Annals of Anatomy 196 (2014) 286–295

Contents lists available at ScienceDirect

Annals of Anatomy journal homepage: www.elsevier.de/aanat

Research article

BDNF and its TrkB receptor in human fracture healing Olaf Kilian a,b , Sonja Hartmann a , Nicole Dongowski c , Srikanth Karnati d , Eveline Baumgart-Vogt d , Frauke V. Härtel e , Thomas Noll e , Reinhard Schnettler a,f , Katrin Susanne Lips a,∗ a

Laboratory of Experimental Trauma Surgery, University of Giessen, Kerkraderstr. 9, 35392 Giessen, Germany Department of Orthopedics and Trauma, Zentralklinik Bad Berka, Robert-Koch-Allee 9, 99437 Bad Berka, Germany Department of Trauma Surgery, Asklepios Klinik Lich, Goethestr. 4, 35423 Lich, Germany d Institute for Anatomy and Cell Biology II, Division of Medical Cell Biology, University of Giessen, Aulweg 123, 35392 Giessen, Germany e Institute of Physiology, University of Giessen, Aulweg 129, 35392 Giessen, Germany f Department of Trauma Surgery, University of Giessen, Rudolf-Buchheim-Str. 7, 35385 Giessen, Germany b c

a r t i c l e

i n f o

Article history: Received 7 February 2014 Received in revised form 5 June 2014 Accepted 6 June 2014 Keywords: Neurotrophin Brain-derived neurotrophic factor Tropomyosin-related kinases Osteoblast Bone Blood vessel Angiogenesis

s u m m a r y Fracture healing is a physiological process of repair which proceeds in stages, each characterized by a different predominant tissue in the fracture gap. Matrix reorganization is regulated by cytokines and growth factors. Neurotrophins and their receptors might be of importance to osteoblasts and endothelial cells during fracture healing. The aim of this study was to examine the presence of brain-derived neurotrophic factor (BDNF) and its tropomyosin-related kinase B receptor (TrkB) during human fracture healing. BDNF and TrkB were investigated in samples from human fracture gaps and cultured cells using RT-PCR, Western blot, and immunohistochemistry. Endothelial cells and osteoblastic cell lines demonstrated a cytoplasmic staining pattern of BDNF and TrkB in vitro. At the mRNA level, BDNF and TrkB were expressed in the initial and osteoid formation phase of human fracture healing. In the granulation tissue of fracture gap, both proteins – BDNF and TrkB – are concentrated in endothelial and osteoblastic cells at the margins of woven bone suggesting their involvement in the formation of new vessels. There was no evidence of BDNF or TrkB during fracture healing in chondrocytes of human enchondral tissue. Furthermore, BDNF is absent in mature bone. Taken together, BDNF and TrkB are involved in vessel formation and osteogenic processes during human fracture healing. The detection of BDNF and its TrkB receptor during various stages of the bone formation process in human fracture gap tissue were shown for the first time. The current study reveals that both proteins are up-regulated in human osteoblasts and endothelial cells in fracture healing. © 2014 Elsevier GmbH. All rights reserved.

1. Introduction Fracture healing is a complex physiological process that involves a sequence of coordinated activity at the cellular and

∗ Corresponding author. Tel.: +49 6419939870; fax: +49 6419939879. E-mail addresses: [email protected] (O. Kilian), [email protected] (S. Hartmann), [email protected] (N. Dongowski), [email protected] (S. Karnati), [email protected] (E. Baumgart-Vogt), [email protected] (F.V. Härtel), [email protected] (T. Noll), [email protected] (R. Schnettler), [email protected] (K.S. Lips). http://dx.doi.org/10.1016/j.aanat.2014.06.001 0940-9602/© 2014 Elsevier GmbH. All rights reserved.

molecular level. It contains discrete stages of tissue response including inflammation, recruitment and differentiation of mesenchymal stem cells, angiogenesis, callus and bone formation by recapitulation of embryonic enchondral and intramembranous ossification (Claes et al., 2012). Within this process, invasion of vessels and their stabilization by matrix is required for successful bone repair (Claes et al., 2012; Geris et al., 2008; Keramaris et al., 2008). Bone matrix is permanently reorganized through cell–cell- and cell–matrix-interactions during fracture healing. These processes are regulated by a precise orchestration of cytokines and growth factors. Changes in their expression pattern as well as an increase in reactive oxygen species (ROS) can induce delayed fracture healing (Beckmann et al., 2014). Systemic and metabolic diseases as well as inflammatory processes also have negative effects on fracture healing by disarrangement of the distinct cellular and molecular

O. Kilian et al. / Annals of Anatomy 196 (2014) 286–295

cascade: Reduced levels of oestrogen and vascular endothelial growth factor (VEGF) as determined after menopause are linked to delayed bone regeneration in the systemic condition of osteoporosis (Giannoudis et al., 2007). In addition, it has been reported that, among others, brain-derived neurotrophic factor (BDNF) is increased in the systemic inflammatory profile of patients with chronic obstructive pulmonary disease (Loza et al., 2012). BDNF belongs to the family of neurotrophins that were originally identified as growth factors for neuronal cells and were detected both in central and peripheral neuronal tissue (Hallbook et al., 2006; Hohn et al., 1990; Huang and Reichardt, 2001; Kaplan and Miller, 2000; Kim et al., 2004). Additionally they were synthesized and released from non-neuronal cells such as fibroblasts, osteoblasts, endothelial cells, monocytes and mast cells (Cartwright et al., 1994; Donovan et al., 2000; Kurihara et al., 2003; Labouyrie et al., 1999). With the exception of osteoblasts these cells are involved in formation of a fracture gap haematoma, replacement of the haematoma by granulation tissue as well as angiogenesis during the inflammatory phase whereas osteoblasts are the most important cells of the repair phase of fracture healing (Claes et al., 2012). Development of osteoblasts from differentiating mesenchymal stem cells as well as building and remodelling of bone matrix are processes that are regulated by growth and osteogenic factors. One of these factors essential to normal angiogenesis and appropriate callus architecture is VEGF (Geris et al., 2008). The neurotrophin BDNF is a signalling partner of VEGF in angiogenic tube formation. Donovan et al. (2000) showed that overexpression of BDNF stimulates angiogenesis through a VEGF signalling pathway (Donovan et al., 2000). The coordinated action of these two growth factors was recently verified (Long et al., 2013). In general, BDNF acts through binding to the highly specific tropomyosin-related kinase receptor B (TrkB) leading as well as to the p75-NTR receptor that is able to bind the other members of the neurotrophin family. When binding to the p75-NTR receptor, BDNF induces apoptosis, whereas coupling to TrkB triggers intracellular survival signalling pathways (Huang and Reichardt, 2001; Kaplan and Miller, 2000). Recently, Usui et al. reported that BDNF leads to ROS generation via activation of TrkB (Usui et al., 2014). The phosphorylation of Akt was identified as the next step downstream which raised the assumption that BDNF induces endothelial cell migration through the PI3-kinase/Akt signalling pathway (Au et al., 2009; Kim et al., 2004). In addition to the cooperation of BDNF and VEGF, TrkB and ROS are able to stimulate VEGF receptor-2 (VEGFR-2) expression and activation (Ushio-Fukai, 2007; Usui et al., 2014). This explains the similar and sometimes even greater pro-angiogenic effect of BDNF in correlation with VEGF (Kermani et al., 2005; Long et al., 2013). Besides acting as a survival factor and promoting angiogenesis, it is supposed that BDNF is involved in the cross talk of inflammatory cells. It has been reported that the release of BDNF is enhanced after nerve injury through mechanisms that include activation of mast cells and macrophage-like cells (Skaper et al., 2012). Additionally, BDNF plasma levels were increased in patients with osteoarthritis compared to healthy individuals (Simao et al., 2014). In both contexts it is suspected that BDNF is involved in pain perception (Simao et al., 2014; Skaper et al., 2012). Furthermore, BDNF enhances the expression and synthesis of alkaline phosphatase, osteopontin, bone morphogenetic protein2 in cementoblasts through the TrkB-cRaf-ERK1/2-Elk1 signalling pathway. Cementoblasts produce mineralized cementum that forms the outer covering of the tooth roots (Kajiya et al., 2008). Thus, the functions of cementoblasts are somehow similar to those of the osteoblast. Within this framework, we were interested in the role of BDNF in osteoblast function and therefore in fracture healing.

