescence microscope. g Committed osteoclast precursors were stimulated with AaLPS and MCSF in the presence or absence of HBD3-C15 for 48 h. The expression levels of phospho-cortactin, cortactin, cofilin, phospho-cofilin, vinculin, and β-actin were determined by Western blotting. *P < 0.05. Data are representative of three independent experiments

We also observed that HBD3-C15 prevents RANKL from activating actin-binding proteins. Podosomes consist of an Factin bundle surrounded by integrin and actin-binding proteins such as cofilin, cortactin, and vinculin [8]. Thus, the activation of actin-binding proteins marks a critical step in podosome formation. Deletion or dephosphorylation of cortactin in the podosome core resulted in a total lack of podosome formation and the complete loss of bone resorption activity [34]. Recently, Zalli et al. reported that an osteoclast-specific deletion of cofilin impaired bone resorption as a result of disrupted podosome belt formation [35]. This group also found that podosome belt formation was rescued in cofilin-deficient

osteoclasts when dephosphorylated, active cofilin was overexpressed. On the other hand, vinculin-deficient osteoclasts did not show any actin ring formation or bone resorption activity [36]. Therefore, it seems likely that HBD3-C15 inhibits podosome belt formation by interfering with the activation of actin-binding proteins. In general, antimicrobial peptides including HBD3 are secreted into mucosal tissues in order to support the host innate immunity by protecting against infection by bacteria, viruses, and fungi [11]. Interestingly, accumulating reports suggest that these antimicrobial peptides are also commonly found in bone tissues and are produced by bone cells. For example,

J Mol Med

HBD1, HBD2, and HBD3 have been observed in human oral bone tissues and, to a greater extent, in chronically infected mandibular bone [37]. Additionally, HBD3 is produced by healthy bones and primary osteoblasts [38]. Furthermore, periprosthetic tissue and cancellous bone in periprosthetic joint infection patients have been found to exhibit an increased number of HBD3-positive cells [39]. HBD3 is also found in the synovial membrane of patients with pyogenic arthritis [40]. HBD3 has been reported to modulate the remodeling processes in articular joint infection and bacteriacontaminated wounds [41–43]. Therefore, at least some of these antimicrobial peptides are expected to play a role in the modulation of bone metabolism. In this study, we report that HBD3-C15 effectively regulates bone metabolism by controlling osteoclast differentiation and function. Further study to determine whether there is a relationship between antimicrobial activity and bone metabolism is needed. In conclusion, we demonstrated that HBD3-C15 has an inhibitory effect on osteoclast differentiation. HBD3-C15 appears to reduce or reverse RANKL- and AaLPS-induced bone resorption by disrupting podosome belt formation in osteoclasts. Taken together, these results suggest that the HBD3C15 could be a potential therapeutic agent against bone diseases and that its future study to that end is warranted. Acknowledgements This work was supported by grants from the National Research Foundation of Korea, which is funded by the Korean government (NRF-2015R1A2A1A15055453, NRF2015M2A2A6A01044894, and NRF-2015R1D1A1A09056592), and by the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health and Welfare (HI14C0469 and HI17C1377), Republic of Korea.

4.

5.

6.

7.

8.

9.

10.

11. 12.

13. 14.

15. Author contributions S.H.H. conceived the idea. S.H.H. and O.-J.P. designed the experiments. O.-J.P., J.K., J.Y.L., Y.-J. P., K.-Y.K., K.B.A., and S.H.H. performed the experiments and/or interpreted the data. C.H.Y. provided the critical comments. All authors contributed to the discussion of the results, followed by writing and reviewing the manuscript.

16.

17. Compliance with ethical standards Animal experiments were approved by the Institutional Animal Care and Use Committee of Seoul National University.

18.

Conflict of interest The authors declare that they have no conflict of interests.

19.

References

20.

1. 2. 3.

Long F (2012) Building strong bones: molecular regulation of the osteoblast lineage. Nat Rev Mol Cell Biol 13:27–38 Boyle WJ, Simonet WS, Lacey DL (2003) Osteoclast differentiation and activation. Nature 423:337–342 Takayanagi H, Kim S, Koga T, Nishina H, Isshiki M, Yoshida H, Saiura A, Isobe M, Yokochi T, Inoue J et al (2002) Induction and activation of the transcription factor NFATc1 (NFAT2) integrate

21.

22.

