Fax +41 61 306 12 34E-Mail karger@karger.chwww.karger.com

Original Paper

Cells Tissues Organs 2010;191:489–499 DOI: 10.1159/000276591

Development of Rat Tibia Innervation: Colocalization of Autonomic Nerve Fiber Markers with Growth-Associated Protein 43

Mariusz Gajda a Jan A. Litwin a Zbigniew Tabarowski b Olaf Zagólski c Tadeusz Cichocki a Jean-Pierre Timmermans d Dirk Adriaensen d

a Department of Histology, Jagiellonian University Medical College, b Department of Physiology, Institute of Zoology, Jagiellonian University, and c Diagnostic and Therapeutic Medical Center ‘Medicina’, Kraków , Poland; d Laboratory of Cell Biology and Histology, Department of Veterinary Sciences, University of Antwerp, Antwerp , Belgium

in the intercondylar eminence. In secondary ossification centers, NPY-IR fibers were seen from PD10, and in thebone marrow of the epiphyses from PD14. First VIP-IR and VAChT-IR fibers were observed on PD4 within the perioste-um, bone marrow and patellar ligament. From PD10 on, VIP-positive fibers were seen in the intercondylar eminence, and from PD14 in secondary ossification centers. GAP-43 proved to be superior to PGP 9.5 as marker of growing nerve fibers, mostly due to its earlier appearance. The presence of spe-cific nerve fibers may suggest possible involvement of auto-nomic innervation in regulation of bone development.

Copyright © 2010 S. Karger AG, Basel

Key Words Autonomic innervation � Bone � Development � Immunohistochemistry � Rat

Abstract Development of autonomic innervation of the tibia wasinvestigated in rat fetuses on gestational days (GD) 17–21 and in juvenile animals on postnatal days (PD) 1–28. Double immunofluorescence combined with confocal microscopy was applied to study colocalization of neuronal growth-associated protein 43 (GAP-43) and panneuronal marker protein gene product 9.5 (PGP) with markers of the auto-nomic nervous system: neuropeptide Y (NPY) and dopamine � -hydroxylase (D � H) for adrenergic, as well as vasoactive intes tinal polypeptide (VIP) and vesicular acetylcholine transporter (VAChT) for cholinergic fibers. The first GAP-43-immunoreactive (GAP-IR) nerve fibers were seen on GD17 in the perichondrium of the proximal epiphysis. Further GAP- and PGP-IR innervation appeared in the perichondrium/periosteum of the diaphysis and in the distal epiphysis (GD19), then in the bone marrow and in the intercondylar eminence (GD21). On PD1, NPY-IR and D � H-IR fibers ap-peared within the diaphyseal periosteum and on PD4 within the bone marrow. From PD14, GAP-43 immunoreactivity of NPY-positive fibers decreased. From PD7 on, NPY-IR fibers were observed in cartilage canals of both epiphyses and

Accepted after revision: October 3, 2009 Published online: January 14, 2010

Dr. Mariusz Gajda Department of Histology Jagiellonian University Medical College Kopernika 7, PL–31-034 Kraków (Poland) Tel./Fax +48 12 422 7027, E-Mail mmgajda @ cyf-kr.edu.pl

© 2010 S. Karger AG, Basel1422–6405/10/1916–0489$26.00/0

Accessible online at:www.karger.com/cto

Abbreviations used in this paper

CGRP calcitonin gene-related peptideD�H dopamine �-hydroxylaseGAP-43, GAP growth-associated protein 43GD gestational dayIR immunoreactiveNPY neuropeptide YPD postnatal dayPGP 9.5, PGP protein gene product 9.5SP substance PTH tyrosine hydroxylaseVAChT vesicular acetylcholine transporterVIP vasoactive intestinal polypeptide

Gajda /Litwin /Tabarowski /Zagólski /Cichocki /Timmermans /Adriaensen

Cells Tissues Organs 2010;191:489–499490

Introduction

The relatively dense innervation of bone includes pri-mary sensory afferents and postganglionic sympathetic fibers, mostly associated with blood vessels [Bjurholm et al., 1988a, 1988b; Hill and Elde, 1991; Hara-Irie et al., 1996; Serre et al., 1999; Mach et al., 2002]. The proportion of sensory (calcitonin gene-related peptide-immunoreac-tive; CGRP-IR) to noradrenergic postganglionic sympa-thetic (tyrosine hydroxylase-immunoreactive; TH-IR) nerve fibers differs in distinct regions of the bone, with sensory fibers being denser in the periosteum and sym-pathetic ones more abundant in the bone marrow [Mach et al., 2002]. In mature bone, autonomic nerve fibers be-long to adrenergic and cholinergic populations [Hohm-ann et al., 1986; Bjurholm et al., 1988b]. The sympathetic nervous system controls bone resorption by influencing osteoclasts and osteoblasts as well as regulating osseal blood flow [Trotman and Kelly, 1963; Hill et al., 1991; Takeda et al., 2002]. Vasoactive intestinal polypeptide (VIP) is known as a potent vasodilator, while noradrena-line causes a vasoconstrictory effect on resistance vessels isolated from the cancellous bone [Lundgaard et al., 1997, 2001]. Unloading-induced bone loss occurs through the control mediated via the sympathetic nervous system [Kondo et al., 2005]. The effect of autonomic neurotrans-mitters [noradrenaline, neuropeptide Y (NPY), VIP] on bone cells was confirmed by in vitro studies and denerva-tion experiments [Bjurholm et al., 1992; Cherruau et al., 1999; Hill et al., 1991; Lundberg et al., 2000]. Recent works have supplemented the traditional view that bone model-ing and remodeling are only regulated by paracrine (cy-tokines) and endocrine (parathormone, calcitonin) mechanisms, suggesting also the importance of the influ-ence of higher neuronal pathways [Patel and Elefteriou, 2007]. Fat-derived hormone leptin was found to control bone formation through a hypothalamic relay and these effects were shown to be mediated via the sympathetic nervous system [Takeda et al., 2002].

