SEARCH

SEARCH BY CITATION

Keywords:

  • mammals;
  • long bones;
  • cartilage canals;
  • endochondral bone formation;
  • secondary ossification centre;
  • type I collagen

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

In mammals, the exact role of cartilage canals is still under discussion. Therefore, we studied their development in the distal femoral epiphysis of mice to define the importance of these canals. Various approaches were performed to examine the histological, cellular, and molecular events leading to bone formation. Cartilage canals started off as invaginations of the perichondrium at day (D) 5 after birth. At D 10, several small ossification nuclei originated around the canal branched endings. Finally, these nuclei coalesced and at D 18 a large secondary ossification centre (SOC) occupied the whole epiphysis. Cartilage canal cells expressed type I collagen, a major bone-relevant protein. During canal formation, several resting chondrocytes immediately around the canals were active caspase 3 positive but others were freed into the canal cavity and appeared to remain viable. We suggest that cartilage canal cells belong to the bone lineage and, hence, they contribute to the formation of the bony epiphysis. Several resting chondrocytes are assigned to die but others, after freeing into the canal cavity, may differentiate into osteoblasts. Developmental Dynamics 236:2077–2088, 2007. © 2007 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

A first step, presumably leading to endochondral ossification of the epiphysis, is the formation of vascularized cartilage canals. These canals run through the uncalcified hyaline cartilage prior to the establishment of a secondary ossification centre (SOC) in long bones (e.g., femur, tibia) of avian (Lutfi, 1970; Blumer et al., 2004a, b, 2005, 2006; Eslaminejad et al., 2006) and numerous mammals (Stockwell, 1971; Cole and Wezeman, 1985; Cole and Cole, 1989; Visco et al., 1990; Burkus et al., 1993; Ganey et al., 1995; Roach et al., 1998; Rivas and Shapiro, 2002; Doschak et al., 2003; Álvarez et al., 2005a, b). Similarly, in human short bones without epiphysis (e.g., tarsal bones, vertebrae), the canals are present before the initial site of an ossification nucleus is evident (Agrawal et al., 1986; Chandraraj and Briggs, 1988; Fritsch and Eggers, 1999; Fritsch et al., 2001). However, it should be also noted that cartilage canals are not necessarily a prerequisite for endochondral ossification in the long bone extremities. During early development, they appear to be absent in the femoral head of several marsupials and rats, as well as in metatarsals of mice (Thorp, 1990; Morini et al., 1999; Reno et al., 2006).

A long debate surrounds the role cartilage canals in vertebrates; however, it is generally accepted that they are necessary for nourishment of the growing chondroepiphysis (Wilsman and Van Sickle, 1972). In addition, it has repeatedly been considered that these canals supply long bones (Lutfi, 1970; Kugler et al., 1979; Burkus et al., 1993; Ganey et al., 1995; Rivas and Shapiro, 2002; Álvarez et al., 2005a, b) as well as short bones (Fritsch et al., 2001) with osteogenic cells. Nevertheless, in mammals, clear data in support of this assumption are lacking so far and, therefore, the importance of the canals in endochondral bone formation is still under discussion. In chicken, however, evidence has been provided that cartilage canal cells fully belong to the bone lineage and they seem to differentiate from preosteoblasts into osteocytes on their course into the SOC. Based on these observations, cartilage canals are considered essential for endochondral bone development in the avian epiphysis (Blumer et al., 2005, 2006). Interestingly, it was noted that in several mammals cartilage canals can regress as the animal gets older (Wilsman and Van Sickle, 1972; Cole and Wezeman, 1985; Ytrehus et al., 2004a, b). This event is a physiological age-dependent process described as chondrification in which the vessels degenerate and the mesenchymal cells are converted into matrix-producing chondrocytes, occluding the lumen of the canal. Consequently, these canals are not involved in bone formation.

Within the clubfoot tarsal bones, cartilage canals appear to be altered, their number diminishes, and endochondral bone formation seems to be disturbed (Fritsch et al., 1999; Gilbert et al., 2001). However, these histological aberrations can not be well elucidated as long as the exact role of cartilage canals in normal bone development is not established.

Early formation of vascularized cartilage canals involves the resorption of resting cartilage (Lee et al., 2001; Álvarez et al., 2005b). There is, however, some controversy about the fate of chondrocytes during this process. Álvarez et al. (2005b) reported that several of these cells are released as viable cells into the canal cavity and finally might differentiate into bone-forming cells. Similarly, freeing of chondrocytes from the matrix surrounding cartilage canals was proposed by Cole and Wezeman (1985). In contrast, other studies suggested that elongation of canals causes death of resting chondrocytes (Roach et al., 1998; Lee et al., 2001).

In a search for processes leading to formation of the SOC in mammals, we studied the role of cartilage canals during normal bone development in the distal femoral epiphysis of postnatal mice. We hypothesized that in mice like in chicken, these canals are essential for endochondral bone formation. We first examined the chronology of canal formation, onset of endochondral bone development, and establishment of the SOC. Second, we verified whether canal mesenchymal cells express type I collagen, a major bone-relevant protein. Third, we investigated the fate of resting chondrocytes among others, by analysis of immunostained consecutive semithin resin sections (2 μm).

