Evidence that the canonical Wnt signalling pathway regulates deer antler regeneration

Abstract

Wnt signalling regulates many developmental processes, including the fate specification, polarity, migration, and proliferation of cranial neural crest. The canonical Wnt pathway has also been shown to play an important role in bone physiology and there is evidence for its recapitulation during organ regeneration in lower vertebrates. This study explores the role of the Wnt signalling pathway in deer antlers, frontal bone appendages that are the only mammalian organs capable of regeneration. Immunocytochemistry was used to map the distribution of the activated form of β-catenin (^aβCAT). A low level of ^aβCAT staining was detected in chondrocytes and in osteoblasts at sites of endochondral bone formation. However, ^aβCAT was localised in cellular periosteum and in osteoblasts in intramembranous bone, where it co-localised with osteocalcin. The most intense ^aβCAT staining was in dividing undifferentiated cells in the mesenchymal growth zone. Antler progenitor cells (APCs) were cultured from this region and when the canonical Wnt pathway was inhibited at the level of Lef/TCF by epigallocatechin gallate (EGCG), the cell number decreased. TUNEL staining revealed that this was as a result of increased apoptosis. Activation of the pathway by lithium chloride (LiCl) had no effect on cell number but inhibited alkaline phosphate activity (ALP), a marker of APC differentiation, whereas EGCG increased ALP activity. This study demonstrates that β-catenin plays an important role in the regulation of antler progenitor cell survival and cell fate. It also provides evidence that β-catenin's function in regulating bone formation by osteoblasts may be site-specific. Developmental Dynamics 235:1390–1399, 2006. © 2006 Wiley-Liss, Inc.

INTRODUCTION

Impaired regeneration of cartilage and bone is a feature of many diseases and can limit the repair of defects post surgery, e.g., following tumour removal or craniofacial reconstruction. The underlying problem is that mammals have a limited capacity to regenerate, whereas many lower organisms can replace in their entirety lost or damaged body parts (Slack,1980; Brockes,1997). Understanding the molecular pathways that control organ regeneration is a strategy that will improve the prospect of developing relevant therapies for stimulating cartilage and bone repair in man. In mammals, the only available model is the deer antler, since these large cranial appendages are the only mammalian structures capable of complete epimorphic regeneration (Price et al.,2005). Antlers grow from permanent extensions of the frontal bone (pedicles) in males and in females of a few species. After the previous year's set are shed each spring, an antler bud forms from which the branched antlers elongate at an unparalleled rate by a process of endochondral ossification. Intramembranous ossification also contributes to their growth in diameter. In late summer, the velvet skin covering them is shed and the antler is retained as solid dead bone used for fighting until it is cast the following spring and another round of regeneration begins.

There is increasing evidence that the molecular mechanisms that regulate regeneration in lower vertebrates, and also in antlers, recapitulate those that control embryonic development (Faucheux et al.,2004). A pathway that plays a fundamental role in development is the Wnt signalling system (Rudnicki and Brown,1997; Hartmann and Tabin,2000; Tufan and Tuan,2001; Enomoto-Iwamoto et al.,2002; Yang,2003), and there is accumulating evidence that Wnts play a role during regeneration of lost structures in lower organisms. For example, Wnts 5 and 3a have been identified in the regenerating zebrafish fin (Poss et al.,2000), Wnts 10a, 7a, 5a, and 5b in the regenerating urodele tail (Caubit et al.,1997), and Wnts have also been shown to be involved during head regeneration in hydra (Bode,2003).

Wnts are a large family of secreted glycoproteins that utilise complex signalling systems, of which the canonical pathway acting through β-catenin is currently the best understood (Nusse and Varmus,1992; van Noort et al.,2002). In un-stimulated cells, cytoplasmic β-catenin is phosphorylated by the enzyme glycogen synthase kinase 3 beta (GSK-3β), which targets it for degradation by the proteosome. However, upon binding of Wnts to a cell surface co-receptor complex of Frizzled (FZ) receptors and LRP5 or LRP6, a signalling cascade is initiated, the consequence of which is the inactivation of GSK3β. This results in the accumulation of de-phosphorylated β-catenin, which translocates to the nucleus and, in association with lymphoid enhancing factor (Lef)/T-cell factor (TCF) transcription factors, activates Wnt responsive genes.