RANKL signaling in terminal differentiation of osteoclasts. Dev Cell 3:889–901 Boyce BF, Xing L (2008) Functions of RANKL/RANK/OPG in bone modeling and remodeling. Arch Biochem Biophys 473:139– 146 Lange AW, Yutzey KE (2006) NFATc1 expression in the developing heart valves is responsive to the RANKL pathway and is required for endocardial expression of cathepsin K. Dev Biol 292: 407–417 Kim K, Lee SH, Ha Kim J, Choi Y, Kim N (2008) NFATc1 induces osteoclast fusion via up-regulation of Atp6v0d2 and the dendritic cell-specific transmembrane protein (DC-STAMP). Mol Endocrinol 22:176–185 Saltel F, Chabadel A, Bonnelye E, Jurdic P (2008) Actin cytoskeletal organisation in osteoclasts: a model to decipher transmigration and matrix degradation. Eur J Cell Biol 87:459–468 Georgess D, Machuca-Gayet I, Blangy A, Jurdic P (2014) Podosome organization drives osteoclast-mediated bone resorption. Cell Adhes Migr 8:191–204 Destaing O, Saltel F, Geminard JC, Jurdic P, Bard F (2003) Podosomes display actin turnover and dynamic self-organization in osteoclasts expressing actin-green fluorescent protein. Mol Biol Cell 14:407–416 Choi SW, Yeon JT, Ryu BJ, Kim KJ, Moon SH, Lee H, Lee MS, Lee SY, Heo JC, Park SJ et al (2015) Repositioning potential of PAK4 to osteoclastic bone resorption. J Bone Miner Res 30:1494– 1507 Ganz T (2003) Defensins: antimicrobial peptides of innate immunity. Nat Rev Immunol 3:710–720 Dhople V, Krukemeyer A, Ramamoorthy A (2006) The human beta-defensin-3, an antibacterial peptide with multiple biological functions. Biochim Biophys Acta 1758:1499–1512 Selsted ME, Ouellette AJ (2005) Mammalian defensins in the antimicrobial immune response. Nat Immunol 6:551–557 Schibli DJ, Hunter HN, Aseyev V, Starner TD, Wiencek JM, PB MC Jr, Tack BF, Vogel HJ (2002) The solution structures of the human beta-defensins lead to a better understanding of the potent bactericidal activity of HBD3 against Staphylococcus aureus. J Biol Chem 277:8279–8289 Lin P, Li Y, Dong K, Li Q (2015) The antibacterial effects of an antimicrobial peptide human beta-defensin 3 fused with carbohydrate-binding domain on Pseudomonas aeruginosa PA14. Curr Microbiol 71:170–176 Rozek A, Powers JP, Friedrich CL, Hancock RE (2003) Structurebased design of an indolicidin peptide analogue with increased protease stability. Biochemistry 42:14130–14138 Jenssen H, Hamill P, Hancock RE (2006) Peptide antimicrobial agents. Clin Microbiol Rev 19:491–511 Lee JK, Chang SW, Perinpanayagam H, Lim SM, Park YJ, Han SH, Baek SH, Zhu Q, Bae KS, Kum KY (2013) Antibacterial efficacy of a human beta-defensin-3 peptide on multispecies biofilms. J Endod 39:1625–1629 Lim SM, Ahn KB, Kim C, Kum JW, Perinpanayagam H, Gu Y, Yoo YJ, Chang SW, Han SH, Shon WJ et al (2016) Antifungal effects of synthetic human beta-defensin 3-C15 peptide. Restor Dent Endod 41:91–97 Horibe K, Nakamichi Y, Uehara S, Nakamura M, Koide M, Kobayashi Y, Takahashi N, Udagawa N (2013) Roles of cathelicidin-related antimicrobial peptide in murine osteoclastogenesis. Immunology 140:344–351 Supanchart C, Thawanaphong S, Makeudom A, Bolscher JG, Nazmi K, Kornak U, Krisanaprakornkit S (2012) The antimicrobial peptide, LL-37, inhibits in vitro osteoclastogenesis. J Dent Res 91: 1071–1077 Park OJ, Cho MK, Yun CH, Han SH (2015) Lipopolysaccharide of Aggregatibacter actinomycetemcomitans induces the expression of