Numerous neuropeptides and enzymes, including substance P, CGRP, VIP, NPY and TH, were identified within bone tissues by immunohistochemistry and ra-dioimmunoassays [Bjurholm, 1991]. Growth-associated protein 43 (GAP-43, GAP) is a phosphoprotein located in the plasma membrane of developing and regenerating ax-ons [Gorgels et al., 1989]. GAP-43 was used in studies of developmental bone innervation [Gajda et al., 2000, 2005]. Protein gene product 9.5 (PGP 9.5, PGP), which belongs to a group of ubiquitin hydrolases, is a panneu-ronal marker and it has also been used in such investiga-

tions [Gajda et al., 2005; Sisask et al., 1995, 1996]. In the case of regenerating peripheral nerves, GAP-43-immu-noreactive fibers appeared earlier than fibers immuno-stained for general panneuronal marker PGP 9.5 [Li et al., 2001; Verze et al., 2003]. Sensory and autonomic neuro-mediators were found in nerve fibers of mature bone [Bjurholm et al., 1988a, 1988b]. Receptors for VIP and pituitary adenylate cyclase activating peptide (VPAC 1 , VPAC 2 , PAC 1 ), NPY (Y1, Y2), CGRP, substance P, as well as for classical neurotransmitters such as noradrenaline, serotonin and glutamate have been identified on bone cells (osteoclasts and osteoblasts) and several in vitro studies have shown that activation of these receptors can affect bone metabolism [Bjurholm, 1992; Konttinen et al., 1996; Lundberg et al., 2000; Chenu, 2004; Baldock et al., 2007]. Sympathetic adrenergic nerves containing nor-adrenaline, catecholamine-producing enzymes (TH and dopamine � -hydroxylase; D � H) and NPY, as well as cho-linergic ones utilizing VIP as cotransmitter were ob-served in bone [Bjurholm et al., 1988b; Hill and Elde, 1991; Hohmann et al., 1986; Tabarowski et al., 1996]. Ad-renergic fibers are exclusively associated with blood ves-sels and they are abundant in regions of high osteogenic activity, such as perichondrium/periosteum adjacent to the epiphyseal growth plate and metaphyseal bone mar-row. VIP-immunoreactive nerve fibers in the periosteum are associated with both vascular and nonvascular com-ponents, mostly within the cellular layer closest to the bone. Vesicular acetylcholine transporter (VAChT) car-rying acetylcholine into synaptic vesicles, thus necessary for cholinergic transmission, has recently been used as a marker for mapping of cholinergic terminals [Roghani et al., 1998].

In the present study, we investigated the time of ap-pearance during the pre- and postnatal development, as well as topographic distribution of autonomic nerve fi-bers in the developing long bone (tibia of rat hindlimb), using autonomic nerve fiber markers and neuronal growth marker GAP-43. Although a similar study was previously published by Sisask et al. [1996], we decided to use a wider panel of antibodies and confocal microscopy in order to obtain precise staging of nerve fiber develop-ment. Moreover, we sought confirmation whether in the developing bone GAP-43-IR fibers appear earlier than fibers immunostained for PGP 9.5, as reported for vari-ous regenerating peripheral nerves.

Development of Autonomic Innervation in Rat Tibia

Cells Tissues Organs 2010;191:489–499 491

Materials and Methods

Animals Adult Wistar rats of both sexes and timed pregnant females

were obtained from Charles River Laboratories (Brussels, Bel-gium). Conception was confirmed by the observation of the vag-inal plug, referred as day 0 of gestation (full gestation = 21 days). The animals were housed separately in acrylic cages with wood shavings in an air-conditioned room (22 8 3 ° C, 12 h/12 h dark/light cycle). They had access to water and standard rodent pellets ad libitum. National and international principles of laboratory animal welfare (conforming to NIH publication No. 86-23, re-vised 1985) were followed and the experiments were approved by the local ethics committee of the University of Antwerp.

Pregnant rats were killed by an overdose of sodium pentobar-bital (Nembutal; Sanofi, Brussels, Belgium) administered intra-peritoneally. Fetuses were obtained at gestational days (GD) 16(n = 8 from 2 different mothers), GD17 (n = 8 from 2 different mothers), GD19 (n = 8 from 2 different mothers) and GD21(n = 8 from 2 different mothers). The hindlimbs of fetuses were dissected out for further processing.

