The Long Pentraxin 3 Plays a Role in Bone Turnover and Repair

Pentraxin 3 (PTX3) is an inflammatory mediator acting as a fluid-phase pattern recognition molecule and playing an essential role in innate immunity and matrix remodeling. Inflammatory mediators also contribute to skeletal homeostasis, operating at multiple levels in physiological and pathological conditions. This study was designed to investigate the role of PTX3 in physiological skeletal remodeling and bone healing. Micro-computed tomography (μCT) and bone histomorphometry of distal femur showed that PTX3 gene-targeted female and male mice (ptx3−/−) had lower trabecular bone volume than their wild-type (ptx3+/+) littermates (BV/TV by μCT: 3.50 ± 1.31 vs 6.09 ± 1.17 for females, p < 0.0001; BV/TV 9.06 ± 1.89 vs 10.47 ± 1.97 for males, p = 0.0435). In addition, μCT revealed lower trabecular bone volume in second lumbar vertebra of ptx3−/− mice. PTX3 was increasingly expressed during osteoblast maturation in vitro and was able to reverse the negative effect of fibroblast growth factor 2 (FGF2) on osteoblast differentiation. This effect was specific for the N-terminal domain of PTX3 that contains the FGF2-binding site. By using the closed transversal tibial fracture model, we found that ptx3−/− female mice formed significantly less mineralized callus during the anabolic phase following fracture injury compared to ptx3+/+ mice (BV/TV 17.05 ± 4.59 vs 20.47 ± 3.32, p = 0.0195). Non-hematopoietic periosteal cells highly upregulated PTX3 expression during the initial phase of fracture healing, particularly CD51+ and αSma+ osteoprogenitor subsets, and callus tissue exhibited concomitant expression of PTX3 and FGF2 around the fracture site. Thus, PTX3 supports maintenance of the bone mass possibly by inhibiting FGF2 and its negative impact on bone formation. Moreover, PTX3 enables timely occurring sequence of callus mineralization after bone fracture injury. These results indicate that PTX3 plays an important role in bone homeostasis and in proper matrix mineralization during fracture repair, a reflection of the function of this molecule in tissue homeostasis and repair.

Pentraxin 3 (PTX3) is an inflammatory mediator acting as a fluid-phase pattern recognition molecule and playing an essential role in innate immunity and matrix remodeling. Inflammatory mediators also contribute to skeletal homeostasis, operating at multiple levels in physiological and pathological conditions. This study was designed to investigate the role of PTX3 in physiological skeletal remodeling and bone healing. Micro-computed tomography (μCT) and bone histomorphometry of distal femur showed that PTX3 gene-targeted female and male mice (ptx3 −/− ) had lower trabecular bone volume than their wild-type (ptx3 +/+ ) littermates (BV/TV by μCT: 3.50 ± 1.31 vs 6.09 ± 1.17 for females, p < 0.0001; BV/TV 9.06 ± 1.89 vs 10.47 ± 1.97 for males, p = 0.0435). In addition, μCT revealed lower trabecular bone volume in second lumbar vertebra of ptx3 −/− mice. PTX3 was increasingly expressed during osteoblast maturation in vitro and was able to reverse the negative effect of fibroblast growth factor 2 (FGF2) on osteoblast differentiation. This effect was specific for the N-terminal domain of PTX3 that contains the FGF2-binding site. By using the closed transversal tibial fracture model, we found that ptx3 −/− female mice formed significantly less mineralized callus during the anabolic phase following fracture injury compared to ptx3 +/+ mice (BV/TV 17.05 ± 4.59 vs 20.47 ± 3.32, p = 0.0195). Nonhematopoietic periosteal cells highly upregulated PTX3 expression during the initial phase of fracture healing, particularly CD51 + and αSma + osteoprogenitor subsets, and callus tissue exhibited concomitant expression of PTX3 and FGF2 around the fracture site. Thus, PTX3 supports maintenance of the bone mass possibly by inhibiting FGF2 and its negative impact on bone formation. Moreover, PTX3 enables timely occurring sequence of callus mineralization after bone fracture injury. These results indicate that PTX3 plays an important role in bone homeostasis and in proper matrix mineralization during fracture repair, a reflection of the function of this molecule in tissue homeostasis and repair.
Keywords: pentraxin 3, fibroblast growth factor 2, inflammation, bone turnover, bone healing, osteoblast, osteoclast inTrODUcTiOn The long pentraxin 3 (PTX3) is a member of the evolutionarily conserved pentraxin superfamily, acting as a soluble pattern recognition molecule (1). The mature protein is a complex octameric molecule whose composing protomer subunits comprise a long N-terminal domain, unrelated to any known protein, and a C-terminal domain homologous to the classical pentraxins, C-reactive protein, and serum amyloid P component (SAP), which collectively form the short arm of the superfamily (2). PTX3 is induced by primary pro-inflammatory cytokines, mainly tumor necrosis factor-α (TNF-α) and interleukin-1 (IL-1), and Toll-like receptor (TLR) agonists in different cell types, such as myeloid (i.e., mononuclear phagocytes, neutrophils, and dendritic cells), endothelial, and stromal cells (1,3,4).
A number of animal models and human studies have shown that PTX3 exerts complex non-redundant roles in the innate resistance to pathogens (5)(6)(7), in the regulation of inflammation and tissue repair (8), acting also as an oncosuppressor gene (9). The multifaceted biology of PTX3 likely originates from its capacity to interact with a broad spectrum of ligands, including complement components (C1q, factor H, C4 binding protein, ficolins, and mannose-binding lectin), adhesion molecules (P-selectin), growth factors [fibroblast growth factor 2 (FGF2) and other members of the FGF family], and extracellular matrix components (inter-α-trypsin inhibitor, TNF-inducible gene-6, and fibrin), as well as to recognize late apoptotic cells and microbes (1,4).