The Wnt pathway's role in development and in cancer has been extensively studied (Cadigan and Nusse,1997; Polakis,2000; Logan and Nusse,2004; Reya and Clevers,2005), and in the context of craniofacial development the Wnt pathway is of particular importance in regulating cranial neural crest cell fate specification, polarity, migration, and proliferation (Dorsky et al.,1998; Lee et al.,2004). In general, the formation of cranial neural crest–derived skeletal structures is lacking or greatly perturbed in β-catenin mutant embryos (Brault et al.,2001; Hari et al.,2002). However, the Wnt pathway is now known to control bone development and bone mass acquisition at all skeletal sites (Johnson et al.,2004). Mutations in the Wnt co-receptor LRP5 have produced striking alterations in bone mass in both humans and mice; a gain in function mutation in LRP5 results in a high bone mass phenotype (Little et al.,2002; Babij et al.,2003), whereas a loss of function mutation leads to a loss of bone mass and severe osteoporosis (Gong et al.,2001; Levasseur et al.,2001). LRP5 signals through β-catenin, which is essential for skeletal lineage specification (Roman-Roman et al.,2003; Holmen et al.,2005), with a recent study in transgenic mice demonstrating that targeted deletion of β-catenin in head and limb mesenchyme prevents the trans-differentiation of osteoblasts into chondrocytes (Hill et al.,2005). Another study in mice has shown that stabilisation of β-catenin in osteoblasts leads to higher bone mass, whereas deletion of β-catenin from osteoblasts leads to osteopaenia (Glass et al.,2005). This study also showed that this increase in bone mass is associated with impaired osteoclast differentiation as a result of increased osteoprotogerin synthesis by osteoblasts.

In addition to being involved in the regulation of bone mass, there is also some evidence that the Wnt pathway may be important during bone repair since several components of the Wnt pathway are up-regulated during the early phases of fracture healing, particularly in osteocytes and osteoblasts (Hadjiargyrou et al.,2002; Zhong et al.,2004). It is also known that Wnts play an important role in controlling chondrogenesis (Karen et al.,2005; Tamamura et al.,2005), although evidence for the pathway's role in cartilage repair has, to our knowledge, yet to be documented. The aim of the present study was to use the deer antler to explore the role of the canonical Wnt pathway during cartilage and bone regeneration in an adult mammal. To identify sites at which the canonical Wnt pathway was activated, we immunolocalised the de-phosphorylated (activated) form of β-catenin (^aβCAT) in sections of antler tissue at an early stage of regeneration and also during a phase of rapid longitudinal growth. Because Wnts are known to regulate the fate of many cell types, including bone cells (Canalis et al.,2005), we also investigated the effect of manipulating the canonical pathway on the growth, survival, and differentiation of antler progenitor cells (APCs).

RESULTS

Casting of the previous year's set of antlers is followed by wound repair over the exposed pedicle surface and regeneration of a new antler bud. The anatomy of the early stages of antler growth has recently been described elsewhere (Li et al.,2004, 2005). However, because the different cell types involved have yet to be well characterised, we did not include antlers at less than 2 weeks of growth in the present study. By day 14, the regenerating antler bud is four to five cm high and consists of an outer region in which the primordia of future branches are already visible as raised swellings (Fig. 1A and B). Centrally there is a “scab,” although wound repair is normally complete by this stage. In longitudinal section, a number of distinct tissue regions can be identified that represent the whole spectrum of cellular differentiation involved in endochondral bone development (Fig. 1A,B). These tissue zones are macroscopically and microscopically similar to those in the distal tip of the branches of a fully formed rapidly growing antler (Fig. 1D); underneath the velvet skin is the perichondrium, then a zone of proliferating mesenchymal cells, and below this chondrogenesis and endochondral bone formation takes place leading to the formation of the primary spongiosa. The periosteum is multi-layered and consists of an outer fibrous component and an inner cellular layer adjacent to sites of intramembranous bone formation. The histology of these regions is illustrated alongside illustrations of ^aβCAT localisation in Figures 2–4.

Figure 1. Gross and macroscopic appearance of 14-day and 6-week-old growing antlers. A: The newly formed antler after 14 days of growth. The dotted line indicates the site of the longitudinal section presented in B (the site of the future main branch of the antler). B: Longitudinal section to show macroscopic localisation of tissues in a 14-day antler. Dotted lines indicate boundaries of the following areas; e, epidermis; d, dermis; v, velvet skin; p, perichondrium; m, mesenchyme; cp, chondroprogenitor region; cart, non-mineralised cartilage; mcart, mineralised cartilage; bone^ps, primary spongiosa (bone that forms by endochondral ossification); boneⁱ, bone at the site of intramembranous ossification below the periosteum (po) at the periphery of the antler shaft. C: An antler at ∼6 weeks of growth. The box delineates the growing tip of the main branch and the position of the longitudinal section illustrated in D. These areas illustrated are as described in B. Crossed arrows indicate the anatomical position of the photographs. A, anterior; Po, posterior; M, medial; L, lateral; Pr, proximal; D, distal. Scale bars, A and B = 0.5 cm; C = 3 cm; D = 1 cm.