287

In rodent models (mouse and rat), the distribution patterns of neurotrophins and their receptors were analysed during fracture healing. BDNF as well as other neurotrophins and its receptors were detected in the fracture callus (Asaumi et al., 2000; Grills and Schuijers, 1998). To the best of our knowledge no information is available to date on neurotrophins and their receptors in the process of human fracture healing. Therefore, in this study, we analysed the expression patterns of BDNF and its receptor TrkB at different time points in bone repair of patients undergoing osteosynthesis after bone fractures.

2. Materials and methods 2.1. Tissue samples Permission for the biopsies was obtained from the Ethics committee of the medical Faculty of the Justus-Liebig-University Giessen (138/07). Mature adult bone samples from the iliac crest were obtained from patients (n = 6) undergoing surgical treatment with autologous bone grafting (Table 1). Biopsies from patients with osteosynthesis after fractures at different stages of fracture healing (initial fibrin matrix, granulation tissue, woven bone) were taken from the fracture gaps (n = 6 per group, approximate size of 5 mm3 ). Samples for RNA extraction and RT-PCR were snap frozen in liquid nitrogen and stored at −80 ◦ C or fixed in RNA stabilizer reagent (RNAlater, Ambion Ltd., Huntingdon, UK). For immunohistochemistry, samples of fracture healing and spongeious bone were fixed in 4% PFA for 24 h. After several washes in PBS the samples were decalcified in 10% EDTA (Carl Roth GmbH, Karlsruhe, Germany), embedded in paraffin and cut into 5 ␮m sections with a Leica RM2155 microtome (Leica, Wetzlar, Germany). Positive control groups for BDNF and TrkB detection were prepared from biopsies of human astrocytoma with different grading, including polycystic astrocytoma grade I, astrocytoma stage II and fibril astrocytoma grade III (Wadhwa et al., 2003; Wang et al., 1998; Yamamoto et al., 1996). Assimakopoulou et al. (2007), Chiaretti et al. (2004) showed components of neurotrophin signalling pathways are expressed in astrocytoma tissue. The neurotrophins are structurally related proteins as well as their receptors (Trk) therefore the antibodies for detection of these proteins in astroytoma tissue should be highly specific.

2.2. HUVEC Human umbilical vein endothelial cells (HUVEC) were isolated from human umbilical cords and cultured according to Peters et al. (2005). The harvested cells were cultured in PromoCellTM endothelial cell growth medium (PromoCell, Heidelberg, Germany) supplemented with 10 vol/vol% foetal calf serum (FCS), 100 U/ml penicillin, 100 ␮g/ml streptomycin and Supplement MixTM (final concentration of the components: 0.4% endothelial growth supplement with heparin, 0.1 ng/ml human EGF, 1.0 ng/ml bFGF, 1.0 ␮g/ml hydrocortisone). Confluent cultures of primary endothelial cells were trypsinized in phosphate-buffered saline (PBS, composition in mM: 137 NaCl, 2.7 KCl, 1.5 KH2 PO4 and 8.0 Na2 HPO4 , at pH 7.4, supplemented with 0.05 wt/vol% trypsin and 0.02 wt/vol% EDTA) and seeded at a density of 2.2 × 104 cells/cm2 on 35-mm culture dishes or chamber slides. Confluent endothelial monolayers of passage 1 were used for RNA isolation or immunocytochemistry 3–4 days after seeding. For immunocytochemistry, cells were fixed with 4 wt/vol% paraformaldehyde (PFA) in PBS for 10 min at RT followed by three washing steps with PBS.

288

O. Kilian et al. / Annals of Anatomy 196 (2014) 286–295

Table 1 Immunohistochemical distribution and mRNA detection of BDNF and TrkB at different stages of fracture healing. Fracture localization/donor site

BDNF mRNA

TrkB mRNA

BDNF protein

TrkB protein

2 3 2 4 3 3

Distal humerus Distal tibia Distal femur Proximal humerus Proximal humerus Distal fibula

+ + + + + +

+ + + + + +

2 2 1 2 2 2

2 2 2 2 2 2

Female Female Female Male Male Male

7 6 6 8 6 7

Calcaneus Proximal tibia Proximal humerus Proximal humerus Clavicle Proximal humerus

+ + + + + +

+ + + + + +

3 2 3 3 2 3

4 2 4 3 3 3

59 28 25 53 34 22

Female Male Male Male Male Male

12 11 12 10 13 13

Proximal humerus Distal tibia Calcaneus Calcaneus Distal tibia Distal tibia

+ + + + + +

+ + + + + +

2 1 2 3 2 3

3 3 2 3 3 3

55 38 66 60 35 65

Male Male Female Male Male Female

– – – – – –

Iliac crest Iliac crest Iliac crest Iliac crest Iliac crest Iliac crest

0 0 0 0 0 0

+ + + + + +

0 0 0 0 0 0

1 2 1 1 1 1

Group

No.

Initial

Age

Gender

1

1 2 3 4 5 6

A.Z. M.R. A.S. W.R. R.D. B.H.