Several lines of evidence indicate that pentraxins, through the interplay with various hormones, growth factors, and matrix proteins, are involved in tissue remodeling in both physiological and pathological conditions. It has been shown that ptx3 −/− female mice display a severely impaired fertility due to defective assembly of the viscoelastic hyaluronic acid-rich matrix surrounding the oocyte in the preovulatory cumulus oophorus complex (10,11). In different models of tissue damage (skin wound healing, chemically induced sterile liver and lung injury, arterial thrombosis), PTX3 deficiency has been associated with increased clot formation, fibrin persistence, and enhanced collagen deposition (12). In this context, macrophages and mesenchymal cells release PTX3 in response to TLR activation and IL-1 stimulation, and the secreted protein promotes remodeling of the fibrin-rich inflammatory matrix, ensuring hemostasis and efficient wound healing. This information notwithstanding, no evidence is available regarding the role of PTX3 in bone turnover, an additional example of tissue remodeling.
Skeletal metabolism is regulated by the cooperative action of two major cell lineages, the mesenchymal osteoblast lineage that mediates bone formation and the hematopoietic osteoclast lineage responsible for bone resorption, with additional contributions from monocytes, macrophages, lymphocytes, and mast cells (13,14). Bone remodeling is a highly coordinated process consisting of timely occurring bone resorption and formation, triggered by changes in mechanical forces and/or microdamage and modulated by a number of factors including inflammation, hormonal levels, and local mediators (15). At various stages of bone remodeling and repair, progenitor cells, osteoclasts, and osteoblasts com municate through multiple cell-to-cell and cell-to-matrix interactions via a number of signaling molecules, including tissue growth factors, cytokines, and matrix proteins. Maintenance of skeletal homeostasis during physiological bone turnover and initiation of efficient reparative process after bone injury are crucial to preserve the major biological roles of the bone tissue, i.e., provide support to mechanical load and exert metabolic functions (16)(17)(18).
Bone remodeling is crucial to control the reshaping or replacement of bone during growth and following injuries (15). Fractures are one of the most frequent injuries of the musculoskeletal system, affecting not only the bone itself but also the periosteum, blood vessels, and the surrounding soft tissue. Fracture healing is a sequential process that requires a complex interaction among molecular factors, immune cells, resident tissue cells, and progenitor cells to stimulate inflammation, angiogenesis, mesenchymal progenitor recruitment, cartilage and bone formation, extracellular matrix synthesis, and callus remodeling (19,20). Inflammatory cells (neutrophils and macrophages) infiltrate the injured area, where they produce inflammatory mediators, such as TNF-α, IL-6, and prostaglandin E2. Mesenchymal progenitors recruited to the site of injury undergo endochondral bone formation, initially differentiating into cartilage, which is further replaced by the calcified bone that incorporates the bone marrow. In addition, periosteal mesenchymal progenitors adjacent to the fracture gap differentiate directly into osteoblasts by intramembranous ossification. Therefore, a local acute inflammatory response sets within the first few days from fracture is required to initiate an adequate healing process (21,22).
The aim of this study is to investigate the role of the inflammatory mediator PTX3 during physiological bone remodeling and fracture healing. By using ptx3 −/− mice, we found that PTX3 is important for bone tissue homeostasis and fracture repair. PTX3 is expressed by osteoblast lineage cells, with osteoprogenitor cells being the major source of PTX3 in the early inflammatory stages of fracture healing.

Mice
Ptx3 −/− (knockout) mice, generated as previously described (6), were used on a 129/SvPas or C57BL/6J inbred (backcrossed for 11 generations) genetic background. Wild-type (ptx3 +/+ ) mice were obtained from Charles River (Charles River Laboratories, Calco, Italy) or were cohoused littermates. Femora and lumbar vertebrae were collected from mice housed in the specific pathogen-free animal facility of the Humanitas Clinical and Research Center in individually ventilated cages. Animals of 10-12 weeks old (young mice) or 6-8 months old (old mice) were used. Procedures involving animals handling and care were conformed to protocols approved by the Ethics Committees of the Humanitas Clinical and Research Center (Rozzano, Milan, Italy). For the in vivo fracture experiments, the mice were transported from Humanitas Clinical and Research Center and housed until 14-16 weeks old, in groups of three to four per cage, and maintained at standard conditions with food and water ad libitum at the animal facility of the Croatian Institute for Brain Research, University of Zagreb School of Medicine (Zagreb, Croatia). All experimental procedures were reviewed and approved by the Ethics Committee of the University of Zagreb School of Medicine and conducted in accordance with accepted standards of humane care and use of laboratory animals. All procedures were performed under tribromoethanol (Avertin) anesthesia, and all efforts were made to minimize suffering of mice during experiments. PTX3 in Bone Homeostasis and Repair Frontiers in Immunology | www.frontiersin.org March 2018 | Volume 9 | Article 417

recombinant PTX3
Recombinant human PTX3 and its C-terminal (C-PTX3) and N-terminal (N-PTX3) domains were purified under endotoxinfree conditions from the supernatants of stably transfected Chinese hamster ovary cells as previously described (8).