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Formation of Cartilage Canals and Bone Development in the Distal Epiphysis

Light microscopy (LM) and transmission electron microscopy (TEM).

In the distal femoral epiphysis, cartilage canals appeared the first time at day (D) 5 and three dimensional (3-d) reconstruction demonstrated that during proceeding development several small ossification nuclei were formed that subsequently joined into a large secondary ossification centre (SOC) (Fig. 1A–D). Only five cartilage canals were found at D 5 (Fig. 1A,E). They were not branched and terminated blindly in the resting cartilage. The canals originated from the perichondrium, and were composed of vessels and perivascular mesenchymal cells embedded in a matrix that did not stain with toluidine blue. However, around the canals, a distinct layer was weakly stained with toluidine blue whereas the remaining cartilage matrix stained deeper (Fig. 2A). Immediately in front of the canal blind ends, several resting chondrocytes had both a dark-stained nucleus and cytoplasm suggesting that these cells were degenerating. At the ultrastructural level, the nucleus of these cells showed margination of the chromatin and budded into apoptotic bodies. However, some chondrocytes with a bright nucleus and pale cytoplasm appeared to be viable (Figs. 2A–D, 3A). At the canal tips, macrophages were present. They contained numerous mitochondria and their apical cell membrane was differentiated into short microvilli contacting the cartilage matrix. Several of these cells had elongated cytoplasmic processes, were found to open the lacunae of dark chondrocytes, and were in close contact with electron dense bodies, presumable dying chondrocytes (Figs. 2A–D, 3B–D). Macrophages were not detectable at the lateral walls of cartilage canals where no continuous endothelium was elaborated (Fig. 3E). At D 8, the epiphysis was structurally similar; however, more canals were present, some of which penetrated deeper into the chondroepiphysis exclusively made up from resting cartilage until this point in time. At D 10, the number of cartilage canals had further increased and almost all of them were highly branched within the zone of hypertrophic chondrocytes. The segments within the resting zone remained non-ramified but had irregular contours (Figs. 1F, 4A,B). Chondrocytes organized in a distinct proliferating zone were not distinguishable. Around hypertrophic chondrocytes, the cartilage matrix was calcified and multinucleated chondroclasts were noted in this area (Fig. 4C). They contained numerous clear vesicles and mitochondria; their cell surface was elaborated to numerous short microvilli (ruffled border), which were in intimate contact with the calcified cartilage starting to resorb this matrix. Early bone matrix (osteoid) was deposited onto non-resorbed cartilage spicules and osteoblasts bordered on these scaffold-like structures (Fig. 4D). Thus, the first events of endochondral bone formation were observed around the branched endings of cartilage canals and several small ossification nuclei were noted at D 10. At D 18, these nuclei had joined into a large SOC. Several short canals were still seen.

thumbnail image

Figure 1. LM micrographs (A–D). Histological longitudinal sections through the distal femoral epiphysis showing a gross overview of the temporal sequence of bone formation (A–D). A,B: At D 5 and 8, the epiphysis is exclusively composed of resting cartilage and cartilage canals (cc) are not branched. C: At D 10, several ossification nuclei (arrows) are detectable. D: At D 18, a large secondary ossification centre (soc) is present in the epiphysis. All scale bars = 100 μm. E,F: Shown are 3-d images of the femoral epiphysis. The 3-d models were aligned according to the orientation seen in A and C. E: At D 5, five short cartilage canals are detectable (arrows). F: At D 10, the length of the canals has increased, they are more numerous and several small ossification nuclei are seen (asterisks). Two cartilage canals entering the epiphysis from different surfaces are united (arrowhead). They do not penetrate into the hypertrophic zone.

Download figure to PowerPoint

thumbnail image

Figure 2. LM micrographs (A–D). Consecutive semithin resin sections (2 μm) through the apical tip of a cartilage canal (cc) at D 5. Within the surrounding resting cartilage, light and dark chondrocytes are detectable. Contrary to the cartilage matrix of the resting zone, the canal matrix does not stain with toluidine blue. Note that a distinct layer around the tip limited by the white arrows in A is less stained. One macrophage with an elongated cytoplasmic process (red arrows) appears to resorb the cartilage matrix. Another macrophage (yellow arrowheads) opens the lacuna of a dark, presumably dying, chondrocyte. A–D: ×630 magnification; scale bar = 20 μm.

Download figure to PowerPoint

thumbnail image

Figure 3. TEM micrographs from a 5-day-old mouse. A: A dark chondrocyte in front of the canal tip. The cell is shrunken, its nucleus (n) is irregularly shaped, and the chromatin coalesces at the margin. White arrows point to apoptotic bodies within the cell. Scale bar = 1 μm. B: A macrophage (m) is found near a presumable dying chondrocyte (ch) (compare with Fig. 2D, yellow arrowheads). Scale bar = 1 μm. C: At the apical tip of cartilage canals, a macrophage (m) seems to resorb the cartilage matrix of the resting zone (rz). The macrophage has an elongated cytoplasmic process, which can be also seen in light micrographs (Fig. 2A–D, red arrows). Scale bar = 2 μm. D: The apical cell membrane of a macrophage (m) is elaborated to short cytoplasmic processes invading the resting cartilage (rz). E: Section through the lateral wall of a cartilage canal showing the interface between the canal (cc) and the resting zone (rz). The canal matrix appears electron translucent whereas the adjacent cartilage matrix is electron dense. At the canal border, the cells are flattened but a continuous endothelium is not elaborated. Scale bar = 5 μm.