Offspring from different litters were sacrificed at postnatal day (PD) 1 (day of birth; n = 4), PD2 (n = 4), PD3 (n = 4), PD7(n = 4), PD10 (n = 3), PD14 (n = 3), PD21 (n = 3) and PD28 (n = 4) using an overdose of pentobarbital. Hindlimbs were dissected and the skin was removed to allow better penetration of the fixa-tive.

Deeply anesthetized (as described above) animals older than PD7 were first transcardially perfused with ice-cold Krebs-Ring-er solution followed by 4% phosphate-buffered (0.1 M , pH = 7.4) freshly prepared paraformaldehyde. Subsequently, limbs were postfixed as described below.

Tissue Preparation Dissected hindlimbs were fixed overnight by immersion in the

paraformaldehyde solution at 4 ° C, followed by rinsing in phos-phate-buffered saline (PBS, 0.01 M , pH = 7.4). Hindlimbs from animals older than GD21 were decalcified in 10% EDTA in Tris buffer (0.1 M , pH = 7) at 4°C for 5–14 days. The solution was re-freshed every 2–3 days. The hindlimbs were then rinsed in PBS and immersed overnight in 25% sucrose in PBS with 0.01% sodi-um azide at 4 ° C. Tissue blocks were mounted in TissueTek OCT compound (Sakura, Tokyo, Japan) on cryostat holders and were snap-frozen. Fifteen-micrometer-thick cryosections were cut in the sagittal plane, thaw-mounted on poly- L -lysine-coated slides and air-dried. Three or four serial sections were collected on each

slide. The procedures of material preparation did not influence immunostaining [Bjurholm et al., 1989].

Immunohistochemistry A pre-incubation step with 10% normal goat serum in PBS

containing 0.01% sodium azide, 0.05% thimerosal, 0.1% bovine serum albumin and 0.5% Triton X-100 was applied for 40 min to reduce non-specific binding and to increase penetration of the antibodies. For simultaneous demonstration of 2 antigens, an in-direct double-staining immunofluorescence procedure was ap-plied. The sections were incubated overnight at room tempera-ture in humid chambers with mixtures of primary antibodies in the following combinations: GAP/PGP, GAP/NPY, GAP/D � H, GAP/VIP and GAP/VAChT (see table 1 for the list of the primary antisera used). After rinsing in PBS, sections were incubated for 2 h at room temperature with a mixture of biotinylated sheep anti-mouse serum (RPN1001; diluted 1: 200; Amersham Biosciences, Amersham, UK) and Cy3-conjugated goat anti-rabbit serum (111-165-144 diluted 1: 500; Jackson IR, West Grove, Pa., USA). Follow-ing another rinse in PBS, FITC-conjugated streptavidin (RPN1232 diluted 1: 200; Amersham) was applied for 1 h at room tempera-ture. Primary and secondary antisera were dissolved in a solution used for pre-incubation; streptavidin conjugate was dissolved in PBS. After a final rinse, the sections were mounted with Vecta-shield medium (H-1000; Vector, Burlingame, Calif., USA) to min-imize photobleaching of fluorochromes. In the controls, the pri-mary or secondary antibodies were omitted and replaced by non-immune serum.

Fluorescence Microscopy Sections were examined using an Olympus BX-50 (Olympus,

Tokyo, Japan) epifluorescence microscope equipped with filter sets: U-MNIBA and U-MNG for FITC and Cy3 visualization, re-spectively. Relative densities of nerve fibers in distinct locations were semiquantitatively evaluated in tissue sections by 2 indepen-dent observers. We adopted very strict topographical criteria, i.e., only fibers intimately related to the bone rudiment were consid-ered.

For precise demonstration and colocalization of the examined antigens, the images were registered with a Zeiss LSM 410 (Zeiss, Jena, Germany) confocal laser scanning microscope. An argon laser ( � = 488 nm), a helium-neon laser ( � = 543 nm) as well as appropriate dichroic mirrors and emission filters (FT510, LP515 and FT560, LP570) were used for excitation of the fluorochromes and acquisition of their emission spectra. Stacks of acquired opti-cal sections were stored as graphic files and further processed

Table 1. Primary antibodies used in the study

Antigen Host/type Dilution Vendor; catalogue number

Growth-associated protein 43 mouse/monoclonal 1:1,000 Boehringer Mannheim, Mannheim, Germany; 1379011Protein gene product 9.5 rabbit/polyclonal 1:500 Biogenesis, Poole, UK; 7863-0504Dopamine �-hydroxylase rabbit/polyclonal 1:1,000 Chemicon, Temecula, Calif., USA; AB1585Neuropeptide Y rabbit/polyclonal 1:400 Affiniti, Exeter, UK; NA1233Vasoactive intestinal polypeptide rabbit/polyclonal 1:400 Affiniti, Exeter, UK; VA1285Vesicular acetylcholine transporter rabbit/polyclonal 1:1,000 Sigma, St. Louis, Mo., USA; V5387

Gajda /Litwin /Tabarowski /Zagólski /Cichocki /Timmermans /Adriaensen

Cells Tissues Organs 2010;191:489–499492

with 3D-reconstruction software (Imaris 3.0; Bitplane, Zurich, Switzerland) working on an Indigo 2 workstation (Silicon Graph-ics, Mountain View, Calif., USA). Final images were obtained as a result of ‘extended focus/maximal intensity projection’ trans-formation and presented as TIFF files at a resolution of 512 ! 512 pixels.