Bone cell cultures
For osteoblast generation, bone marrow cells were flushed out from the femoral medullary cavity of ptx3 +/+ or ptx3 −/− mice. Cells adjacent to endosteal and periosteal surfaces of mid-diaphyseal femoral bone shafts were released by homogenization. Harvested cells were plated into 24-well plates at a density of 0.8 × 10 6 cells/ well in 0.8 mL/well of α-MEM/10% fetal calf serum (FCS). Three wells from two to three mice were done for each gender/genotype, and three independent experiments were performed. Osteoblast differentiation was induced from day 7 by addition of 50 µg/mL ascorbic acid (AA) and 8 mM β-glycerophosphate (β-GP) (Sigma-Aldrich for both, Saint Louis, MO, USA). At day 14-17, osteoblast cultures were stained cytochemically for the activity of alkaline phosphatase (ALP) using a commercially available kit (Sigma-Aldrich), and red color ALP + osteoblast surface per well was measured by a custom-made software (23). Total colony surface was assessed by methylene blue staining. For osteoclast generation, bone marrow cells from ptx3 +/+ or ptx3 −/− mice were cultured overnight with 5 ng/mL recombinant mouse M-CSF (R&D Systems, NE Minneapolis, MN, USA) in α-MEM/10% FCS to stimulate the monocyte/macrophage lineage followed by harvesting of non-adherent cells as enriched hematopoietic monocyte/macrophage progenitors. Non-adherent cells were replated into 48-well plates at a density of 0.25 × 10 6 /well in 0.5 mL/well of α-MEM/10% FCS supplemented with 20 ng/mL M-CSF and 40 ng/mL RANKL (R&D Systems). Spleen cells, obtained by smashing the spleen tissue between a pair of frosted microscope slides, were cultured in 48-well plates at a density of 0.5 × 10 6 /well in 0.5 mL/well of α-MEM/10% FCS supplemented with 20 ng/mL M-CSF and 40 ng/mL RANKL. At day 5-9 of culture, tartrate-resistant acid phosphatase (TRAP) + multinucleated cells (≥3 nuclei/cell) were identified using a commercially available kit (Sigma-Aldrich) and counted by light microscopy (24). Four wells from two to three mice were done for each gender/genotype, and three independent experiments were performed.
Osteoblastogenic and osteoclastogenic cultures were treated with a dose range of PTX3 (0-50 nM). In some experiments, either full-length PTX3, its N-or C-terminal domain (50 nM for all), or/ and FGF2 (0.25 nM and 0.5 nM) were added to osteoblastogenic cultures in addition to AA and β-GP. Molarity was expressed based on the molecular weight of the protomer subunits of PTX3 and its N-and C-terminal domains, and the primary sequence of FGF2, that is, 42.500 kDa for PTX3, 18.163 kDa for N-PTX3, 25.000 kDa for C-PTX3, and 17.254 kDa for FGF2 (25).
Fluorescence-activated cell sorting (FACS) of osteoprogenitor cells was performed in a BD FACSAria I instrument. Cells were extracted from the periosteal layer and callus tissue and labeled as described previously (24). Briefly, tibias were dissected, soft tissue was removed, and bone marrow was flushed. Periosteal layer of diaphyseal bone fragments/fracture callus were briefly crushed using tissue homogenizer and further digested in PBS containing 0.1% collagenase A (Roche, Mannheim, Germany) for 30 min at 37°C with agitation. For phenotyping of osteoblast lineage cells, the following antibodies were used: anti-CD51 biotinylated (clone RMV-7; with streptavidin coupled to PE-Cy7), anti-CD105 PE or PE-Cy7 (clone MJ7/18), anti-CD140b APC (clone APB5), anti-α-smooth muscle actin (αSma) Alexa Fluor 488 (clone 1A4), and anti-Sca-1 FITC or Pacific Blue (clone D7), according to previous studies (26). Samples were acquired at a speed of approximately 3,000 cells/s to separate hematopoietic/ endothelial-negative (CD45 − Ter119 − CD31 − ) and hematopoieticpositive fractions (CD45 + Ter119 + CD31 + ). Sorting parameters and experimental set-up were optimized for high purity sorting. Sorting purity was determined by a reanalysis of fractioned populations and was greater than 98% for all samples.
For flow cytometric analysis of PTX3 expression, purified goat immunoglobulin G (IgG) anti-mouse PTX3 (R&D Systems) or purified IgG from pre-immune goat serum (Agilent-Dako, Santa Clara, CA, USA) were used with anti-goat antibody coupled to PE (eBiosciences) as previously described (9). Cells were first stained for surface marker expression using the previously described antibodies to hematopoietic and osteoprogenitor lineages, then permeabilized with BD Cytofix/Cytoperm kit (BD Bio sciences), according with manufacturer's instructions, and, finally, stained for intracellular expression of PTX3. Gates were defined using unlabeled cells and isotype controls. Specificity of anti-PTX3 labeling was also confirmed by parallel staining in ptx3 +/+ and ptx3 −/− mice. Samples were acquired using BD FACSAria I instrument and analyzed by FlowJo software.

Fractures
Closed transverse diaphyseal fractures of tibias were created in 14-to 16-week-old female ptx3 +/+ and ptx3 −/− mice. Anesthesia was induced and maintained using tribromoethanol (Avertin), and 0.08 mg/kg buprenorphine hydrochloride was administered subcutaneously for analgesia. Prior to fracture, a tract was created through the tibial plateau into the medullary canal using a 25-gauge needle, through which a 0.4-mm diameter stainless steel pin was positioned to fix the subsequent fracture (26). Fractures were created 1-2 mm proximal to the distal tibia-fibula junction using the modified, limited displacement three-point contact device (27). Unfractured bones and bones with inserted pins without fracture were used as controls.