Download figure to PowerPoint

thumbnail image

Figure 4. LM (A,B) and TEM micrographs (C,D) from a mouse aged 10 days are shown. A,B: In this stage, cartilage canals (cc) are intensively branched within the hypertrophic zone (hz). The segment within the resting zone (rz) remains non-ramified. Scale bar = 50 μm (A) and 20 μm (B). C,D: Branched tip of a canal within the hypertrophic zone. C: A chondroclast (chc) with numerous clear vesicles and mitochondria resorbs the calcified matrix. Scale bar = 5 μm. D: An osteoblast (ob), rich in rough endoplasmic reticulum, lines the non-resorbed calcified matrix (asterisks). Arrows point to an electron-dense layer interpreted as early bone matrix (osteoid). Scale bar = 2 μm. Inset shows a higher magnification of the osteoid, which is composed of cross-banded fibrils. Scale bar = 0.5 μm.

Download figure to PowerPoint

Light microscopical examination of complete series of semithin sections (2 μm) through the whole epiphysis of 5-, 8-, and 10-day-old mice revealed that occlusion of the cartilage canals' lumen was never observed and no ultrastructural features of degenerating cells within the canals were found either.

Immunohistohemistry (IHC).

To identify the role of cartilage canal cells in bone formation, the localization of type I collagen, a major bone-relevant protein (Franz-Odendall et al., 2006), was examined. Immunodetection was performed on both histological paraffin (6 μm) and semithin resin sections (2 μm) and the staining pattern was identical in both approaches; however, the thinner sections allowed a more detailed analysis. At D5 and 8, a faint labeling was visible within cartilage canals (Fig. 5A). From D 10, this labeling pattern increased and in addition a clear layer of type I collagen (= osteoid) lined the canal lateral walls in the resting zone. This layer could be traced back to the hypertrophic zone where canal mesenchymal cells, now interpreted as osteoblasts, bordered on it (Fig. 5B,C). Type I collagen was also found in the lacunae of several hypertrophic chondrocytes but examination of consecutive semithin sections demonstrated that these lacunae were then opened by chondroclasts (Fig. 5C, inset). Short positive labeled cartilage canals were still present at D 18 (Fig. 5D). Positive staining for type I collagen was also observed in the primary ossification centre (POC) and the perichondral bone lamellae.

thumbnail image

Figure 5. IHC on histological paraffin (6 μm) and semithin resin sections (2 μm) showing the localization of type I collagen during different developmental stages. A: Longitudinal paraffin section through a cartilage canal (cc) within the reserve zone (rz). The canal originates from the perichondrium (p) and stains weakly whereas the perichondrium labels more intensely (D 5). Scale bar = 50 μm. Inset shows a low amount of type I collagen within the canal. Scale bar = 20 μm. B: Longitudinal resin section through cartilage canal at D 10. An overview is shown. A thin continuous layer of type I collagen (early bone matrix) is elaborated at the lateral wall of the canal (cc) and can be traced back into the hypertrophic zone (hz) where the canals are intensively branched and an ossification nuclei originates. Scale bar = 50 μm. Inset: An increased labeling for type I collagen is visible within the non-branched segment of the reserve zone (rz) in comparison with D5. Scale bar = 20 μm. C: Same section. Higher magnification of the canal branched segment. Osteoblasts (arrows) line the early bone matrix, which is laid down on calcified cartilage matrix (asterisks). Note that type I collagen is also present within the opened lacuna of a hypertrophic chondrocyte (Inset, same magnification). Scale bar = 20 μm. D: Longitudinal paraffin section through the epiphysis. At D 18, a large secondary ossification centre (soc) has developed. Short cartilage canals are still present, which show the same labeling pattern as for D 10. Scale bar = 20 μm.

Download figure to PowerPoint

In Situ hybridization (ISH).

