11, 275-279 (1990) Printed in the USA . All rights reserved .
Bone,
8756-3282/90 $3 .00 + .00 Copyright © 1990 Pergamon Press plc
Distribution of Type I and Type II Collagen Gene Expression during the Development of Human Long Bones S . MUNDLOS, H . ENGEL, I . MICHEL-BEHNKE and B . ZABEL Department of Pediatrics, Johannes Gutenberg Universitat Mainz. Federal Republic of Germany Address for correspondence and reprints :
Dr . Stefan Mundlos, Kinderklinik der Universitat Mainz, Langenbeckstr . 1, D-6500 Mainz, Federal
Republic of Germany . Abstract
differentiates in an outer fibroblastic and an inner osteoblastic layer of cells, which then form a thin periosteal bone around the diaphysis . Concurrently with the formation of a bone collar, the cartilagenous matrix of this zone becomes calcified . The next step is the resorption of hypertrophied chondrocytes and the invasion of the newly formed cavity by blood vessels . This marks the end of the embryonic period . The rapid protrusion of the bone marrow cavity towards the epiphysis goes along with the deposition of enchondral bone onto the calcified cartilage . The growth of the bone now takes place at the epiphysis, where spicules of calcified cartilage covered by enchondral bone are constantly removed and newly generated at the growth plate . A number of different collagen and other components of the extracellular matrix are produced at different stages during this complex process . Type I and type II collagen are considered the major collagens of bone and cartilage respectively . Analysis of spatial and temporal changes in extracellular matrix during embryogenesis have been carried out using biochemical methods (Stanescu et al . 1973), immunohistochemistry (von der Mark 1976, 1980 ; Schmid and Linsenmayer 1985) and various in vitro models of chondrogenesis (Kravis & Upholt 1985 ; Kosher et al. 1986) . Recently, the gene expression of type I and type 11 collagen has been studied at mRNA level using in situ hybridization by several authors during limb development of the chick (Kosher et al . 1986 ; Hayashi et al . 1986; Swalla et al . 1988 ; Devlin et al . 1988) and in human fetal tissue (Sandberg and Vuorio 1987) . These studies mainly concentrated on the very early development of the chick limb bud or were carried out in fetal bone of one developmental stage . In this study we report the localization of type I and type II collagen gene expression during the development of the human long bone from the stage of the cartilagenous bone anlage to the formation of a regular growth plate in the newborn .
The temporal and spatial gene expression of collagen type I and type II during the development of the human long bones was studied by the technique of in situ hybridization covering the period from the cartilagenous bone anlage to the formation of a regular growth plate in the newborn . Analysis of the early stages around the seventh week of gestation revealed for type II collagen a strong hybridization signal limited to the chondrogenic tissue . The surrounding connective tissue and the perichondrium showed weak type I collagen expression, while the zones of dermal ossification like the clavicle gave a strong signal . Beginning with the eighth week of gestation, type I collagen mRNA was detectable in newly formed osteoblasts at the diaphysis and appeared along with the formation bone marrow, in the areas of enchondral ossification . Parallel to the development of the different zones of cartilage differentiation, a specific pattern of type II expression could be observed : type II was mainly found in the chondrocytes of the hypertrophic zone and to a lesser degree in the zone of proliferation, while the resting zone and the zone of provisional calcification showed little activity . This segregation of type II expression was most pronounced in the early stages of cartilage calcification and in the growth plate of the newborn . Key Words : Collagen-Type I-Type II-Gene expression .
Introduction The development of the human long bones involves a number of transient stages with changes in morphological appearance and in the composition of extracellular matrix (Stanescu et al . 1973 ; Gardner 1956) . All bones are indicated first as a condensation of mesenchyme, which appears in the early embryonic period . As prechondrogenic mesenchyme cells develop to chondrocytes, a dramatic increase in the cytoplasmatic volume, the rough endoplasmatic reticulum and the Golgi apparatus takes place . This is paralleled by the switch from collagen type 1, the predominant collagen of fibroblasts, to collagen type 11, the major collagen found in cartilage (von der Mark 1980) . The newly formed cartilage grows by interstitial and appositional growth to form the cartilagenous anlage of the bone . With further development . the chondrocytes at the diaphysis change their appearance and become so-called hypertrophic . Meanwhile, the perichondrium
Materials and Methods Selection and preparation of tissue
Limbs of human fetuses between the 7th and 15th menstrual weeks were obtained from therapeutic abortions . In addition, autopsy material from later stages and newborns was analyzed . The tissue was fixed in 4% paraformaldehyde in PBS for 12 hours and was then placed into 0 .5 M sucrose for 6-12 hours to prevent freeze artifacts . Tissue was stored in liquid nitrogen until further use . 5 µm sections were cut on a Leitz cryostat and were 275
276
S . Mundlos et al . : Type I and II collagen gene expression
mounted to triethoxyethylamine treated microscopic slides . We used the developmental landmarks of osteogenesis as described by Gardner for the humerus (Gardner, 1956) and the femur (Gardner and Gray 1970) to give an approximate staging of our specimens, which was expressed as menstrual age in weeks . For better comparison of the different stages of the developing bone, only sections through humerus or femur were used . Preparation of hybridization probes Two cDNA plasmids were used . pHF32 coding for a2(I) (Bernard et al . 1983) and pHCAR3 coding for al(II) (Elima et al . 1987) collagen with inserts of 2200 base pairs (bp) and 1470 by respectively . Both probes code for the carboxyterminal and part of the helical domain of the two different collagens . CARS additionally covers a 400 by segment of 3'-nontranslated region, which shows little homology to other mRNAs (Stoker et al . 