Figure 2. Immunolocalisation of ^aβCAT in the perichondrium and mesenchyme of the 14-day antler. A: H&E stained undecalcified paraffin section of the perichondrium. B: ^aβCAT immunostaining (brown stain) in a proportion of perichondrium cells. C: PCNA staining (nuclei are stained black) shows a small number of dividing cells (small arrows) in the perichondrium. D: H&E stained undecalcified paraffin section of mesenchyme. E: ^aβCAT immunlocalises in the majority of cells in this region. Inset shows negative control section. F: A significant proportion of mesenchymal cells are PCNA positive. Scale bars = 80 μm.

Figure 3. Immunolocalisation of ^aβCAT in cartilage of the 14-day antler. A: H&E stained undecalcified paraffin section of the chondroprogenitor region illustrating chondroblasts arranged in columns surrounding vascular channels (v). Arrowheads indicate perivascular tissue, a site of osteoblast and osteoclast differentiation. B: ^aβCAT immunostaining (brown stain) in perivascular cells. Staining is barely detectable in cartilage cells. C: PCNA staining shows a small number of dividing cells in perivascular tissue. D: H&E stained undecalcified paraffin section of non-mineralised cartilage illustrating hypertrophic chondrocytes (**) adjacent to vascular spaces (v). E: No significant ^aβCAT immunostaining in hypertrophic chondrocytes or perivascular cells. F: PCNA staining showing a small number of dividing cells in perivascular tissue. Scale bars = 80 μm.

Figure 4. Immunolocalisation of ^aβCAT and osteocalcin at sites of endochondral and intramembranous bone formation in the 14-day antler. A–D: Primary spongiosa (b); formed endochondrally. A: H&E stained undecalcified paraffin section illustrating osteoblasts (arrowheads) on the surface of trabecular bone (b) within which are recently embedded osteocytes (small arrows). B: There is no ^aβCAT immunostaining in osteoblasts or osteocytes at this site. C: PCNA staining shows a small number of dividing osteoblast lineage cells. D: Immunolocalisation of osteocalcin in differentiated osteoblasts. E–H: Intramembranous bone formation. E: H&E stained undecalcified paraffin section of intramembranous bone (*). F: Immunolocalisation of ^aβCAT in osteoblasts on the bone surface and in cells in the surrounding tissue. G: PCNA staining shows that there are very few dividing cells at this site, which indicates that they are more differentiated. H: Immunolocalisation of osteocalcin in osteoblasts on the bone surface and in surrounding cells. I-L: The periosteum. I: H&E stained undecalcified paraffin section of fibrous (f) and cellular (c) periosteum. J: Immunolocalisation of ^aβCAT in a number of cells in the cellular periosteum. K: PCNA staining shows that a number of cells in the cellular periosteum are dividing. L: There is no osteocalcin staining in the cellular periosteum indicating that cells at this site are not differentiated osteoblasts. Scale bars = 80 μm.

Immunolocalisation of ^aβCAT

In the rapidly growing antler (Fig. 1C), the pattern of expression of ^aβCAT was identical to that seen in the 14-day antler bud. Therefore, only results for the 14-day antler are shown.

Localisation of ^aβCAT in perichondrium and in mesenchyme.

In the perichondrium, which contains very few proliferating cells (Fig. 2C), ^aβCAT immunoreactivity was only observed in a small number of cells (Fig. 2B). In contrast, there was intense staining for ^aβCAT in cells of the underlying mesenchyme (Fig. 2E). This has been described previously as the antlers “growth zone” (Colitti et al.,2005) because a large proportion of cells are PCNA positive (Fig. 2F).

Immunolocalisation of ^aβCAT in cartilage.

Below the mesenchyme is a zone of chondroprogenitors that differentiate into non-proliferating chondrocytes arranged in vertical columns separated by vascular spaces. Chondroprogenitors, in the main, were negative for ^aβCAT and in this region staining was mainly confined to the endothelium of the blood vessels and to small numbers of proliferating perivascular cells (Fig. 3B,C). More proximally, in mineralised cartilage, very little ^aβCAT immunoreactivity was observed in hypertrophic chondrocytes (Fig. 3E), and there was no significant staining of proliferating perivascular cells in this region (Fig. 3F).