Results

A schematic drawing presenting topography of the de-veloping rat tibia is shown in figure 1 .

Fetal Period (GD16–GD21) The earliest GAP-immunoreactive (GAP-IR) nerve fi-

bers related to the tibial rudiment were found on GD17 in the perichondrium of the proximal epiphysis ( fig. 2 ). At this stage of development the expression of panneuronal marker PGP was weaker than that of GAP. From GD19 on, GAP-IR and PGP-IR fibers were clearly seen in the perichondrium/periosteum of the diaphysis and in the distal epiphysis. From GD21 they also appeared in the bone marrow cavity. Nonvascular fibers were seen run-ning free between hematopoietic cells but most of medul-lary fibers accompanied blood vessels ( fig. 3 ). From GD21 on, nerve fibers immunoreactive for GAP-43 and PGP 9.5 were located in the region of the intercondylar eminence. During the fetal period, no fibers expressing any of the autonomic markers could be observed in the close vicin-ity of the tibial rudiment.

Postnatal Period (PD1–PD28): Adrenergic Innervation (NPY, D � H) From PD1 onwards, nerve fibers containing NPY were

observed within the periosteum of the diaphysis ( fig. 4 ) and from PD4 within the bone marrow of the shaft ( fig. 5 ). Initially, they were not numerous but their number and intensity of immunostaining increased until PD14. All NPY-IR fibers accompanied blood vessels of different siz-es, forming intramural plexuses ( fig. 6 ). Especially dense innervation was observed in large nutritious vessels pen-etrating into the bone ( fig. 7 ). They were frequently seen in vascular canals ( fig. 8 ). Already among early detected NPY-IR fibers there were ones not co-expressing GAP-43 ( fig. 5 ). From PD14 onwards, GAP-43 immunoreactivity of NPY-positive fibers markedly decreased ( fig. 5 , 6 , 8 ). From PD7 on, NPY-positive fibers could be observed in the cartilage canals of both epiphyses and in the intercon-dylar eminence ( fig. 9 ). NPY-IR fibers were present in the secondary ossification centers from PD10 and in the bone marrow of the epiphyses from PD14. After PD14, the

number and localization of NPY-positive fibers remained unchanged. NPY-IR fibers displayed a characteristic var-icose morphology. In bone marrow areas, megakaryo-cytes/megakaryoblasts revealed cross-immunoreactivity with antibodies raised against NPY antigen ( fig. 5 , 6 ), as already reported [Ericsson et al., 1987].

The temporal pattern of appearance and localization of D � H-IR fibers was identical to those immunostained for NPY but the immunostaining intensity of this mark-er was inferior to NPY (not shown).

Postnatal Period (PD1–PD28): Cholinergic Innervation (VIP, VAChT) First VIP-IR fibers could be observed on PD4 in the

periosteal region of the shaft ( fig. 10 ) as well as in the

Proximalsecondaryossificationcenter

Primaryossificationcenter

Distalsecondaryossificationcenter

Fig. 1. Schematic drawing of a sagittal section of the tibia showing locations of primary and secondary ossification centers.

Development of Autonomic Innervation in Rat Tibia

Cells Tissues Organs 2010;191:489–499 493

2a 2b

3a

3b

4b4a

50 μm 50 μm

50 μm

50 μm

30 μm 30 μm

Fig. 2. GD17. Weakly PGP-immunopositive ( a ) and GAP-IR ( b ) fibers in the perichondrium of the proximal epiphysis of the cartilagineous primordium. Some GAP-IR and PGP-immunonegative fibers are in evidence (arrows). PE = Proximal epiphysis; PC = perichondrium. Fig. 3. PD1. PGP-IR ( a ) and GAP-IR ( b ) fibers accompanying blood vessels (arrows) in the bone marrow of the diaphysis. Fig. 4. PD1. NPY-IR ( a ) and GAP-IR ( b ) fibers (arrows) in the periosteum (PO) of the diaphysis (D). M = Skeletal muscle.

Gajda /Litwin /Tabarowski /Zagólski /Cichocki /Timmermans /Adriaensen

Cells Tissues Organs 2010;191:489–499494

5a 5b

6a

6b

7a 7b 8a 8b

40 μm 40 μm 40 μm

40 μm

40 μm 40 μm50 μm 50 μm

Fig. 5. PD7. NPY-IR ( a ) and GAP-IR ( b ) fibers in the bone marrow of the diaphysis. Some of NPY-IR fibers lack GAP-43 immunostaining (arrows). Note NPY-like immunoreactivity in megakaryocyte (arrowhead). Fig. 6. PD28. NPY-IR ( a ) and GAP-immunonegative ( b ) fibers (arrows) in a plexus of fibers around the large blood vessel of the diaphyseal bone marrow. Note almost complete disappearance of GAP-IR fibers. Megakary-ocytes are marked with arrowheads. Fig. 7. PD7. NPY-IR ( a ) and GAP-IR ( b ) fibers accompanying the large nutritional artery penetrating from the periosteum (PO) to the diaphysis (D). Fig. 8. PD28. NPY-IR ( a ) and GAP-immunonegative ( b ) fibers (arrow) around blood vessel in bone canal of the diaphysis.

