Morphogenesis of the Manubrium of Sternum in Human Embryos: A New Concept†
Article first published online: 19 NOV 2012
Copyright © 2012 Wiley Periodicals, Inc.
The Anatomical Record
Volume 296, Issue 2, pages 279–289, February 2013
How to Cite
Rodríguez-Vázquez, J. F., Verdugo-López, S., Garrido, J. M., Murakami, G. and Kim, J. H. (2013), Morphogenesis of the Manubrium of Sternum in Human Embryos: A New Concept. Anat Rec, 296: 279–289. doi: 10.1002/ar.22623
This article was published online on 19 November 2012. An error was subsequently identified. This notice is included in the online and print versions to indicate that both have been corrected 29 November 2012.
- Issue published online: 25 JAN 2013
- Article first published online: 19 NOV 2012
- Manuscript Accepted: 21 SEP 2012
- Manuscript Received: 25 MAY 2012
- manubrium of the sternum;
- sternal bands;
- human embryology
To revisit many theories on fetal development of the manubrium of the sternum, we examined 25 mid-term fetuses at 6–9 weeks of gestation. The initial developmental stage of the manubrium was characterized by a distinct interclavicular mesenchyme that was continuous with the developing clavicles. Because parts of the clavicle in which endochondral ossification occurs originate from the neural crest, the interclavicular mesenchyme seems to be of the same origin. The sternal bands, possibly of the lateral plate origin, were restricted at the anterior ends of the ribs in the paired thoracic walls. The interclavicular mesenchyme extended caudally and laterally to reach the anterior ends of the first ribs, and thus the interclavicular mesenchyme expanded into the intercostoclavicular mesenchyme. Then, the primitive manubrium was delimited by the sternoclavicular joint and its related ligaments, all of which developed from the interclavicular and intercostoclavicular mesenchymes. Although the first ribs were attached to the intercostoclavicular mesenchyme, the former was vimentin-negative in contrast to the latter, positive mesenchyme. Soon afterwards, the small upper end of the sternal bands was integrated into the intercostoclavicular mesenchyme to form the primitive manubrium. The infrahyoid muscles and their supplying nerves maintained a close topographical relation to the interclavicular or intercostoclavicular mesenchyme, whereas the pectoralis major muscle kept attachments to the sternal bands. Consequently, the manubrium of sternum appeared to develop in a complex way at a junction area between derivatives of the neural crest, lateral plate, and somite. Anat Rec, 2013. © 2012 Wiley Periodicals, Inc.
The manubrium of sternum corresponds to the wider, cranial part of the superior segment of the sternum, being attached to the body of sternum by a thin layer of cartilage (Testut and Latarjet, 1975; Williams and Warwick, 1985). It is well known that, in humans (Paterson, 1900; Hamilton and Mossman, 1975; Doskocil, 1993) as well as in higher vertebrates (Gladstone and Wakeley, 1932), the xiphoid process and the body of the sternum develop from a pair of mesenchymal tissue bands (i.e. the so-called “sternal band”) near the ventral end of anlagen of the ribs. These sternal bands develop independently of the ribs and fuse secondarily with the latter (Hanson, 1919; Gladstone and Wakeley, 1932; Fell, 1939; Seno, 1961; Klíma, 1968). The fusion of the sternal bands was demonstrated in cultured mouse embryos (Chen, 1952).
Although some of them may overlap, different theories have been proposed to explain development of the manubrium of sternum: (a) the manubrium originated from the sternal bands themselves (Hamilton and Mossman, 1975; Fig. 1A); (b) the medial ends of the clavicles bond with one another by a dense episternal band that is similar to the episternum of lower vertebrates and that is later joined to the sternal bands (Keibel and Mall, 1912; Fig. 1B); (c) the sternal bands are joined by a pair of mesenchymal condensations between the clavicles (Corliss, 1979; Larsen, 2003; Fig. 1C); (d) this scenario is similar to that of theory c but a mid-sagittal, ventral unpaired structure joins the sternal bands together (Klíma, 1968; Fig. 1D); (e) a single “interclavicular blastema” joins the sternal bands (O'Rahilly and Müller, 1986; Fig. 1E). Therefore, the different points between theories appear to coincide on (1) whether the manubrium originates from the sternal bands themselves or from another mesenchymal condensation and; (2) whether or not parts of the clavicles contribute to the manubrium formation. Depending on the variety of theories and according to knowledge from the comparative anatomy of vertebrates (Gladstone and Wakeley, 1932; Cobb, 1937), various names have been given to rare variations in the human manubrium, such as the suprasternum or the episternum. Those theories of the fetal development may be based on a concept of homology found in the comparative anatomy.