Immunolocalisation of ^aβCAT in bone.

Antlers grow by a process of modified endochondral ossification and proximal to mineralised cartilage in the centre of the antler shaft is primary spongiosa. Surprisingly, no positive staining for ^aβCAT was observed in fully differentiated osteoblasts in this region (Fig. 4B). The differentiated osteoblast phenotype was confirmed by osteocalcin immunolocalisation (Fig. 4D) and, as expected, only a small proportion of these cells are proliferating (Fig. 4C). However, at sites of intramembranous bone formation at the periphery of the antler shaft, ^aβCAT was localised in osteocalcin positive osteoblasts that line “islands” of newly formed bone (Fig. 4F,H). Immunofluorescent staining and con-focal microscopy revealed that ^aβCAT co-localised with osteocalcin at focal sites within the cytoplasm of these osteoblasts and, as expected, was also localised in nuclei (Fig. 5A–F).

Figure 5. Co-localisation of ^aβCAT and osteocalcin in osteoblasts at sites of intramembranous bone formation. Localisation of ^aβCAT (A and D, green stain), osteocalcin (B and E, red stain) and their co-localisation (C and F, yellow stain). A–C: At sites of intramembranous bone formation, ^aβCAT and osteocalcin co-localise in osteoblasts and in recently embedded osteocytes. Scale bar = 50 μm. D–F: Higher magnification images showing both cytoplasmic and nuclear localisation of ^aβCAT (arrowheads). In the cytoplasm, ^aβCAT and osteocalcin show focal co-localisation. The white dotted lines delineate the periphery of cell and the outline of the nucleus. Scale bar = 8 μm.

^aβCAT staining was identified in only a small proportion of cells in the fibrous periosteum, whereas staining was seen in the majority of cells in the cellular periosteum (Fig. 4J). PCNA staining revealed that a number of cells in the periosteum are dividing (Fig. 4K) and were negative for osteocalcin (Fig. 4L). No staining was seen in negative control sections.

The Effect of Wnt Signalling on Antler Progenitor Cells (APCs)

A series of in vitro studies were undertaken to establish the potential role of the Wnt pathway in regulating growth, apoptosis, and differentiation in APCs cultured from the mesenchyme region of the distal antler tip. Treatment of sub-confluent cultures of APCs with EGCG (25 μM) for 24 h significantly decreased the mean total cell number per well (control: 1831.3 ± 82.0, EGCG: 751 ± 213.4, P < 0.001), whereas LiCl (10 mM) had no effect (Fig. 6A). This result reflected an increase in programmed cell death as EGCG treatment significantly increased the number of TUNEL-positive APCs (control: 2.6 ± 1.2%;EGCG: 8.8% ± 1.9%; P < 0.001). Treatment with LiCl had no effect on apoptosis (Fig. 6B).

Figure 6. The effect of LiCl and EGCG on cell number and on apoptosis in monolayer cultures of antler progenitor cells (APCs). A: Treatment of monolayer cultures with EGCG (25 μM) for 24 hr significantly decreased cell number (*P < 0.001), whereas LiCl (10 mM) had no effect. NaCl (10 mM) was the negative control. B: EGCG (25 μm) for 24 hr significantly increased the percentage of TUNEL positive cells (*P < 0.001), whereas LiCl (10 mM) had no effect. Results show the mean ± SD for 4 wells. Results shown are a representative of 3 experiments.

To establish the effect of the Wnt pathway on APC differentiation, confluent cultures of APCs were incubated with LiCl or EGCG for 5–7 days and ALP activity, a marker of osteoblast differentiation, was determined by histochemical stain (data not shown) or biochemical assay. ALP activity was reduced in cultures treated with LiCl but increased in cultures treated with EGCG (Fig. 7). These results were the same for 5 and 7 days post-treatment.

Figure 7. The effect of LiCl and EGCG on alkaline phosphatase (ALP) activity in antler progenitor cells (APCs). Treatment of monolayer cultures with EGCG (25 μM) for 5 days increased (*P < 0.001) ALP activity compared to controls (NaCl 10 mM), whereas ALP activity was decreased by LiCl (10 mM) (**P < 0.01). Results show the mean ± SD for 4 samples. Results shown are representative examples of 4 experiments.