1985) . pBR-DNA served as control . Since it is known that both probes have a certain degree of homology, especially in the helical part, we have used different fragments of CAR3 to rule out cross-hybridization . For this purpose, the complete 1470 by fragment was compared to the 550 by Eco RI-Eco RI fragment that covers the noncoding region . Probes were labeled with 35 S-CTP (1000 mCi/mmol, Amersham) with a standard nick tanslation procedure (Boehringer Mannheim) to a specific activity of 1 .5 - 2 x 10s cpm/p.g DNA . The nick-translated probes were separated from unincorporated nucleotides by Sephadex G 50 columns . Hybridization The hybridization procedure was basically the same as described by Hafen et al . (1983) with a few modifications . Freshly cut sections were baked onto the slides at 42°C on a hot plate . The slides were then immersed in paraformaldehyde 4% in PBS for 10 min, followed by PBS for five min . To remove basic proteins, the slides were treated with 0 .2 N HCI for 20 min at room temperature, washed in H=O and incubated in 2 x SSC at 60 °C for 30 min . The sections were next dehydrated in graded series of ethanol and air dried . The hybridization mixture contained the labeled DNA probe (130 p.g/ml), formamide 50%, dextrane sulfate 10%, 0 .3 M NaCl, 10 mM Tris/HCI pH 7, 10 mM Na-P0, pH 5, 5 mM EDTA, 0 .02% Denhardt's, t-RNA 0 .5 mg/ml (Serva), SS-DNA (80 µg/ml) . The complete hybridization mixture was heated for 10 min at 80°C to denature the labeled DNA probe and chilled quickly in ethanol/dry ice . Depending on the size of the section, 20-50 xl of hybridization mixture was spread over the pretreated dry sample, immediately covered with a siliconized cover slip and sealed with rubber cement . Hybridization was carried out at 45°C in a humidified chamber for 15-18 hours . Following hybridization, the coverslips were removed in 2 x SSC at room temperature, followed by three washes in 2 X SSC at room temperature, once in 0 .5 x SSC and three times in 0 .1 x SSC at 50°C for 10 min each . Afterwards, the slides were dehydrated in graded series of ethanol and air dried . The slides were immersed in Kodak NTB-2 emulsion diluted 1 :1 with Hz 0 and air dried . 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 and to minimize cross-reactivity, the probes were hybridized under
Flg. 1- Localization of collagen type I (c and f) and II (b and e) mRNAs by in situ hybridization using 35 S labeled cDNA probes . (b and e) Dark field image of sections through the humerus and the clavicle of a seven-to-eight week embryo with a bright light image given in a . (e,f) Dark field image of sections through the femur of a 10 week embryo with a bright light image given in d (van Kossa stain) . Arrows indicate the area of calcified matrix . Note uneven distribution of grains in b, and type 11 collagen negative cells in the calcified area in e . a,b,d,e,f : 2 .5x ; c: 20x .
various conditions to sections through limbs of human embryos of the ninth week of gestation . At this stage of development, cells that are known to produce type I collagen (fibroblasts, osteoblasts) are found on the same section in close contact to chondrocytes that are the source of type II collagen . A section through a developing bone with beginning calcification is therefore a good control for the specificity of the probes used . Using the procedure described here, virtually no cross-reactivity between the probes for type I and type II collagen was found . At no time did mature chondrocytes show any type I activity, and type 11 was found solely over chondrogenic tissue . The Eco RI fragment of Car3 showed the same distribution as the 1470 by probe . Background activity was minimal . No hybridization signal was obtained using the pBR probe . An autoradiograph of a frontal section through a sevento-eight week humerus hybridized to the type II collagen probe is illustrated in Fig . 1b with a bright light image shown in Fig, la . At this stage, the future bone consists of a cartilagenous model without any histological signs of calcification . The chondrocytes have a uniform appearance, ate small, densely
S . Mundlos et al . : Type I and II collagen gene expression
2 77
Fig . 2. Localization of collagen type I (c) and type H (b,e,l) mRNAs by in situ hybridization using "S labeled cDNA probes . Dark field images of sections through the femoral growth plate of a 15-week fetus (b,c) and at the age of three months (e) are shown . Corresponding bright light images are given in a and d . Arrows indicate the area of magnification . (f) Bright light image of hypertrophic cartilage with arrows indicating type II collagen negative cells . Note the uneven distribution of grains in e with little expression of type Ii collagen in the resting cartilage . a,b,c : 5x ; e: 10x ; f:20x . packed, show little extracellular matrix, and lack a specific arrangement. At the diaphysis ; however, a band of larger cells that represents the developing hypertrophic zone can be seen . In the dark field, the distribution of silver grains revealed a strong expression of type II collagen in the cells of the hypertrophic zone, with decreasing intensity toward the ends of the bone model . At this stage of development, type I expression is found to a small degree in the perichondreal sheet surrounding the bone anlage (not shown) . In the zones of desmal ossification, an intense hybridization signal to the type I probe was observed . Fig . Ic shows a cross section through a clavicle of the same developmental stage, hybridized to the type I probe . Type I positive osteoblasts are found in large numbers, irregularly arranged and densely packed at the periphery of the bone . With further development, secondary cartilage develops at the end and in the center of the clavicle . This tissue, which morphologically resembles early cartilage, also showed type II activity (not shown) . With further growth of the bone, the hypertrophic zone
enlarges and a small rim of periosteal bone is built around the diaphysis . At this point we find a segregation of chondrocyte activity in those that still express type II collagen and those that have ceased producing mRNA coding for type II collagen . Fig . le shows a band of type II negative cells at the diaphysis that is sharply separated from type II positive cells on both sides . This area of type II negative cells resembles the zone of the calcified cartilagenous matrix as shown by the van Kossa stain (Fig . ld) . Gene expression of type I collagen at this stage of development is shown in Fig . If. Osteoblasts at the newly formed bone collar showed strong type I activity corresponding to the calcification seen in Fig . Id (van Kossa stain) . To a lesser degree, type I positive cells can be found in fibroblasts of the perichondrium and in tendons . Type I positive cells were only found at the outside of the newly formed bone matrix . No type I activity was found in chondrogenic tissue . Around the 11th week of gestation, hypertrophic chondrocytes at the diaphysis are removed and replaced by the bone marrow cavity . At this stage, type II expression is found in all chondrocytes except the layer of cells adjacent to osseochondral
27 8
S . Mundlos et al . : Type I and II collagen gene expression
Collagen I II - ---------------a
b
C
-------------------Fig . 3 . Diagram showing the distribution of type I and type II collagen gene expression during the early development of the human long bones . (a) Sixth to seventh week, cartilagenous bone anlage; (b) ninth to 11th week, periosteal bone formation and calcification of cartilage at diaphysis ; (c) > 12th week, formation of a growth plate, desnial and enchondral ossification .