J Mol Med chemokines MCP-1, MIP-1alpha, and IP-10 via similar but distinct signaling pathways in murine macrophages. Immunobiology 220: 1067–1074 23. Kim J, Yang J, Park OJ, Kang SS, Kim WS, Kurokawa K, Yun CH, Kim HH, Lee BL, Han SH (2013) Lipoproteins are an important bacterial component responsible for bone destruction through the induction of osteoclast differentiation and activation. J Bone Miner Res 28:2381–2391 24. Yoon H, Yoon YS, Kim MS, Chung MJ, Yum DY (2014) A probiotic preparation duolac-gold ameliorates dextran sulphate sodium-induced mouse colitis by downregulating the expression of IL-6. Toxicol Res 30:27–32 25. Hoover DM, Wu Z, Tucker K, Lu W, Lubkowski J (2003) Antimicrobial characterization of human beta-defensin 3 derivatives. Antimicrob Agents Chemother 47:2804–2809 26. Gil-Henn H, Destaing O, Sims NA, Aoki K, Alles N, Neff L, Sanjay A, Bruzzaniti A, De Camilli P, Baron R et al (2007) Defective microtubule-dependent podosome organization in osteoclasts leads to increased bone density in Pyk2(-/-) mice. J Cell Biol 178:1053–1064 27. Fuller K, Ross JL, Szewczyk KA, Moss R, Chambers TJ (2010) Bone is not essential for osteoclast activation. PLoS One 5:e12837 28. Prakasam A, Elavarasu SS, Natarajan RK (2012) Antibiotics in the management of aggressive periodontitis. J Pharm Bioallied Sci 4: S252–S255 29. Madeira MF, Queiroz-Junior CM, Cisalpino D, Werneck SM, Kikuchi H, Fujise O, Ryffel B, Silva TA, Teixeira MM, Souza DG (2013) MyD88 is essential for alveolar bone loss induced by Aggregatibacter actinomycetemcomitans lipopolysaccharide in mice. Mol Oral Microbiol 28:415–424 30. Linder S, Aepfelbacher M (2003) Podosomes: adhesion hot-spots of invasive cells. Trends Cell Biol 13:376–385 31. Matsumoto T, Nagase Y, Hirose J, Tokuyama N, Yasui T, Kadono Y, Ueki K, Kadowaki T, Nakamura K, Tanaka S (2013) Regulation of bone resorption and sealing zone formation in osteoclasts occurs through protein kinase B-mediated microtubule stabilization. J Bone Miner Res 28:1191–1202 32. Vives V, Cres G, Richard C, Busson M, Ferrandez Y, Planson AG, Zeghouf M, Cherfils J, Malaval L, Blangy A (2015) Pharmacological inhibition of Dock5 prevents osteolysis by

affecting osteoclast podosome organization while preserving bone formation. Nat Commun 6:6218 33. Ray BJ, Thomas K, Huang CS, Gutknecht MF, Botchwey EA, Bouton AH (2012) Regulation of osteoclast structure and function by FAK family kinases. J Leukoc Biol 92:1021–1028 34. Tehrani S, Faccio R, Chandrasekar I, Ross FP, Cooper JA (2006) Cortactin has an essential and specific role in osteoclast actin assembly. Mol Biol Cell 17:2882–2895 35. Zalli D, Neff L, Nagano K, Shin NY, Witke W, Gori F, Baron R (2016) The actin-binding protein cofilin and its interaction with cortactin are required for podosome patterning in osteoclasts and bone resorption in vivo and in vitro. J Bone Miner Res 31:1701–1712 36. Fukunaga T, Zou W, Warren JT, Teitelbaum SL (2014) Vinculin regulates osteoclast function. J Biol Chem 289:13554–13564 37. Warnke PH, Springer IN, Russo PA, Wiltfang J, Essig H, Kosmahl M, Sherry E, Acil Y (2006) Innate immunity in human bone. Bone 38:400–408 38. Varoga D, Wruck CJ, Tohidnezhad M, Brandenburg L, Paulsen F, Mentlein R, Seekamp A, Besch L, Pufe T (2009) Osteoblasts participate in the innate immunity of the bone by producing human beta defensin-3. Histochem Cell Biol 131:207–218 39. Liu GD, Yu HJ, Ou S, Luo X, Ni WD, Huang XK, Chen JY, Wang Y, Javard P, Fei J (2014) Human beta-defensin-3 for the diagnosis of periprosthetic joint infection and loosening. Orthopedics 37:e384–e390 40. Paulsen F, Pufe T, Conradi L, Varoga D, Tsokos M, Papendieck J, Petersen W (2002) Antimicrobial peptides are expressed and produced in healthy and inflamed human synovial membranes. J Pathol 198:369–377 41. Tohidnezhad M, Varoga D, Podschun R, Wruck CJ, Seekamp A, Brandenburg LO, Pufe T, Lippross S (2011) Thrombocytes are effectors of the innate immune system releasing human beta defensin-3. Injury 42:682–686 42. Varoga D, Pufe T, Mentlein R, Kohrs S, Grohmann S, Tillmann B, Hassenpflug J, Paulsen F (2005) Expression and regulation of antimicrobial peptides in articular joints. Ann Anat 187:499–508 43. Varoga D, Pufe T, Harder J, Schroder JM, Mentlein R, MeyerHoffert U, Goldring MB, Tillmann B, Hassenpflug J, Paulsen F (2005) Human beta-defensin 3 mediates tissue remodeling processes in articular cartilage by increasing levels of metalloproteinases and reducing levels of their endogenous inhibitors. Arthritis and Rheum 52:1736–1745