DISCUSSION

The Wnt signalling pathway's role in controlling cell growth and fate determination during embryonic development is well established. More recently, the importance of this pathway in cartilage biology and in the regulation of bone mass in mice and in humans has also been recognised. In this study, we used the deer antler as a model to identify whether Wnt signalling plays a role when cranial bone and cartilage regenerates in an adult mammal, since there is already evidence that regeneration in lower vertebrates is controlled by this pathway (Caubit et al.,1997; Poss et al.,2000). Immunohistochemistry was used to identify those cells in the antler in which β-catenin, a critical component of the canonical Wnt pathway, was activated. Recent gain and loss of function studies in transgenic mice have shown that studying the function of β-catenin provides a powerful tool for investigating the function of the canonical Wnt signalling pathway in different contexts (Glass et al.,2005; Hill et al.,2005).

We found that during early stages of regeneration (∼14 days) and in fully formed, rapidly growing antlers, activated β-catenin was localised in skin (data not shown), in undifferentiated mesenchymal cells, in the periosteum, and in osteoblasts at sites of intramembranous bone formation. It is likely that β-catenin is being activated by Wnts secreted by antler cells. However, it must be considered that the activity of β-catenin can be influenced by other, Wnt-independent, mechanisms (Monga et al.,2002; Haq et al.,2003). Studies are in progress aimed at identifying the expression profile of Wnts in antlers, and preliminary studies have shown that LEF/TCFs, the transcriptional targets of the canonical pathway, are highly expressed in the mesenchymal region. In fact, the most intense staining for ^aβCAT was identified in this region where the processes of cell death and cell growth are closely associated (Colitti et al.,2005), consistent with this being the antler's “growth zone” (similar to that present in the regenerating crab claw; Huxley,1932). Our previous studies have shown that these mesenchymal cells are not highly differentiated; they express only very low levels of alkaline phosphatase and do not express type II collagen, although they subsequently differentiate into chondrocytes in vivo (Price et al.,1994, 1996). In culture, these cells have an extended lifespan and can be induced to differentiate along the adipocyte lineage and to express alkaline phosphatae, a marker of the osteoblast lineage, which indicates a degree of multipotentiality (Faucheux et al.,2001). Furthermore, these mesenchymal cells also synthesise PTHrP, which appears to be a marker of progenitor cells in both the antler bud and rapidly growing antler (Price and Allen,2004). In fact, we have preliminary evidence that PTHrP and ^aβCAT co-localise in mesenchymal cells (unpublished observations).

Localisation of ^aβCAT in the majority of antler progenitor cells indicates that the Wnt pathway may play a significant role in controlling their lineage specification. Our results suggest that activation of the canonical Wnt pathway represses their differentiation of chondrocytes, which is consistent with the situation during early embryogenesis when β-catenin activity is required to repress differentiation of early mesenchymal cells along the chondrocytes lineage (Hill et al.,2005). Our results are also consistent with the finding that β-catenin expression is high in prechondrogenic mesenchymal cells from chick wing buds (Ryu et al.,2002). Furthermore, the ectopic expression of canonical Wnt genes, including Wnt 1 and Wnt 7a, inhibits chondrogenesis in the developing limb (Rudnicki and Brown,1997; Stott et al.,1999; Tufan and Tuan,2001) and in the developing craniofacial mesenchyme (Rudnicki and Brown,1997). Similarly, in mice in which the β-catenin gene was inactivated by Wnt1-Cre-mediated deletion, it was found that most of the skeletal structures derived from cranial neural crest were missing (Brault et al.,2001; Hari et al.,2002).

The absence of significant staining for β-catenin in differentiated antler chondrocytes supports the hypothesis that it acts as a negative regulator of chondrocyte differentiation. However, in a recent study that used the TOPGAL model to study sites of β-catenin activation in late stage embryos, neonatal and post natal mice, canonical Wnt signalling was found to be generally active in chondrocytes, particularly in growth plates of endochondral bones and in costal cartilage (Hens et al.,2005). This apparent inconsistency with our study may reflect different requirements for Wnt signalling during the life cycle of the chondrocyte and between early and late stages of development. Wnts inhibit the differentiation of cartilage at an early stage as it is formed by chondroblasts (Rudnicki and Brown,1997), whereas they are required for maturation and/or hypertrophy of cartilage (Kitagaki et al.,2003).