junction . Again, these cells do not show any type II expression and are located in the zone of provisional calcification . With further development, the area of type II negative cells becomes smaller and smaller . However, type II negative chondrocytes were always found at the osseochondral junction . Fig . 2b and Fig . 2c show the type I and type II collagen gene expression in a section of a humerus of the 15th week of gestation with a bright light image given in Fig . 2a . Parallel to the formation of the bone marrow, zones of enchondral ossification develop where osteoblasts that surround resting cartilagenous matrix show type I activity . By birth, a regular growth plate with hypertrophied cartilage arranged in columns has been formed (Fig . 2d) . Again, type II negative chondrocytes are found at the osseochondral junction (Fig. 2f) . At this stage, type II expression is mainly found in chondrocytes of the hypertrophic and to a lesser degree of the proliferative zone, while the resting cartilage shows little activity (Mg . 2e) . If attempts are made to quantify the type II collagen gene expression by counting the number of grains/cell, the chondrocytes in the embryonic and fetal tissue showed a three to four times higher number of grains/cell in the hypertrophic zone than in the resting zone. This difference was even more pronounced in the growth plate of the newborn, where the hypertrophic cartilage showed a 10 to 15 times higher number of grainslcell than the resting cartilage . Figure 3 summarizes the results . Discussion The developmental patterns of collagen distribution have been extensively studied in the chick by the use of specific antibodies and immunohistochemistry (von der Mark et al . 1976; von der Mark 1980) . Type I collagen was found to be present in tendons,
fibroblasts, perichondrium, and bone, whereas type II collagen was present in cartilage matrix . Using immunohistochetnical methods, a rather homogeneous distribution of type II collagen throughout the cartilage zones was observed, reflecting the uniform appearance of the extracellular matrix . The technique of in situ hybridization offers the advantage of detecting switches in gene expression before changes in extracellular matrix result . This is of special importance for investigations of the developing skeleton and the growth plate, because this tissue is subject to rapid transition of collagen types during growth . In the present study, we were able to identify the particular type of cell involved in production of type I or type II collagen during different stages of the development of the human skeleton using in situ hybridization with cDNA probes . Naturally, the reliability of the method rests on the specificity of the probes used . Type I and type II collagen are known to have a certain degree of homology . Consequently, a number of authors have selected certain parts of the collagen cDNA that show a lesser degree of homology and have used these fragments for in situ hybridization . Specificity was usually tested by northern blot or dot blot analysis . This procedure, however, does not rule out cross-reactivity using the conditions of in situ hybridization . The best control in our opinion is a section through developing bone where cartilage (containing no type I collagen) is next to newly formed bone (known to contain no type II Collagen) . These tissue sections served as internal controls during optimization of the hybridization conditions . Using the above described method, practically no cross-hybridization was observed . It should be emphasized that hybridization and washes were performed under high stringency conditions as described by other authors to prevent cross-hybridization between different collagen probes (Hayashi et al . 1986) . mRNA coding for type II collagen was exclusively found in chondrogenic tissue, and chondrocytes were always negative for type I expression . This is in contrast to the findings of Devlin and coworkers (1988), who were able to show significant levels of type I collagen mRNA in mature cartilage and type 11 expression of the mesenchyme in the developing chick . Similar results were obtained during in vitro chondrogenesis (Kosher et al . 1986) and in total RNA extracted from epiphyseal cartilage (Elima et al . 1987) . Since type II collagen is not present in nonchondrogenic tissue, and cartilage does not contain detectable amounts of type I collagen, the authors explain this phenomenon by regulation of the protein synthesis at the level of translation . However, cross-reactivity of the two collagen probes is not ruled out, especially if RNA probes are used (Hayashi et al . 1986) . Our results support the transcriptional level as the main control point of collagen synthesis . Figure la shows the cartilagenous model of the humerus with well developed perichondrium . Growth takes place by apposition from the perichondrium and through multiplication of cartilage cells already formed together with an increase in extracellular matrix . In the middle of the bone a layer of hypertrophied cells can be seen that shows a larger accumulation of gainslcell than the rest of the cartilage . This diffc.. .uce of type II collagen gene expression between the resting and the hypertrophied cartilage was observed throughout the development of the growth plate. In contrast, von der Mark (1980) describes a less intense staining with anti-type H collagen antibody of hypertrophied cartilage in the chick at a comparable stage of development . Possibly, these cells also produce large amounts of other extracellular matrix components that mask type II collagen antibody reaction . Concurrently with the formation of periosteal bone, an extraordinary switch of chondrocyte gene expression takes place . Chondrocytes at the diaphysis, which produced large amounts of type II collagen in earlier developmental stages, suddenly cease expressing the type II collagen gene . This sudden switch is not
S . Mundlos et al . : Type I and II collagen gene expression
paralled by any morphological change of the cells, whereas the extracellular matrix shows signs of calcification and is reduced to a small rim around the cells . The phenomenon just described was observed in all the developmental stages of the growth plate thereafter, and is in agreement with the findings of Sandberg & Vuorio (1987), who also describe Type II negative chondrocytes in this location in fingers of 15 week fetuses . With further growth, the band of type II negative cells becomes smaller and smaller and has almost disappeared by the 15th week . It seems that these cells are selectively removed by the protruding bone marrow . Nevertheless, type II expression in the zone of provisional calcification remains at a low level throughout further development, and type II negative cells can always be found at the osseochondral border . The reason for this switch remains obscure . Do these cells start producing type X collagen instead of type II? In the chick tibiotarsus, type X collagen was shown to occur only in chondrocytes of the lower hypertrophic zone that were undergoing calcification (Schmid and Conrad, 1982) . Type X collagen is assumed to play some role in the calcification of cartilagenous matrix . In vitro experiments could demonstrate a coordinate regulation of alkaline phosphatase and type X collagen production by embryo tibial chondrocytes, suggesting that their role in calcification may be coupled (Habuchi et al . 1985) . Thus, chondrocytes regenerated in the proliferative zone (Kember 1978) change their pericellular matrices on the way through the different zones of the growth plate until they reach the osseochondral junction . The composition of hyaline cartilage is modified in a progressive fashion to yield a tissue that is functionally distinct and able to undergo calcification . Concurrently with the development of hypertrophied chondrocytes, we observed a change in the amount of type II coding mRNA per cell, as judged by grains/cell . Chondrocytes of the upper hypertrophic zone and the zone of proliferation showed the largest number of grains/cell, whereas the resting zone showed a uniform but much lesser degree of staining . This phenomenon again was observed throughout the further steps of development of the growth plate . In the newborn, type II activity in the resting cartilage is even less pronounced, and type II expression is almost exclusively found in the hypertrophic zone . In the mature growth plate, growth takes place by cell division in the upper proliferative zone . In contrast to earlier stages in development, there is little appositional growth . The higher degree of collagen gene expression in the hypertrophic and proliferative zone when compared to the resting zone is most likely caused by the larger number of cell divisions in this area resulting in increased production of extracellular matrix . No type I activity was found in the zone of provisional calcification . In contrast, von der Mark (1980) describes staining with anti-type I collagen antibodies in the core of the calcified cartilage in the chick . Immunoreactive type I collagen was recently described by Horton et al . (1988) in the zone of provisional calcification of the human growth plate of the ileac crest . According to our results, a significant production of type I collagen by hypertrophied chondrocytes seems unlikely . In summary, in situ hybridization was shown to be a powerful tool for the investigation of developmental processes, especially if proteins of the extracellular matrix with low turnover rates are involved . We were able to complement earlier immunohistological studies and show that type II collagen is expressed in cartilage to a very different degree, depending on the developmental stage . Monitoring these patterns of gene expression during development will provide a basis for understanding normal growth and the pathophysiology of osteochondrodysplasias .