To elucidate the differences in type I collagen localization inside cartilage canals noted during advancing development, expression for type I collagen mRNA was examined in mice aged 5 and 10 days, respectively. Our results demonstrated that 5 days after birth, only a small number of canal cells showed a weak signal for type I collagen (Fig. 6A). In 10-day-old mice, however, according to our IHC observations, the number of positively labeled cells had increased and examination of numerous profiles indicated that almost all canal cells expressed type I collagen (Fig. 6B,D). Furthermore, type I collagen mRNA was observed within osteoblasts of the ossification nuclei and the cells of the perichondrium (Fig. 6C,D). Weak expression appeared to be present in a few hypertrophic chondrocytes within the non-resorbed cartilage spicules of the ossification nuclei (data not shown). Osteogenic precursor cells and osteoblasts of the POC also labeled positively. Osteocytes clearly expressed lower levels of type I collagen mRNA. Random controls served as negative controls for validating the specificity of the signal and showed no labeling (Fig. 6E).

thumbnail image

Figure 6. ISH showing the localization of type I collagen mRNA in a 5- and 10-day-old mouse. The sections are not counterstained. A,B: Cross-section through a cartilage canal (cc) within the resting zone of a mouse aged 5 and 10 days, respectively. A: At D 5, only a few cells reveal weak expression (arrowheads) whereas 10 days after birth (B) the signal for type I collagen is stronger (dark blue staining) and detectable within almost all canal cells. C: Section through an ossification nucleus. Osteoblasts (arrowheads) express type I collagen. They border on the early bone matrix, which is laid down on the calcified cartilage matrix labeled by asterisks (compare with Fig. 4D and 5C). D: Longitudinal section though the epiphysis shows an overview. The cells within both the secondary ossification centre (soc) and the cartilage canal (cc) express type I collagen. E: Same magnification as F (×200). The image shows the following section. No signal is observed when sections are hybridized with the random control. Scale bars = 50 μm.

Download figure to PowerPoint

The Fate of Resting Chondrocytes During the Formation of Cartilage Canals

LM and TEM.

As mentioned before, several resting chondrocytes directly in front of the canal apical tips revealed signs of degeneration whereas others remained viable. Light microscopic viewing of consecutive semithin sections through the canal lateral walls also showed that the surrounding chondrocytes had a different appearance. The typical normal chondrocyte of this zone had a pale cytoplasm and a bright nucleus. In contrast, in the atypical chondrocyte both appeared dark. Occasionally, doublets were found consisting of a light as well as a dark chondrocyte within one lacuna (Fig. 7A–D). Observations with the TEM demonstrated that light chondrocytes had a relatively electron-translucent cytoplasm, with great amounts of rough endoplasmic reticulum and a round nucleus (Fig. 7E). In contrast, dark chondrocytes appeared to shrink; their nucleus was irregularly shaped showing a patchy condensation of chromatin. Furthermore, these cells contained abundant rough endoplasmic reticulum, the ribosomes of which were hardly visible (Fig. 7F). These structural findings immediately around the canal lateral walls were observed throughout all developmental stages. Within the hypertrophic zone, the majority of chondrocytes were shrunken and pale, showing a morphology that indicated the terminal stage in their life cycle.

thumbnail image

Figure 7. LM (A–D) and TEM micrographs (E–F). A–D: Consecutive semithin resin sections (2 μm) through the lateral wall of a cartilage canal from a mouse aged 10 days. The panels show the interface of the canal (cc) and the adjacent resting zone (rz). Two types of chondrocytes, occasionally both within one lacuna, are found and their course is shown in the following sections. The black arrows point to a light chondrocyte and the asterisks indicate a dark chondrocyte. The white arrowheads label a doublet. A–D, ×630 magnification; scale bar = 20 μm. E,F: The ultrastructural features of these two types of resting chondrocytes (D 10). E: The light chondrocyte has an oval shape and contains a spherical nucleus as well as great amounts of rough endoplasmic reticulum. Scale bar = 5 μm. F: The dark chondrocyte appears shrunken and the nucleus displays a patchy arrangement of the chromatin. Abundant rough endoplasmic reticulum with hardly detectable ribosomes is distributed throughout the cell (inset). Scale bars = 1 μm.

Download figure to PowerPoint

IHC and TEM.

To specifically identify cells undergoing programmed cell death, active caspase 3 was used as a marker. Different developmental stages were examined (D5, 8, and 10). Positively stained chondrocytes were located in close proximity to the apical tip and lateral wall of cartilage canals within the resting cartilage. However, only several cells display labeling, suggesting that not all chondrocytes underwent programmed cell death (Fig. 8A,B). In hypertrophic chondrocytes, active caspase 3 immunostaining could also be detected (Fig. 8C). Within cartilage canals, positive cells were never seen.