Micro-computed Tomography
The distal metaphyses of femora and second lumbar vertebrae were scanned using a micro-computed tomography (μCT) system (1172 SkyScan, SkyScan, Kontich, Belgium) at 50 kV and 200 mA with a 0.5 aluminum filter using a detection pixel size of 4.3 µm (24). Images were captured every 0.7° through 180° [second lumbar (L2) vertebrae] and every 0.7° through 360° (femora) rotation. The scanned images were reconstructed using the SkyScan Recon software and analyzed using SkyScan CTAnalyser. Three-dimensional analysis and reconstruction of trabecular bone were performed on the bone region 1-2.3 mm distal to the growth plate. The trabecular bone compartment was manually delineated from the cortical bone, and the following parameters were measured: trabecular bone volume fraction (BV/TV, %), trabecular number (Tb.N/μm), trabecular thickness (Tb.Th; μm), and trabecular separation (Tb.Sp; μm).
In fracture experiments, tibial fractured areas were scanned and three-dimensional reconstructions were performed to measure mineralized callus bone volume (BV/TV, %).
immunohistochemistry Bone sections were prepared as described and deparaffinized. Fractured bones had the intramedullary pin removed prior to embedding. Sections were incubated overnight at 65°C with 10 mM sodium citrate (pH 6) for antigen retrieval. Endogenous peroxidases were inactivated with 0.03% H2O2 in methanol for 30 min at room temperature. Serial sections were blocked with Rodent block M (Bio care Medical, Concord, CA, USA) for PTX3, osterix (Osx), and runt-related transcription factor 2 (Runx2) staining or 5% BSA in 0.5% Triton/PBS for FGF2 staining and then incubated with the primary antibody at 4°C overnight. As primary antibodies, we used an affinity purified rabbit IgG against human PTX3 (5 µg/mL), a monoclonal rabbit IgG against mouse/ rat/human Osx (1:1,000; Abcam, Cambridge, UK), a monoclonal rabbit IgG against mouse/rat/human Runx2 (1:2,000; Abcam), or polyclonal goat IgG against human FGF2 (5 µg/mL; R&D Systems) (28). Secondary reagents were Envision anti-rabbit horse radish peroxidase (HRP) (Dako, Glostrup, Denmark) for PTX3, Osx, and Runx2 staining or goat probe with goat HRP polymer (Bio care Medical) for FGF2 staining. Reactions were visualized using 3.3′-diaminobenzidine as a chromogen (Vector laboratories Inc. Burlingham, CA, USA) and stopped with distilled water. Sections were counterstained with hematoxylin and covered using Hydromount (National Diagnostics, Atlanta, GA, USA).
For quantitative analysis of PTX3 expression, images of three serial sections from four mice per group were digitally captured at 200× magnification. PTX3 immunoreactive areas were measured using ImageJ software (version 1.45; NIH, MD, USA). Immunoreactive area was automatically selected on the basis of immunoreactive sample intensity, and results are expressed as median percentage of the immunoreactive pixels in total pixels with interquartile range (IQR).

gene expression analysis
Total RNA from cultures and callus-derived cells was extracted using TRIzol Reagent (Applied Biosystems, Thermo Fisher Scientific, Waltham, MA, USA), reversely transcribed (1 µg) to cDNA, and amplified (20 ng/well in duplicates) by quantitative (q)PCR using an AB7500 (Applied Biosystems) instrument. Each sample was prepared by pooling three to four culture wells or callus tissue from two to three bones. RNA quality was verified on a Bioanalyzer RNA PicoChip (Agilent Technologies, Santa Clara, CA, USA) and quantified using Nanodrop spectrophotometer (Thermo Scientific, Thermo Fisher Scientific). Expression of osteoblast-specific genes: osteocalcin (OC), bone sialoprotein (BSP), Osx, ALP, α2-chain of type I collagen (Col1); osteoclast specific genes: colony stimulating factor 1 receptor (cFms), receptor activator of nuclear factor-κB (RANK), and calcitonin receptor (CalcR); inflammatory markers: C-C chemokine ligand 2 (CCL2), TNF-α, IL-1α, and IL-6; and regulatory molecules: PTX3 and FGF2 was determined using commercially available TaqMan Gene Expression Assays (Applied Biosystems). The expression of a specific gene was calculated according to the relative standard curve of gene expression in the calibrator sample (cDNA from ptx3 +/+ callus or cultures) and then normalized to the expression level of the β-actin gene as an endogenous control (24). elisa Blood was collected from the retro-orbital plexus of control and fractured mice at day 2 and 6 postfracture; serum was stored at −20°C until cytokine analysis. Murine PTX3, CCL2, TNF-α, and IL-6 levels were measured in serum by ELISA (DuoSet ELISA Development System, R&D Systems) according to manufacturer's instructions.
statistics Normality of data distribution was verified using Kolmogorov-Smirnov test. Number of independent experiments performed for each setting with total number of analyzed mice per group is indicated in figure legends. Results are presented as mean ± SD for normally distributed data or median with IQR for notnormally distributed data. Statistical analysis of the group difference was performed using the parametric Student's t-test (for group-to-group comparison) and analysis of variance with Student-Newman-Keuls post hoc test (for multiple group comparisons) or the non-parametric Kruskal-Wallis test followed by Mann-Whitney test. For all experiments, α-level was set at 0.05. Statistical analysis was performed using the software packages MedCalc (version 12.5, MedCalc Inc., Mariakerke, Belgium).