Development of Autonomic Innervation in Rat Tibia

Cells Tissues Organs 2010;191:489–499 495

50 μm

50 μm

40 μm

40 μm 40 μm 40 μm

40 μm40 μm

9a

9b

11a

11b

10a

12a 12b

10b

Fig. 9. PD14. NPY-IR ( a ) and GAP-IR ( b ) vascular fibers in the intercondylar eminence (arrows). Fig. 10. PD4. VIP-IR ( a ) and GAP-IR ( b ) fibers in the periosteum (PO) of the diaphysis. Fig. 11. PD14. VIP-IR ( a ) and GAP-IR ( b ) fibers in the region of patellar ligament (PL). P = Patella; IE = inter-condylar eminence. Fig. 12. PD28. VIP-IR ( a ) and GAP-IR ( b ) fibers (arrow) in the intercondylar eminence (IE).

Gajda /Litwin /Tabarowski /Zagólski /Cichocki /Timmermans /Adriaensen

Cells Tissues Organs 2010;191:489–499496

bone marrow. However, they were significantly fewer in number than were NPY-positive fibers and they were particularly numerous in the neighborhood of the patel-lar ligament ( fig. 11 ). From PD10 onwards, numerous VIP-positive fibers were observed within soft tissues of the knee joint and the intercondylar eminence ( fig. 12 ). They reached the area of the secondary ossification cen-ter through cartilage canals around PD14. VIP-positive fibers had varicose morphology and disclosed no partic-ular predisposition to perivascular location.

The temporal pattern of appearance and localization of VAChT-IR fibers was identical to those immuno stained for VIP but the immunostaining intensity of this marker was inferior to VIP (not shown).

Immunoreactivity for GAP-43 did not disappear even in the most mature individuals studied; GAP-IR fibers are presented in figures 9 b, 12 b (PD14) and in figures 6 b, 8 b, 12 b (PD28).

Relative densities of nerve fibers in distinct areas of the developing rat tibia are summarized in table 2 .

Discussion

This study demonstrated the development of auto-nomic innervation in long bone using for the first time combined double immunohistochemistry for autonomic and developmental markers and confocal microscopy. In our hands, GAP-43 appeared to be a superior marker of growing nerve fibers since it allowed earlier detection of bone-related nerve fibers than PGP 9.5 [Gajda et al., 2005; Li et al., 2001]. Moreover, the disappearance of GAP-43-immunoreactive fibers in later stages of the development may reflect nerve fiber maturation (PGP 9.5 remains). Only a very limited number of papers have been pub-lished so far in which the authors systematically followed

Table 2. Sequence of appearance of nerve fibers immunoreactive for various neural markers and their relative densities in selected regions of rat tibia

Region of tibia Marker Stage of development

GD17 GD19 GD21 PD1 PD4 PD7 PD10 PD14 and later

Perichondrium/periosteum GAP-43 + ++ ++ +++ +++ +++ +++ ++PGP 9.5 + a ++ ++ ++ +++ +++ +++ +++NPY/D�H + ++ ++ ++ ++VIP/VAChT + + + +

Bone marrow (diaphysis) GAP-43 + + ++ +++ ++ ++PGP 9.5 + + ++ ++ +++ +++NPY/D�H + ++ ++ ++VIP/VAChT + + + +

Intercondylar eminence GAP-43 + ++ +++ +++ ++ ++PGP 9.5 + + ++ +++ +++ +++NPY/D�H + ++ ++VIP/VAChT ++ b +++ b

Secondary ossification centers GAP-43 + c ++ ++PGP 9.5 + c ++ ++NPY/D�H + c + ++VIP/VAChT + c +

Bone marrow (epiphysis) GAP-43 + ++PGP 9.5 + ++NPY/D�H ++VIP/VAChT +

Relative densities of nerve fibers: + = single, ++ = sparse, +++ = numerous.a Very low intensity. b Patellar ligament. c Cartilage canals.

Development of Autonomic Innervation in Rat Tibia

Cells Tissues Organs 2010;191:489–499 497

the innervation in long bone during the fetal and postna-tal periods using immunolabeling for neuropeptides and general neuronal markers [Gajda et al., 2000, 2005, 2009; Sisask et al., 1995, 1996; Hara-Irie et al., 1996]. Only a single study [Sisask et al., 1996] has described the devel-opment of autonomic nerve fibers in bone.

Our results are mostly consistent with those reported by Sisask et al. [1996] but provide precise temporal and spatial description of growing autonomic nerve fibers in the developing rat tibia. We found adrenergic fibers in the secondary ossification centers from PD14, while Sisask et al. [1996] observed them much later (PD24). Also, cholin-ergic (VIP-IR) fibers in the secondary ossification centers were seen earlier in our study. This discrepancy can be explained by more sensitive immunohistochemical pro-cedure and type of primary antibodies used.

In this study, autonomic nerve fibers were demon-strated in the joint-associated tissues: NPY-positive fibers in the intercondylar eminence from PD7 onwards and VIP-positive fibers in the knee joint and the intercondy-lar eminence from PD10 on. Abundant VIP-positive fi-bers were seen in the neighborhood of the patellar liga-ment. Sisask et al. [1996] did not describe presence of au-tonomic nerve fibers in these locations. The distribution of autonomic neuropeptides, especially NPY and VIP, may suggest their specific roles in the local regulation of bone physiology, such as blood flow, bone formation and resorption. NPY and VIP are supposed to be involved in bone growth and remodeling [Bjurholm et al., 1988b; Hill and Elde, 1991; Lundberg et al., 2000; Chenu, 2004; Bal-dock et al., 2007].