Having identified high levels of ^aβCAT in antler progenitor cells in vivo, a series of in vitro studies were then undertaken to try and establish the role of the Wnt pathway in regulating their growth, apoptosis, and differentiation. Induction of β-catenin accumulation by inhibition of GSK-3β with LiCl, had no effect on cell number whereas inhibition of the Wnt pathway with EGCG significantly decreased cell numbers. These results indicate that while Wnts may not function as mitogens for mesenchymal growth in antlers, β-catenin signalling may play a role in maintaining the size of the precursor cell pool, similar to that seen in neuronal progenitor cells in the mouse (Zechner et al.,2003). Interestingly, this result also demonstrates that antler progenitors do not respond to activation of the canonical Wnt pathway in the same way as bone marrow stromal cells; the latter proliferate when treated with LiCl while retaining pluripotency (de Boer et al.,2004). Different responses to LiCl between antler and bone marrow stem cells probably reflect their different embryological origin; antlers are almost certainly derived from cranial neural crest since they are frontal bone appendages.

Since a large proportion of cells in the antler's growth zone are TUNEL positive (Colitti et al.,2005), we explored whether the decrease in APC cell number following treatment with EGCG was due to apoptosis. In other systems, there is evidence that Wnt antagonists regulate apoptosis; e.g., in developing mouse and chicken limbs, DKK 1, a potent antagonist of the Wnt pathway, promotes apoptosis in the interdigital mesenchyme (Grotewold and Ruther,2002). Our results are entirely consistent with these findings since EGCG increased the number of apoptotic cells in APC cultures whereas LiCl had no effect. Although we used a dose of EGCG that has been shown to inhibit β-catenin/TCF (Dashwood et al.,2002) in previous studies, the specificity of EGCG remains open to question. Therefore, more specific inhibitors of the pathway (e.g., DKK 1 and SFRP) now need to be tested in our model.

We also studied the effect of canonical Wnt signalling on APC differentiation using ALP as a marker since previously we have shown in vivo and in vitro that ALP activity increases as antler cells differentiate (Price et al.,1994). Although ALP it is normally used as marker of the osteoblast phenotype, its activity has also been shown to increase with chondrocyte terminal differentiation (Church et al.,2002). These studies revealed that LiCl inhibited ALP activity whereas EGCG increased differentiation and supports our in vivo evidence, which suggests that β-catenin inhibits differentiation of early antler progenitor cells. However, caution must be made when extrapolating from in vitro studies to the in vivo situation and further studies are required to further investigate the effect of the canonical Wnt pathway on cell lineage specification.

Our study has also shown that the canonical Wnt pathway is activated in osteoblasts during adult bone regeneration. There is now a considerable body of evidence that β-catenin regulates osteoblastogenesis (Gong et al.,2001; Bodine and Komm,2002; Bain et al.,2003; Rawadi et al.,2003; Day et al.,2005; Hill et al.,2005) and there is also evidence that components of the pathway are expressed in repairing fractures (Hadjiargyrou et al.,2002; Otto and Rao,2004). A finding of particular interest was that the canonical Wnt pathway was apparently not activated in osteoblasts on trabecular bone surfaces at sites of endochondral bone formation. In contrast, there was strong staining for ^aβCAT in the cellular layer of the periosteum and in osteoblasts in newly formed intramembranous bone. In keeping with this, murine periosteal cells stain intensely for β-catenin (Hill et al.,2005) and β-gal staining was observed in the dome of the skull of neonatal TOPGAL mice (Hens et al.,2005), which forms through intramembranous ossification. Interestingly, again in neonatal TOPGAL mice, while activation of Wnt signalling was found in osteoblasts at both endosteal and periosteal surfaces of developing skull bones, osteoblasts on the surface of trabeculae in vertebrae did not show β-galactosidase activity (Hens et al.,2005). A more recent study has shown that in the developing frontal bone of mice, an up-regulation of β-catenin precedes early osteoblast differentiation during intramembranous ossification, whereas at sites of endochondral ossification in the limb β-catenin expression is down regulated (Day et al.,2005). This suggests that the mechanisms controlling the differentiation of osteoblasts during intramembranous and endochondral bone development are different. However, the functional significance of this observation remains unclear and clearly deserves further investigation.

Although β-catenin does not appear to control osteoblast differentiation at sites of endochondral bone formation, it may regulate osteoclast differentiation since positive staining was detected in perivascular tissue in antler cartilage and bone, the site of osteoclast differentiation (Faucheux et al.,2001). In fact, a recent study has shown that mice over-expressing ^aβCAT in osteoblasts have high bone mass as a consequence of a defect in osteoclast differentiation (Glass et al.,2005). Furthermore, a recent study in our laboratory has shown that activation of the canonical Wnt pathway with LiCl inhibits osteoclast differentiation in cultures of antler osteoclast-like cells, while inhibition of the pathway promotes differentiation (Althnaian et al.,2004).