Ptx3 −/− Female Mice have lower Bone Mass compared to ptx3 +/+ Mice
In the first set of experiments, we aimed to investigate the role of PTX3 in maintenance of skeletal homeostasis. Both axial (at the level of L2 vertebra) and appendicular (at the level of distal femoral metaphysis) skeleton of ptx3 +/+ and ptx3 −/− mice were assessed by μCT. Distal femora from young adult ptx3 −/− females (10-12 weeks old) on B6 background showed highly significant decrease in trabecular bone volume (p < 0.0001), which was paralleled by decreased trabecular number and trabecular thickness, as well as increased trabecular separation compared to ptx3 +/+ mice (Figure 1). Difference in distal femoral trabecular bone volume was marginally significant between ptx3 −/− and ptx3 +/+ males (p = 0.0435), with lower trabecular number in ptx3 −/− mice. Moreover, trabecular bone volume, trabecular number, and trabecular thickness were decreased in lumbar vertebrae from ptx3 −/− compared to ptx3 +/+ female and male mice (Figure 1). Histomorphometry of distal femoral histological sections confirmed lower bone mass in ptx3 −/− female and male mice found by μCT ( Figure S1 in Supplementary Material). Finally, we analyzed bone mass in aged female ptx3 −/− mice (6-8 months old) on B6 background and found significantly lower distal femoral trabecular bone volume and trabecular thickness compared to ptx3 +/+ mice ( Figure S2 in Supplementary Material).
In addition, μCT analysis performed in young adult mice on S129 background revealed the decrease in trabecular bone volume in distal femora of both female and male ptx3 −/− mice as well as in axial skeleton of ptx3 −/− females ( Figure S3 in Supplementary Material). These findings point to strain-independent alteration in bone remodeling, in particular in female mice lacking the PTX3 gene.

PTX3 Deficiency Does not affect the Osteoclastogenic and Osteoblastogenic Potential of Progenitor cells In Vitro
To clarify the mechanisms underlying the bone phenotype in ptx3 −/− mice, we analyzed the osteoclastogenic and osteoblastogenic potential of bone progenitor cells. Osteoclasts were differentiated from bone marrow and spleen cells in the presence of RANKL and M-CSF. Number of differentiated osteoclasts, detected as TRAP + multinucleated cells, did not significantly differ between ptx3 +/+ and ptx3 −/− mice in both females and males (Figure 2A). In addition, frequency of putative osteoclast progenitors (29,30), defined as CD3 − B220 − CD11b −/lo cKit + cFms + subset for bone marrow ( Figure S4A in Supplementary Material) and CD3 − B220 − NK1.1 − CD11b + Ly6C + cFms + subset for spleen (not shown), was comparable between ptx3 +/+ and ptx3 −/− mice.
Osteoblasts were differentiated in vitro from bone marrow and bone adjacent (endosteal and periosteal) cells by the addition of AA and β-GP. Ptx3 −/− mice did not show significant difference in the percentage of ALP + surface area ( Figure 2B) as well as frequency of osteoprogenitor cells (31,32), defined as CD45 − Ter119 − CD31 − CD166 + Sca-1 + subset ( Figure S4B  Since reduction in bone mass arises from unbalanced bone resorption (by osteoclasts) and formation (by osteoblasts), we assessed in vivo activity of osteoclasts and osteoblasts on histology specimens from distal femora. Ptx3 −/− mice did not differ from ptx3 +/+ mice in the number of active trabecular and endosteal osteoclasts, counted as multinucleated TRAP + cells adjacent to the bone surface ( Figure 3A). However, dynamic histomorphometry, measuring distance of calcein labels injected in 5-day interval, revealed defective osteoblast activity and decreased BFR on endosteal and trabecular bone surface in both female and male ptx3 −/− mice ( Figure 3B). These findings suggest that the lower bone volume in ptx3 −/− mice might be ascribed to suppressed osteoblast function and prompted us to exploit functional model of bone injury requiring intensive new bone formation in vivo.  ptx3 −/− mice. Three weeks after tibial fracture injury, during the reparative phase of the fracture healing process marked by intensive endochondral and intramembranous ossification, both percentage of mineralized callus tissue and expression of the bone-specific marker Col1 were lower in female ptx3 −/− mice than in ptx3 +/+ controls (Figures 4A,B). Immunohistochemistry revealed that PTX3 is present in the callus tissue of ptx3 +/+ mice during the mineralizing phase of fracture healing ( Figure 4C). Areas of newly formed mineralized tissue (detected as GT green stain) produce PTX3, including cuboidal active osteoblasts, mostly apparent on high magnification. Therefore, we concluded that PTX3 has a role during fracture repair and that inactivation of the PTX3 gene is responsible for the reduced osteoblast function and incomplete bone formation observed in ptx3 −/− mice.