thumbnail image

Figure 8. IHC and a TEM micrograph. A–C: Localization of active caspase 3 is shown on paraffin sections. A: Several chondrocytes (arrows) of the resting zone (rz) around the apical tip of a cartilage canal label positively. The arrowhead indicates a macrophage that opens the lacuna of a resting chondrocyte (D 5). B: Caspase 3 positive resting chondrocytes (arrows) near the later wall of a cartilage canal (D 8). C: Active caspase 3 reaction is found in hypertrophic chondrocytes (arrow). All scale bars = 20 μm. D–M: The Localization of type II collagen within the resting zone (rz) is shown on semithin resin as well as histological paraffin sections. D–G: Consecutive resin sections through the lateral wall of a cartilage canal. The sections have a thickness of 2 μm. A chondrocyte (arrow) of a 10-day-old mouse is released into the canal cavity where it is still surrounded by thin layer of type II collagen. A–D: ×630 magnification; scale bar = 20 μm. H: Semithin resin section. A chondrocyte with a pericellular rim of type II collagen (arrow) is located inside the canal cavity at its blind end (D 5). Scale bar, 20 μm. I: TEM micrograph. A resting chondrocyte within the canal cavity. Arrows limit an electron dense layer around the cell which corresponds to the type II collagen layer seen in G and H, respectively. Scale bar = 2 μm. J: The micrographs (E–G) are set semi-transparent, stacked into a composite photograph, and the so-formed image now matches a histological paraffin section with a thickness of 6 μm. It shows a cell (arrow) with a pericellular rim of type II collagen at the border of the canal; however, the origin of the cell can not be traced. Scale bar = 20 μm. K–M: Analysis of consecutive paraffin sections (6 μm) do not demonstrate that resting chondrocytes are released into the cartilage canal. In M (arrow), a situation similar to J is shown but the following section (inset) does not show that this cell is released into the canal cavity. K–M: ×400 magnification. Scale bar = 20 μm. The inset in M: ×630 magnification. Scale bar = 20 μm.

Download figure to PowerPoint

In order to examine whether resting chondrocytes were released into the cartilage canals, the distribution for type II collagen was studied. Immunohistochemistry was performed on consecutive semithin as well as histological sections and mice aged 5 and 10 days, respectively, were examined. Both approaches displayed the same staining pattern with intense labeling in the resting zone and faint labeling in the hypertrophic zone. However, ribbons of consecutive semithin resin sections made with a histo-jumbo-diamond knife had the following advantages over paraffin sections cut with a steel knife: the sections had a thickness of only 2 μm, their quality was improved, and the sections had the same alignment. This facilitated tracing of labeled structures, yielding a more detailed analysis (Blumer et al., 2002). Examination of labeled consecutive semithin sections clearly demonstrated that resting chondrocytes near the canal apical tip as well as the lateral wall were released into the cavity where they were surrounded by a distinct layer of type II collagen (Fig. 8D–H). On the ultrastructural level, these cells had a pericellular portion of electron-dense material; they contained large amounts of rough endoplasmic reticulum and appeared viable (Fig. 8I). However, examination of several cartilage canals revealed that only a small number of chondrocytes were released into the canal cavity. Analysis of consecutive histological sections (6 μm) never demonstrated similar results when compared with a series of semithin sections (Fig. 8J–M). Hypertrophic chondrocytes were never freed into cartilage canals.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Formation of Cartilage Canals and Bone Development in the Distal Epiphysis

In the proximal tibia and distal femur of murine species, cartilage canals originate in early postnatal stages and it was shown that they enable the formation of an SOC within the epiphysis (Kugler et al., 1979; Cole and Cole, 1989; Holmbeck et al., 1999; Davoli et al., 2001; Lee et al., 2001; Álvarez et al., 2005a, b). However, it was never demonstrated that cartilage canals supply the SOC with osteogenic cells and hence the exact role of cartilage canals in normal bone development is still not well characterized.

In the present study, we examined the distal femoral epiphysis in an age series of mice postnatal stages and our observations show that several cartilage canals appear the first time at D 5. The canals start off as invaginations of the perichondrium and bear vessels as well as perivascular mesenchymal cells. As in chicken (Blumer et al., 2004a, 2005), our ultrastructural data clearly demonstrate that no endothelium borders cartilage canals contradicting the light microscopical observations in the rat humerus where a continuous endothelial cell lining has been described (Morini et al., 2004). During canal formation, macrophages are detectable at the apical tips. These cells resorb the cartilaginous matrix and open the lacunae of adjacent chondrocytes. Previous studies have reported similar results where mononuclear or TRAP-positive cells (tartrate-resistent acid phosphatase) are present in the same position (Cole and Wezeman 1985; Chappard et al., 1986; Lee et al., 2001; Álvarez et al., 2005a, b). The resting chondrocytes immediately in front of the apical tips show morphological evidence of cell death and also display active caspase 3 reactivity, thus confirming previous observations in the rat tibia where chondrocytes seemed to degenerate during canal formation (Lee et al., 2001). Here, it is concluded that the macrophages enable the advancement of the canals within the chondroepiphysis and this process is coupled with programmed cell death of several chondrocytes. This early step of excavation is induced by the expression of several matrix metalloproteinases (MMPs) that cleave components of the hyaline cartilage matrix (Davoli et al., 2001; Lee et al., 2001; Álvarez et al., 2005a; Melton et al., 2006) and this is illustrated by a weakly toluidine blue stained layer immediately in front of the canal tips. Expression of MMPs is a key event during early canal formation (Álvarez et al., 2005a) and mice, deficient in MMPs, show both severely impaired vascularization and delayed ossification of chondroepiphysis (Holmbeck et al., 1999; Zhou et al., 2000).