During development of long bone innervation, we ob-served a characteristic sequence of marker appearance. Initially, nerve fibers revealed the presence of develop-mental markers, then panneuronal markers appeared, followed by sensory or autonomic markers. In case of au-tonomic fibers, the adrenergic population appeared first, followed by the cholinergic one.

The earliest nerve fibers in developing tibia (observed on GD17) expressed only GAP-43 immunoreactivity [Gajda et al. 2000, 2005]. The next marker was PGP 9.5, being widely present from GD19. Neurochemically deter-mined fibers appeared later (GD21) and displayed expres-sion of sensory neuropeptides CGRP and substance P [Gajda et al., 2005]. The first autonomic adrenergic and cholinergic fibers were discerned on PD1 and PD4, re-spectively. A similar sequence has also been reported by others [Sisask et al., 1995, 1996; Hara-Irie et al., 1996]. Interestingly, in our studies GAP-43 persisted in nerve fibers of the most mature animals. This observation could

suggest that this protein is not only a developmental marker but is also present in the mature nervous system [Del et al., 1994; Vento and Soinila, 1999].

An analogous sequence of nerve fiber appearance has been observed in fracture healing and in ectopic bone formation. First, GAP-43-IR nerve fibers can be found in the fracture hematoma and in the periosteum as early as 3 days after experimental fracture of rat tibia [Li et al., 2001]. PGP 9.5-IR fibers were present 1 week post-trau-ma, and within 14 days CGRP-containing fibers were ob-served in the periosteum [Hukkanen et al., 1993]. NPY-positive fibers were only occasionally observed during fracture healing and very weak immunoreactivity for VIP was seen in scanty fibers present in soft tissues sur-rounding the fracture callus [Hukkanen et al., 1995].

In another experimental model of bone formation, an allogenic demineralized bone matrix implanted into rat abdominal wall, CGRP-IR, substance P-immunoreac-tive, NPY-IR, VIP-IR and TH-IR fibers appear in the con-nective tissue between chondroblastoid cells 10 days after implantation [Bjurholm et al., 1990]. Over the next 3 weeks, they gradually attain a shape and distribution re-sembling normal osseal nerves. NPY-containing nerve fibers can be seen after 8 weeks, particularly in the bone marrow, and these fibers remain abundant at 16 weeks. VIP-immunoreactive fibers are only observed in the sur-rounding periosteum-like fibrous tissue 4–6 weeks after implantation.

The time of appearance of the particular neuromedia-tors in nerve fibers in peripheral tissues does not neces-sarily reflect the sequence of appearance of the same markers in corresponding ganglia. In the rat, VIP is first detectable at embryonic day 14.5 in one third of all cate-cholaminergic neurons of the stellate ganglion (some of which are still dividing). Thereafter, the VIP content in the neurons decreases by 95% until birth [Tyrrell and Landis, 1994]. In postnatal life, the proportion of VIP-positive neurons increases till PD10 [Masliukov and Tim-mermans, 2004]. Similarly, the expression of other neu-ropeptides (NPY, galanin, somatostatin) and enzymes involved in neurotransmitter production (tyrosine hy-droxylase, choline acetyltransferase) vary during early postnatal development of stellate ganglia in rats and mice [Masliukov and Timmermans, 2004].

In adult rats, sympathetic cholinergic postganglionic neurons innervating the periosteum are one of the three neuron classes, apart from sudomotor and vasodilator neurons [Anderson et al., 2006]. They contain VIP and lack NPY. Asmus et al. [2000, 2001] proved that during de-velopment, axons growing from sympathetic ganglia, dis-

Gajda /Litwin /Tabarowski /Zagólski /Cichocki /Timmermans /Adriaensen

Cells Tissues Organs 2010;191:489–499498

References

Anderson, C.R., A. Bergner, S.M. Murphy (2006) How many types of cholinergic sympathetic neuron are there in the rat stellate ganglion? Neuroscience 140: 567–576.

Asmus, S.E., S. Parsons, S.C. Landis (2000) De-velopmental changes in the transmitter properties of sympathetic neurons that in-nervate the periosteum. J Neurosci 20: 1495–1504.

Asmus, S.E., H. Tian, S.C. Landis (2001) Induc-tion of cholinergic function in cultured sym-pathetic neurons by periosteal cells: cellular mechanisms. Dev Biol 235: 1–11.

Baldock, P.A., S.J. Allison, P. Lundberg, N.J. Lee, K. Slack, E.J. Lin, R.F. Enriquez, M.M. Mc-Donald, L. Zhang, M.J. During, D.G. Little, J.A. Eisman, E.M. Gardiner, E. Yulyaning-sih, S. Lin, A. Sainsbury, H. Herzog (2007) Novel role of Y1 receptors in the coordinated regulation of bone and energy homeostasis. J Biol Chem 282: 19092–19102.

Bjurholm, A. (1991) Neuroendocrine peptides in bone. Int Orthop 15: 325–329.