PTX3 is expressed during the early inflammatory Phase of Bone healing revealed weakly positive osteocytes incorporated in the cortical bone matrix even in the unfractured bone tissue ( Figure 5A). Moreover, expression of PTX3 is induced in the soft callus tissue early after the bone fracture injury, as evidenced by the significant increase of immunopositive area at day 6 postfracture (median, 5.3; IQR, 3.6-9.4 vs 0.4, IQR, 0.2-0.6 in control; p = 0.0012) ( Figure 5A). Specifically, in areas of endochondral ossification, hypertrophying chondrocytes highly express PTX3 ( Figure 5B). In parallel, serum level of PTX3 was increased at day 2 postfracture compared to unfractured ptx3 +/+ mice ( Figure  S5A in Supplementary Material). Furthermore, we tested whether mice lacking the inflammatory mediator PTX3 show alterations in systemic and local fracture-induced inflammatory response at early postfracture days. Serum levels of the inflammatory chemokine CCL2 were comparable between ptx3 +/+ and ptx3 −/− mice, with the peak value at day 2 postfracture (Figure S5A in Supplementary Material). Levels of TNF-α and IL-6 were undetectable in serum    Figure S5B in Supplementary Material), possibly because their increase occurs much earlier following injury. However, no obvious differences were observed in the expression of pro-inflammatory cytokines (CCL2 and IL-1α, Figure S5B in Supplementary Material; IL-6, not shown) or the frequency of inflammatory cell subsets (Ly6C + , Ly6G + , and CD11b + ) between ptx3 +/+ and ptx3 −/− mice ( Figure S5C in Supplementary Material). The fracture model used is based on the intramedullary insertion of a stainless steel pin that, per se, can induce an inflam ma tory response. Thus, we assessed in both ptx3 +/+ and ptx3 −/− mice the contribution of injury made by the intramedullary pin insertion to the inflammatory response associated with the fracture model applied in this set of experiments. Expression of inflammatory mediators did not significantly differ between control (intact) animals and animals with inserted pins. Although pin insertion probably causes a certain level of intramedullary inflammatory response, bone injury and subsequent periosteal reaction seem to be necessary to produce increased expression of inflammatory markers in serum and callus tissue ( Figure S5 in Supplementary Material).

PTX3 expression increases with Bone cell Differentiation In Vitro
Next, we asked whether PTX3 could act as an autocrine growth factor for bone cells and characterized the kinetics of PTX3 expression (along with that of selected differentiationspecific genes) in osteoclastogenic and osteoblastogenic cultures ( Figure 7A). Osteoblast differentiation (as induced by AA and β-GP) was confirmed by a large increase in the expression of both early and late osteoblast-specific genes (Osx, ALP, OC). In these conditions, PTX3 expression was highly induced at later time points. Osteoclastogenic cultures derived from bone marrow cells and stimulated with RANKL and M-CSF showed an increased expression of differentiation genes (cFms, RANK, and CalcR). PTX3 was also expressed in the course of osteoclast maturation, however, at lower levels (6-to 10-fold less) than those observed in the osteoblastogenic cultures (PTX3 was normalized using the same relative standard curve to allow comparison of the two cultures).
Although we observed a marked increase in PTX3 expression in mature osteoblasts in vitro and could expect, from previous findings, that PTX3 may directly stimulate osteoblast maturation, exogenously added PTX3 did not change the percentage of surface covered by ALP + osteoblast ( Figure S6A in Supplementary  Material). Similarly, addition of PTX3 into osteoclastogenic cultures did not alter the number of mature TRAP + multinucleated osteoclasts. Therefore, we proposed that PTX3 action on bone cells may be indirect, through interaction with some other bone matrix regulatory protein.

PTX3 Binds to FgF2 and reverses its inhibitory role in Osteoblast Differentiation
To elucidate the mechanisms by which PTX3 is required for physiological bone formation and bone tissue regeneration after injury, we performed combined treatment of osteoblastogenic cultures with PTX3 and FGF2, a known regulator of osteoblastogenesis and also a ligand of PTX3 (34). As shown in Figures 7B,C, addition of FGF2 markedly reduced the percentage of surface covered by ALP + osteoblast after culture treatment in a dose-dependent manner. This effect is only seen if FGF2 is added during the differentiation phase (day 7-14) of osteoblastogenic cultures and not during the proliferative phase (day 0-7), as described in Materials and Methods ( Figure S6B in Supplementary Material). Addition of PTX3 to FGF2 treatment was able to reverse the negative effect of FGF2 on osteoblast differentiation (Figures 7B,C). The N-terminal domain of PTX3 [that contains the FGF2-binding site (35)] recapitulates the FGF2-neutralizing activity of the full lengths protein, whereas no effect was observed with the C-terminal domain (that does not interact with FGF2). In addition, FGF2 strongly suppressed the expression of the differentiation markers Osx, Col1, and BSP, whereas PTX3 rescued their expression toward control levels ( Figure 7D).
Finally, FGF2 was detected (both protein and mRNA) at early days from fracture and showed a similar pattern of expression as PTX3 in bone sections and extracted callus cells of ptx3 +/+ mice (Figures 8A,D). Moreover, immunohistochemical staining of serial bone sections indicated concomitant high expression of Osx and Runx2 by infiltrating osteoprogenitor cells in the areas of periosteal reaction and bone induction (Figure 8A). The cellular source of PTX3 was determined by intracellular staining of PTX3 in parallel to the staining of surface hematopoietic (data not shown) and mesenchymal/osteoprogenitor markers using flow cytometry (Figures 8B,C). The percentage of PTX3 + cells increased following fracture (Figure 8B), and among them, the most abundant subsets were αSma + early osteoprogenitors and CD51 + more mature preosteoblast (Figure 8C), whereas hematopoietic cells exhibited almost no PTX3 expression at day 2 and 6 postfracture (data not shown). There were no differences in PTX3 expression between control (intact) animals and animals with inserted pins (Figure 8B). A similar pattern of expression for FGF2, Osx, and Runx2 was observed in bone sections of ptx3 −/− mice, with no PTX3 immunoreactivity ( Figure S7 in Supplementary Material). In addition, the percentages of hematopoietic/endothelial-negative (CD45 − Ter119 − CD31 − ) osteo progenitor subpopulations (CD51 + and CD140b + , Figure S8 in Supplementary Material; CD105, data not shown) were comparable between ptx3 +/+ and ptx3 −/− mice except for the αSma + subset that was lower in ptx3 −/− mice ( Figure S8 in Supplementary Material).