27 43 40 68 19 47

Male Male Female Male Male Male

2

1 2 3 4 5 6

B.C. G.M. G.A. H.S. Z.B. Z.S.

47 53 35 58 65 20

3

1 2 3 4 5 6

S.S. L.S. S.S. F.W. M.M. D.M.

4

1 2 3 4 5 6

W.F. W.S. E.F. O.K. P.F. T.F.

Time after fracturing (day)

Group 1: fracture haematoma/fibrin matrix, group 2: granulation tissue, group 3: osteoid/woven bone, group 4: mature bone/spongiosa. PCR level: 0 = no detection, + = positive detected. Immunohistochemical staining level: 0 = no staining, 1 = focal, 2 = faint, 3 = moderate, and 4 = intense staining.

2.3. Human osteoblastic cell line The osteoblastic cell line HOS-58 from a 23-years-old male was established and characterized by the Institute of Pathology, University Giessen. Cells (Siggelkow et al., 1998) were maintained in DMEM medium (Invitrogen) supplemented with 5 vol/vol% foetal calf serum (FCS, Invitrogen), 100 U/ml of penicillin and 100 ␮g/ml of streptomycin in a humidified atmosphere of 95% air, and 5% CO2 at 37 ◦ C. Cells were cultivated at a starting density of 104 cells/cm2 . Culture medium was replaced every third day. Cells were examined dependent upon the specific growth curve of each cell line (beginning of exponential growth phase after 54 h on average). For immunocytochemistry, the cells were fixed within the chamber slides with 4% PFA for 10 min at room temperature followed by three washing steps with PBS. 2.4. TrkB Western blot analysis Snap-frozen astrocytoma (Grade III) (Siggelkow et al., 1998; Wadhwa et al., 2003; Wang et al., 1998; Yamamoto et al., 1996) was lysed and boiled in Laemmli sample buffer (Laemmli, 1970). The protein content was determined using the Micro BCATM protein assay kit (Pierce, Rockford, IL, USA). SDS/PAGE was carried out using an 8% polyacrylamide gel according to Laemmli (Laemmli, 1970). Sixteen ␮g protein of different samples were loaded and subjected to electrophoretic separation alongside a Precision Plus protein dual-colour standard molecular mass marker (BioRad Laboratories, Munich, Germany). Proteins were transferred electrophoretically onto ImmobilonTM -P polyvinylidene difluoride membranes (Millipore, Bedford, MA, USA). After blocking with 1x Roti® -Block (Carl Roth, Karlsruhe, Germany) in PBS (pH 7.4) the blots were incubated with the polyclonal rabbit anti-TrkB antibody (Santa Cruz, Heidelberg, Germany), diluted 1:4.000 in PBS containing 1× Roti® -Block. Blots were washed in PBS containing 0.1% Tween 20 and bound primary rabbit antibodies were detected by peroxidase-conjugated goat anti-rabbit IGs (Dako, Glostrup, Denmark), diluted 1:5000 in PBS containing 2.5% skimmed milk

powder and 0.1% Tween 20. Bound peroxidase was visualized by Lumi-Light PLUS Western Blotting Substrate (Roche Diagnostics, Mannheim, Germany). 2.5. BDNF Western blot analysis Two g of tissue from each astrocytoma (3 different biopsies) were homogenized with a Potter-Elvehjem homogenizer (B. Braun Biotech International, Melsungen, Germany) at 1000 rpm (1 stroke, 60 s) in 2 ml ice-cold homogenization buffer (HMB, pH 7.4: 0.25 M sucrose and 5 mM MOPS, 1 mM EDTA, 0.1% ethanol, 0.2 mM DTT, 1 mM aminocaproic acid and 100 ␮l cocktail of protease inhibitors # 39102, Serva, Germany). The quality of the homogenization process was controlled through trypan blue staining of 5 ␮l of the tissue homogenates under a light microscope. Clumps of tissue were separated by centrifugation of the homogenates at 100 × g for 10 min at 4 ◦ C. The supernatant was processed further through centrifugation at 2.500 × g for 10 min at 4 ◦ C to pellet large organelles. The 2.500 g supernatant, containing small organelles and cytosolic proteins, was used for Western blot analysis. Protein concentrations of different samples were determined using the Bradford assay (Bradford, 1976) with Bio-Rad solutions (Bio-Rad, Munich, Germany) according to the manufacturer’s instructions. SDS-PAGE was performed using a Bio-Rad Trans-Blot SD electrophoresis apparatus. One hundred ␮g each of different 2500 g – supernatant fractions were separated by SDS-PAGE on a 15% resolving gel. As molecular weight markers, 5 ␮l of colour-stained and 5 ␮l of biotinylated Precision Markers from Bio-Rad were used. Protein transfer was done by semi-dry blotting onto polyvinylidene difluoride membranes (PVDF) with a semi-dry Trans-Blot SD apparatus (parameters: 55 min, 120 mA). The non-specific protein binding sites on the PVDF membrane were blocked overnight at 4 ◦ C in blocking solution consisting of 10% fat-free milk powder in TBST (Bio-Rad, München, Germany). The blocked PVDF membrane was incubated with the primary antibody against rabbit anti-human BDNF (Santa Cruz, Heidelberg, Germany), diluted to 1:100 in 5% blocking solution (10 mM Tris/HCl, 0.15 M NaCl, 0.05%