In mice 5 and 8 days of age, few cartilage canals are present and type I collagen expression is low, presuming that the canal cells mainly contribute to elongation of the vessels up to 10 days after birth. At this point in time, the number of canals has increased, they are highly branched within the hypertrophic cartilage, and the first features of endochondral bone formation are obvious. Furthermore, our results demonstrate that canal mesenchymal cells highly express type I collagen at this developmental stage and this labeling pattern can be traced back to the sites of the ossification nuclei where probably the canal cells have transformed into osteoblasts. Similarly, in the human, calcaneus type I collagen was distributed in those canals that are connected with the ossification centre (Fritsch et al., 2001). In contrast, this labeling pattern has not been noted within the normal as well as clubfoot calcaneus (Gilbert et al., 2001). However, in the chicken femur cartilage canal, mesenchymal cells unambiguously express various bone-relevant markers and differentiate from preosteoblasts into osteocytes on their course into the SOC (Blumer et al., 2006). In avian, the canals are more numerous than in mice and are arranged in a complex network. They appear early, before hatching, and the time interval between canal formation and the first signs of endochondral ossification (after hatching) corresponds to approximately 13 days. Furthermore, cartilage canals are barely branched within the hypertrophic zone and only one ossification nucleus is formed, expanding slowly in a radial pattern (Lutfi, 1970; Doménech-Ratto et al., 1999; Blumer et al., 2004a, b, 2005; Eslaminejad et al., 2006). In mice, however, cartilage canals are formed after birth, the enlargement of the marrow cavity proceeds quickly with the first features of ossification 5 days after the canals' development. Initially, several small ossification nuclei are present, which finally coalesce, and at 18 days of age, a large SOC is established in the femoral epiphysis. These results are similar to other studies on murine long bones (Cole and Cole, 1989; Holmbeck et al., 1999; Álvarez et al., 2005a, b); however, the formation of small ossification nuclei prior to the establishment of the SOC is novel. Taken together, former investigations as well as the present findings indicate that the epiphysis in murine species ossifies more quickly than in avian. Furthermore, as in chicken, the canal mesenchymal cells express type I collagen, consistent with the notion that they might be a source of bone-forming cells. Accordingly, in both species, regression of cartilage canals is never observed and thus all canals can contribute to the ossification process of the epiphysis.

The Fate of Resting Chondrocytes During the Formation of Cartilage Canals

Conflicting results have been published regarding the fate of these chondrocytes in the rat tibia. Whereas Lee et al. (2001) assumed that at the blind end of canals a distinct layer of the resting cartilage is characterized by dying chondrocytes, Álvarez et al. (2005b) could not find signs of cell death either in the same position nor around the canal lateral walls. However, we have identified two morphologic-distinguishable types of resting chondrocytes immediately around cartilage canals: on the one hand, a light chondrocyte that seems viable and, on the other hand, a dark chondrocyte (sometimes both types within one lacuna) showing unambiguous ultrastructural evidence for degeneration. Both cell types, frequently forming a doublet, were also found in the epiphysis of chicken and rabbit long bones and this was interpreted as asymmetric cell division where one daughter cell remains viable whereas the other appears to disintegrate (Roach et al., 1995; Roach and Clarke, 1999).

Our results, derived from immunohistochemistry using active caspase 3 as a usual marker for programmed cell death demonstrate that several resting chondrocytes label positively, matching the morphological data. Until D 10, only several chondrocytes die. However, we cannot exclude that during advancing development their number increases. Roach et al. (2004) regarded dying chondrocytes as “chondroptotic” cells that are dependent on complete self-distraction rather then phagocytosis. In the mouse femur, several degenerating chondrocytes around the canal lateral walls are trapped in their lacunae and macrophages are not located in these areas thus excluding phagocytosis. These cells show patches of condensed chromatin throughout the nucleus and additionally abundant rough endoplasmic reticulum with indistinguishable ribosomes is encountered. These morphological features are considered by Roach et al. (2004) as evidence for chondroptosis. However, resting chondrocytes immediately in front of the canal tips show margination and coalescence of the nuclear chromatin, they contain apoptotic bodies and are in close proximity with macrophages and thus their remnants can be phagocytosed. Phagocytosis is an essential requirement for classical apoptosis (Roach and Clark, 1999) and summarizing both the present and previous findings, we conclude that the dark chondrocytes surrounding cartilage canals can undergo two pathways of programmed cell death, namely chondroptosis and apoptosis, respectively. In both processes, caspase 3 finally seems to be the executor of cell death.