Matsuoka et al. (2005) demonstrated that, in mice, the endochondral medial part of the clavicle as well as a part of the sternum originates from the post-otic neural crest and that the latter is attached to the lateral plate-derived parts of the sternum. The fact that the sternal abnormality is most likely to accompany the heart abnormality also suggests the common cue or origin from the neural crest (reviewed by Weston et al., 2006). By contrast, Valasek et al. (2011) demonstrated a common factor (Tbx5) required for development of not only all limb elements but also the sternum. Thus, the sternum is likely to carry dual origins. Simultaneously, development of the sternum is known to be regulated by Hox genes such as axial somite derivatives (Barrow et al., 1996; Manley et al., 2001; McIntyre et al., 2007). Therefore, the manubrium of sternum appears to be positioned at a border shared by three different types of mesenchymal tissues, that is, those originating from the lateral plate, the somite, and the neural crest. The variety in classical interpretations of the fetal development (see above) seems to be a result of the multiple possibilities of the origins and regulation.
Although a demarcation between mesenchymal condensations does not always indicate a difference in origin, we believe that a precise description on the early stage of the fetal manubrium is first needed for the further discussion, including classical theories as well as molecular biology with its recent advances. The topographical anatomy between a mesenchymal condensation and a specific muscle anlage, such as that of the infrahyoid muscles, would provide a clue for a better understanding (Saunders, 1948; Seno, 1961; Matsuoka et al., 2005; Valasek et al., 2011). Consequently, the aim of this study was to revisit the development of the manubrium of the sternum to clarify the confusion. Here is a hotspot at which the head and neck, trunk, and limbs interact, as well as trying to establish the corresponding homologies with the sternal apparatus of vertebrates, since this terminology has been constantly used to explain the development of the sternum in human.
MATERIALS AND METHODS
The study was performed in accordance with the Declaration of Helsinki 1995 (as revised in Edinburgh, 2000). We examined the paraffin-embedded histology of 25 mid-term embryos and fetuses at 6–9 weeks of gestation (greatest length or GL, 12.5–36 mm). The parameters used to determine the postconception age were the GL and external and internal criteria (O'Rahilly and Müller, 2010). The present paraffin blocks contained entire parts of the neck and thorax. All specimens, from a large collection of the Embryology Institute of the University Complutense of Madrid, were obtained from miscarriages and ectopic pregnancies from Department of Obstetrics of the same university. Approval for the study was granted by the ethics committee of the university. After routine procedure for paraffin-embedded histology, sections were cut horizontally with a thickness of 10 μm and at intervals of 50 μm. Most sections were stained with hematoxylin and eosin (HE) or azan staining, while some sections of larger specimens (9 weeks) were subjected to immunohistochemistry.
Vimentin, one of intermediate filament proteins, is known to be a marker of osteoprogenitor cells (Watanabe et al., 1993; Shapiro et al., 1995; Xian et al., 2004) and is expressed in the cell body and processes of osteoblasts and osteocytes (Shapiro et al., 1995). Our recent study (Abe et al., 2012) demonstrated ligament-specific distributions of vimentin-positive mesenchyme in the developing atlanto-occipital junction. To examine vimentin expression at and around the primitive manubrium, we conducted immunohistochemistry using some sections of the fetuses at 9 weeks. The primary antibodies used were mouse monoclonal anti-human vimentin (dilution, 1:10; Dako, Glostrup, Denmark) and mouse monoclonal anti-human striated-muscle myosin heavy chain (dilution, 1:100; Dako). Pretreatment with an autoclave was avoided because of the fragile nature of the fetal tissues. The secondary antibody (Dako Chem Mate Envison Kit; Dako) was labeled with horseradish peroxidase (HRP), and antigen–antibody reactions were detected via an HRP-catalyzed reaction with diaminobenzidine. Counterstaining with hematoxylin was performed on the same samples.