279
Acknowledgments : We are grateful to Dr . Vuorio for generously provid-
ing us with the probe for al(II) collagen . The clone for a2(I) collagen was kindly supplied by Drs . Prockop and Ramirez . References Bernard, M.P. ; Myers, J .C . ; Chu, M.L . ; Ramirez, F . ; Eikenberry, E .F. ; Prockop, D .J . Structure of a cDNA for proa2 chain of human type I procollagen . Biochemistry 22 :1139-1145 ; 1983 . Devlin, C.J. ; Brickell, P.M . ; Taylor, E .R. ; Hombruch, R .K . ; Craig, R.K . ; Wolpert, L . In situ hybridisation reveals differential spatial distribution of mRNAs for type I and type II collagen in the chick limb bud . Development 103:111-118, 1988 . Elima, K. ; Vuorio, T . ; Vuorio, E . Determination of the single polyadenylation site of the human pros 1(11) collagen . Nucleic Acids Res . 15 :9499-9504 ; 1987 . Gardner, E . Osteogenesis in the human embryo and fetus . Boume, G.H ., ed . The Biochemistry and Physiology of Bone . New York : Academic Press ; 1956 : 359-399 . Gardner, E . ; Gray, D.J. The prenatal development of the human femur . Am . J . Anat . 129:121-140 ; 1970. Habuchi, H . ; Conrad, E .H . ; Glaser, J .H . Coordinate regulation of collagen and alkaline phosphatase levels in chick embryo chondrocytes . J . Biol . Chem . 260 :13029-13034; 1985. Hafen, E . ; Levine, M . ; Garber, R.L . ; Gehring, W .J. An improved in situ hybridisation method for the detection of cellular RNAs in Drosophila tissue sections and its application for localizing transcripts of the homeotic Antennapedia gene complex. EMBO J . 2:617-623 : 1983 . Hayashi, M. ; Ninomiya, Y . ; Parsons, J . ; Hayashi, K . ; Olsen, B .R . ; Trelstad, R .L . Differential localization of mRNAs of collagen types I and II in chick fibroblasts, chondrocytes, and corneal cells by in situ hybridisation using cDNA probes. J Cell Biol. 102 :2302-2309; 1986 . Horton, W .A . ; Machado, M .M . Extracellular matrix alterations during enchondral ossification in humans . J. Orthop . Res. 6:793-803 ; 1988 . Kember, N .F. Cell kinetics and the control of growth in long bones . Cell Tissue Kinet . 11 :477-485 ; 1978 . Kosher, R.A . ; Kulyk, W .M . ; Gay, S .W. Collagen gene expression during limb cartilage differentiation . J. Cell Biol. 102:1151-1156 ; 1986. Kravis, D . ; Upholt, W .B . quantitation of type II procollagen mRNA levels during chick limb cartilage differentiation . Dev . Biol. 108:164-172 ; 1985 . Sandberg, M . ; Vuorio, E . Localisation of types I, II, and III collagen mRNAs in developing human skeletal tissues by in situ hybridization . J. Cell Biol. 104 :1077-1084 ; 1987 . Schmid, T .M . ; Conrad, H .E. Metabolism of low molecular weight collagen by chondrocytes obtained from histologically distinct zones of the chick embryo tibiotarsus . J . Biol . Chem . 257 :12451-12457 ; 1982 . Schmid, T .M . ; Linsenmayer, T .F . Immunohistochemical localization of short chain cartilage collagen (type X) in avian tissues . J . Cell Biol . 100 :598-605; 1985 . Stanescu, V . ; Stanescu, R . ; Maroteaux, P. Chemical studies on the human growth cartilage in fetuses and newborns . Biol. Neonate 23:432-445 ; 1973 . Stoker, N .G . ; Cheah, K.S .E . ; Griffin, R . ; Pope, F .M . ; Solomon, E. A highly polymorphic region 3' to human type II collagen gene . Nucleic Acids Res. 13 :4613-4622; 1985 . Swalla, B .J . ; Upholt, W .B . ; Solursh, M . Analysis of type II collagen RNA localisation in chick wing buds by in situ hybridisation . Dev . Biol . 125 :51-58 ; 1988 . von der Mark, H . ; von der Mark, K . ; Gay, S. Study of differential collagen synthesis during development of the chick embryo by immunofluorescence . I . Preparation of collagen type I and type II specific antibodies and their application to early stages of the chick embryo . Devl. Biol . 18 :237-249; 1976. von der Mark, K . Immunological studies on collagen type transition in chondrogenesis. Monroy, A . ; Moscona, A .A ., eds. Current topics in developmental biology, Vol . 14 . New York : Academic Press; 1980:199-225 . von der Mark, K . ; von der Mark, H . ; Gay, S. Study of differential collagen synthesis during development of the chick embryo by immunofluorescence . II. Localization of type I and type II collagen during bone development . Dev. Biol . 53:153-170 ; 1976 .