In summary, this study is the first to use immunocytochemistry to localise the activated form of β-catenin, a critical component of the canonical Wnt pathway, in regenerating mammalian cartilage and bone. Its localisation suggests that this pathway may be involved in a number of biological processes during regeneration, including chondrogenesis, as well as osteoblast differentiation. Its pattern of localisation and function also recapitulate that of the canonical Wnt pathway during differentiation of cranial neural crest into skeletal tissue, the likely source of the cells that form the antler. A particularly novel, and potentially important, finding is that Wnt signalling may have a site-specific function in osteoblasts during bone regeneration. Mesenchymal progenitor cells in the antler's growth zone appear to be a particularly important target, and in vitro studies showed that inhibition of the canonical Wnt signalling increased apoptosis and also stimulated differentiation of these cells. Future studies should be directed at elucidation of the mechanisms by which the Wnt signalling pathway interacts with other local, as well as hormonal, stimuli, to regulate antler cell growth, survival, and differentiation. In the long term, this research could be applied clinically to improve the repair of cranial skeletal defects.

EXPERIMENTAL PROCEDURES

Harvesting of Antler Tissue

Antlers were harvested at post mortem from red deer (Cervus elaphus) stags (∼2 years old) at two different stages of regeneration; 14 days after the previous set of antlers had been cast or between 4 and 8 weeks after casting (period of rapid longitudinal growth). The distal tip of the main branch of the antler was removed aseptically and was either placed in culture medium for transport to the laboratory [DMEM supplemented with 10% FBS and (Penicillin/Streptomycin) (100 IU–100 mg/ml)] or cut into 0.5-cm³ blocks, fixed for 2–3 days in 4% paraformaldehyde (PFA; pH 7.4), and then embedded in paraffin wax for sectioning.

Immunohistochemistry

Sections were de-waxed in xylene, rehydrated in graded ethanols, then washed in distilled water (dH₂O) and phosphate-buffered saline (PBS). The primary antibodies, a mouse monoclonal IgG raised against ^aβCAT (Upstate), a mouse anti- human PCNA (Dako, Cambridge, UK), and a polyclonal rabbit antibody raised against osteocalcin (courtesy of L. Fisher), were diluted in PBS-containing 0.05% Triton X-100 and 10% horse serum (PBS/TX-100/HS) at 1:200. Endogenous peroxidase activity was blocked with methanol containing 3% hydrogen peroxide (H₂O₂) for 30 min. For ^aβCAT, antigen was retrieved by boiling for 20 min in 0.01M sodium citrate buffer (pH 6.0). Sections were then pre-blocked with PBS/TX-100/HS (30′) followed by incubation with the primary antibody for 2 hr. Mouse IgG was substituted for the primary antibody in negative control sections. The sections were then incubated with a biotin conjugated horse anti rabbit/mouse polyclonal antibody (Dako) for 30 min. Detection was carried out using the vector ABC method with 3,3′-diaminobenzidine tetrachloride (DAB) as substrate (Sigma, Poole, UK). After each of the above steps, the sections were washed twice in PBS/TX. The sections were counterstained with 1% haematoxylin solution for 30 sec, dehydrated in graded ethanols, cleared with xylene, and mounted using VectaMount (Vector Labs, Peterborough, UK). Sections were viewed and photographed using a Leica Q550IW light microscope with a DC 500 Leica digital camera (Leica, Solms, Germany).

Immunofluorescence

Sections were placed in permeabilisation buffer for 15 min. Non-specific binding was blocked using PBS with 10% Newborn Calf Serum for 30 min at room temperature and samples sequentially incubated with primary antibodies (as described above) for 2 hr at room temperature. Sections were washed in PBS, incubated 2 hr at room temperature with Fluorescein isothiocyanate (FITC) conjugated goat anti-mouse and Tetramethylrhodamine isothiocyanate (TRITC) conjugated anti-rabbit secondary Abs (1:100; Dako, Cambridge, UK). Imaging was performed with Zeiss LSM 410 laser scanning confocal microscope (Carl Zeiss, Inc., Thornwood, NY).