It was chosen the specimen with the highest degree of preservation for the presentation of the histological sections. The relative comparison of the level of sections of each figure during the development are shown in Table 1.
|I - CL - IL - M||A (left side) B(right side)||A||A||A||Panels of the sections|
|IC - M||C (right side)||B||B||B,D||A, B, C|
|IC - R1 - M||C||C||C, E|
|SB -R1||D (right side)||D||D||D|
|R1 - R2||D (left side)||E||F||E|
|R1 - R2 - R3||F|
Interclavicular Mesenchyme and Its Lateral Extension into the Thoracic Wall
The first subsection will describe observations at 6 weeks of gestation. At the level of the neck fold, a midline unpaired mesenchymal condensation appeared, this being continuous with the bilateral anlagen of the clavicle (Figs. 2 and 7A). We called this mesenchymal condensation the “interclavicular mesenchyme.” The anlagen of the sternothyroid and sternohyoid muscles as well as branches of the ansa cervicalis nerve were adjacent to the posterior aspect of the interclavicular mesenchyme. At this stage, the first and second ribs were restricted in the bilateral thoracic walls attaching to the lateral aspects of the developing pericardial cavity (Fig. 2). The sternal bands were not evident and identified as a spotty mesenchymal condensation attaching to the anterior ends of each of the ribs. In the immediately caudal side of the infrahyoid muscles, the pectoralis muscle anlage appeared, making an insertion into the sternal band (Fig. 2D).
Slightly afterwards, the interclavicular mesenchyme extended caudally and laterally, approached the anterior ends of the first ribs, and faced the pericardial cavity: thus, the interclavicular mesenchyme expanded into the intercostoclavicular mesenchyme (Figs. 3 and 7A). However, the caudal extensions of the midline mesenchyme were still separate from either the first rib or the sternal band by a loose tissue. The infrahyoid muscle anlagen maintained the close topographical relation to the intercostoclavicular mesenchyme, whereas the pectoralis muscle anlage to the sternal bands. The intercostoclavicular mesenchyme and the infrahyoid muscle anlagen were arranged as a semicircle on the top of the pericardial cavity.
Midline Fusion of the Intercosclavicular Mesenchyme in the Anterior Thoracic Wall
The second subsection will describe observations at 7 weeks. The endochondral ossification started in the medial ends of the clavicles. Between the medial ends of the clavicles, the interclavicular mesenchyme was more condensed and better delimited. Now the original interclavicular mesenchyme joined the medial ends of the first ribs by the intercostoclavicular mesenchyme (Figs. 4 and 7B). Thus, the mesenchyme did not face the pericardial cavity but rather was connected between the bilateral first ribs. Below the first rib, the sternal band connected between the anterior rib ends to make the bilateral chain-like arrangement. With the exception of the first ribs, the bilateral chains of the rib ends were separated by a loose tissue that was immediately beneath the surface ectoderm.
However, between lower rib ends, the midline area was still open and occupied by an ectoderm-like tissue. The border became unclear between the intercostoclavicular mesenchyme and the first rib in the midline area and both structures appeared to make a complex, although at the lateral sites the rib was covered by a primitive perichondrium. Thus, the sternal bands were identified as bilateral continuous bands connecting the anterior rib ends and reaching the intercostoclavicular mesenchyme-first rib complex (Figs. 4D and 7B). The intercostoclavicular mesenchyme-first rib complex maintained a close relation to the infrahyoid muscle anlagen, whereas the sternal bands kept a tight attachment to the pectoralis major muscle.
Delimitation of the Sternocostoclavicular Joint in a Homogeneous Interzone Phase and Delimitation of the Primitive Manubrium
The third subsection will describe observations at 8 and 9 weeks. The manubrium of sternum became clearly delimited by the developing sternoclavicular joint (Figs. 5 and 7C). Simultaneously, the interclavicular ligament (Fig. 5A) as well as the posterior sternoclavicular ligament (Fig. 5D) was identified as an independent mesenchymal condensation. The former was located among the anlage of the primitive manubrium and the clavicles and the first ribs. The posterior sternoclavicular ligament separated the primitive manubrium from the infrahyoid muscles. Caudally to the anlage of the interclavicular ligament, the primitive manubrium maintained a unique, rounded morphology and was connected to the first ribs by the intercostoclavicular mesenchyme (Fig. 5). The sternocostal head of the pectoralis major muscle did not attach to the primitive manubrium of sternum. Likewise, the clavicular and sternocostal heads of this muscle were clearly separate (Fig. 5A,C). However, the sternocostal head of the pectoralis muscle attached to the anterior aspect of the first rib. The first ribs were connected with the second ribs by the bilateral sternal bands which were still separate along the midline.