On the basis of these findings we propose that PTX3, produced by osteoblast lineage cells, serves as a bone protective factor, important to unlock osteoblast maturation by neutralizing the FGF2 anti-differentiation effect during bone formation.

DiscUssiOn
In this study, we identified a new role for the long pentraxin PTX3 as a local regulator of bone metabolism. We observed that young adult ptx3 −/− mice (2.5 months of age) have lower trabecular bone mass in long bones and axial skeleton, indicating that genetic deficiency of PTX3 alters the balance between bone formation and resorption in both B6 and S129 strain. Skeletal sexual dimorphism is present in ptx3 −/− mice, since more obvious bone phenotype compared to ptx3 +/+ mice has been observed for young adult females, known to have lower BFR compared to male mice at the analyzed age (36). Levels of several hormones, cytokines, and growth factors important for osteoblast function may be different in females vs males (including sex hormones, growth hormone, and RANKL), making them susceptible to specific conditions, i.e., PTX3 deficiency (37). Compared to 2.5 months, the bone phenotype was less pronounced in females at 6-8 months when age-related decline in osteoblast functions and trabecular bone loss have already begun (38,39). Adult bone is a dynamic tissue that requires constant remodeling to maintain structural integrity and fulfill its metabolic roles. In addition to hormonal signals and mechanical stimuli, paracrine effects of local mediators, including tissue growth factors, cytokines, and bone matrix proteins are important for physiological and pathological bone turnover (15). The trabecular bone is particularly susceptible to modulatory signals and may undergo a negative bone balance in the adult organism due to high metabolic rate and great surface area-to-volume ratio (21,38,40,41). The presence of lower bone mass in the spine of ptx3 −/− mice, which in small quadrupeds is loaded by much lower forces compared to the femur, suggests that PTX3 is required for bone homeostasis not only in weight bearing, rapidly remodeled bones, but also in the axial skeleton.
Bone matrix incorporates a large number of growth factors, including insulin-like growth factors, transforming growth factor-β, BMPs, platelet-derived growth factor, and parathyroid hormone-related protein and FGFs, which play a role in the physiologic bone development and in the remodeling associated with skeletal repair. Matrix proteins mostly produced by the bone cells themselves, such as Col1, ALP, osteonectin, fibronectin, and OC, may act as growth factor modulators (16). They participate in the regulation of proliferation, migration, and differentiation of osteoblastic cells, especially during the resorption phase of bone remodeling when osteoclasts degrade bone matrix, releasing the incorporated proteins. Moreover, a number of inflammatory mediators including cytokines, chemokines, prostaglandins, and leukotrienes are produced by bone and immune cells and, through autocrine and paracrine loops, regulate bone cell functions (13,22). However, these inflammatory mediators have been mostly studied in the context of pathological inflammation (i.e., arthritis, periodontitis, osteomyelitis), whereas our study determined the role of PTX3 in physiological bone remodeling and repair. We found that PTX3 is expressed by differentiating osteoblasts in vitro and matrix-embedded osteocytes in vivo and that PTX3-deficient mice have significantly lower bone formation and bone repair rate than ptx3 +/+ controls. Nevertheless, frequency and differentiation potential of bone marrow-derived osteoprogenitors in vitro were similar in ptx3 +/+ and ptx3 −/− mice, indicating that an in situ microenvironment is required for the bone-spearing effect of PTX3. In addition, culture treatment with exogenous PTX3 did not directly affect osteoclast or osteoblast differentiation, similar to what observed by Lee et al. (42). This study also showed that PTX3 enhanced the RANKL expression in primary mouse calvarial osteoblasts and stimulated osteoclastogenesis under high TNF-α conditions. Other authors have investigated the role of PTX3 in chronic inflammatory diseases with aggravated bone resorption, including periodontitis (43)(44)(45), arthritis (46)(47)(48)(49)(50)(51)(52), and tumor-associated inflammation (53,54), showing increased PTX3 levels, especially in areas of inflammation. However, in these models, it is difficult, if not impossible, to define whether PTX3 is the causal factor of bone loss or it is just a byproduct, downstream of pro-inflammatory cytokines that per se induce bone resorption.