O. Kilian et al. / Annals of Anatomy 196 (2014) 286–295

Tween 20. pH 8.0) for 1 h (h) at room temperature on a rotor shaker. After 3 times washing for 10 min with TBST to remove excess primary antibody the blot was incubated for 1 h with an alkaline phosphatase-conjugated goat anti-rabbit secondary antibody. After a final washing (3 × 10 min) in TBST immunoreactive bands were visualized with the ImmunostarTM – AP detection kit from BioRad with a chemiluminescent substrate for alkaline phosphatase. BioMax MR-films (Kodak, Stuttgart, Germany) were exposed for ca. 5 min to visualize BDNF immunoreactive bands. The biotinylated marker proteins on membranes were visualized through incubation with an alkaline phosphatase-labelled streptavidin conjugate. 2.6. Isolation of RNA Total RNA was isolated out of 100 mg of tissue and confluently grown monolayers of HUVEC as well as the osteoblastic cell line HOS-58 cells in 10 cm2 culture flasks by the use of TRIzol® reagent (Invitrogen) according to the manufacturer’s protocol. The yield of total RNA was determined in a WPA Biowave spectrophotometer (Biochrom, Cambridge, England). 2.7. Reverse transcription One ␮g of RNA was treated with 1 U/ml of DNase I (Invitrogen) at 25 ◦ C for 15 min in 10 ␮l reaction volume prior to reverse transcription. Reverse transcription was carried out in an iCycler (Bio-Rad Laboratories, Munich, Germany) as follows: 330 ng of DNase Itreated RNA were incubated with a final concentration of 50 U/ml MultiscribeTM Reverse Transcriptase, 1× RT-Buffer, 5 mM MgCl2 , 2.5 mM random hexamers, 1 mM dNTPs (all Applied Biosystems, Darmstadt, Germany) and 1 U/ml RNase Inhibitor (Roche) in a volume of 20 ␮l at 25 ◦ C for 5 min followed by 42 ◦ C for 45 min. The reverse transcriptase was inactivated through incubation at 95 ◦ C for 5 min. A negative control of every RNA preparation was generated by omitting the reverse transcriptase in the RT-approach. 2.8. PCR reaction Commercially synthesized primers (Eurofins MWG Operon, Munich, Germany) used to amplify specific mRNA-transcripts are described in Table 2. As positive control the housekeeping gene GapDH (glyceraldehyde-3-phosphate dehydrogenase) was amplified. All PCRs were performed in an iCycler (Bio-Rad). A volume of 2 ␮l of each transcribed cDNA-preparation and 1 U/ml of HotMaster Taq DNA Polymerase (5Prime GmbH, Hamburg, Germany) was used for the PCR reactions. Amplicons were generated with an initial denaturation step (95 ◦ C, 2 min) and 37 cycles of denaturation at 95 ◦ C for 20 s, annealing at the specific annealing temperature for 20 s (see Table 2), extension at 65 ◦ C for 25 s and followed by a final extension step of 7 min at 65 ◦ C. The PCR products were analysed through gel electrophoresis in a 2% agarose gel stained with 1× SYBR® Green I (Sigma–Aldrich, Hamburg, Germany). As negative controls 2 ␮l of the RT-approaches without reverse transcriptase were used for PCRs of distinct samples. 2.9. Immunohistochemical staining (single staining) After deparaffinization of tissue sections the endogenous peroxidase was blocked through incubation in 3% H2 O2 for 5 min at RT. After intensive washing the samples were incubated with the primary antibody (rabbit anti-human BDNF, diluted 1:100; rabbit anti-TrkB, diluted 1:100; both from Santa Cruz, Heidelberg, Germany) in antibody diluent with background reducing components (Dako, Glostrup, Denmark) overnight at 4 ◦ C. After washing,

289

the samples were incubated with a biotinylated secondary goat anti-rabbit antibody (Dako, diluted 1:800) in 3% FCS and 12% human serum for 30 min at RT. After washing, the biotinylated secondary antibody was detected with horseradish peroxidase labelled streptavidin (Dako) for 30 min. The chromogen Nova Red (Vector Laboratories, Burlingame, CA, USA) was used for visualization of the peroxidase activity. Nuclei were counterstained with haematoxylin (Shandon Scientific Ltd., Cheshire, UK). For semi-quantitative grading of the amount of stained cells and its labelling intensity a score of 0–4 was used which was determined as follows: 0 = no staining, 1 = faint staining of some only focal situated cells, 2 = faint staining of approximately 30% of the cells, 3 = moderate staining of approximately 50% of the cells, and 4 = intense staining of approximately 70% of the cells found in the fracture gap tissue.

2.10. Immunohistochemical staining (double staining) After deparaffinization, the slides were incubated in 1% H2 O2 for 30 min at RT to block endogenous peroxidase. After washing in PBS, pH 7.2, the sections were incubated for 30 min in PBS with 1% bovine serum albumin (BSA, Serva, Electrophoresis GmbH, Heidelberg, Germany) and 0.1% NaN3 (Merck, Darmstadt, Germany), followed by overnight incubation with a 1:50 dilution of the rabbit anti-human BDNF antibody at 4 ◦ C diluted in the same solution. Bound primary antibodies were detected using the rabbit EnVision peroxidase system according to the manufacturer’s recommendation (Dako) as well as 3,3 -diaminobenzidine (DAB, Sigma–Aldrich). After washing with PBS the slides were preincubated in the BSA/NaN3 solution, followed by incubation with a 1:100 dilution of the rabbit anti-human TrkB antibody at 4 ◦ C diluted in the same solution. Bound anti-TrkB antibodies were detected using the anti-mouse EnVision alkaline phosphatase system according to the manufacturer’s recommendation (Dako). Fast Blue (Sigma–Aldrich) was used as chromogen in a solution containing 2 mM Levamisol to inhibit the endogenous alkaline phosphatase activity.

3. Results In Table 1, an overview was given over mRNA and immunohistochemical findings of BDNF and TrkB in different tissue. In physiologic bone remodelling, there was no evidence of mRNA transcription of BDNF, whereas BDNF mRNA was detectable in initial fibrin matrix, in fracture healing tissue (osteoid phase), cultured human endothelial cells (HUVEC), in cultured osteoblasts, and in astrocytoma tissue. TrkB mRNA has been detected in physiologic bone remodelling, in initial fibrin matrix, in osteoid tissue, in endothelial cells, and osteoblasts culture (Fig. 1). In astrocytoma tissue, which served as a positive control, BDNF (Fig. 2A) und TrkB (Fig. 2C) expression were demonstrated by immunohistochemistry. The specificity of the BDNF and TrkB antibodies used were proven by Western blot analysis of astrocytoma tissue (Fig. 2B and D). In human endothelial cell cultures, BDNF had a cytoplasmatic distribution (Fig. 3A), whereas TrkB showed a spot-like pattern and was enriched in areas near the nucleus most probably being the Golgi apparatus (Fig. 3B). In an osteoblast culture, i.e. the HOS-58 cell line BDNF, an intensive staining of BDNF (Fig. 3D) and TrkB (Fig. 3E) was seen which was reticular located in cytoplasm. In mature human bone tissue, BDNF protein could not be detected by immunohistochemistry (Fig. 4A), whereas TrkB antibodies labelled blood vessels (Fig. 4B).

290

O. Kilian et al. / Annals of Anatomy 196 (2014) 286–295

Table 2 Primers used for RT-PCR. Primer sequences

Acc. No.

Fragment length

Annealing temp.

GapDH Forward Reverse

5 -tggtatcgtggaaggactcatga-3 5 -atgccagtgagcttcccgttca-3

NM 002046

189 bp

60 ◦ C

BDNF Forward Reverse

5 -atagaagagctgttggatgag-3 5 -ctgcagccttcttttgtgtaa-3

NM 170731

392 bp

53 ◦ C

TrkB Forward Reverse

5 -tggtgcattccattcactgt-3 5 -cgtggtactccgtgtgattg-3

NM 006180

130 bp

56 ◦ C

Fig. 1. RT-PCR detection of BDNF (392 bp), TrkB (130 bp), and housekeeping gene GapDH (189 bp). BDNF and TrkB mRNA was determined in fracture healing tissue of initial fibrin matrix (lane 2), osteoid formation phase (lane 3), HUVEC (lane 4), and HOS-58 cells (lane 5). In spongious bone (lane 1) only TrkB but no BDNF mRNA is expressed. Astrocytoma tissue (lane 6) was used as positive control. The housekeeping gene GapDH was expressed in every sample. M: 100 bp marker.