The primitive manubrium of the sternum was in a cartilaginous phase with a wide cranial part and its lateral elevations. The intercostoclavicular parts of the primitive manubrium were completely fused on its caudal part, precisely in the area of continuity with the first ribs (Fig. 5E). The contact of the cranial ends of the sternal band was clearly discriminated from the second ribs by the perichondrium of the latter (Figs. 5F and 6F). The band showed a smooth and round margin facing to the rib in contrast to finger-like small protrusions at the anterior ends of the ribs (Fig. 6F). The sternocostal head of the pectoralis major muscle inserted into the anterior aspect of each sternal band (Fig. 5F).
The intercostoclavicular parts of the primitive manubrium appeared to correspond to the vimentin-positive area (Fig. 6B,C) in contrast to the vimentin-negative ribs. The vimentin-positive mesenchymal condensation made a pair of posterior protrusions possibly corresponding to the putative posterior sternoclavicular ligaments (Fig. 6A). In the caudal site, the intercostoclavicular mesenchyme-first rib complex of the primitive manubrium was involved in the cranially developing sternal bands. Thus, the sternal bands appeared to reach the first rib (Fig. 6E,F). Subsequently, the first and second ribs were tightly connected. Muscle fibers of the pectoralis major muscle were vimentin-positive at and near the attachment to the primitive manubrium and the sternal band.
This study demonstrates that the initial and major part of the manubrium of sternum developed from the interclavicular mesenchymal condensation continuous with the bilateral clavicles (Fig. 7A). This fact was consistent with O'Rahilly and Müller (1986), although they described an interruption of the tissue along the midline. The present interclavicular mesenchyme was characterized by having (1) no connection with either the first ribs or the sternal bands in the initial stage and (2) a close topographical relationship with the developing infrahyoid muscles and their supplying nerves. Because parts of the clavicle in which endochondral ossification occurs originated from the post-otic neural crest (Matsuoka et al., 2005) and because of the close relation with the infrahyoid muscle anlagen, the interclavicular mesenchyme appears to be of the neural crest origin. Therefore, major parts of the manubrium and clavicle, these together showing a semicircular ring-like arrangement on the top of the pericardial cavity, seem to develop in relation with the neural crest migration into the heart. The midline cleft of the interclavicular mesenchyme (O'Rahilly and Müller, 1986; see above) may be combined with a failure of midline fusion of the branchial arches (Seyhan and Kýlýnr, 2002; Miynarek et al., 2003). Subsequently, the caudal and lateral part of the interclavicular mesenchyme approached the first ribs before the anterior body-wall closure. Thus, on both sides of the pericardial cavity, the interclavicular mesenchyme expanded into the intercostoclavicular mesenchyme (Fig. 7A). At this early stage, the sternal bands were still separate from the uppermost thoracic-wall mesenchyme. The intercostoclavicular mesenchyme also maintained a close relationship with the infrahyoid muscles dorsally, but not with the pectoralis muscle, which attached to the sternal bands.
The second stage (Fig. 7B) was characterized by a midline fusion of the intercostoclavicular mesenchyme. Manley et al. (2001) demonstrated an unsuccessful fusion of the bilateral anlagen of the sternum in hoxb4 mutant mice. Likewise, in AP-2alpha transcription factor knock-out mice, the unsuccessful midline fusion occurs because the abdominal band (rectus muscle anlagen) as well as the covering abdominal epidermis fails to develop (Brewer and Williams, 2004). These references suggest that a midline fusion of the sternal body is regulated by somite derivatives. The first rib, of the somite origin, is likely to regulate the development of the manubrium possibly via mechanical support for the intercostoclavicular mesenchyme, since from our observations the intercostoclavicular formations continued with the first ribs laterally and there is no a demarcation or discontinuity which separates them. Actually, even in adults, the attachment morphology of the first rib to the manubrium differs from that of the other ribs to the sternal body (Testut and Latarjet, 1975; Williams and Warwick, 1985; Doskocil, 1993). Likewise, Barchilon et al. (1996) showed that the ossification of the first costal cartilage chronologically differed from that of others.