Programmed cell death, on the other hand, appears not to be the fate of light resting chondrocytes. Immunohistochemical analysis of type II collagen distribution shows that these cells are occasionally released from the adjacent cartilage matrix into the canal cavity. This could be found near the canal walls as well as the blind ends and these results were only obtained when labeled consecutive resin sections (2 μm) were examined. We, however, were unable to obtain similar results from consecutive paraffin sections with a thickness of 6 μm and, thus, the former approach allows a more detailed analysis. The present findings are consistent with the observations in the rat tibia (Álvarez et al., 2005b); however, these authors suggested that freeing of chondrocytes occurs only at the canal apical tips during a short period of time (D 5) and, hence, this process seems to be both temporally and regionally highly restricted. In contrast, in mice the release of chondrocytes is also detectable near the canal lateral walls and additionally occurs in later developmental stages (D 10). Considering the fate of these cells, it was speculated that they may re-enter an undifferentiated cell cycle and subsequently develop into bone-forming cells (Álvarez et al., 2005b). In the present study, we show that type I collagen is detectable within those canal segments of the resting cartilage from which several chondrocytes are freed into the canal cavity and, moreover, in situ hybridization provided evidence that all canal cells express type I collagen. Additionally, our ultrastructural observations suggest that released chondrocytes remain viable. Therefore, our data obviously strengthen the assumptions of Álvarez et al. (2005b). Similarly, it has repeatedly been assumed that hypertrophic chondrocytes in close distance to bone-forming areas have the capacity to synthesize type I collagen and may differentiate into osteoblasts (Roach and Shearer, 1989; Galatto et al., 1994; Roach et al., 1995; Roach 1997, 1999). We also noted that a few hypertrophic chondrocytes expressed type I collagen mRNA and a distinct layer of this molecule is present within their lacuna. However, these lacunae are then opened by chondroclasts and, therefore, we conclude that the osteoblast-like stage of hypertrophic chondrocytes is only a short-time event confirming our previous observations in chicken (Blumer et al., 2005).

In conclusion, in the distal femoral epiphysis of mice, bone development advances quickly and at first several small ossification nuclei originate, which then coalesce into a large SOC. We demonstrate that cartilage canals are formed clearly prior to the establishment of the SOC, their mesenchymal cells have an osteogenic potential, and thus these canals play a pivotal role in endochondral bone formation. Furthermore, all canals seem to be involved in the ossification process. Finally, canal advancement is coupled with the death of several resting chondrocytes but others are released into its cavity and may enter an osteogenic pathway. We believe that the present data on normal bone formation may be helpful for future histological studies on disturbed bone development such as the human clubfoot.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Mice were obtained from the central laboratory animal facilities of the Innsbruck Medical University. Developmental stages day (D) 4, 5, 8, 10, and 18 after birth were used and two animals per age group were investigated in this study. Mice were anesthetized with CO2, and killed by cervical displacement. Subsequently, legs were amputated and the femur was isolated from the tibia. The distal epiphysis of the femur was examined.

Light and Transmission Electron Microscopy

The bones of mice were fixed in 2.5% glutaraldehyde, 2% paraformaldehyde buffered in sodium cacodylate (0.1 M), pH 7.4, for 4 hr at room temperature and rinsed in the same buffer. They were postfixed in 0.5% osmium tetroxide, 1% potassium hexacyanoferrat III in distilled water overnight at 4°C, rinsed in distilled water and decalcified in 3% ascorbic acid in sodium chloride (0.15 M) for 12 hr at 4°C. This was followed by dehydration in graded ethanol series and embedding in Spurr's epoxy resin. Ribbons of consecutive semithin sections (2 μm) were cut on a Reichert Ultracut S microtome (Leica Microsystem, Wetzlar, Germany) with a histo-jumbo-diamond knife (Diatome, Biel, Switzerland) (Blumer et al., 2002) and stained with toluidine blue for 20 sec at 60°C. A complete series of ultrathin sections (80 nm) were cut on a Reichert Utracut S microtome with an ultra-diamond knife, mounted on dioxan-formvar coated copper slot-grids, and stained with an aqueous solution of uranyl acetate (1%) for 30 min at 20°C followed by lead citrate for 3 min at 20°C. Ultrathin sections were examined with an electron microscope 10A (Zeiss, Oberkochen, Germany).

3-D Reconstruction

The distal femoral epiphysis of a 5- and 10-day-old mouse, respectively, was reconstructed. Ribbons of complete series of semithin cross sections were made and stained as described before. Each tenth section was photographed as a color image using a Zeiss AxioCam HR (Zeiss, Oberkochen, Germany) and AxioVision 4.1. software running on a Pentium 4 (Intel Inc., Santa Cruz, CA) with WindowsXP (Microsoft Inc., Redmond, WA). For 3-d reconstruction, SurfDriver software (Surfdriver, Kailua, HI) was used, running on an Apple Macintosh (Apple, Cupertino, CA). The high-definition TIFF files were downscaled to 1,024×768pixel/inch, converted into PICT with GraphicConverter (Lemke Software, Peine, Germany), and imported into SurfDriver. The perichondrium was used as reference point. First, the contour of the epiphysis was outlined, and then canals were reconstructed. In 10-day-old mice, “contour selector” was used to outline the ossification nuclei. After adjusting and rendering, the 3-d models were aligned according to the orientation of the histological longitudinal sections through the epiphysis. After re-conversion of the 3-d models from PICT to TIF with GraphicConverter, Adobe PhotoShop (Adobe Inc, San Jose, CA) was used to perform post-processing and to generate high-resolution images.

Tissue Preparation for Immunohistochemistry and In Situ Hybridization

Bones of mice were fixed with 4% paraformaldehyde in phosphate buffer saline (PBS, 0.1 M), rinsed in the same buffer, and decalcified as described before. Subsequently, the bones were dehydrated in graded isopropanol and xylene series and embedded in paraffin or Spurr's epoxy resin. Paraffin serial sections (6 μm) were made on a HM 355S microtome (Microm, Walldorf, Germany) and for histology several sections were stained with Masson's trichrome. Ribbons of consecutive semithin sections (2 μm) of resin-embedded samples were made as mentioned before. All sections were collected on SuperFrost®Plus slides.