Received : July 28, 1989 Revised: December 20, 1989 Accepted: December 20, 1989
Bone,
8756-3282/90 $3 .00 + .00 Copyright © 1990 Pergamon Press plc
Distribution of Type I and Type II Collagen Gene Expression during the Development of Human Long Bones S . MUNDLOS, H . ENGEL, I . MICHEL-BEHNKE and B . ZABEL Department of Pediatrics, Johannes Gutenberg Universitat Mainz. Federal Republic of Germany Address for correspondence and reprints :
Dr . Stefan Mundlos, Kinderklinik der Universitat Mainz, Langenbeckstr . 1, D-6500 Mainz, Federal
Republic of Germany . Abstract
differentiates in an outer fibroblastic and an inner osteoblastic layer of cells, which then form a thin periosteal bone around the diaphysis . Concurrently with the formation of a bone collar, the cartilagenous matrix of this zone becomes calcified . The next step is the resorption of hypertrophied chondrocytes and the invasion of the newly formed cavity by blood vessels . This marks the end of the embryonic period . The rapid protrusion of the bone marrow cavity towards the epiphysis goes along with the deposition of enchondral bone onto the calcified cartilage . The growth of the bone now takes place at the epiphysis, where spicules of calcified cartilage covered by enchondral bone are constantly removed and newly generated at the growth plate . A number of different collagen and other components of the extracellular matrix are produced at different stages during this complex process . Type I and type II collagen are considered the major collagens of bone and cartilage respectively . Analysis of spatial and temporal changes in extracellular matrix during embryogenesis have been carried out using biochemical methods (Stanescu et al . 1973), immunohistochemistry (von der Mark 1976, 1980 ; Schmid and Linsenmayer 1985) and various in vitro models of chondrogenesis (Kravis & Upholt 1985 ; Kosher et al. 1986) . Recently, the gene expression of type I and type 11 collagen has been studied at mRNA level using in situ hybridization by several authors during limb development of the chick (Kosher et al . 1986 ; Hayashi et al . 1986; Swalla et al . 1988 ; Devlin et al . 1988) and in human fetal tissue (Sandberg and Vuorio 1987) . These studies mainly concentrated on the very early development of the chick limb bud or were carried out in fetal bone of one developmental stage . In this study we report the localization of type I and type II collagen gene expression during the development of the human long bone from the stage of the cartilagenous bone anlage to the formation of a regular growth plate in the newborn .
The temporal and spatial gene expression of collagen type I and type II during the development of the human long bones was studied by the technique of in situ hybridization covering the period from the cartilagenous bone anlage to the formation of a regular growth plate in the newborn . Analysis of the early stages around the seventh week of gestation revealed for type II collagen a strong hybridization signal limited to the chondrogenic tissue . The surrounding connective tissue and the perichondrium showed weak type I collagen expression, while the zones of dermal ossification like the clavicle gave a strong signal . Beginning with the eighth week of gestation, type I collagen mRNA was detectable in newly formed osteoblasts at the diaphysis and appeared along with the formation bone marrow, in the areas of enchondral ossification . Parallel to the development of the different zones of cartilage differentiation, a specific pattern of type II expression could be observed : type II was mainly found in the chondrocytes of the hypertrophic zone and to a lesser degree in the zone of proliferation, while the resting zone and the zone of provisional calcification showed little activity . This segregation of type II expression was most pronounced in the early stages of cartilage calcification and in the growth plate of the newborn . Key Words : Collagen-Type I-Type II-Gene expression .
Introduction The development of the human long bones involves a number of transient stages with changes in morphological appearance and in the composition of extracellular matrix (Stanescu et al . 1973 ; Gardner 1956) . All bones are indicated first as a condensation of mesenchyme, which appears in the early embryonic period . As prechondrogenic mesenchyme cells develop to chondrocytes, a dramatic increase in the cytoplasmatic volume, the rough endoplasmatic reticulum and the Golgi apparatus takes place . This is paralleled by the switch from collagen type 1, the predominant collagen of fibroblasts, to collagen type 11, the major collagen found in cartilage (von der Mark 1980) . The newly formed cartilage grows by interstitial and appositional growth to form the cartilagenous anlage of the bone . With further development . the chondrocytes at the diaphysis change their appearance and become so-called hypertrophic . Meanwhile, the perichondrium
Materials and Methods Selection and preparation of tissue
Limbs of human fetuses between the 7th and 15th menstrual weeks were obtained from therapeutic abortions . In addition, autopsy material from later stages and newborns was analyzed . The tissue was fixed in 4% paraformaldehyde in PBS for 12 hours and was then placed into 0 .5 M sucrose for 6-12 hours to prevent freeze artifacts . Tissue was stored in liquid nitrogen until further use . 5 µm sections were cut on a Leitz cryostat and were 275
276
S . Mundlos et al . : Type I and II collagen gene expression
mounted to triethoxyethylamine treated microscopic slides . We used the developmental landmarks of osteogenesis as described by Gardner for the humerus (Gardner, 1956) and the femur (Gardner and Gray 1970) to give an approximate staging of our specimens, which was expressed as menstrual age in weeks . For better comparison of the different stages of the developing bone, only sections through humerus or femur were used . Preparation of hybridization probes Two cDNA plasmids were used . pHF32 coding for a2(I) (Bernard et al . 1983) and pHCAR3 coding for al(II) (Elima et al . 1987) collagen with inserts of 2200 base pairs (bp) and 1470 by respectively . Both probes code for the carboxyterminal and part of the helical domain of the two different collagens . CARS additionally covers a 400 by segment of 3'-nontranslated region, which shows little homology to other mRNAs (Stoker et al . 1985) . pBR-DNA served as control . Since it is known that both probes have a certain degree of homology, especially in the helical part, we have used different fragments of CAR3 to rule out cross-hybridization . For this purpose, the complete 1470 by fragment was compared to the 550 by Eco RI-Eco RI fragment that covers the noncoding region . Probes were labeled with 35 S-CTP (1000 mCi/mmol, Amersham) with a standard nick tanslation procedure (Boehringer Mannheim) to a specific activity of 1 .5 - 2 x 10s cpm/p.g DNA . The nick-translated probes were separated from unincorporated nucleotides by Sephadex G 50 columns . Hybridization The hybridization procedure was basically the same as described by Hafen et al . (1983) with a few modifications . Freshly cut sections were baked onto the slides at 42°C on a hot plate . The slides were then immersed in paraformaldehyde 4% in PBS for 10 min, followed by PBS for five min . To remove basic proteins, the slides were treated with 0 .2 N HCI for 20 min at room temperature, washed in H=O and incubated in 2 x SSC at 60 °C for 30 min . The sections were next dehydrated in graded series of ethanol and air dried . The hybridization mixture contained the labeled DNA probe (130 p.g/ml), formamide 50%, dextrane sulfate 10%, 0 .3 M NaCl, 10 mM Tris/HCI pH 7, 10 mM Na-P0, pH 5, 5 mM EDTA, 0 .02% Denhardt's, t-RNA 0 .5 mg/ml (Serva), SS-DNA (80 µg/ml) . The complete hybridization mixture was heated for 10 min at 80°C to denature the labeled DNA probe and chilled quickly in ethanol/dry ice . Depending on the size of the section, 20-50 xl of hybridization mixture was spread over the pretreated dry sample, immediately covered with a siliconized cover slip and sealed with rubber cement . Hybridization was carried out at 45°C in a humidified chamber for 15-18 hours . Following hybridization, the coverslips were removed in 2 x SSC at room temperature, followed by three washes in 2 X SSC at room temperature, once in 0 .5 x SSC and three times in 0 .1 x SSC at 50°C for 10 min each . Afterwards, the slides were dehydrated in graded series of ethanol and air dried . The slides were immersed in Kodak NTB-2 emulsion diluted 1 :1 with Hz 0 and air dried . 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 and to minimize cross-reactivity, the probes were hybridized under
Flg. 1- Localization of collagen type I (c and f) and II (b and e) mRNAs by in situ hybridization using 35 S labeled cDNA probes . (b and e) Dark field image of sections through the humerus and the clavicle of a seven-to-eight week embryo with a bright light image given in a . (e,f) Dark field image of sections through the femur of a 10 week embryo with a bright light image given in d (van Kossa stain) . Arrows indicate the area of calcified matrix . Note uneven distribution of grains in b, and type 11 collagen negative cells in the calcified area in e . a,b,d,e,f : 2 .5x ; c: 20x .
various conditions to sections through limbs of human embryos of the ninth week of gestation . At this stage of development, cells that are known to produce type I collagen (fibroblasts, osteoblasts) are found on the same section in close contact to chondrocytes that are the source of type II collagen . A section through a developing bone with beginning calcification is therefore a good control for the specificity of the probes used . Using the procedure described here, virtually no cross-reactivity between the probes for type I and type II collagen was found . At no time did mature chondrocytes show any type I activity, and type 11 was found solely over chondrogenic tissue . The Eco RI fragment of Car3 showed the same distribution as the 1470 by probe . Background activity was minimal . No hybridization signal was obtained using the pBR probe . An autoradiograph of a frontal section through a sevento-eight week humerus hybridized to the type II collagen probe is illustrated in Fig . 1b with a bright light image shown in Fig, la . At this stage, the future bone consists of a cartilagenous model without any histological signs of calcification . The chondrocytes have a uniform appearance, ate small, densely
S . Mundlos et al . : Type I and II collagen gene expression
2 77
Fig . 2. Localization of collagen type I (c) and type H (b,e,l) mRNAs by in situ hybridization using "S labeled cDNA probes . Dark field images of sections through the femoral growth plate of a 15-week fetus (b,c) and at the age of three months (e) are shown . Corresponding bright light images are given in a and d . Arrows indicate the area of magnification . (f) Bright light image of hypertrophic cartilage with arrows indicating type II collagen negative cells . Note the uneven distribution of grains in e with little expression of type Ii collagen in the resting cartilage . a,b,c : 5x ; e: 10x ; f:20x . packed, show little extracellular matrix, and lack a specific arrangement. At the diaphysis ; however, a band of larger cells that represents the developing hypertrophic zone can be seen . In the dark field, the distribution of silver grains revealed a strong expression of type II collagen in the cells of the hypertrophic zone, with decreasing intensity toward the ends of the bone model . At this stage of development, type I expression is found to a small degree in the perichondreal sheet surrounding the bone anlage (not shown) . In the zones of desmal ossification, an intense hybridization signal to the type I probe was observed . Fig . Ic shows a cross section through a clavicle of the same developmental stage, hybridized to the type I probe . Type I positive osteoblasts are found in large numbers, irregularly arranged and densely packed at the periphery of the bone . With further development, secondary cartilage develops at the end and in the center of the clavicle . This tissue, which morphologically resembles early cartilage, also showed type II activity (not shown) . With further growth of the bone, the hypertrophic zone
enlarges and a small rim of periosteal bone is built around the diaphysis . At this point we find a segregation of chondrocyte activity in those that still express type II collagen and those that have ceased producing mRNA coding for type II collagen . Fig . le shows a band of type II negative cells at the diaphysis that is sharply separated from type II positive cells on both sides . This area of type II negative cells resembles the zone of the calcified cartilagenous matrix as shown by the van Kossa stain (Fig . ld) . Gene expression of type I collagen at this stage of development is shown in Fig . If. Osteoblasts at the newly formed bone collar showed strong type I activity corresponding to the calcification seen in Fig . Id (van Kossa stain) . To a lesser degree, type I positive cells can be found in fibroblasts of the perichondrium and in tendons . Type I positive cells were only found at the outside of the newly formed bone matrix . No type I activity was found in chondrogenic tissue . Around the 11th week of gestation, hypertrophic chondrocytes at the diaphysis are removed and replaced by the bone marrow cavity . At this stage, type II expression is found in all chondrocytes except the layer of cells adjacent to osseochondral
27 8
S . Mundlos et al . : Type I and II collagen gene expression
Collagen I II - ---------------a
b
C
-------------------Fig . 3 . Diagram showing the distribution of type I and type II collagen gene expression during the early development of the human long bones . (a) Sixth to seventh week, cartilagenous bone anlage; (b) ninth to 11th week, periosteal bone formation and calcification of cartilage at diaphysis ; (c) > 12th week, formation of a growth plate, desnial and enchondral ossification .
junction . Again, these cells do not show any type II expression and are located in the zone of provisional calcification . With further development, the area of type II negative cells becomes smaller and smaller . However, type II negative chondrocytes were always found at the osseochondral junction . Fig . 2b and Fig . 2c show the type I and type II collagen gene expression in a section of a humerus of the 15th week of gestation with a bright light image given in Fig . 2a . Parallel to the formation of the bone marrow, zones of enchondral ossification develop where osteoblasts that surround resting cartilagenous matrix show type I activity . By birth, a regular growth plate with hypertrophied cartilage arranged in columns has been formed (Fig . 2d) . Again, type II negative chondrocytes are found at the osseochondral junction (Fig. 2f) . At this stage, type II expression is mainly found in chondrocytes of the hypertrophic and to a lesser degree of the proliferative zone, while the resting cartilage shows little activity (Mg . 2e) . If attempts are made to quantify the type II collagen gene expression by counting the number of grains/cell, the chondrocytes in the embryonic and fetal tissue showed a three to four times higher number of grains/cell in the hypertrophic zone than in the resting zone. This difference was even more pronounced in the growth plate of the newborn, where the hypertrophic cartilage showed a 10 to 15 times higher number of grainslcell than the resting cartilage . Figure 3 summarizes the results . Discussion The developmental patterns of collagen distribution have been extensively studied in the chick by the use of specific antibodies and immunohistochemistry (von der Mark et al . 1976; von der Mark 1980) . Type I collagen was found to be present in tendons,
fibroblasts, perichondrium, and bone, whereas type II collagen was present in cartilage matrix . Using immunohistochetnical methods, a rather homogeneous distribution of type II collagen throughout the cartilage zones was observed, reflecting the uniform appearance of the extracellular matrix . The technique of in situ hybridization offers the advantage of detecting switches in gene expression before changes in extracellular matrix result . This is of special importance for investigations of the developing skeleton and the growth plate, because this tissue is subject to rapid transition of collagen types during growth . In the present study, we were able to identify the particular type of cell involved in production of type I or type II collagen during different stages of the development of the human skeleton using in situ hybridization with cDNA probes . Naturally, the reliability of the method rests on the specificity of the probes used . Type I and type II collagen are known to have a certain degree of homology . Consequently, a number of authors have selected certain parts of the collagen cDNA that show a lesser degree of homology and have used these fragments for in situ hybridization . Specificity was usually tested by northern blot or dot blot analysis . This procedure, however, does not rule out cross-reactivity using the conditions of in situ hybridization . The best control in our opinion is a section through developing bone where cartilage (containing no type I collagen) is next to newly formed bone (known to contain no type II Collagen) . These tissue sections served as internal controls during optimization of the hybridization conditions . Using the above described method, practically no cross-hybridization was observed . It should be emphasized that hybridization and washes were performed under high stringency conditions as described by other authors to prevent cross-hybridization between different collagen probes (Hayashi et al . 1986) . mRNA coding for type II collagen was exclusively found in chondrogenic tissue, and chondrocytes were always negative for type I expression . This is in contrast to the findings of Devlin and coworkers (1988), who were able to show significant levels of type I collagen mRNA in mature cartilage and type 11 expression of the mesenchyme in the developing chick . Similar results were obtained during in vitro chondrogenesis (Kosher et al . 1986) and in total RNA extracted from epiphyseal cartilage (Elima et al . 1987) . Since type II collagen is not present in nonchondrogenic tissue, and cartilage does not contain detectable amounts of type I collagen, the authors explain this phenomenon by regulation of the protein synthesis at the level of translation . However, cross-reactivity of the two collagen probes is not ruled out, especially if RNA probes are used (Hayashi et al . 1986) . Our results support the transcriptional level as the main control point of collagen synthesis . Figure la shows the cartilagenous model of the humerus with well developed perichondrium . Growth takes place by apposition from the perichondrium and through multiplication of cartilage cells already formed together with an increase in extracellular matrix . In the middle of the bone a layer of hypertrophied cells can be seen that shows a larger accumulation of gainslcell than the rest of the cartilage . This diffc.. .uce of type II collagen gene expression between the resting and the hypertrophied cartilage was observed throughout the development of the growth plate. In contrast, von der Mark (1980) describes a less intense staining with anti-type H collagen antibody of hypertrophied cartilage in the chick at a comparable stage of development . Possibly, these cells also produce large amounts of other extracellular matrix components that mask type II collagen antibody reaction . Concurrently with the formation of periosteal bone, an extraordinary switch of chondrocyte gene expression takes place . Chondrocytes at the diaphysis, which produced large amounts of type II collagen in earlier developmental stages, suddenly cease expressing the type II collagen gene . This sudden switch is not
S . Mundlos et al . : Type I and II collagen gene expression
paralled by any morphological change of the cells, whereas the extracellular matrix shows signs of calcification and is reduced to a small rim around the cells . The phenomenon just described was observed in all the developmental stages of the growth plate thereafter, and is in agreement with the findings of Sandberg & Vuorio (1987), who also describe Type II negative chondrocytes in this location in fingers of 15 week fetuses . With further growth, the band of type II negative cells becomes smaller and smaller and has almost disappeared by the 15th week . It seems that these cells are selectively removed by the protruding bone marrow . Nevertheless, type II expression in the zone of provisional calcification remains at a low level throughout further development, and type II negative cells can always be found at the osseochondral border . The reason for this switch remains obscure . Do these cells start producing type X collagen instead of type II? In the chick tibiotarsus, type X collagen was shown to occur only in chondrocytes of the lower hypertrophic zone that were undergoing calcification (Schmid and Conrad, 1982) . Type X collagen is assumed to play some role in the calcification of cartilagenous matrix . In vitro experiments could demonstrate a coordinate regulation of alkaline phosphatase and type X collagen production by embryo tibial chondrocytes, suggesting that their role in calcification may be coupled (Habuchi et al . 1985) . Thus, chondrocytes regenerated in the proliferative zone (Kember 1978) change their pericellular matrices on the way through the different zones of the growth plate until they reach the osseochondral junction . The composition of hyaline cartilage is modified in a progressive fashion to yield a tissue that is functionally distinct and able to undergo calcification . Concurrently with the development of hypertrophied chondrocytes, we observed a change in the amount of type II coding mRNA per cell, as judged by grains/cell . Chondrocytes of the upper hypertrophic zone and the zone of proliferation showed the largest number of grains/cell, whereas the resting zone showed a uniform but much lesser degree of staining . This phenomenon again was observed throughout the further steps of development of the growth plate . In the newborn, type II activity in the resting cartilage is even less pronounced, and type II expression is almost exclusively found in the hypertrophic zone . In the mature growth plate, growth takes place by cell division in the upper proliferative zone . In contrast to earlier stages in development, there is little appositional growth . The higher degree of collagen gene expression in the hypertrophic and proliferative zone when compared to the resting zone is most likely caused by the larger number of cell divisions in this area resulting in increased production of extracellular matrix . No type I activity was found in the zone of provisional calcification . In contrast, von der Mark (1980) describes staining with anti-type I collagen antibodies in the core of the calcified cartilage in the chick . Immunoreactive type I collagen was recently described by Horton et al . (1988) in the zone of provisional calcification of the human growth plate of the ileac crest . According to our results, a significant production of type I collagen by hypertrophied chondrocytes seems unlikely . In summary, in situ hybridization was shown to be a powerful tool for the investigation of developmental processes, especially if proteins of the extracellular matrix with low turnover rates are involved . We were able to complement earlier immunohistological studies and show that type II collagen is expressed in cartilage to a very different degree, depending on the developmental stage . Monitoring these patterns of gene expression during development will provide a basis for understanding normal growth and the pathophysiology of osteochondrodysplasias .
279
Acknowledgments : We are grateful to Dr . Vuorio for generously provid-
ing us with the probe for al(II) collagen . The clone for a2(I) collagen was kindly supplied by Drs . Prockop and Ramirez . References Bernard, M.P. ; Myers, J .C . ; Chu, M.L . ; Ramirez, F . ; Eikenberry, E .F. ; Prockop, D .J . Structure of a cDNA for proa2 chain of human type I procollagen . Biochemistry 22 :1139-1145 ; 1983 . Devlin, C.J. ; Brickell, P.M . ; Taylor, E .R. ; Hombruch, R .K . ; Craig, R.K . ; Wolpert, L . In situ hybridisation reveals differential spatial distribution of mRNAs for type I and type II collagen in the chick limb bud . Development 103:111-118, 1988 . Elima, K. ; Vuorio, T . ; Vuorio, E . Determination of the single polyadenylation site of the human pros 1(11) collagen . Nucleic Acids Res . 15 :9499-9504 ; 1987 . Gardner, E . Osteogenesis in the human embryo and fetus . Boume, G.H ., ed . The Biochemistry and Physiology of Bone . New York : Academic Press ; 1956 : 359-399 . Gardner, E . ; Gray, D.J. The prenatal development of the human femur . Am . J . Anat . 129:121-140 ; 1970. Habuchi, H . ; Conrad, E .H . ; Glaser, J .H . Coordinate regulation of collagen and alkaline phosphatase levels in chick embryo chondrocytes . J . Biol . Chem . 260 :13029-13034; 1985. Hafen, E . ; Levine, M . ; Garber, R.L . ; Gehring, W .J. An improved in situ hybridisation method for the detection of cellular RNAs in Drosophila tissue sections and its application for localizing transcripts of the homeotic Antennapedia gene complex. EMBO J . 2:617-623 : 1983 . Hayashi, M. ; Ninomiya, Y . ; Parsons, J . ; Hayashi, K . ; Olsen, B .R . ; Trelstad, R .L . Differential localization of mRNAs of collagen types I and II in chick fibroblasts, chondrocytes, and corneal cells by in situ hybridisation using cDNA probes. J Cell Biol. 102 :2302-2309; 1986 . Horton, W .A . ; Machado, M .M . Extracellular matrix alterations during enchondral ossification in humans . J. Orthop . Res. 6:793-803 ; 1988 . Kember, N .F. Cell kinetics and the control of growth in long bones . Cell Tissue Kinet . 11 :477-485 ; 1978 . Kosher, R.A . ; Kulyk, W .M . ; Gay, S .W. Collagen gene expression during limb cartilage differentiation . J. Cell Biol. 102:1151-1156 ; 1986. Kravis, D . ; Upholt, W .B . quantitation of type II procollagen mRNA levels during chick limb cartilage differentiation . Dev . Biol. 108:164-172 ; 1985 . Sandberg, M . ; Vuorio, E . Localisation of types I, II, and III collagen mRNAs in developing human skeletal tissues by in situ hybridization . J. Cell Biol. 104 :1077-1084 ; 1987 . Schmid, T .M . ; Conrad, H .E. Metabolism of low molecular weight collagen by chondrocytes obtained from histologically distinct zones of the chick embryo tibiotarsus . J . Biol . Chem . 257 :12451-12457 ; 1982 . Schmid, T .M . ; Linsenmayer, T .F . Immunohistochemical localization of short chain cartilage collagen (type X) in avian tissues . J . Cell Biol . 100 :598-605; 1985 . Stanescu, V . ; Stanescu, R . ; Maroteaux, P. Chemical studies on the human growth cartilage in fetuses and newborns . Biol. Neonate 23:432-445 ; 1973 . Stoker, N .G . ; Cheah, K.S .E . ; Griffin, R . ; Pope, F .M . ; Solomon, E. A highly polymorphic region 3' to human type II collagen gene . Nucleic Acids Res. 13 :4613-4622; 1985 . Swalla, B .J . ; Upholt, W .B . ; Solursh, M . Analysis of type II collagen RNA localisation in chick wing buds by in situ hybridisation . Dev . Biol . 125 :51-58 ; 1988 . von der Mark, H . ; von der Mark, K . ; Gay, S. Study of differential collagen synthesis during development of the chick embryo by immunofluorescence . I . Preparation of collagen type I and type II specific antibodies and their application to early stages of the chick embryo . Devl. Biol . 18 :237-249; 1976. von der Mark, K . Immunological studies on collagen type transition in chondrogenesis. Monroy, A . ; Moscona, A .A ., eds. Current topics in developmental biology, Vol . 14 . New York : Academic Press; 1980:199-225 . von der Mark, K . ; von der Mark, H . ; Gay, S. Study of differential collagen synthesis during development of the chick embryo by immunofluorescence . II. Localization of type I and type II collagen during bone development . Dev. Biol . 53:153-170 ; 1976 .
Received : July 28, 1989 Revised: December 20, 1989 Accepted: December 20, 1989