Bjurholm, A., A. Kreicbergs, E. Brodin, M. Schultzberg (1988a) Substance P- and CGRP-immunoreactive nerves in bone. Peptides 9: 165–171.

Bjurholm, A., A. Kreicbergs, L. Dahlberg, M. Schultzberg (1990) The occurrence of neuro-peptides at different stages of DBM-induced heterotopic bone formation. Bone Miner 10: 95–107.

Bjurholm, A., A. Kreicbergs, M. Schultzberg (1989) Fixation and demineralization of bone tissue for immunohistochemical stain-ing of neuropeptides. Calcif Tissue Int 45: 227–231.

Bjurholm, A., A. Kreicbergs, M. Schultzberg, U.H. Lerner (1992) Neuroendocrine regula-tion of cyclic AMP formation in osteoblastic cell lines (UMR-106-01, ROS 17/2.8, MC3T3-E1, and Saos-2) and primary bone cells. J Bone Miner Res 7: 1011–1019.

Bjurholm, A., A. Kreicbergs, L. Terenius, M. Goldstein, M. Schultzberg (1988b) Neuro-peptide Y-, tyrosine hydroxylase- and vaso-active intestinal polypeptide-immunoreac-tive nerves in bone and surrounding tissues. J Auton Nerv Syst 25: 119–125.

Chenu, C. (2004) Role of innervation in the con-trol of bone remodeling. J Musculoskelet Neuronal Interact 4: 132–134.

Cherruau, M., P. Facchinetti, B. Baroukh, J.L. Saffar (1999) Chemical sympathectomy im-pairs bone resorption in rats: a role for the sympathetic system on bone metabolism. Bone 25: 545–551.

Del, F.M., M. Quartu, J.V. Priestley, M.D. Setzu, M.L. Lai (1994) GAP-43 persists in adult-hood and coexists with SP and CGRP in hu-man trigeminal sensory neurones. Neurore-port 5: 2349–2352.

Ericsson, A., M. Schalling, K.R. McIntyre, J.M. Lundberg, D. Larhammar, K. Seroogy, T. Hökfelt, H. Persson (1987) Detection of neu-ropeptide Y and its mRNA in megakaryo-cytes: enhanced levels in certain autoim-mune mice. Proc Natl Acad Sci USA 84: 5585–5589.

Gajda, M., D. Adriaensen, T. Cichocki (2000) Development of the innervation of long bones: expression of the growth-associated protein 43. Folia Histochem Cytobiol 38: 103–110.

Gajda, M., J.A. Litwin, T. Cichocki, J.P. Timmer-mans, D. Adriaensen (2005) Development of sensory innervation in rat tibia: co-localiza-tion of CGRP and substance P with growth-associated protein 43 (GAP-43). J Anat 207: 135–144.

Gajda, M., J.A. Litwin, O. Zagólski, G.J. Lis, T. Cichocki, J.P. Timmermans, D. Adriaensen (2009) Development of galanin-containing nerve fibres in rat tibia. Anat Histol Embryol 38: 112–117.

Gorgels, T.G., C.M. Van Lookeren, A.B. Oest-reicher, A.A. Gribnau, W.H. Gispen (1989) B-50/GAP43 is localized at the cytoplasmic side of the plasma membrane in developing and adult rat pyramidal tract. J Neurosci 9: 3861–3869.

Hara-Irie, F., N. Amizuka, H. Ozawa (1996) Im-munohistochemical and ultrastructural lo-calization of CGRP-positive nerve fibers at the epiphyseal trabecules facing the growth plate of rat femurs. Bone 18: 29–39.

Hill, E.L., R. Elde (1991) Distribution of CGRP-, VIP-, D � H-, SP-, and NPY-immunoreactive nerves in the periosteum of the rat. Cell Tis-sue Res 264: 469–480.

Hill, E.L., R. Turner, R. Elde (1991) Effects of neonatal sympathectomy and capsaicin treatment on bone remodeling in rats. Neu-roscience 44: 747–755.

Hohmann, E.L., R.P. Elde, J.A. Rysavy, S. Einzig, R.L. Gebhard (1986) Innervation of perios-teum and bone by sympathetic vasoactive in-testinal peptide-containing nerve fibers. Sci-ence 232: 868–871.

Hukkanen, M., Y.T. Konttinen, S. Santavirta, L. Nordsletten, J.E. Madsen, R. Almaas, A.B. Oestreicher, T. Rootwelt, J.M. Polak (1995) Effect of sciatic nerve section on neural in-growth into the rat tibial fracture callus. Clin Orthop 311: 247–257.

playing initially noradrenergic properties, are down-regu-lated and they acquire cholinergic and peptidergic pheno-type. In periosteum transplanted to a site supplied by noradrenergic fibers, noradrenergic to cholinergic/peptid-ergic shift was observed [Asmus et al., 2000]. These find-ings suggest that nerve fibers acquire their final physiolog-ical phenotype after contact with target tissue. Periosteal cells influence sympathetic neuron phenotype by releasing a soluble cholinergic phenotype-promoting factor that is neither leukemia inhibitory factor nor ciliary neurotroph-ic factor [Asmus et al., 2001]. In our study, VIP-IR nerve fibers were discerned later than NPY-IR ones. This obser-vation could be explained by possible change from NPY-IR to VIP-IR fibers in the developing bone.

Conclusions

Relatively dense autonomic innervation of long bone rudiment may suggest its involvement in the control of long bone development. Changes in expression of the particular neural markers reflect maturation of the os-seal innervation. GAP-43 proved to be a superior nerve fiber marker to PGP 9.5 due to its earlier appearance in bone-related nerve fibers and its (at least partial) disap-pearance in maturing fibers.

Development of Autonomic Innervation in Rat Tibia

Cells Tissues Organs 2010;191:489–499 499

Hukkanen, M., Y.T. Konttinen, S. Santavirta, P. Paavolainen, X.H. Gu, G. Terenghi, J.M.Polak (1993) Rapid proliferation of calcito-nin gene-related peptide-immunoreactive nerves during healing of rat tibial fracture suggests neural involvement in bone growth and remodelling. Neuroscience 54: 969–979.

Kondo, H., A. Nifuji, S. Takeda, Y. Ezura, S.R. Rittling, D.T. Denhardt, K. Nakashima, G. Karsenty, M. Noda (2005) Unloading induc-es osteoblastic cell suppression and osteo-clastic cell activation to lead to bone loss via sympathetic nervous system. J Biol Chem 280: 30192–30200.

Konttinen, Y., S. Imai, A. Suda (1996) Neuropep-tides and the puzzle of bone remodeling: state of the art. Acta Orthop Scand 67: 632–639.

Li, J., T. Ahmad, M. Spetea, M. Ahmed, A.Kreicbergs (2001) Bone reinnervation after fracture: a study in the rat. J Bone Miner Res 16: 1505–1510.

Lundberg, P., A. Lie, A. Bjurholm, P.P. Lehen-kari, M.A. Horton, U.H. Lerner, M. Ransjo (2000) Vasoactive intestinal peptide regu-lates osteoclast activity via specific binding sites on both osteoclasts and osteoblasts. Bone 27: 803–810.

Lundgaard, A., C. Aalkjaer, A. Bjurholm, M.J. Mulvany, E.S. Hansen (1997) Vasorelaxation in isolated bone arteries: vasoactive intesti-nal peptide, substance P, calcitonin gene-re-lated peptide, and bradykinin studied in pigs. Acta Orthop Scand 68: 481–489.

Lundgaard, A., C. Aalkjaer, C. Bunger, E.S. Han-sen (2001) Adrenergic responses in human small arteries isolated from the femoral neck. J Orthop Res 19: 104–112.

Mach, D.B., S.D. Rogers, M.C. Sabino, N.M. Luger, M.J. Schwei, J.D. Pomonis, C.P. Key-ser, D.R. Clohisy, D.J. Adams, P. O’Leary, P.W. Mantyh (2002) Origins of skeletal pain: sensory and sympathetic innervation of the mouse femur. Neuroscience 113: 155–166.

Masliukov, P.M., J.P. Timmermans (2004) Im-munocytochemical properties of stellate ganglion neurons during early postnatal de-velopment. Histochem Cell Biol 122: 201–209.

Patel, M.S., F. Elefteriou (2007) The new field of neuroskeletal biology. Calcif Tissue Int 80: 337–347.

Roghani, A., A. Shirzadi, L.L. Butcher, R.H. Ed-wards (1998) Distribution of the vesicular transporter for acetylcholine in the rat cen-tral nervous system. Neuroscience 82: 1195–1212.

Serre, C.M., D. Farlay, P.D. Delmas, C. Chenu (1999) Evidence for a dense and intimate in-nervation of the bone tissue, including gluta-mate-containing fibers. Bone 25: 623–629.

Sisask, G., A. Bjurholm, M. Ahmed, A. Kreic-bergs (1995) Ontogeny of sensory nerves in the developing skeleton. Anat Rec 243: 234–240.

Sisask, G., A. Bjurholm, M. Ahmed, A. Kreic-bergs (1996) The development of autonomic innervation in bone and joints of the rat. J Auton Nerv Syst 59: 27–33.

Tabarowski, Z., K. Gibson-Berry, S.Y. Felten (1996) Noradrenergic and peptidergic inner-vation of the mouse femur bone marrow. Acta Histochem 98: 453–457.

Takeda, S., F. Elefteriou, R. Levasseur, X. Liu, L. Zhao, K.L. Parker, D. Armstrong, P. Ducy, G. Karsenty (2002) Leptin regulates bone for-mation via the sympathetic nervous system. Cell 111: 305–317.

Trotman, N.M., W.D. Kelly (1963) The effect of sympathectomy on blood flow to bone. JAMA 183: 121–122.

Tyrrell, S., S.C. Landis (1994) The appearance of NPY and VIP in sympathetic neuroblasts and subsequent alterations in their expres-sion. J Neurosci 14: 4529–4547.

Vento, P., S. Soinila (1999) Quantitative compar-ison of growth-associated protein GAP-43, neuron-specific enolase, and protein gene product 9.5 as neuronal markers in mature human intestine. J Histochem Cytochem 47: 1405–1416.

Verze, L., A. Paraninfo, C. Viglietti-Panzica, G.C. Panzica, G. Ramieri (2003) Expression of neuropeptides and growth-associated protein 43 (GAP-43) in cutaneous and mu-cosal nerve structures of the adult rat lower lip after mental nerve section. Ann Anat 185: 35–44.