Immunohistochemistry

For type I collagen, rabbit anti-human collagen type I (1:100 in antibody diluents) (LF-67 from Prof. L. Fisher, National Institute of Health, Bethesda, MD) and for type II collage, rabbit anti-human collagen type II (1:100 in antibody diluents) (catalogue no. CL50211AP, Cedarlane, Ontario, Canada) were used. For both types of collagen, immunohistochemistry was performed on histological paraffin (6 μm) as well as semithin resin sections (2 μm). Previous studies have demonstrated that both types of collagen can be traced on resin sections (Mizoguchi et al., 1992; Blumer et al., 2004a, 2005). Histological sections were deparaffinized and semithin sections were treated with 3% NaOH in 100% ethanol for 2 min to dissolve the resin. The following procedure was the same. Sections were rinsed in PBS and endogenous peroxidase activity was blocked with 0.5% H2O2 in 30% methanol for 15 min in the dark. The subsequent protocol comprised a proteolytic digestion step using protease 1 (Ventana, Strasbourg, France) for 10 min, primary antibody (type I and type II collagen antibody) incubation overnight at 4°C, and saturation of unspecific sites with 10% normal goat serum (NGS). This was followed by application of a secondary antibody (goat anti-rabbit IgG/HRP-conjugated 1:500 in PBS) (catalogue no. P0448, DakoCytomation, Glostrup, Denmark) for 3 hr at room temperature, washing in PBS and chromogenic detection of the antigen-antibody complex with 0.05% diaminobenzidine (DAB), 0.01% H202 in distilled water in the dark for approximately 10 min.

To evaluate programmed cell death, goat anti-human active caspase 3 (catalogue no. sc-1226 Santa Cruz Biotechnology, Inc., Santa Cruz, CA) was used. The sections were deparaffinized, rinsed in PBS, and endogenous peroxidase activity was blocked as described before. Sections were then washed in PBS and heat-induced epitope/antigen retrieval with citric buffer (1:10, pH = 6.0) was performed for 5 min in a microwave oven (400 Watt). After a cooling step for 20 min, sections were washed in PBS, incubated with caspase 3 antibody (1:50, 0.1% Triton X in antibody diluents) overnight at 4°C, rinsed, and then exposed to a secondary antibody (rabbit anti-goat IgG/HRP-conjugated 1:1,000 in PBS) (catalogue no. P0449, DakoCytomation, Glostrup, Denmark) for 3 hr at room temperature. Antibody binding sites were detected as described before.

All sections were counterstained with Gils' haematoxylin. Negative controls on both paraffin and resin sections were obtained by substituting the primary antibody with antibody diluents. These sections yielded no labeling.

mRNA In Situ Hybridization

The following HybriProbe was used: mouse alpha 1 type I collagen (Colla1). The HybriProbe was composed of three target-specific ready 5′ and 3′ double FITC-labelled single-stranded phosphodiester DNA oligonucleotides (Biognostik, Göttingen, Germany) each of which had a length of 30 bases. The oligonucleotide sequence was based on the GenBank accession no. 007742.

Longitudinal sections (6 μm) from paraffin-embedded tissue were deparaffinized and heat-induced target retrieval was carried out as described before and in a recent study (Blumer et al., 2006). This was followed by prehybridization in a ready-to-use HybriBuffer ISH (Biognostik, Göttingen, Germany) for 3 hr at 30°C.

Subsequently, sections were hybridized overnight in a humid chamber at 30°C and, after that, a stringent wash treatment in 0.5× SSC for 30 min at 40°C was performed. The hybridized probe was detected by incubation with a goat anti-fluorescein/alkaline-phosphatase conjugated antibody (1:200 in PBS) (catalogue no. A 4843, Sigma-Aldrich, Taufkirchen, Germany) for 3 hr at room temperature and chromogenic reaction was obtained by using a ready-to-use nitroblue tetrazolium/5-bromo-4-chloro-indolyl phosphate substrate (NBT/BCIP) (DakoCytomation, Glostrup, Denmark) for 20 min at room temperature. Sections were not counterstained. Random control HybriProbes served as negative controls and revealed no labeling.

All sections were examined with a Zeiss Axioplan 2 (Zeiss, Oberkochen, Germany) and photographed as color images using Zeiss AxioCam HR and AxioVision 4.1. software running on a Pentium 4 (Intel Inc., Santa Cruz, CA) with WindowsXP (Microsoft Inc., Redmond, WA).

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

The authors thank Prof. H. Dietrich for kindly providing the material for this study. We gratefully acknowledge the provision of anti-type I collagen (LF-67) by Prof. L. Fisher (Bethesda, MD). We especially thank E. Richter for assistance in the laboratory, Prof. K. Pfaller for his valuable discussion, and T. Pérez for carefully reading and correcting the manuscript.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES