Matrix Vol. 1111991, pp. 339-346 © 1991 by Gustav Fischer Verlag, Stuttgart
Distribution of Cartilage Proteoglycan (Aggrecan) Core Protein and Link Protein Gene Expression during Human Skeletal Development S. MUNDLOS 1 , R. MEYER 1 , Y. YAMADA 2 and B. ZABEL 1 I
2
Department of Pediatrics, Johannes-Gutenberg-Universitat, Mainz, F. R. G. and Laboratory of Developmental Anomalies, National Institute of Health, Bethesda, MD, USA.
Abstract The distribution of cartilage proteoglycan core protein (aggrecan) and cartilage proteoglycan link protein was investigated by in situ hybridization during different stages of human skeletal development. Aggrecan and link protein expression were confined to chondrocytes of the developing skeleton and other cartilaginous structures. Distribution and intensity of the signal was identical with aggrecan as compared to link protein probes. Parallel to the calcification of cartilaginous matrix, chondrocytes of this area lost the expression of aggrecan and link protein specific mRNA and stayed negative throughout the following stages of skeletal development. Highest expression was found in the lower proliferative and upper hypertrophic zone whereas the resting zone showed less expression. Aggrecan gene expression was additionally investigated in iliac crest biopsies of 3 patients with pseudo achondroplasia and compared to age-matched controls. Distribution and intensity of staining revealed no abnormalities. Thus, the phenotypic changes during chondrocyte maturation are accompanied by distinct changes in aggrecan and link protein gene expression. This pattern was maintained in the growth plate of patients with pseudoachondroplasia. Key words: aggrecan, cartilage proteoglycan, core protein, link protein, skeletal development. Introduction The development of the human skeleton involves a number of transient stages characterized by specific histological findings and the production of stage-specific components of extracellular matrix (ECM). The cartilaginous anlage of the future bone results from a transient cellular condensation process and the subsequent switch of gene expression from mesenchymal to cartilage specific ECMmolecules (von der Mark, 1980). During the following developmental stages, the chondrocyte undergoes a developmental change in phenotypic expression. The proliferation and differentiation into maturing and degenerating chondrocytes is accompanied by progressive alterations in the composition and organisation of ECM. Little is known about the associated transcriptional events and their regulation. In order to understand the mechanisms of
chondrocyte differentiation during human osteogenesis, the temporal and spatial patterns of chondrocyte gene expression have to be defined. Cartilage proteoglycan core protein (aggrecan) and cartilage proteoglycan link protein are some of the major noncollagenous components of cartilage ECM. Biochemical studies performed mainly on bovine tissue have established many of the properties of aggrecan, including the formation of large aggregates with hyaluronic acid via an amini-terminal globular region and its extensive modification with chondroitin sulfate, keratan sulfate and other oligosaccharides (Hascall and Hascall, 1981). Link protein binds simultaneously to the hyaluronic acid binding region of the proteoglycan molecule and hyaluronate and by this stabilizes the aggregate against dissociative forces (Hardingham and Muir, 1974). Both proteins are essential for the structural integrity of cartilage and may be involved
340
S. Mundlos et a1.
Fig. 1. Localization of aggrecan and link protein gene expression by in situ hybridization. Dark field images of sections through a 6-week humerus (b), a 7-week humerus (d) and an 8-week femur (f) with corresponding bright light images in a, c and e. Note lack of expression in areas of calcification (arrows).
Core/Link Protein Gene Expression in the formation of birth defects (Leonhard et aI., 1989) and/or heritable chondrodysplasias (Stanescu et aI., 1982). In this study we have used recently cloned cDNAs (Doege et aI., 1991) to investigate the temporal and spatial gene expression of aggrecan and link protein during human skeletal development from the early cartilaginous bone anlage to the formation of a regular growth plate in the newborn. In addition, biopsy specimens of apophyseal growth plate cartilage were investigated for aggrecan expression in patients with pseudoachondroplasia and compared to age matched controls.
Materials and Methods
Selection and preparation of hybridization probes
The cDNA for link protein for CPLP had a length of 1. 7 kb coding for the complete protein (Y. Yamada, unpublished data). Link protein contains a tandem repeat that functions as the hyaluronic acid binding site (Perin et aI., 1987). These domains share a certain degree of homology with the hyaluronic acid binding region of aggrecan (amino-terminal end of the molecule) at the amino acid level in the rat (Neame et aI., 1986; Doege et aI., 1987). Consequently, sub clones of the aggrecan cDNA (Doege et aI., 1991) coding for the carboxy-terminal end (1.4 kb) and for the serine/glycine repeat domain (2.7 kb) were chosen for the in situ experiments in order to rule out cross-reactivity. cDNA fragments of aggrecan and link protein were subcloned in Bluescript vectors and transcribed with high efficiency from the T 3 or T 7 promotors to make sense and antisense cRNA. Linearized plasmid DNA was transcribed in the presence of [35 Sj_UTP 800 !-lCi/mmol (Amersham, UK). After polymerization, the DNA templates were digested with DNAse for 15 min at 37°C. The RNA transcripts were hydrolyzed for 30 min in 0.2 N NaOH on ice to a final length of 50-200 bases. After neutralization with acetic acid, the probes were purified by several ethanol precipitation steps. Analytical acrylamide gel electrophoresis was used to check the transcript size before and after hydrolysis. The probes had a specific activity of 2 - 3 8 X 10 dpmhlg. Hybridization
Selection and preparation of tissue was performed as described earlier (Mundlos et aI., 1990). Biopsies of the apophyseal growth plate cartilage (iliac crest) were obtained from three patients with pseudoachondroplasia undergoing surgical procedures. Informed consent was obtained prior to surgery. The hybridization procedure was basically the same as described by Hogan (Hogan et aI., 1986) with a few modifications. The sections were baked onto the slides at 45°C on
341
a hot plate for 15 min followed by fixation in 4 % paraformaldehyde in PBS, 0.2 N HCI for 20 min, 2 x SSC 60°C for 20 min, digestion with pronase (Boehringer Mannheim, F.R.G.) at 0.125mglml for 10min at room temperature, acetylation in acetic anhydride 1:400 in 0.1 M triethanolamine for 10 min, followed by dehydration. The hybridization mixture contained the labelled RNA probe (0.5 !-lg/ml) in formamide 50%, dextran sulfate 10%, NaCl 0.3 M, 10mM Tris/HCI pH7, 10mM Na-P0 4 , pH5.5, EDTA 5 mM, Denhardt's 0.02 %, tRNA 0.5 mg/ml. Hybridization was carried out over night at 45°C in a humidified chamber. Post-hybridization washings were performed at 50 °C in 50% formamide and salt concentration as described above for 2 x 1 hour, followed by RNAse digestion at 37°C (20 [lg/ml RNAse, Sigma) for 30 min and 2 x SSC at 37°C for 2 x 1 hour. Slides were immersed in Kodak NTB-2 foto emulsion diluted 1:1 with water. Autoradiography was performed at 4°C in a dry chamber. Exposure time varied between 10 and 20 days. The exposed slides were developed, fixed, stained and examined using the Zeiss system for epipolarization.
Results
In order to optimize the ratio of signal to background, the probes were hybridized under various conditions to sections through 10-week limbs reflecting the different stages of chondrocyte maturation. No difference was observed in aggrecan expression with the probe coding for the serine/ glycine repeat domain as compared to the carboxy-terminal domain. No signal was obtained using the aggrecan or link protein sense probes. Aggrecan and link protein gene expression was identical throughout the different stages of skeletal development. No difference was observed in either intensity or distribution of the signal. The following figures therefore reflect aggrecan and link protein expression. An autoradiograph of a frontal section through a 6-week humerus hybridized to the aggrecan probe is illustrated in Fig. 1 b with a bright light image given in Fig. 1 a. At this stage, no calcification of cartilage or bone is present. At the diaphysis, a small band of hypertrophied cells can be seen. In the dark field, these cells show the largest number of grains/cell with decreasing intensity towards the ends of the bone model. Parallel to the development of a diaphyseal bone collar, the cartilaginous matrix of this area begins to calcify. Fig. 1 c shows a van Kossa stain of a 7 -week humerus with a small area of diaphyseal calcification (arrows). In contrast to earlier stages of development, a segregation of chondrocyte activity can be observed in those that express aggrecan and link protein and those that do not show expression. Fig. 1 d shows a band of negative chondrocytes in the area of calcification sharply separated from intensely positive
342
S. Mundlos et al.
Fig.2. Localization of aggrecan and link protein gene expression by in situ hybridization. Dark field images through a 13-week femur (b) and the distal femura! growth plate of a new born male (d). Bright light images are given in a and c. Note lack of expression in areas of calcification as marked by arrows. "Cartilage canal.
cells in the hypertrophic and proliferative zones. As in the earlier stages, a decreasing intensity of the signal towards the ends of the bone (resting cartilage) is apparent. With further growth, the area of calcified cartilage enlarges
(Fig. 1 e) and the area of negative chondrocytes greatly increases (Fig. 1£). Beginning with the 9th week of development, calcified cartilage is constantly removed and replaced by vasculate
Core/Link Protein Gene Expression
343
Fig. 3. Localization of aggrecan and link protein gene expression by in situ hybridization. Longitudinal section of the developing spine and cross-section of a vertebra in the lower right part. Note lack of expression in areas of calcification (large arrows) and expression in cartilaginous tissue of the trachea (small arrows).
344
S. Mundlos et al. and marrow. Fig. 2 a shows a section through the distal part of a 13-week femur. The area of calcified cartilage is reduced to a few cell layers as indicated by arrows. At this stage, we found aggrecan and link protein negative chondrocytes in and adjacent to the area of calcification (Fig. 2 b). The largest number of grains/cell was present in the upper hypertrophic/lower proliverative zone. At birth, a regular growth plate with resting cartilage as well as proliferative and hypertrophied chondrocytes arranged in columns has been formed at the metaphysis (Fig. 2 c). At this stage, gene expression of aggrecan and link protein was limited to the small area of hypertrophied and proliferative chondrocytes. Resting cartilage and cartilage close to the osseochondral junction (zone of calcification, arrow) showed no expression (Fig. 2 d). A similar picture was seen at the epiphyseal growth zone (not shown). Besides in developing long bones, aggrecan and link protein expression was also found in other cartilaginous tissues. Chondrocytes of the developing spine gave a strong signal. Fig.3 shows a longitudinal section through the developing vertebrae of a 8-week embryo with a transverse section given in the lower right part. Heavy labeling was found throughout the vertebrae except for the developing intervertebral discs and for the areas of beginning calcification (arrows). Fig. 3 also shows expression in the cartilaginous tissue of the trachea (arrows). The iliac crest biopsies of three patients with pseudoachondroplasia had a very similar picture with a severely distorted growth plate, almost complete loss of columnization and lack of regular areas of calcification at the osseo chondral border. The hypertrophic cells formed clusters that contained few cells and were separated by an excessive amount of ECM. Fig. 4 b shows the distribution of aggrecan gene expression in an iliac crest biopsy of a seven year old patient with the bright light image given in Fig. 4 a. In spite of the absence of the normal chondrocyte architecture, the pattern of aggrecan gene expression was identical to the control tissue (Fig.4 c, d). Again, hypertrophied and proliferating chondrocytes gave the strongest signal; whereas the resting zone showed less expression and calcified chondrocytes were negative.
Discussion
Fig. 4. Localization of aggrecan expression by in situ hybridization in the apophyseal growth plate (iliac crest) of a seven-year old patient with pseudo achondroplasia (a, b) and an age-matched control (c, d). Note strong expression in hypertrophied chondrocytes (arrows) and little expression in resting cartilage.
This study describes the temporal and spatial pattern of aggrecan and link protein gene expression during human skeletal development. We were able to demonstrate distinct differences in chondrocyte gene expression which correlate to specific phenotypes. Pseudo achondroplasia is among the most common skeletal dysplasias affecting both the metaphyseal and epiphyseal growth plate. Growth cartilage of patients with pseudoachondroplasia shows peculiar, finger-print-like inclusion bodies that were proposed to be proteoglycan
Core/Link Protein Gene Expression remnants (Stanescu et al., 1982). However, the primary defect is unknown. In a recent study, Finkelstein et al. (1990) were able to show that mutations in the aggrecan gene locus are not likely to be the primary defect causing pseudoachondroplasia. The present study supports this finding in that the distribution of aggrecan gene expression in iliac crest biopsies of patients with pseudoachondroplasia was identical to the pattern found in controls in spite of the distorted chondrocyte architecture. In the different tissues and developmental stages investigated, no difference in gene expression of aggrecan as compared to link protein was observed. The amount of specific mRNA as indicated by the number of grains/cell as well as the cell specific distribution of grains was not significantly different. Aggrecan and link protein have common binding sites for hyaluronic acid which share a 40% and 33 % homology at the amino acid level in the rat (Doege et al., 1987). Although there is a significant homology at the protein level, there is divergence at the nucleotide level (Doege et al., 1991) so that cross-hybridization can be ruled out under the stringent conditions used. The identical pattern of aggrecan and link protein gene expression is in agreement with the findings of Stirpe and Goetinck who showed that the expression of both mRNAs is initiated at the same stage of limb development in the chick (Stirpe et al., 1989). However, the synthesis of both matrix molecules seems not to be coordinately regulated. Recently, link protein specific transcripts were detected in the chick mesonephros (Stirpe et al., 1990) and in a human colon tumor cell line (Osborne-Lawrence et al., 1990). Besides in cartilage of the early bone anlage and the growth plate, aggrecan and link protein expression was found in cartilaginous tissue of the trachea, the larynx and the mid-face. Areas of desmal ossification like the calvaria showed no expression whereas Meckel's cartilage was positive. Parallel to the formation of hypertrophied chondrocytes, a change in the amount of specific mRNA per cell was observed. Chondrocytes of the upper hypertrophic and lower proliferative zone gave the strongest signal whereas the resting zone showed a uniform but much lesser degree of staining. This finding correlates to the rate of cell divisions which is highest in the proliferative zone (Kember, 1978). Cell division and growth is linked to the production of ECM, in this case aggrecan and link protein. Fig. 1 a shows the early cartilaginous anlage of the future bone with an even distribution of expression. Concurrently with the formation of the bone collar, the ECM of the newly formed hypertrophic chondrocytes begins to calcify. As Fig. 1 d shows, these cells have stopped to express aggrecan and link protein specific mRNA. The chondrocytes of the upper hypertrophic and proliferative zone, however, are strongly positive whereas the remaining resting cartilage shows less expression. This phenomenon of loss of expression in areas of calcification was found throughout the following stages of skeletal development. Similar results
345
were obtained when studying type II collagen gene expression (Mundlos et al., 1990). With the formation of hypertrophied chondrocytes, matrix vesicles are found in clusters in the ECM. These vesicles are thought to function as the initial site of calcification (Anderson, 1989). The large aggregating proteoglycan of cartilage is known to inhibit mineralization (Blumenthal et al., 1979) and thus may prevent calcification in the proliferative and upper hypertrophic zone. It was proposed that proteolytic degradation of proteoglycans occurs in the upper hypertrophic zone resulting in the transformation of non-calcifiable to calcifiable matrix. However, large amounts of aggrecan and link protein can be detected in calcified matrix immunohistochemically (Poole et al., 1982) or by x-ray microprobe analysis for sulfur (Shephard and Mitchell, 1985). Consequently, proteoglycan aggregates have to be structurally modified to lose their capacity to inhibit calcification and/ or other molecules are involved in the process of crystal formation. We recently suggested that chondrocytes of the zone of calcification selectively switch to the expression of genes that promote calcification. Type X collagen (Schmidt and Linsenmayer, 1985) and alkaline phosphatase (S. Mundlos, unpublished observations) are both expressed selectively in hypertrophied chondrocytes undergoing calcification. Both are assumed to playa role in the calcification of cartilaginous matrix. Our present finding that cessation of aggrecan/link protein gene expression is followed by mineralization supports this hypothesis. Accordingly, changes in the phenotype of maturing chondrocytes are accompanied by changes in gene expression. The regulatory events that promote such drastic alterations in ECM production remain uncertain and will be subject of further studies. Acknowledgements This work was supported by the Deutsche Forschungsgemeinschaft.
References Anderson, H. c.: Mechanism of mineral formation in bone. Lab. Invest. 60: 320-330, 1989. Blumenthal, N. c., Posner, A. S., Silverman, L. D. and Rosenberg, L. c.: Effect of proteoglycans on in vitro hydroxyapatite formation. Calcif Tissue Int. 27: 75 -82,1979. Doege, K., Sasaki, M., Horigan, E., Hassell,J.R. and Yamada, Y.: Complete primary structure of the rat cartilage proteoglycan core protein deduced from cDNA clones. J. Cell. BioI. 262: 17757-17767,1987. Doege, K., Sasaki, M., Kimura, T. and Yamada, Y.: Complete coding sequence and deduced primary structure of the human cartilage large aggregating proteoglycan, aggrecan: Humanspecific repeats and additional alternatively-spliced forms. J. BioI. Chern., in press, 1991. Finkelstein, J.E., Doege, K., Yamada, Y., Pyeritz, R.E., Graham, J.M., Moeschler, J.B., Pauli, R.M., Hecht, J.T. and Fran-
346
S. Mundlos et al.
comano, CA.: Analysis of the Chondroitin sulfate proteoglycan core protein (CSPGCP) gene in achondroplasia and pseudoachondroplasia. Am. I Hum. Genet. 48: 97-102, 1991. Hardingham, T.E. and Muir, H.: Hyaluronic acid in cartilage and proteoglycan aggregation. Biochem. I 139: 565 - 581,1974. Hascall, V. C and Hascall, G. K.: Proteoglycans. In: Cell Biology of Extracellular Matrix, ed. by Hay, E.D., Plenum, New York, 1981, pp. 39-63. Hogan, B., Constantini, F. and Lacy, E.: Manipulating the Mouse Embryo. Cold Spring Harbor Laboratory, New York, 1986. Kember, N.F.: Cell kinetics and the control of growth in long bones. Cell Tissue Kinet. 11: 477-485, 1978. Leonhard, CM., Bergman, M., Frenz, D.A., Macreery, L.A. and Newman, A.: Abnormal ambient glucose levels inhibit proteoglycan core protein gene expression and reduce proteoglycan accumulation during chondrogenesis: possible mechanism for teratogenic effects of maternal diabetes. Proc. Natl. Acad. Sci. USA 86: 10113-10117,1989. Mundlos, S., Engel, H., Michel-Behnke, I. and Zabel, B.: Distribution of type I and type II collagen gene expression during development of human long bones. Bone 11: 275-279, 1990. Neame, P.]., Christner, ]. E. and Baker,]. R.: The primary structure of link protein from rat chondrosarcoma proteoglycan aggregate. I Bioi. Chem. 261: 3519-3535, 1986. Osborne-Lawrence, S. L., Sinclair, A. K., Hicks, R. C, Lacey, S. W., Eddy, R. L., Byers, M. C, Shows, T. B. and Duby, A. D.: Complete amino acid sequence of human cartilage link protein (CRTLl) deduced from cDNA clones and chromosomal assignment of the gene. Genomics 8: 562-567, 1990.
Perin, ]. P., Bonnet, F. and Jolles, P.: Structural relationships between link proteins and proteoglycan monomers. FEBS Lett. 206: 73-77, 1986. Poole, A. R., Pidoux, I. and Rosenberg, L.: Role of proteoglycans in enchondral ossification: Immunofluorescent localization of link protein and proteoglycan monomer in bovine fetal epiphyseal growth plate. I Cell Bioi. 92: 249-260, 1982. Schmid, T.M. and Linsenmayer, T.F.: Immunohistochemical localization of short chain cartilage collagen (type X) in avian tissues. I Cell Bioi. 100: 598-605, 1985. Shephard, N. and Mitchell, N.: Ultrastructural modifications of proteoglycans coincident with mineralization in local regions of rat growth plate. I Bone ft. Surg. 67-A: 455 -464,1985. Stanescu, V., Maroteaux, P. and Stanescu, R.: The biochemical defect in pseudoachondroplasia. Eur. I Ped. 138: 221-225, 1982. Stirpe, N. S. and Goetinck, P. F.: Gene regulation during cartilage differentiation: temporal and spatial expression of link protein and cartilage matrix protein in the developing limb. Development 107: 23-33, 1989. Stirpe, N. S., Dickerson, K. T. and Goetinck, P. F.: The chicken mesonephros synthesizes link protein, an extracellular matrix molecule usually found in cartilage. Dev. Bioi. 137: 419-424, 1990. von der Mark, K.: Immunological studies on collagen type transition in chondrogenesis. Current Topics in Developmental Biology 14: 199-225, 1980. Dr. Stefan Mundlos, Universitats-Kinderklinik, Langenbeckstr. 1, 6500 Mainz, F. R. G.
Distribution of Cartilage Proteoglycan (Aggrecan) Core Protein and Link Protein Gene Expression during Human Skeletal Development S. MUNDLOS 1 , R. MEYER 1 , Y. YAMADA 2 and B. ZABEL 1 I
2
Department of Pediatrics, Johannes-Gutenberg-Universitat, Mainz, F. R. G. and Laboratory of Developmental Anomalies, National Institute of Health, Bethesda, MD, USA.
Abstract The distribution of cartilage proteoglycan core protein (aggrecan) and cartilage proteoglycan link protein was investigated by in situ hybridization during different stages of human skeletal development. Aggrecan and link protein expression were confined to chondrocytes of the developing skeleton and other cartilaginous structures. Distribution and intensity of the signal was identical with aggrecan as compared to link protein probes. Parallel to the calcification of cartilaginous matrix, chondrocytes of this area lost the expression of aggrecan and link protein specific mRNA and stayed negative throughout the following stages of skeletal development. Highest expression was found in the lower proliferative and upper hypertrophic zone whereas the resting zone showed less expression. Aggrecan gene expression was additionally investigated in iliac crest biopsies of 3 patients with pseudo achondroplasia and compared to age-matched controls. Distribution and intensity of staining revealed no abnormalities. Thus, the phenotypic changes during chondrocyte maturation are accompanied by distinct changes in aggrecan and link protein gene expression. This pattern was maintained in the growth plate of patients with pseudoachondroplasia. Key words: aggrecan, cartilage proteoglycan, core protein, link protein, skeletal development. Introduction The development of the human skeleton involves a number of transient stages characterized by specific histological findings and the production of stage-specific components of extracellular matrix (ECM). The cartilaginous anlage of the future bone results from a transient cellular condensation process and the subsequent switch of gene expression from mesenchymal to cartilage specific ECMmolecules (von der Mark, 1980). During the following developmental stages, the chondrocyte undergoes a developmental change in phenotypic expression. The proliferation and differentiation into maturing and degenerating chondrocytes is accompanied by progressive alterations in the composition and organisation of ECM. Little is known about the associated transcriptional events and their regulation. In order to understand the mechanisms of
chondrocyte differentiation during human osteogenesis, the temporal and spatial patterns of chondrocyte gene expression have to be defined. Cartilage proteoglycan core protein (aggrecan) and cartilage proteoglycan link protein are some of the major noncollagenous components of cartilage ECM. Biochemical studies performed mainly on bovine tissue have established many of the properties of aggrecan, including the formation of large aggregates with hyaluronic acid via an amini-terminal globular region and its extensive modification with chondroitin sulfate, keratan sulfate and other oligosaccharides (Hascall and Hascall, 1981). Link protein binds simultaneously to the hyaluronic acid binding region of the proteoglycan molecule and hyaluronate and by this stabilizes the aggregate against dissociative forces (Hardingham and Muir, 1974). Both proteins are essential for the structural integrity of cartilage and may be involved
340
S. Mundlos et a1.
Fig. 1. Localization of aggrecan and link protein gene expression by in situ hybridization. Dark field images of sections through a 6-week humerus (b), a 7-week humerus (d) and an 8-week femur (f) with corresponding bright light images in a, c and e. Note lack of expression in areas of calcification (arrows).
Core/Link Protein Gene Expression in the formation of birth defects (Leonhard et aI., 1989) and/or heritable chondrodysplasias (Stanescu et aI., 1982). In this study we have used recently cloned cDNAs (Doege et aI., 1991) to investigate the temporal and spatial gene expression of aggrecan and link protein during human skeletal development from the early cartilaginous bone anlage to the formation of a regular growth plate in the newborn. In addition, biopsy specimens of apophyseal growth plate cartilage were investigated for aggrecan expression in patients with pseudoachondroplasia and compared to age matched controls.
Materials and Methods
Selection and preparation of hybridization probes
The cDNA for link protein for CPLP had a length of 1. 7 kb coding for the complete protein (Y. Yamada, unpublished data). Link protein contains a tandem repeat that functions as the hyaluronic acid binding site (Perin et aI., 1987). These domains share a certain degree of homology with the hyaluronic acid binding region of aggrecan (amino-terminal end of the molecule) at the amino acid level in the rat (Neame et aI., 1986; Doege et aI., 1987). Consequently, sub clones of the aggrecan cDNA (Doege et aI., 1991) coding for the carboxy-terminal end (1.4 kb) and for the serine/glycine repeat domain (2.7 kb) were chosen for the in situ experiments in order to rule out cross-reactivity. cDNA fragments of aggrecan and link protein were subcloned in Bluescript vectors and transcribed with high efficiency from the T 3 or T 7 promotors to make sense and antisense cRNA. Linearized plasmid DNA was transcribed in the presence of [35 Sj_UTP 800 !-lCi/mmol (Amersham, UK). After polymerization, the DNA templates were digested with DNAse for 15 min at 37°C. The RNA transcripts were hydrolyzed for 30 min in 0.2 N NaOH on ice to a final length of 50-200 bases. After neutralization with acetic acid, the probes were purified by several ethanol precipitation steps. Analytical acrylamide gel electrophoresis was used to check the transcript size before and after hydrolysis. The probes had a specific activity of 2 - 3 8 X 10 dpmhlg. Hybridization
Selection and preparation of tissue was performed as described earlier (Mundlos et aI., 1990). Biopsies of the apophyseal growth plate cartilage (iliac crest) were obtained from three patients with pseudoachondroplasia undergoing surgical procedures. Informed consent was obtained prior to surgery. The hybridization procedure was basically the same as described by Hogan (Hogan et aI., 1986) with a few modifications. The sections were baked onto the slides at 45°C on
341
a hot plate for 15 min followed by fixation in 4 % paraformaldehyde in PBS, 0.2 N HCI for 20 min, 2 x SSC 60°C for 20 min, digestion with pronase (Boehringer Mannheim, F.R.G.) at 0.125mglml for 10min at room temperature, acetylation in acetic anhydride 1:400 in 0.1 M triethanolamine for 10 min, followed by dehydration. The hybridization mixture contained the labelled RNA probe (0.5 !-lg/ml) in formamide 50%, dextran sulfate 10%, NaCl 0.3 M, 10mM Tris/HCI pH7, 10mM Na-P0 4 , pH5.5, EDTA 5 mM, Denhardt's 0.02 %, tRNA 0.5 mg/ml. Hybridization was carried out over night at 45°C in a humidified chamber. Post-hybridization washings were performed at 50 °C in 50% formamide and salt concentration as described above for 2 x 1 hour, followed by RNAse digestion at 37°C (20 [lg/ml RNAse, Sigma) for 30 min and 2 x SSC at 37°C for 2 x 1 hour. Slides were immersed in Kodak NTB-2 foto emulsion diluted 1:1 with water. Autoradiography was performed at 4°C in a dry chamber. Exposure time varied between 10 and 20 days. The exposed slides were developed, fixed, stained and examined using the Zeiss system for epipolarization.
Results
In order to optimize the ratio of signal to background, the probes were hybridized under various conditions to sections through 10-week limbs reflecting the different stages of chondrocyte maturation. No difference was observed in aggrecan expression with the probe coding for the serine/ glycine repeat domain as compared to the carboxy-terminal domain. No signal was obtained using the aggrecan or link protein sense probes. Aggrecan and link protein gene expression was identical throughout the different stages of skeletal development. No difference was observed in either intensity or distribution of the signal. The following figures therefore reflect aggrecan and link protein expression. An autoradiograph of a frontal section through a 6-week humerus hybridized to the aggrecan probe is illustrated in Fig. 1 b with a bright light image given in Fig. 1 a. At this stage, no calcification of cartilage or bone is present. At the diaphysis, a small band of hypertrophied cells can be seen. In the dark field, these cells show the largest number of grains/cell with decreasing intensity towards the ends of the bone model. Parallel to the development of a diaphyseal bone collar, the cartilaginous matrix of this area begins to calcify. Fig. 1 c shows a van Kossa stain of a 7 -week humerus with a small area of diaphyseal calcification (arrows). In contrast to earlier stages of development, a segregation of chondrocyte activity can be observed in those that express aggrecan and link protein and those that do not show expression. Fig. 1 d shows a band of negative chondrocytes in the area of calcification sharply separated from intensely positive
342
S. Mundlos et al.
Fig.2. Localization of aggrecan and link protein gene expression by in situ hybridization. Dark field images through a 13-week femur (b) and the distal femura! growth plate of a new born male (d). Bright light images are given in a and c. Note lack of expression in areas of calcification as marked by arrows. "Cartilage canal.
cells in the hypertrophic and proliferative zones. As in the earlier stages, a decreasing intensity of the signal towards the ends of the bone (resting cartilage) is apparent. With further growth, the area of calcified cartilage enlarges
(Fig. 1 e) and the area of negative chondrocytes greatly increases (Fig. 1£). Beginning with the 9th week of development, calcified cartilage is constantly removed and replaced by vasculate
Core/Link Protein Gene Expression
343
Fig. 3. Localization of aggrecan and link protein gene expression by in situ hybridization. Longitudinal section of the developing spine and cross-section of a vertebra in the lower right part. Note lack of expression in areas of calcification (large arrows) and expression in cartilaginous tissue of the trachea (small arrows).
344
S. Mundlos et al. and marrow. Fig. 2 a shows a section through the distal part of a 13-week femur. The area of calcified cartilage is reduced to a few cell layers as indicated by arrows. At this stage, we found aggrecan and link protein negative chondrocytes in and adjacent to the area of calcification (Fig. 2 b). The largest number of grains/cell was present in the upper hypertrophic/lower proliverative zone. At birth, a regular growth plate with resting cartilage as well as proliferative and hypertrophied chondrocytes arranged in columns has been formed at the metaphysis (Fig. 2 c). At this stage, gene expression of aggrecan and link protein was limited to the small area of hypertrophied and proliferative chondrocytes. Resting cartilage and cartilage close to the osseochondral junction (zone of calcification, arrow) showed no expression (Fig. 2 d). A similar picture was seen at the epiphyseal growth zone (not shown). Besides in developing long bones, aggrecan and link protein expression was also found in other cartilaginous tissues. Chondrocytes of the developing spine gave a strong signal. Fig.3 shows a longitudinal section through the developing vertebrae of a 8-week embryo with a transverse section given in the lower right part. Heavy labeling was found throughout the vertebrae except for the developing intervertebral discs and for the areas of beginning calcification (arrows). Fig. 3 also shows expression in the cartilaginous tissue of the trachea (arrows). The iliac crest biopsies of three patients with pseudoachondroplasia had a very similar picture with a severely distorted growth plate, almost complete loss of columnization and lack of regular areas of calcification at the osseo chondral border. The hypertrophic cells formed clusters that contained few cells and were separated by an excessive amount of ECM. Fig. 4 b shows the distribution of aggrecan gene expression in an iliac crest biopsy of a seven year old patient with the bright light image given in Fig. 4 a. In spite of the absence of the normal chondrocyte architecture, the pattern of aggrecan gene expression was identical to the control tissue (Fig.4 c, d). Again, hypertrophied and proliferating chondrocytes gave the strongest signal; whereas the resting zone showed less expression and calcified chondrocytes were negative.
Discussion
Fig. 4. Localization of aggrecan expression by in situ hybridization in the apophyseal growth plate (iliac crest) of a seven-year old patient with pseudo achondroplasia (a, b) and an age-matched control (c, d). Note strong expression in hypertrophied chondrocytes (arrows) and little expression in resting cartilage.
This study describes the temporal and spatial pattern of aggrecan and link protein gene expression during human skeletal development. We were able to demonstrate distinct differences in chondrocyte gene expression which correlate to specific phenotypes. Pseudo achondroplasia is among the most common skeletal dysplasias affecting both the metaphyseal and epiphyseal growth plate. Growth cartilage of patients with pseudoachondroplasia shows peculiar, finger-print-like inclusion bodies that were proposed to be proteoglycan
Core/Link Protein Gene Expression remnants (Stanescu et al., 1982). However, the primary defect is unknown. In a recent study, Finkelstein et al. (1990) were able to show that mutations in the aggrecan gene locus are not likely to be the primary defect causing pseudoachondroplasia. The present study supports this finding in that the distribution of aggrecan gene expression in iliac crest biopsies of patients with pseudoachondroplasia was identical to the pattern found in controls in spite of the distorted chondrocyte architecture. In the different tissues and developmental stages investigated, no difference in gene expression of aggrecan as compared to link protein was observed. The amount of specific mRNA as indicated by the number of grains/cell as well as the cell specific distribution of grains was not significantly different. Aggrecan and link protein have common binding sites for hyaluronic acid which share a 40% and 33 % homology at the amino acid level in the rat (Doege et al., 1987). Although there is a significant homology at the protein level, there is divergence at the nucleotide level (Doege et al., 1991) so that cross-hybridization can be ruled out under the stringent conditions used. The identical pattern of aggrecan and link protein gene expression is in agreement with the findings of Stirpe and Goetinck who showed that the expression of both mRNAs is initiated at the same stage of limb development in the chick (Stirpe et al., 1989). However, the synthesis of both matrix molecules seems not to be coordinately regulated. Recently, link protein specific transcripts were detected in the chick mesonephros (Stirpe et al., 1990) and in a human colon tumor cell line (Osborne-Lawrence et al., 1990). Besides in cartilage of the early bone anlage and the growth plate, aggrecan and link protein expression was found in cartilaginous tissue of the trachea, the larynx and the mid-face. Areas of desmal ossification like the calvaria showed no expression whereas Meckel's cartilage was positive. Parallel to the formation of hypertrophied chondrocytes, a change in the amount of specific mRNA per cell was observed. Chondrocytes of the upper hypertrophic and lower proliferative zone gave the strongest signal whereas the resting zone showed a uniform but much lesser degree of staining. This finding correlates to the rate of cell divisions which is highest in the proliferative zone (Kember, 1978). Cell division and growth is linked to the production of ECM, in this case aggrecan and link protein. Fig. 1 a shows the early cartilaginous anlage of the future bone with an even distribution of expression. Concurrently with the formation of the bone collar, the ECM of the newly formed hypertrophic chondrocytes begins to calcify. As Fig. 1 d shows, these cells have stopped to express aggrecan and link protein specific mRNA. The chondrocytes of the upper hypertrophic and proliferative zone, however, are strongly positive whereas the remaining resting cartilage shows less expression. This phenomenon of loss of expression in areas of calcification was found throughout the following stages of skeletal development. Similar results
345
were obtained when studying type II collagen gene expression (Mundlos et al., 1990). With the formation of hypertrophied chondrocytes, matrix vesicles are found in clusters in the ECM. These vesicles are thought to function as the initial site of calcification (Anderson, 1989). The large aggregating proteoglycan of cartilage is known to inhibit mineralization (Blumenthal et al., 1979) and thus may prevent calcification in the proliferative and upper hypertrophic zone. It was proposed that proteolytic degradation of proteoglycans occurs in the upper hypertrophic zone resulting in the transformation of non-calcifiable to calcifiable matrix. However, large amounts of aggrecan and link protein can be detected in calcified matrix immunohistochemically (Poole et al., 1982) or by x-ray microprobe analysis for sulfur (Shephard and Mitchell, 1985). Consequently, proteoglycan aggregates have to be structurally modified to lose their capacity to inhibit calcification and/ or other molecules are involved in the process of crystal formation. We recently suggested that chondrocytes of the zone of calcification selectively switch to the expression of genes that promote calcification. Type X collagen (Schmidt and Linsenmayer, 1985) and alkaline phosphatase (S. Mundlos, unpublished observations) are both expressed selectively in hypertrophied chondrocytes undergoing calcification. Both are assumed to playa role in the calcification of cartilaginous matrix. Our present finding that cessation of aggrecan/link protein gene expression is followed by mineralization supports this hypothesis. Accordingly, changes in the phenotype of maturing chondrocytes are accompanied by changes in gene expression. The regulatory events that promote such drastic alterations in ECM production remain uncertain and will be subject of further studies. Acknowledgements This work was supported by the Deutsche Forschungsgemeinschaft.
References Anderson, H. c.: Mechanism of mineral formation in bone. Lab. Invest. 60: 320-330, 1989. Blumenthal, N. c., Posner, A. S., Silverman, L. D. and Rosenberg, L. c.: Effect of proteoglycans on in vitro hydroxyapatite formation. Calcif Tissue Int. 27: 75 -82,1979. Doege, K., Sasaki, M., Horigan, E., Hassell,J.R. and Yamada, Y.: Complete primary structure of the rat cartilage proteoglycan core protein deduced from cDNA clones. J. Cell. BioI. 262: 17757-17767,1987. Doege, K., Sasaki, M., Kimura, T. and Yamada, Y.: Complete coding sequence and deduced primary structure of the human cartilage large aggregating proteoglycan, aggrecan: Humanspecific repeats and additional alternatively-spliced forms. J. BioI. Chern., in press, 1991. Finkelstein, J.E., Doege, K., Yamada, Y., Pyeritz, R.E., Graham, J.M., Moeschler, J.B., Pauli, R.M., Hecht, J.T. and Fran-
346
S. Mundlos et al.
comano, CA.: Analysis of the Chondroitin sulfate proteoglycan core protein (CSPGCP) gene in achondroplasia and pseudoachondroplasia. Am. I Hum. Genet. 48: 97-102, 1991. Hardingham, T.E. and Muir, H.: Hyaluronic acid in cartilage and proteoglycan aggregation. Biochem. I 139: 565 - 581,1974. Hascall, V. C and Hascall, G. K.: Proteoglycans. In: Cell Biology of Extracellular Matrix, ed. by Hay, E.D., Plenum, New York, 1981, pp. 39-63. Hogan, B., Constantini, F. and Lacy, E.: Manipulating the Mouse Embryo. Cold Spring Harbor Laboratory, New York, 1986. Kember, N.F.: Cell kinetics and the control of growth in long bones. Cell Tissue Kinet. 11: 477-485, 1978. Leonhard, CM., Bergman, M., Frenz, D.A., Macreery, L.A. and Newman, A.: Abnormal ambient glucose levels inhibit proteoglycan core protein gene expression and reduce proteoglycan accumulation during chondrogenesis: possible mechanism for teratogenic effects of maternal diabetes. Proc. Natl. Acad. Sci. USA 86: 10113-10117,1989. Mundlos, S., Engel, H., Michel-Behnke, I. and Zabel, B.: Distribution of type I and type II collagen gene expression during development of human long bones. Bone 11: 275-279, 1990. Neame, P.]., Christner, ]. E. and Baker,]. R.: The primary structure of link protein from rat chondrosarcoma proteoglycan aggregate. I Bioi. Chem. 261: 3519-3535, 1986. Osborne-Lawrence, S. L., Sinclair, A. K., Hicks, R. C, Lacey, S. W., Eddy, R. L., Byers, M. C, Shows, T. B. and Duby, A. D.: Complete amino acid sequence of human cartilage link protein (CRTLl) deduced from cDNA clones and chromosomal assignment of the gene. Genomics 8: 562-567, 1990.
Perin, ]. P., Bonnet, F. and Jolles, P.: Structural relationships between link proteins and proteoglycan monomers. FEBS Lett. 206: 73-77, 1986. Poole, A. R., Pidoux, I. and Rosenberg, L.: Role of proteoglycans in enchondral ossification: Immunofluorescent localization of link protein and proteoglycan monomer in bovine fetal epiphyseal growth plate. I Cell Bioi. 92: 249-260, 1982. Schmid, T.M. and Linsenmayer, T.F.: Immunohistochemical localization of short chain cartilage collagen (type X) in avian tissues. I Cell Bioi. 100: 598-605, 1985. Shephard, N. and Mitchell, N.: Ultrastructural modifications of proteoglycans coincident with mineralization in local regions of rat growth plate. I Bone ft. Surg. 67-A: 455 -464,1985. Stanescu, V., Maroteaux, P. and Stanescu, R.: The biochemical defect in pseudoachondroplasia. Eur. I Ped. 138: 221-225, 1982. Stirpe, N. S. and Goetinck, P. F.: Gene regulation during cartilage differentiation: temporal and spatial expression of link protein and cartilage matrix protein in the developing limb. Development 107: 23-33, 1989. Stirpe, N. S., Dickerson, K. T. and Goetinck, P. F.: The chicken mesonephros synthesizes link protein, an extracellular matrix molecule usually found in cartilage. Dev. Bioi. 137: 419-424, 1990. von der Mark, K.: Immunological studies on collagen type transition in chondrogenesis. Current Topics in Developmental Biology 14: 199-225, 1980. Dr. Stefan Mundlos, Universitats-Kinderklinik, Langenbeckstr. 1, 6500 Mainz, F. R. G.