Fig. 2. Detection of BDNF and TrkB in human astrocytoma tissue. BDNF (A) and TrkB (C) labelling was found in vessels (arrow) and astrocytoma cells (arrow head). Negative controls without incubation of primary antibody showed no staining (E). Western blot analyses of BDNF (B) and TrkB (D) determined specific bands of an approximately molecular mass of 20 kDa and 95 kDa. M: molecular mass marker. Scale bars correspond to 50 ␮m.

During the initial phase of fracture healing, i.e. haematoma and fibrin matrix formation, only hematopoietic cells such as granulocytes stained positively for BDNF (Fig. 5A) and TrkB (Fig. 5B). In the granulation tissue of fracture healing BDNF was detected only sporadically in vessel walls (Fig. 5D), while TrkB protein (Fig. 5E) appeared with intensive staining in the endothelium of newly formed vessels. Stromal fibroblasts were also BDNF- and TrkB-positive. In the osteoid phase and during woven bone formation of fracture healing, osteoblasts at the osteoid rim of newly formed bone were BDNF- (Fig. 5G) and TrkB-positive (Fig. 5H). No chondrocytes were labelled through antibodies against BDNF (Fig. 5L) or TrkB (Fig. 5M) whereas some lining cells in samples with strong TrkB labelling intensity showed a faint anti-TrkB immunoreaction (Fig. 5O).

In double staining, active osteoblasts at the osteoid rim seem to express BDNF and TrkB simultaneously (Fig. 6A and B), while vessel walls primarily showed the BDNF protein (Fig. 6C). 4. Discussion In our study, we investigated mRNA expression and protein distribution of BDNF and its specific receptor TrkB in tissue samples obtained from fracture gaps at different phases of human diaphyseal fracture healing. BDNF and TrkB expression were determined in hematopoietic cells in early fracture hematomas, in fibroblastlike cells in granulation tissue and in active osteoblasts in later stages of fracture healing (enchondral ossification, osteoid formation and woven bone formation). Similar to our study, Asaumi et al. showed the presence of BDNF in osteoblast-like cells in the vicinity

O. Kilian et al. / Annals of Anatomy 196 (2014) 286–295

291

Fig. 3. Immunohistochemistry of cultured cells. BDNF in HUVEC (A) showed a mainly cytoplamic staining, whereas TrkB (C) could be found in distinct areas corresponding to the golgi apparatus of HUVEC. In the osteoblastic cell-line HOS-58 BDNF (D) and TrkB (E) were localized in the cytoplasm. Negative controls where primary antibody was omitted (C: HUVEC, F: HOS-58) showed no staining. Nuclei were counterstained with haematoxylin. Scale bar corresponds to 50 ␮m in A–F and 20 ␮m in inset of B.

Fig. 4. Immunohistochemistry of BDNF and TrkB in physiological bone remodelling. Whereas BDNF (A) could not be detected, TrkB (B) was found in vessels (arrow) of healthy bone. C: negative control. Scale bars: 50 ␮m.

292

O. Kilian et al. / Annals of Anatomy 196 (2014) 286–295

Fig. 5. Immunohistochemistry of human fracture healing. BDNF (A) and TrkB (B) labelled hematopoetic cells, monocytes and granulocytes (grey arrow) in initial fibrin matrix phase of fracture healing. In granulation tissue within the fracture gap BDNF (D) and TrkB (E) were detected in newly formed vessels (white arrows) and in spindle-shaped cells, most probably fibroblasts or osteoprogenitor cells. In the woven bone formation phase of fracture healing BDNF (G) and TrkB (H) occurred in osteoblasts (thick arrows). Additionally TrkB labelled endothelial cells (thin arrow). Chondrocytes lacked BDNF (L) and TrkB (M) labelling. Lining cells showed no staining for BDNF (N) whereas samples with strong TrkB staining intensity within the osteoid phase revealed a faint TrkB immunoreaction (O). Negative controls (primary antibody was omitted) in various stages of fracture healing showed no labelling (C, F and K). Scale bars: 50 ␮m.

O. Kilian et al. / Annals of Anatomy 196 (2014) 286–295

293

Fig. 6. Double-immunohistochemistry of BDNF (brown staining) and TrkB (blue staining). BDNF (thick arrow) and TrkB (arrow heads) were both localised in osteoblast (A and B) and newly formed vessels (D) in the later phase of fracture healing. No staining was detected in the nuclei (thin arrows) and the negative controls (C: osteoblasts, E: blood vessels) or woven bone (*). Scale bars: 50 ␮m.

of enchondral ossification and osteoid formation in a mice model of fractured rib bone. In contrast to our investigations of human samples Asaumi et al. were unable to detect expression of TrkB receptor in fracture callus between days 2 and 20 in the mouse model (Asaumi et al., 2000). In our study TrkB and BDNF revealed a similar expression pattern during fracture healing. Thus we guess that BDNF binds to TrkB in the fracture gap tissue. In addition to the fracture gap tissue we detected BDNF as well as TrkB in osteoblasts in vitro. The main duties of osteoblast are synthesis of the collagen bone matrix and initiation of its mineralisation. Besides bone, teeth are also built up by mineralised tissue that is generated amongst others by cementoblasts. BDNF enhances the expression of bone/cementum-related proteins (e.g. alkaline phosphatase, osteopontin, bone morphogenetic protein-2) after application in vitro (Kajiya et al., 2008). Furthermore, evidence has been provided for the involvement of the BDNF-TrkB-cRaF-ERK1/2Elk-1 signalling pathway by performing gene silencing experiments with small interfering RNA (siRNA) for TrkB (Kajiya et al., 2008; Yamashiro et al., 2001). For osteoblasts, it has already been shown that the ERK1/2 cascade promotes osteoblast differentiation and enhances expression of bone matrix proteins (Xiao et al., 2000). Stimulated synthesis of bone matrix proteins (e.g. collagen) in vitro as well as regeneration of alveolar bone of dogs in vivo were determined after application of BDNF in a dose dependent manner and in combination with a synthetic scaffold providing hyaluronic acid (Jimbo et al., 2014; Kurihara et al., 2003; Takeda et al., 2005). Besides BDNF, other members of the neurotrophin family were also found in fracture callus tissue by means of immunohistochemistry (Asaumi et al., 2000; Grills and Schuijers, 1998). Application of NGF increased fracture repair as well as the catecholamine concentration in a rat rib fracture model (Grills et al., 1997). Catecholamines transmitted from the sympathetic nervous system were associated with increased bone reabsorption and substantial bone loss (Yirmiya et al., 2006). In fracture healing, bone resorbing osteoclasts and forming osteoblasts are interacted by mutual stimulation. Elevated levels of nerve growth factor (NGF) during enchondral ossification seem to be correlated with an increased amount of nerve fibres and hyperalgesia (Yasui et al., 2012). In our study, we could not observe BDNF in nerve fibres in the fracture gap tissue. BDNF plasma levels were enhanced in patients with knee osteoarthritis compared to healthy controls whereas BDNF concentration in synovial fluid was not changed (Simao et al., 2014). Thus it was suggested that systemic but not local BDNF levels might be associated with modulation of pain and joint inflammatory process (Simao et al., 2014). BDNF and TrkB are expressed by immune cells and regulated during inflammation (Loza et al., 2012; Maroder