The most difficult point was to show how the intercostoclavicular mesenchyme was involved into the sternal band. In contrast to Keibel and Mall (1912) and O'Rahilly and Müller (1986), we found no direct contribution of the sternal bands to a formation of the manubrium (see Fig. 1 B,E). Likewise, in contrast to Hamilton and Mossman (1975), our results do not support a classical theory that the sternal bands themselves extend cranially to provide the interclavicular mesenchyme (Fig. 1A). The intercostoclavicular mesenchyme was different from the sternal bands: the bands curved dorsally in a hook-like shape and were likely to appear after establishment of the interclavicular mesenchyme. As described by Paterson (1900) and Gladstone and Wakley (1932), and Hanson (1919), the sternal bands seemed to be formed by the joining of mesenchymal condensations adjacent to the anterior ends of the ribs. The sternal bands are characterized by their attachment to anlage of the sternocostal portion of the pectoral major muscle (Gladstone and Wakeley, 1932): this morphology has been used as evidence to support the appendicular origin of the sternum in experimental embryology (Saunders, 1948; Seno, 1961; Valasek et al., 2011). The fact that the sternal bands and pectoralis muscle did not reach the interclavicular mesenchyme in the initial stage also seems to suggest different origins of the manubrium and body of the sternum. In addition, we did not detect either an independent, midline anlage of the manubrium (Gladstone and Wakeley, 1932; Klíma, 1968; Corliss, 1979; Larsen, 2003; Carlson, 2009; Fig. 1) or an intermediate structure between the clavicles and the first ribs (O'Rahilly and Müller, 1986; Fig. 1E).
The third developmental stage of the manubrium had a clear demarcation of the sternoclavicular joint from the anlage of the manubrium. This is not consistent with Gladstone and Wakeley (1932), who stated that the sternoclavicular joint develops from two lateral condensations adjacent to the medial ends of the clavicles. At this stage, the interclavicular ligament as well as the posterior sternoclavicular ligament become evident. We considered the primitive sternoclavicular joint to be a homogeneous interzone phase (O'Rahilly and Müller, 1986) surrounded by the manubrium, clavicles, and first rib. When the sternoclavicular joint becomes clear, the fusion reaches the caudal margin of the intercostoclavicular mesenchyme of the primitive manubrium Therefore, the small part of the cranial end of the sternal bands, that joins the first and second ribs, seem to be integrated into the final manubrium. A junction between the adult manubrium and body of the sternum seems not to correspond to a border between the intercostoclavicular mesenchyme and sternal band but rather to a line across the cranial part of the latter. The present discussion proposes the way in which partial absence of the sternum occurs (Steiner et al., 1976; Cottrill, 1998).
From a phylogenetic perspective, the interclavicular mesenchyme was very similar to that in rodents and probably represents the vestige of the interclavicle (Gladstone and Wakeley, 1932). The rodent manubrium might correspond to the lower part of the human manubrium because of the interclavicle independent of the sternum. The interclavicular mesenchyme might be different from the so-called episternum (Keibel and Mall, 1912; Klíma, 1968). Otherwise, the topographical anatomy of the human intercostoclavicular mesenchyme seems to be analogous to the bilateral epicoracoid structures of the sternal apparatus of the lowest mammals (Gladstone and Wakeley, 1932; Cobb, 1937). In vertebrates with an independent interclavicular or suprasternal bone, the neural crest may not contribute to the formation of the manubrium itself. A remodeling of the superior margin of the manubrium has been hypothesized to have given rise to suprasternalis structures in humans (Klíma, 1968; Carlson, 2009; Fig. 1). However, according to the present observations, they were not independent but a part of the manubrium of the interclavicular mesenchymal origin. In this study, the small upper part of the sternal bands was integrated into the final manubrium (see above paragraph). In the sternal apparatus in lower mammals, this upper end of the bands appears to correspond to the so-called presternum that receives the first and second ribs (Gladstone and Wakeley, 1932). Taking all the evidence together, we conclude that, if phylogenetic terms are available for embryologic descriptions, the development of the manubrium starts from a phase of the “interclavicle,” expands the area into the “epicoracoid,” and joins the “presternum” to constitute the final morphology.
- 2012. Fetal development of the transversel atlantis and alar ligaments at the craniovertebral junction. Clin Anat 25: 714–721. , , , , , , .
- 1996. Factors affecting the rate and pattern of the first costal cartilage ossification. Am J Forensic Med Pathol 17: 239–247. , , , , , , , .