As opposed to chronic inflammatory diseases or excessive inflammation associated with massive trauma, when the inflammatory response has destructive effects on bone tissue, inflammation plays an immediate beneficial role in promoting regeneration after bone fracture or injury (19,20). A number of inflammatory cytokines (TNF-α, IL-1α, IL-1β, and IL-6) are expressed during the acute inflammatory phase within a week from fracture, which, in cooperation with growth factors and matrix proteins released from injured tissues, recruit inflammatory cells, promote angiogenesis, and guide mesenchymal progenitor cell migration and differentiation (19). In our fracture model, the pro-inflammatory chemokine CCL2 was particularly useful as an indicator of early fracture injury in serum and at the fracture site. Later on, remodeling of the mineralized callus by the resorptive action of osteoclasts is paralleled by a second peak of inflammatory cytokines and matrix proteins. Although the recruited inflammatory cells, mainly neutrophils and macrophages, are major immediate producers of cytokines and growth factors, in the regenerative callus, these mediators are primarily produced by bone cells, mainly osteoblasts and chondrocytes (55). Here, we showed that PTX3 is expressed in both the early (inflammatory) and the late (regenerative) phases of fracture healing. Moreover, it is highly induced in non-hematopoietic/non-endothelial population of early callus tissue, which is mainly composed of osteoprogenitor cells expressing mesenchymal/osteoblast lineage markers (i.e., αSma, CD51, CD140b). In particular, PTX3 is produced by a substantial proportion of CD51 + and αSma + subsets of osteoprogenitor cells. Moreover, the proportion of αSma + subset, which label osteoprogenitor cells induced at early postfracture days (26), was significantly decreased in ptx3 −/− mice. In a model of wound healing, Doni et al. also showed that PTX3 was mainly produced by non-hematopoietic mesenchymal lineage cells (9). Balanced cytokine milieu is crucial for the beneficial effect of acute inflammation on bone regeneration, so alterations in the local concentration of growth factors and inflammatory mediators may have significant impacts on the healing process. In mice lacking PTX3, callus mineralization is delayed with lower expression of Col1, the most abundant osteoblast-specific matrix protein. Despite lacking PTX3, the overall profile of systemic and fracture-associated inflammatory indicators (CCL2, IL-1α, IL-6) and inflammatory cell subsets (Ly6G, Ly6C, CD11b) in ptx3 −/− mice showed no significant difference compared to ptx3 +/+ mice. To the best of our knowledge, there are no other studies investigating the role of PTX3 during bone regeneration. However, in the model of orthodontic tooth movement, characterized by a rapid bone remodeling, PTX3, together with other inflammatory mediators, was significantly overexpressed throughout the periodontal ligament of the tension zone (56). Other in vivo studies have demonstrated that mice lacking pro-inflammatory cytokines, such as IL-6 or TNF-α, or treated with anti-inflammatory drugs, such as cyclooxygenase-2 inhibitors, have disturbed bone formation and delayed fracture healing (19,20,22). Osteoblasts taken from human patients with fracture non-unions have abnormal expression of genes associated with inflammatory response, suggesting that inflammatory mediators are critical for bone healing (57).
Bone remodeling and regeneration relies on tight spatiotemporal interactions between growth factors and extracellular matrix molecules, which, in an autocrine and paracrine manner, promote osteoblast development and differentiation. Our results and previous studies showed that human and animal stromal/ osteoblast lineage cells express PTX3 (9,26,58,59). Moreover, a recent study by Scimeca et al. indicated importance of PTX3 for human osteoblast differentiation and showed reduced PTX3 expression in osteoblasts from osteoporotic patients (60). The multifunctional properties of PTX3 are based on its capacity to interact with several ligands, including complement components, adhesion molecules, growth factors, and matrix components (9,25). In particular, PTX3 binds with high affinity to FGF2 (34,35) and to other members of the FGF family (28,61) via the N-terminal domain and prevents FGF2 from interacting with its receptor. The FGF family comprises 22 members, highly conserved in evolution (62), exerting various biological activities both in vivo and in vitro, including roles in angiogenesis, mitogenesis, cellular differentiation, cell migration, and tissue injury repair. In addition, FGF2 is expressed during the early stages of bone formation and is abundantly accumulated in bone matrix, participating to osteoblastogenesis and skeletal remodeling (63)(64)(65). Therefore, we propose that PTX3 might exert a bone regulatory effect in repair and remodeling through the interaction with locally produced FGF2. At early stages following fracture, non-hematopoietic/non-endothelial cells comprising the osteoprogenitor population showed concomitant induction in PTX3 and FGF2 gene expression. The FGF2 protein was detected by immunohistochemistry in close proximity to the fracture gap, overlapping with the areas of PTX3 staining and osteoprogenitor cells expressing Osx and Runx2. FGF2 is a key mitogen able to expand the pool of various bone cells, including osteoprogenitors. However, its effects in the later stages of matrix maturation and mineralization are controversial, with both positive and negative impacts described (66,67). Genetic studies in animals and etiology of human skeletal dysplasias showed that FGF2 overexpression as well as mutation of FGF2 or the corresponding receptors cause bone abnormalities (65,68). Similar to that described by Kalajzic et al. (69), in our in vitro conditions, FGF2 blocked osteoblast differentiation from bone marrow stromal cells, whereas PTX3 (and the PTX3 N-terminal domain that contains the FGF2-binding sites) was able to reverse the inhibitory effect of FGF2. Similarly, by sequestering FGF2, PTX3 can inhibit FGF2-dependent endothelial cell proliferation in vitro and angiogenesis in vivo (33,35). Several extracellular matrix molecules, besides PTX3, are able to trap FGF2 in the extracellular environment (such as thrombospondin-1 and cleaved syndecan) and modulate its effects during processes such as inflammation, wound healing, atherosclerosis, and neoplasia (70,71).
In conclusion, we observed that PTX3-deficient mice have impaired bone formation during physiological remodeling as well as regeneration following fracture injury. In particular, PTX3 is important for adequate osteoblast differentiation and proper matrix mineralization, and this effect is achieved, at least in part, by sequestering FGF2 in the extracellular matrix of newly remodeled bone. We propose that the role of PTX3 in the regulation of bone remodeling is a part of its general biological function in the orchestration of tissue repair. aUThOr cOnTriBUTiOns DG, IK, AM, and BB designed the study; DG, LD, HC, VK, and TK performed the experiments; DG, MS, SV, LD, HC, NK, VK, and TK acquired and analyzed data; DG, MS, SV, AI, NK, IK, AM, and BB interpreted the results; DG, AI, AM, and BB prepared the manuscript. All authors critically revised the manuscript and approved the final version.