et al., 1996; Skaper et al., 2012). In central nervous system injury neurotrophins act as anti-inflammatory neuroprotective factors (Tabakman et al., 2004) whereas others reported an up-regulation of BDNF and TrkB in dorsal root ganglia during inflammatory pain (Lin et al., 2011). In bone regeneration, inflammatory processes are necessary to start the initial stage of healing. Thus, an up-regulation of BDNF and TrkB during the early phase of fracture healing would improve bony repair. However, this needs to be proven in future analysis. In our study, BDNF and TrkB occurred in human osteoblasts in vitro. Previous studies demonstrated the expression of neurotrophins and their receptors in murine osteoblasts in vitro (Nakanishi et al., 1994) and discussed their role as survival factor. Since the anti-apoptotic effect of BDNF has already been confirmed for cementoblasts (Kajiya et al., 2009), we suppose that BDNF also act as survival factor for osteoblasts and perhaps also for endothelial cells. Considerable studies determined the expression of BDNF and TrkB receptors in endothelial cell cultures where one of their functions might be the stabilization of newly formed vessels (Donovan et al., 2000; Wagner et al., 2005). This role of BDNF is supported by its stimulating effect on endothelial cell–cell contacts (Hu et al., 2005; Sun et al., 2006). In our study we determined the expression of BDNF and TrkB in endothelial cell cultures as well as in vessels in the granulations tissue of physiological fracture healing. Recently, an intracellular angiogenesis signalling cascade downstream of BDNF was identified (Usui et al., 2014). It is well known that BDNF binding to TrkB leads to receptor dimerization and phosphorylation. This is associated with early activation of intracellular signalling cascades. One of the BDNF signalling pathways involves downstream of TrkB the activation of gp91phox/(NOX2)-dependent NADPH oxidase subsequently followed by ROS generation and finally stimulation of the PI3kinase/Akt pathway (Kim et al., 2004; Ushio-Fukai, 2007; Usui et al., 2014). This pathway promotes migration of endothelial cells which is one of the necessary steps in the process of capillary sprouting. Evidence for the pro-angiogenic effect of BDNF was also given by Sun et al. who stimulated HUVEC cultures with BDNF (25–400 ng/ml) and measured a dose-dependent increases of matrix metalloproteinase (MMP)2 and MMP9 mRNA transcription (Sun et al., 2006). MMPs are pro-angiogenetic factors controlling vascular tube formation. The same authors reported a timedependent effect of BDNF on the expression and activity of plasminogen activator inhibitor-1 (PAI-1) and urokinase-type plasminogen activator (uPA) in HUVEC cultures (Sun et al., 2006). These

294

O. Kilian et al. / Annals of Anatomy 196 (2014) 286–295

results suggest that BDNF affects microvascularisation and sprouting or intussusceptive angiogenesis. By linking to TrkB, BDNF also stimulates the migration of the smooth muscle cells which contribute to vascular differentiation. Both, Donovan et al. and Wagner et al. demonstrated that in BDNF deficient mice TrkB expression was inhibited in endothelial cells. Furthermore the lack of BDNF impairs endothelial cell survival, leads to declined differentiation and stabilization of newly formed vessels (Donovan et al., 2000; Wagner et al., 2005). In our study we determined scattered BDNF stained vessels in the inflammatory phase of fracture healing when granulation tissue replaces the fibrin network in the fracture gap. Not yet fully differentiated vessels in the tissue matrix were strikingly positive for BDNF and TrkB, while vessels in later stages of fracture healing did not stain for BDNF. No evidence of BDNF expression was found in the spongy bone during physiological remodelling. This is in accordance with the studies published by Labouyrie et al. where no BDNF was expressed in the spongiosa during physiological bone remodelling (Labouyrie et al., 1999). In our study BDNF and TrkB did not appear in chondroblasts in the fracture callus. Yamashiro et al. found BDNF and TrkB in chondroblasts in the mature zone of the epiphyseal growth plate of Wistar rats as well as in the vicinity of the osteoid rim (Yamashiro et al., 2001). The presence of BDNF in osteoblast-like cells in the vicinity of enchondral ossification and osteoid formation was similar to our study with human samples. However, we did not investigate the epiphyseal growth plates which are more important for bone length than for bone healing. Evidence for the involvement of BDNF in epiphyseal growth plate function was reported recently (Camerino et al., 2012). Camerino et al. (2012) showed that depletion of central nervous BDNF leads to an increase in femur length, bone mineral density, and content as well as white adipose tissue as shown in conditional BDNF knockout mice which leads to the suggestion that neuronal BDNF exerts influence on bone remodelling. In conclusion, our investigation reveals BDNF and TrkB expression in the active tissue during fracture healing. In addition, BDNF and its receptor appear in fracture haematoma, fibroblast-like cells in granulation tissue and in osteoblasts that are in charge of matrix production. BDNF was observed only during formation processes of new vessels, whereas TrkB also appeared in mature vessels in physiological bone remodelling. In healthy mature bone tissue BDNF is absent. Thus BDNF and TrkB were up-regulated in tissue forming processes during human fracture healing. To our knowledge, this is the first time that BDNF and its receptor TrkB were demonstrated in human fracture gap tissue with respect to the healing phases. The study reveals that BDNF and TrkB are involved in matrix reorganisation. Further investigations will be required to determine the precise role of BDNF and TrkB in human fracture healing.

Acknowledgements This study was supported by the German Research Foundation (Transregional Collaborative Research Centre SFB/TRR 79). The authors thank A. Hild and I. Schütz for excellent technical assistance and G. Fuchs-Moll and S. Wilker (Laboratory of Experimental Surgery, Department of General and Thoracic Surgery, University of Giessen) for performing double immunohistochemical staining.

Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at http://dx.doi.org/10.1016/j.plantsci. 2004.08.011.

References Asaumi, K., Nakanishi, T., Asahara, H., Inoue, H., Takigawa, M., 2000. Expression of neurotrophins and their receptors (TRK) during fracture healing. Bone 26, 625–633. Assimakopoulou, M., Kondyli, M., Gatzounis, G., Maraziotis, T., Varakis, J., 2007. Neurotrophin receptors expression and JNK pathway activation in human astrocytomas. BMC Cancer 7, 202. Au, C.W., Siu, M.K., Liao, X., Wong, E.S., Ngan, H.Y., Tam, K.F., Chan, D.C., Chan, Q.K., Cheung, A.N., 2009. Tyrosine kinase B receptor and BDNF expression in ovarian cancers – effect on cell migration, angiogenesis and clinical outcome. Cancer Lett. 281, 151–161. Beckmann, R., Tohidnezhad, M., Lichte, P., Wruck, C.J., Jahr, H., Pape, H.C., Pufe, T., 2014. New from old: relevant factors for fracture healing in aging bone. Orthopade 43, 298–305. Bradford, M.M., 1976. Rapid and sensitive method for quantitation of microgram quantities of protein utilizing principle of protein-dye binding. Anal. Biochem. 72, 248–254. Camerino, C., Zayzafoon, M., Rymaszewski, M., Heiny, J., Rios, M., Hauschka, P.V., 2012. Central depletion of brain-derived neurotrophic factor in mice results in high bone mass and metabolic phenotype. Endocrinology 153, 5394–5405. Cartwright, M., Mikheev, A.M., Heinrich, G., 1994. Expression of neurotrophin genes in human fibroblasts: differential regulation of the brain-derived neurotrophic factor gene. Int. J. Dev. Neurosci. 12, 685–693. Chiaretti, A., Aloe, L., Antonelli, A., Ruggiero, A., Piastra, M., Riccardi, R., Tamburrini, G., Di Rocco, C., 2004. Neurotrophic factor expression in childhood low-grade astrocytomas and ependymomas. Childs Nerv. Syst. 20, 412–419. Claes, L., Recknagel, S., Ignatius, A., 2012. Fracture healing under healthy and inflammatory conditions. Nat. Rev. Rheumatol. 8, 133–143. Donovan, M.J., Lin, M.I., Wiegn, P., Ringstedt, T., Kraemer, R., Hahn, R., Wang, S., Ibanez, C.F., Rafii, S., Hempstead, B.L., 2000. Brain derived neurotrophic factor is an endothelial cell survival factor required for intramyocardial vessel stabilization. Development 127, 4531–4540. Geris, L., Gerisch, A., Sloten, J.V., Weiner, R., Oosterwyck, H.V., 2008. Angiogenesis in bone fracture healing: a bioregulatory model. J. Theor. Biol. 251, 137–158. Giannoudis, P., Tzioupis, C., Almalki, T., Buckley, R., 2007. Fracture healing in osteoporotic fractures: is it really different? A basic science perspective. Injury 38 (Suppl. 1), S90–S99. Grills, B.L., Schuijers, J.A., 1998. Immunohistochemical localization of nerve growth factor in fractured and unfractured rat bone. Acta Orthop. Scand. 69, 415–419. Grills, B.L., Schuijers, J.A., Ward, A.R., 1997. Topical application of nerve growth factor improves fracture healing in rats. J. Orthop. Res. 15, 235–242. Hallbook, F., Wilson, K., Thorndyke, M., Olinski, R.P., 2006. Formation and evolution of the chordate neurotrophin and Trk receptor genes. Brain Behav. Evol. 68, 133–144. Hohn, A., Leibrock, J., Bailey, K., Barde, Y.A., 1990. Identification and characterization of a novel member of the nerve growth factor/brain-derived neurotrophic factor family. Nature 344, 339–341. Hu, Y., Sun, C.Y., Wang, Y.D., Wei, W.N., Wu, T., He, W.J., Zhao, S., 2005. Study on the high expression of brain-derived neurotrophic factor in multiple myeloma patients and its possible mechanism. Zhongguo shi yan xue ye xue za zhi/Zhongguo bing li sheng li xue hui 13, 104–109. Huang, E.J., Reichardt, L.F., 2001. Neurotrophins: roles in neuronal development and function. Annu. Rev. Neurosci. 24, 677–736. Jimbo, R., Tovar, N., Janal, M.N., Mousa, R., Marin, C., Yoo, D., Teixeira, H.S., Anchieta, R.B., Bonfante, E.A., Konishi, A., Takeda, K., Kurihara, H., Coelho, P.G., 2014. The effect of brain-derived neurotrophic factor on periodontal furcation defects. PLOS ONE 9, e84845. Kajiya, M., Shiba, H., Fujita, T., Ouhara, K., Takeda, K., Mizuno, N., Kawaguchi, H., Kitagawa, M., Takata, T., Tsuji, K., Kurihara, H., 2008. Brain-derived neurotrophic factor stimulates bone/cementum-related protein gene expression in cementoblasts. J. Biol. Chem. 283, 16259–16267. Kajiya, M., Shiba, H., Fujita, T., Takeda, K., Uchida, Y., Kawaguchi, H., Kitagawa, M., Takata, T., Kurihara, H., 2009. Brain-derived neurotrophic factor protects cementoblasts from serum starvation-induced cell death. J. Cell. Physiol. 221, 696–706. Kaplan, D.R., Miller, F.D., 2000. Neurotrophin signal transduction in the nervous system. Curr. Opin. Neurobiol. 10, 381–391. Keramaris, N.C., Calori, G.M., Nikolaou, V.S., Schemitsch, E.H., Giannoudis, P.V., 2008. Fracture vascularity and bone healing: a systematic review of the role of VEGF. Injury 39 (Suppl. 2), S45–S57. Kermani, P., Rafii, D., Jin, D.K., Whitlock, P., Schaffer, W., Chiang, A., Vincent, L., Friedrich, M., Shido, K., Hackett, N.R., Crystal, R.G., Rafii, S., Hempstead, B.L., 2005. Neurotrophins promote revascularization by local recruitment of TrkB+ endothelial cells and systemic mobilization of hematopoietic progenitors. J. Clin. Invest. 115, 653–663. Kim, H., Li, Q., Hempstead, B.L., Madri, J.A., 2004. Paracrine and autocrine functions of brain-derived neurotrophic factor (BDNF) and nerve growth factor (NGF) in brain-derived endothelial cells. J. Biol. Chem. 279, 33538–33546. Kurihara, H., Shinohara, H., Yoshino, H., Takeda, K., Shiba, H., 2003. Neurotrophins in cultured cells from periodontal tissues. J. Periodontol. 74, 76–84. Labouyrie, E., Dubus, P., Groppi, A., Mahon, F.X., Ferrer, J., Parrens, M., Reiffers, J., de Mascarel, A., Merlio, J.P., 1999. Expression of neurotrophins and their receptors in human bone marrow. Am. J. Pathol. 154, 405–415. Laemmli, U.K., 1970. Cleavage of structural proteins during assembly of head of bacteriophage-T4. Nature 227, 680-&.