- 1996. Targeted disruption of the Hoxb-2 locus in mice interferes with expression of Hoxb-1 and Hoxb-4. Development 122: 3817–3828. , .
- 2004. Loss of AP-2alpha impacts multiple aspects of ventral body wall development and closure. Dev Biol 267: 399–417. , .
- 2009. Embriología humana y bilogía del desarrollo. 4a ed. Barcelona: Elsevier España S.L. .
- 1952. Studies on the morphogenesis of the mouse sternum. I. Normal embryonic development. J Anat 86: 373–386. .
- 1937. The ossa suprasternalia in whites and American negroes and the form of the superior border of manubrium sterni. J Anat 71: 245–291. .
- 1979. Embriología Humana de Patten. Fundamentos del Desarrollo Clínico. Buenos Aires: El Ateneo. .
- 1998. Sternal defects associated with congenital pericardial and cardiac defects. Cardiol Young 8: 100–104. , , .
- 1993. Contribution to the study of the development and ossification of human sternum. Funct Dev Morphol 3: 251–257. .
- 1939. The origin and development mechanics of the avian sternum. Phil Trans R Soc Lond 229: 407–463. .
- 1932. The morphology of the sternum and its relation to the ribs. J Anat 66: 508–564. , .
- 1975. Embriología Humana. 4a ed. Buenos Aires: Intermédica. , .
- 1919. The ontogeny and phylogeny of the sternum. Am J Anat 26: 41–115. .
- 1912. Manual of human embryology. Philadelphia: JB Lippincott Company. Vol. I. , .
- 1968. Early development of the human sternum and the problem of homologization of the so-called suprasternal structures. Acta Anat 69: 473–484. .
- 2003. Embriología Humana. Madrid: Elsevier España S.A. .
- 2001. Hoxb2 and hoxb4 act together to specify ventral body wall formation. Dev Biol 237: 130–144. , , , .
- 2005. Neural crest origins of the neck and shoulder. Nature 436: 347–355. , , , , , , , .
- 2007. Hox patterning of the vertebrate rib cage. Development 134: 2981–2989. , , , , , , .
- 2003. Congenital mid-line cervical cleft: case report and review of literature. Int J Pediatr Otorhinolaryngol 67: 1243–1249. , , , .
- 1986. Human embryology and teratology. 2nd ed. New York: Wiley-Liss. , .
- 2010. Developmental stages in human embryos: revised and new measurements. Cells Tissue Organs 192: 73–84. , .
- 1900. The sternum: its early development and ossification in man and mammals: preliminary communication. J Anat Physiol 35: 21–32. .
- 1948. The proximo-distal sequence of origin of the parts of the chick wing and the role of the ectoderm. J Exp Zool 108: 364–403. .
- 1961. The origin and evolution of the sternum. Anat Anz 110: 97–101. .
- 2002. Median cleft of the lower lip: report of two new cases and review of the literature. Ann Otol Rhinol Laryngol 111: 217–221. , .
- 1995. Transmisson electron microscopic demonstration of vimentin in rat osteoblast and osteocyte cell bodies and processes using the immunogold technique. Anat Rec 241: 39–48. , , , .
- 1976. Absent mesosternum in congenital heart disease. Am J Roentgenol 127: 923–925. , , .
- 1975. Tratado de Anatomía Humana. 9 th ed. Barcelona: Salvat Editores. , .
- 2011. Cellular and molecular investigations into the development of the pectoral girdle. Dev Biol 357: 108–116. , , , , , , , , , , , , , , .
- 1999. A hypomorphic myogenin allele reveals distinct myogenin expression levels required for viability, skeletal muscle development, and sternum formation. Dev Biol 208: 44–55. , , , .
- 1993. Immunohistochemical study of the cytoskeleton of osteoblasts in the rat calvaria. Intermediate filaments and microfilaments as demonstrated by detergent perfusion. Acta Anat 147: 14–23. , , .
- 2006. Thoracic skeletal defects and cardiac malformations: a common epigenetic link? Birth Defects Res C Embryo Today 78: 354–370. , , .
- 1985. Gray Anatomía. Barcelona: Salvat Editores. , .
- 2004. Intramembranous ossification mechanism for bone bridge formation at the growth plate cartilage injury site. J Orthop Res 22: 417–426. , , , .