Cell Culture

Cell digests were prepared from the zone of mesenchymal tissue in the distal antler tip that contains undifferentiated antler progenitor cells (APCs). Tissue was diced into 1–2 mm³ pieces, washed with Hanks balanced salt solution (HBSS) containing penicillin and streptomycin (PS, 100 IU–100 mg/ml) and fungizone (F, 100 mcg/ml) (HBSS/PS/F). The tissue was digested with 0.25% type I bacterial collagenase (Sigma, Poole, UK) in HBSS/PS/F at 37°C for 2 hr, filtered through a nylon mesh (pore size, 125 μm), and then the cell suspension centrifuged (200g) for 10 min. The cell pellet was washed twice in HBSS/PS/F, and then re-suspended in culture medium [Fitton Jackson's modification of BGJb medium, supplemented with 10% foetal bovine serum (FBS), PS (100 IU–100 mg/ml) and F (100 mcg/ml)].

For all experiments, cells were used at first passage and seeded into 6- or 12-well plates and maintained in culture medium at 37°C in 5% CO₂/95% air. Medium containing test factors or vehicle control was changed every 2 days. Lithium chloride (LiCl), at 10-mM concentration, was used to activate the Wnt signalling pathway, which gave maximum response (data not shown) (de Boer et al.,2004). Inhibition of signalling was achieved with epigallocatechin-3-gallate (EGCG), which was used at 25 μM, the lowest dose, which gives maximum inhibition of β-catenin/TCF activity (Dashwood et al.,2002). Unless otherwise stated, for each treatment group four wells were used and experiments were repeated at least twice using cells from two different deer. For experiments investigating the effects of Wnt signalling on cell number and apoptosis, cells were grown until sub-confluent, then treated for 24 hr. Cultures were then fixed with 4% PFA for 20 min at 4°C, then washed three times in PBS. For experiments investigating the effect of Wnt signalling on alkaline phosphatase (ALP) activity, cells were grown until confluent, then treated for 5–7 days. Following treatment, cells were either washed in PBS and solubilized in 200 μl of 0.1% Triton (for the biochemical assay) or fixed in PFA (for histochemical staining).

TUNEL Staining

Apoptosis was assayed in situ using the terminal deoxynucleotidyl transferase (TdT) mediated dUTP Nick-End Labeling (TUNEL) method (Gavrieli et al.,1992), using a DeadEnd™ Clorimetric TUNEL System (Promega, Madison, WI), according to the manufacturer's instructions. In brief, PFA-fixed cells were permeabilised by immersing in 0.2% Triton X-100 in PBS for 5 min at room temperature. Cultures were equilibrated in 100 μl of buffer [200 mM Potassium cacodylate, 10 mM NaCl, 6 mM MgCl₂ and 10 mM CaCl₂] at room temperature for 5–10 min, then incubated for 1 hr at 37°C in a 100 μl mixture of terminal deoxynucleotidyl transferase rTdT enzyme and biotinylated nucleotides. The reaction was stopped by immersion in 2× sodium saline citrate for 15 min and blocked in methanol containing 3% H₂O₂. The sections were then incubated for 30 min with a solution of streptavidin-peroxidase conjugate (diluted 1:500 in PBS) before developing with H₂O₂ and 3,3′-diaminobenzidine (DAB; Sigma, Poole, UK).

Quantification of Cell Number and Percentage of Apoptotic Cells

Cell number and the percentage of TUNEL-positive cells were determined in 5 microscopic fields (area 0.24 × 0.24 mm) from each well at 40× magnification. The fields were photographed using an inverted light microscope (Leica Q550IW) with a digital camera (DC 500 Leica) and cells were counted using a computer programme (Leica Qwin). The apoptotic index was calculated as the mean ratio of the number of positive cells to total cell number.

Alkaline Phosphatase Staining and Assay

Assay protocols have been described previously (Price et al.,1994). Briefly, for histochemical staining, fixed cells were stained for 1 hr in a mixture of 0.1M Tris/HCl (pH 9), 0.1% Fast Red TR as coupler and 0.02% napthol AS-MX phosphate as artificial substrate. ALP of cell lysates was assayed in a diethanolamine buffer (pH 10.5) using para-nitrophenyl phosphate as substrate. Hydrolysis of substrate to para-nitrophenol was measured by spectroscopy as a change in absorbance at 410 nm. Total cell protein was measured by a modification of the Lowry method using bovine serum albumin (BSA) as standard (Lowry et al.,1951), and ALP specific activity was then expressed as μmol/μg cell protein/minute.

Data was analysed using SPSS 12.0 for Windows using One Way ANOVA. Results were considered statistically significant at P < 0.05, unless otherwise specified. Results shown represent the mean ± one standard deviation of the mean.

Acknowledgements

The authors thank Prof. Mike Horton, Head of the Bone and Mineral Centre, University College London, for the use of the scanning confocal microscope.