O. Kilian et al. / Annals of Anatomy 196 (2014) 286–295 Lin, Y.T., Ro, L.S., Wang, H.L., Chen, J.C., 2011. Up-regulation of dorsal root ganglia BDNF and TrkB receptor in inflammatory pain: an in vivo and in vitro study. J. Neuroinflammation 8, 126. Long, B.L., Rekhi, R., Abrego, A., Jung, J., Qutub, A.A., 2013. Cells as state machines: cell behavior patterns arise during capillary formation as a function of BDNF and VEGF. J. Theor. Biol. 326, 43–57. Loza, M.J., Watt, R., Baribaud, F., Barnathan, E.S., Rennard, S.I., 2012. Systemic inflammatory profile and response to anti-tumor necrosis factor therapy in chronic obstructive pulmonary disease. Respir. Res. 13, 12. Maroder, M., Bellavia, D., Meco, D., Napolitano, M., Stigliano, A., Alesse, E., Vacca, A., Giannini, G., Frati, L., Gulino, A., Screpanti, I., 1996. Expression of TrkB neurotrophin receptor during T cell development. Role of brain derived neurotrophic factor in immature thymocyte survival. J. Immunol. 157, 2864–2872. Nakanishi, T., Takahashi, K., Aoki, C., Nishikawa, K., Hattori, T., Taniguchi, S., 1994. Expression of nerve growth factor family neurotrophins in a mouse osteoblastic cell line. Biochem. Biophys. Res. Commun. 198, 891–897. Peters, S.C., Reis, A., Noll, T., 2005. Preparation of endothelial cells from microand macrovascular origin. In: Dhein, S., Mohr, F.W., Delmar, M. (Eds.), Practical Methods in Cardiovascular Research. Springer, Berlin. Siggelkow, H., Niedhart, C., Kurre, W., Ihbe, A., Schulz, A., Atkinson, M.J., Hufner, M., 1998. In vitro differentiation potential of a new human osteosarcoma cell line (HOS 58). Differentiation 63, 81–91. Simao, A.P., Mendonca, V.A., de Oliveira Almeida, T.M., Santos, S.A., Gomes, W.F., Coimbra, C.C., Lacerda, A.C., 2014. Involvement of BDNF in knee osteoarthritis: the relationship with inflammation and clinical parameters. Rheumatol. Int., http://dx.doi.org/10.1007/s00296-013-2943-5 [Epub ahead of print], Print ISSN 0172-8172, Online ISSN 1437-160X. Skaper, S.D., Giusti, P., Facci, L., 2012. Microglia and mast cells: two tracks on the road to neuroinflammation. FASEB J. 26, 3103–3117. Sun, C.Y., Hu, Y., Wang, H.F., He, W.J., Wang, Y.D., Wu, T., 2006. Brain-derived neurotrophic factor inducing angiogenesis through modulation of matrix-degrading proteases. Chin. Med. J. (Engl.) 119, 589–595. Tabakman, R., Lecht, S., Sephanova, S., Arien-Zakay, H., Lazarovici, P., 2004. Interactions between the cells of the immune and nervous system: neurotrophins as neuroprotection mediators in CNS injury. Prog. Brain Res. 146, 387–401.

295

Takeda, K., Shiba, H., Mizuno, N., Hasegawa, N., Mouri, Y., Hirachi, A., Yoshino, H., Kawaguchi, H., Kurihara, H., 2005. Brain-derived neurotrophic factor enhances periodontal tissue regeneration. Tissue Eng. 11, 1618–1629. Ushio-Fukai, M., 2007. VEGF signaling through NADPH oxidase-derived ROS. Antioxid. Redox Signal. 9, 731–739. Usui, T., Naruo, A., Okada, M., Hayabe, Y., Yamawaki, H., 2014. Brain-derived neurotrophic factor promotes angiogenic tube formation through generation of oxidative stress in human vascular endothelial cells. Acta Phys.. Wadhwa, S., Nag, T.C., Jindal, A., Kushwaha, R., Mahapatra, A.K., Sarkar, C., 2003. Expression of the neurotrophin receptors Trk A and Trk B in adult human astrocytoma and glioblastoma. J. Biosci. 28, 181–188. Wagner, N., Wagner, K.D., Theres, H., Englert, C., Schedl, A., Scholz, H., 2005. Coronary vessel development requires activation of the TrkB neurotrophin receptor by the Wilms’ tumor transcription factor Wt1. Genes Dev. 19, 2631–2642. Wang, Y., Hagel, C., Hamel, W., Muller, S., Kluwe, L., Westphal, M., 1998. Trk A, B, and C are commonly expressed in human astrocytes and astrocytic gliomas but not by human oligodendrocytes and oligodendroglioma. Acta Neuropathol. (Berl.) 96, 357–364. Xiao, G., Jiang, D., Thomas, P., Benson, M.D., Guan, K., Karsenty, G., Franceschi, R.T., 2000. MAPK pathways activate and phosphorylate the osteoblast-specific transcription factor, Cbfa1. J. Biol. Chem. 275, 4453–4459. Yamamoto, M., Sobue, G., Yamamoto, K., Terao, S., Mitsuma, T., 1996. Expression of mRNAs for neurotrophic factors (NGF, BDNF, NT-3, and GDNF) and their receptors (p75NGFR, trkA, trkB, and trkC) in the adult human peripheral nervous system and nonneural tissues. Neurochem. Res. 21, 929–938. Yamashiro, T., Fukunaga, T., Yamashita, K., Kobashi, N., Takano-Yamamoto, T., 2001. Gene and protein expression of brain-derived neurotrophic factor and TrkB in bone and cartilage. Bone 28, 404–409. Yasui, M., Shiraishi, Y., Ozaki, N., Hayashi, K., Hori, K., Ichiyanagi, M., Sugiura, Y., 2012. Nerve growth factor and associated nerve sprouting contribute to local mechanical hyperalgesia in a rat model of bone injury. Eur. J. Pain 16, 953–965. Yirmiya, R., Goshen, I., Bajayo, A., Kreisel, T., Feldman, S., Tam, J., Trembovler, V., Csernus, V., Shohami, E., Bab, I., 2006. Depression induces bone loss through stimulation of the sympathetic nervous system. Proc. Natl. Acad. Sci. U.S.A. 103, 16876–16881.