Bone Morphogenetic Protein Signals Are Required for Cartilage Formation and Differently Regulate Joint Development During Skeletogenesis

Authors


  • The authors have no conflict of interest

Abstract

The bone morphogenetic protein (BMP) family consists of a large number of members and has diverse biological activities during development. Various tissues express pleural BMP family members, which seem to cooperatively regulate developmental events. Here, multiple BMP signals were inactivated in chondrocytes to clarify the function of BMPs during skeletogenesis. To obtain tissue-specific inactivation, Noggin gene (Nog) was overexpressed in cartilage under the control of α2(XI) collagen gene (Col11a2) promoter/enhancer sequences. The resultant transgenic mice lacked most of their cartilaginous components, suggesting that cartilage does not develop without BMP signals. These effects seem to be mediated through down-regulation of Sox9 expression. Conversely, specific BMP signals were activated in the skeleton by targeted expression of Bmp4 in cartilage and the resultant phenotype was compared with that of transgenic mice expressing growth and differentiation factor-5 (GDF-5), another BMP family member. Overactivity of Bmp4 in the skeleton caused an increase of cartilage production and enhanced chondrocyte differentiation, as GDF5 expression did, but it did not disturb joint formation as GDF5 did. During skeletogenesis, unique roles of each BMP may reside in the regulation of joint development. Together with the common effect on the cartilage overproduction by Bmp4 and GDF5 overactivation, loss of cartilage by inactivation of multiple BMPs in Noggin transgenic mice indicates that signals for cartilage production are reinforced by multiple BMPs exclusively. These conclusions may account for the reason why multiple BMPs are coexpressed in cartilage.

INTRODUCTION

CARTILAGE SERVES as the template for the development of skeletal components. Formation of the skeleton is initiated by mesenchymal cell condensation, forming primordial cartilage followed by endochondral ossification. This process includes proliferative and hypertrophic chondrocytes. As a final step in endochondral bone formation, the hypertrophic cartilage is invaded by blood vessels and osteoprogenitor cells, and the calcified cartilage is subsequently replaced by bone.

Bone morphogenetic proteins (BMPs) were originally identified as secreted signaling molecules that could induce endochondral bone formation.(1) Subsequent molecular cloning studies(2) have revealed that the BMP family consists of various molecules, including members of the growth and differentiation factor (GDF) subfamily. BMP family members have diverse biological activities during the development of various organs and tissues, as well as during embryonic axis determination.(3) Expression analysis of BMP family members has revealed that each protein has a unique tissue distribution, with several BMPs being coexpressed in the same tissues. Bmp2, Bmp4, Bmp7, and Gdf5 are expressed in perichondrium and are believed to regulate cartilage formation and development.(4–6) It has been reported that mice lacking Noggin, an antagonist for BMPs, showed oversized cartilage and impairment of joint development.(7) In Noggin-deficient mice, multiple BMPs, including BMP-2, BMP-4, BMP-7, and GDF-5, seem to be overactivated in cartilage, indicating that these BMPs as a group expand cartilage. However, the unique role of each BMP during skeletogenesis still remains obscure.

To unravel the in vivo function of each BMP, knockout mice have been created and analyzed. Bmp2 and Bmp4 show early expression in postimplantation embryos, and inactivation of these genes results in death at an early stage of gestation, before the onset of chondrogenesis.(8,9) Embryonic mice lacking BMP receptor type IA (BMPRIA)(10) or type II (BMPRII)(11) also fail to form the mesoderm and die by 9.5 days postcoitus (d.p.c.), before the onset of skeletogenesis, probably caused by impairment of the BMP signal transduction. In contrast, homozygous loss-of-function mutants for Bmp5,(12)Bmp6,(13)Bmp7,(14,15) and Gdf5 (also known as Cdmp1 and Bmp14)(16) are viable and exhibit limited malformation of skeletal components. Mice with inactivated BMP receptor type IB gene (BmprIB) are viable and exhibit defects that are largely restricted to the appendicular skeleton.(17) The relatively minor nature of these developmental defects suggests that other coexpressed BMP family members or their receptors may functionally compensate for the absence of a protein normally expressed in the same tissue. This concept is supported by the early embryonic death of Bmp5; Bmp7 double-knockout mice.(18) Therefore, function of BMPs during mammalian skeletogenesis is not well understood.

To examine the physiological role of BMPs during skeletogenesis in vivo, we inactivated multiple BMP signals in the skeleton by targeted expression of Noggin, a BMP antagonist, in cartilage. The resultant transgenic mice lacked most of their skeletal components, suggesting that BMP signaling is required for cartilage formation. Conversely, we activated specific BMP signals in skeleton by targeted expression of BMP-4 in mouse cartilage and compared the phenotype with that of transgenic mice expressing GDF-5,(19) another BMP family member. Excess BMP-4 activity in the skeleton caused an increase of cartilage and enhanced chondrocyte differentiation as GDF-5 overactivation did, but did not disturb joint formation as GDF-5 did. This difference indicates that each BMP has a unique role in joint formation during skeletogenesis. We conclude that formation of cartilage is mainly dependent on BMPs and that each BMP has a unique role, possibly explaining the known existence of multiple BMP expression in cartilage.

MATERIALS AND METHODS

Construction of transgenes

The Col11a2-based expression vector, Col11a2-LacZ, is identical with the 742lacZInt that have been described previously.(20)Col11a2-LacZ contains the Col11a2 promoter (−742 bp to +380 bp), an SV40 RNA splice site, the β-galactosidase reporter gene, and the SV40 polyadenylation signal, as well as 2.3 kilobases (kb) of the first intron sequence of Col11a2 as an enhancer (Fig. 1A).

Figure FIG. 1..

The structure of each transgene, cartilage-specific expression of the transgenes and the skeletal abnormalities in Col11a2-Nog and Col11a2-Bmp4 transgenic mice. (A) For preparation of the transgenes, a 742-bp promoter sequence and 2.3-kb intron enhancer sequence from the type XI collagen α2 chain gene (Col11a2) were ligated to the LacZ gene, Nog cDNA, Bmp4 cDNA, and GDF5 cDNA respectively. Brackets show the SV40 intron cassette. (B) Transgenic mice bearing Col11a2-LacZ were stained with X-gal to detect LacZ activity at 13.5 d.p.c. Transgene expression directed by the Col11a2 promoter/enhancer sequences was restricted to the primordial cartilage of the long bones of the limbs and to that of the ribs. Whole skeletons from (C) normal, (D) Col11a2-Nog, and (E) Col11a2-Bmp4 transgenic mice at 16.5 d.p.c. were stained with alcian blue (for cartilage) and alizarin red (for bone). Only traces of cartilage were recognizable in Col11a2-Nog transgenic mice, whereas cartilage was expanded in Col11a2-Bmp4 transgenic mice. The 7th ribs shown are from (F) normal, (G) Col11a2-Nog, and (H) Col11a2-Bmp4 transgenic mice. Scale bars, 2 mm.

A 0.7-kb DNA fragment covering the entire coding region of the mouse Nog complementary (c)DNA(21) was generated by polymerase chain reaction (PCR) using a forward primer tagged with NotI site (ATAAGAAGCGGCCGCTAGAGTCATTCAGCGGCTGGTCAGAGGATGGAGCGCTGCCCCAGCCTG) and a reverse primer with NotI site (ATAGTTTGCGGCCGCGAGTTCTAGCAGGAACACTTACACTC). A 1.2-kb DNA fragment covering the entire coding region of the mouse Bmp4 cDNA(22) was also generated by PCR using a forward primer tagged with NotI site (ATAAGAAGCGGCCGCTAGAGTCATTCAGCGGCTGGTCAGAGGATGATTCCTGGTAACCGAATGCTG) and a reverse primer with NotI site (ATAGTTTGCGGCCGCTCAGCGGCATCCACACCCCTCTAC). After digestion with NotI, these PCR fragments were cloned into the NotI site of Col11a2-LacZ by replacing the β-galactosidase gene to create Noggin-expression vector, Col11a2-Nog, and Bmp4-expression vector, Col11a2-Bmp4, respectively (Fig. 1A).

The previously reported GDF5 (CDMP1) expression vector, 742-CDMP1-Int,(19) was designated as Col11a2-GDF5, containing complete human GDF5 cDNA coding sequences ligated to the promoter and enhancer sequences of the Col11a2 and SV40 RNA splice site, which are identical with those of Col11a2-LacZ, Col11a2-Nog, and Col11a2-Bmp4 (Fig. 1A).

Generation of transgenic mice

The plasmids containing transgene constructs were digested with restriction emzymes to release the inserts from their vector sequences. Transgenic mice were produced by microinjecting each of the inserts into the pronuclei of fertilized eggs from F1 hybrid mice (C57BL/6 × C3H) as described previously.(20) Transgenic embryos were identified by PCR or Southern assays of genomic DNA extracted from the placenta or skin. Previously reported transgenic embryos bearing Col11a2-GDF5 (742-CDMP1-Int)(19) were reproduced for comparison of phenotype.

Staining of skeleton and histology

Cartilage and bones of embryos and newborn mice were stained as previously described.(23) After skin and internal organs were removed, samples were fixed in 96% ethanol for 2 days followed by staining with alcian blue (Sigma-Aldrich, St. Louis, MO, USA) solution (80 ml 96% ethanol, 20 ml acetic acid, and 15 mg alcian blue) for 2 days. The samples were dehydrated in 100% ethanol for 5 days and immersed in 1% KOH for 2 days. The samples were stained with 0.001% alizarin red S (Sigma-Aldrich) solution in 1% KOH for 2 days, dehydrated in graded solutions of glycerin, and stored in 100% glycerin. For histological analysis, embryos were dissected with a stereomicroscope, fixed in 4% paraformaldehyde, processed, and embedded in paraffin. Serial sections were prepared and stained with hematoxylin and eosin, safranin O-fast green-iron hematoxylin (Sigma-Aldrich).

In situ hybridization and probes

Digoxigenin-11-UTP-labeled single-strand RNA probes were prepared using a DIG RNA labeling kit (Boehringer Mannheim, Indianapolis, IN, USA) according to the manufacturer's instructions. cDNAs described below were used to generate antisense and sense probes. Hybridization was performed as described previously.(24) Briefly, after deparaffinization, the sections were treated with 10 μg/ml of proteinase K for 15 minutes at room temperature and subjected to 0.2N HCl to inactivate endogenous alkaline phosphatase. Hybridization was performed at 50°C in 50% formamide, and washes were carried out at a stringency of 2× SSC containing 50% formamide at 55°C. The slides were subjected to 10 μg/ml of RNAse A in TNE (10 mM Tris-HCl [pH 8.0], 500 mM NaCl, and 1 mM EDTA) at 37°C for 30 minutes for digestion of nonhybridized transcripts and washed. A Genius Detection System (Boehringer Manheim) was used to detect signals according to the manufacturer's instructions.

Col2a1 and rat Sox9 cDNAs were obtained from Y. Yamada (National Institute of Health, Bethesda, MD, USA).(25) Mouse type IIA procollagen cDNA (exon 2) was from L.J. Sandell (Washington University, St. Louis, MO, USA).(26) Mouse Col10a1 cDNA (pRK26) was provided by K.S.E. Cheah (University of Hong Kong, Hong Kong).(27)

RESULTS

Targeted expression of transgene and transgenic mice expressing Noggin in cartilage

Four DNA constructs were prepared to generate transgenic mice (Fig. 1A). We first expressed β-galactosidase reporter gene (LacZ) under the control of the promoter and first intron enhancer sequences derived from α2(XI) collagen gene (Col11a2). As reported,(20) the Col11a2 promoter/enhancer sequences started to direct expression to mesenchymal condensation in limbs at 12.5 d.p.c. At 13.5 d.p.c., the transgenic mice showed clear X-gal staining specifically in the primordial cartilage of the long bones of the limbs and ribs (Fig. 1B), when cells in mesenchymal condensation differentiated into chondrocytes. Cranial components undergoing membranous ossification did not show X-gal staining.

We next tried to inactivate BMP signals in a tissue-specific manner to study the role of BMPs in skeletal development. For this purpose, we expressed Nog in cartilage to effectively block multiple BMP signals under the control of the promoter and first intron enhancer sequences derived from Col11a2. Noggin has been reported to antagonize the activities of BMP-2, BMP-4, BMP-7, and GDF-5 by binding to these proteins and preventing interaction with their receptors.(21,28) The transgene construct, Col11a2-Nog, was introduced into the pronuclei of fertilized eggs to generate transgenic mice. The Col11a2-Nog transgenic mice were stillborn due to respiratory failure; therefore, we analyzed pleural generation zero (G0) embryos (Table 1). We obtained seven mice with abnormal appearances. Average size of crown-rump length of the mice was 10% smaller than that of normal littermates. These mice showed striking skeletal defects, whereas other tissues (including the viscera, muscle, and skin) were histologically normal because transgene expression was restricted to the skeletal tissues (not shown). Alcian blue staining of cartilage showed that all the cartilage components of Col11a2-Nog transgenic mice were severely hypoplastic compared with those of normal mice (Figs. 1C and 1D). Col11a2-Nog transgenic mice only showed traces of rib cartilage, unlike the rodlike normal rib cartilage (Figs. 1F and 1G). Primordial cartilage of long bones of limbs was very hypoplastic in Col11a2-Nog transgenic mice compared with that of normal mice (Figs. 2A and 2B). Alizarin red staining of skeleton showed that the rib bones and the long bones of the limbs were severely hypoplastic in Col11a2-Nog transgenic mice compared with those of normal mice. To our analysis, every skeletal components expressing transgene was affected and minimally formed. Alizarin red staining of bone showed that skull was nearly normal, probably because the transgene was not expressed in bone undergoing membranous ossification.

Table Table 1.. Production Frequency of Transgenic Mice
original image
Figure FIG. 2..

Hindlimb skeletons of Col11a2-Nog, Col11a2-Bmp4, and Col11a2-GDF5 transgenic mice. Hindlimb skeletons from (A) normal, (B) Col11a2-Nog, (C) Col11a2-Bmp4, and (D) Col11a2-GDF5 transgenic mice at 16.5 d.p.c. were stained with alcian blue (for cartilage) and alizarin red (for bone). (B) Only traces of cartilage were recognizable in Col11a2-Nog transgenic mice. (C) Col11a2-Bmp4 transgenic mice showed well-formed joints, whereas (D) Col11a2-GDF5 transgenic mice exhibited skeleton with completely fused joints. Scale bar, 1 mm.

Transgenic mice expressing Bmp4 and GDF5 in cartilage

We next attempted to activate single BMP signals in cartilage to characterize the action of each BMP ligand during skeletal development. We generated transgenic mice expressing Bmp4 in cartilage under the control of the Col11a2 promoter/enhancer sequences. The Col11a2-Bmp4 transgenic mice were dead at birth due to respiratory failure; therefore, pleural G0 embryos were analyzed. Activation of BMP-4 signals in cartilage led to an increase of cartilage. Alcian blue staining showed that the entire cartilaginous skeleton was enlarged and thickened when compared with normal mice (Figs. 1C and 1E). The rib cartilage of these Col11a2-Bmp4 transgenic embryos was three times thicker in diameter on average than normal rib cartilage (Figs. 1F and 1H), resulting in loss of the intercostal spaces in the transgenic mice (Fig. 1E). We obtained 12 G0 transgenic embryos with such changes in their cartilage. Ten of the 12 embryos showed well-formed joints (Fig. 2C), while 2 embryos exhibited partially fused joints (Table 1). To elucidate functional difference between each BMPs, we generated mice expressing GDF5 under the control of the identical Col11a2 promoter/enhancer sequences as reported(19) and compared their phenotype with Col11a2-Bmp4 transgenic mice. The skeletal phenotype of the Col11a2-Bmp4 transgenic mice was very similar to Col11a2-GDF5 transgenic mice. Both transgenic mice had a very similar chodrodysplasia-like skeletal phenotype with kyphosis and expansion of cartilage (Fig. 1E).(19) However, joint formation was different between these two types of transgenic mice. For Col11a2-GDF5 transgenic mice, we obtained nine G0 embryos with expanded skeleton, and seven of them showed completely fused joints (Table 1, Fig. 2D). These results suggest that BMP-4 and GDF-5 have similar roles in cartilage formation and different functions in joint formation during development.

Cartilage formation and BMP signals

Considering that the final target of BMP signaling in cartilage should be the genes encoding cartilage matrix components, we examined gene expression in the limb cartilage of transgenic mice using in situ hybridization. Transcriptional factor Sox9 binds to the regulatory sequences of the type II collagen gene (Col2a1) and Col11a2 gene to activate their expression.(25,29,30) Therefore, we analyzed expression of Sox9 in cartilage. The number of Sox9-positive cells in Col11a2-Nog transgenic mice was dramatically decreased compared with normal mice (Figs. 3D and 3E). In accordance with depletion of Sox9-positive cells, expression of Col2a1 was also decreased in Col11a2-Nog transgenic mice compared with normal mice (Figs. 4D and 4E). In addition, histological analysis showed that the staining intensity of cartilage with safranin O was dramatically reduced in Col11a2-Nog transgenic mice (Figs. 3A, Figs. 3B, Figs. 4A, and Figs. 4B), suggesting low content of glycosaminoglycan, a component of proteoglycan in cartilage.

Figure FIG. 3..

Expression of Sox9 and type IIA collagen mRNA in developing cartilage. Sections of the distal phalanges from (A) normal, (B) Col11a2-Nog, and (C) Col11a2-Bmp4 transgenic mice at 16.5 d.p.c. were stained with safranin O to detect the extracellular matrix of cartilage (red). Semiserial sections were used to assess (D-F) Sox9 expression and (G-I) type IIA collagen mRNA expression by in situ hybridization. Sox9 is believed to regulate the transcription of genes encoding various extracellular matrix components of cartilage, including types II and XI collagen. The number of cells expressing Sox9 was decreased in (E) Col11a2-Nog transgenic mice and was increased in (F) Col11a2-Bmp4 transgenic mice when compared with (D) normal mice. (I) Expression of type IIA mRNA (TypeIIA), the longer form of type II collagen mRNA generated by alternative splicing, was increased in cells around the primordial cartilage of Col11a2-Bmp4 transgenic mice, indicating the enhanced commitment of mesenchymal cells to the chondrocytic lineage. Scale bar, 100 μm.

Figure FIG. 4..

Influence of the modification of BMP signaling on chondrocyte differentiation. Sections of metacarpals from (A) normal, (B) Col11a2-Nog, and (C) Col11a2-Bmp4 transgenic mice at 16.5 d.p.c. were stained with safranin O to detect the extracellular matrix of cartilage (p, proliferating chondrocytes; h, hypertrophic chondrocytes). Semiserial sections were used to assess (D-F) type II collagen gene (Col2a1) and (G-I) type X collagen gene (Col10a1) expression by in situ hybridization. In wild-type metacarpals, both ends of each skeletal component consisted of proliferating chondrocytes marked by (D) Col2a1 expression and the cartilage had (A) abundant extracellular matrix. Differentiation proceeds from the central part of each skeletal component, with chondrocytes becoming (A) hypertrophic and changing from production of (D) type II collagen to (G) type X collagen. In Col11a2-Nog transgenic mice, (B) the primordial cartilage was hypoplastic and contained little extracellular matrix with (E) weak Col2a1 expression. The cartilage only contained Col2a1-positive cells and expression of (H) the Col10a1 was not detected, suggesting impairment of the differentiation process. On the other hand, the primordial cartilage was expanded in Col11a2-Bmp4 transgenic mice and contained (C) abundant matrix with (F) Col2a1-positive cells at both ends of each skeletal component. At the central part of each component, (C) the zone of hypertrophic chondrocytes was markedly wider than normal and (I) expression of the Col10a1 was prominent indicating the acceleration of chondrocyte differentiation. Scale bar, 100 μm.

Conversely, there was a marked increase of cells expressing Sox9 in Col11a2-Bmp4 transgenic mice compared with normal mice (Fig. 3, D and F). Primordial cartilage of Col11a2-Bmp4 transgenic mice was wider than that of normal mice (Figs. 3A,3C, Fig. 4A, and Fig. 4C) and filled with chondrocytes expressing Col2a1 intensely at epiphyseal regions (Fig. 4F). Widening of primordial cartilage in Col11a2-Bmp4 transgenic mice may be attributed to increased number of chondroprogenitor cells around cartilage. This idea was tested by expression analysis of type II collagen mRNA. In the early stage of chondrocyte development, two forms of type II procollagen are generated by alternative splicing of the exon 2 sequence.(31) The longer form (type IIA) containing the exon 2 sequence is predominantly expressed by immature chondroprogenitor cells.(26) Expression of type IIA mRNA was greater in the cells around Col11a2-Bmp4 transgenic cartilage than in cells in wild-type cartilage (Figs. 3G and 3I), suggesting the increased number of chondroprogenitor cells in BMP-4 transgenic cartilage.

Chondrocyte differentiation in transgenic mice

After commitment of mesenchymal cell to chondrocytic lineage, proliferating chondrocytes produce the short form of type II collagen(26) (Fig. 4D) and build an extracellular matrix for cartilage that contains abundant glycosaminoglycans, which can be stained with safranin O (Fig. 4A). Along with differentiation to mature hypertrophic chondrocytes, the expression of Col2a1 ceases, and the cells begin to express type X collagen gene (Col10a1; Fig. 4G), so that the matrix architecture becomes suitable for subsequent bone formation. Compared with normal mice, Col11a2-Nog transgenic mice had hypoplastic cartilage with weak safranin O staining (Figs. 4A and 4B). In situ hybridization showed that the cartilage of these mice contained cells expressing the Col2a1 gene (Fig. 4E) but lacked Col10a1-positive cells (Fig. 4H). In addition, type IIA mRNA, a marker of immature chondroprogenitor cells, was not only expressed by peripheral cells but also by cells at the center of the primordial cartilage (Fig. 3H). These findings suggest that Col11a2-Nog transgenic cartilage was composed of relatively immature chondrocytes.

On the other hand, Col11a2-Bmp4 transgenic mice showed expansion of cartilage, with an increase in the thickness of the zone of hypertrophic chondrocytes (Fig. 4C). The number of Col10a1-positive cells was also increased in Col11a2-Bmp4 transgenic mice compared with normal mice (Figs. 4G and 4I).

DISCUSSION

To examine the physiological function of BMPs during skeletogenesis in vivo, we generated transgenic mice expressing Noggin or BMPs under the control of the same cartilage-specific promoter/enhancer sequences derived from Col11a2. The resultant transgenic mice showed distinct skeletal abnormalities, although their other tissues developed normally because of the high tissue specificity of the promoter/enhancer sequences. These results provide convincing evidence that BMPs play an important role in mammalian skeletogenesis.

BMPs are required for cartilage formation

A striking feature of the Col11a2-Nog transgenic mice was the absence of nearly all cartilage. In these mice, Noggin was overexpressed in the mesenchymal condensation at 12.5 d.p.c. and subsequently was overexpressed in the proliferating chondrocytes of all primordial cartilage. Noggin has been reported to antagonize the activities of BMP-2, BMP-4, BMP-7, and GDF-5 by binding to these proteins and preventing interaction with their receptors.(21,28) Therefore, an excess amount of Noggin might have dramatically depressed the activities of these BMPs in the cartilage of Col11a2-Nog transgenic mice.

It has been shown that overactivity of BMPs causes expansion of the cartilage in Noggin knockout mice(7) and GDF-5 transgenic mice.(19) In addition to these findings, a definitive assessment of the importance of BMP in normal skeletogenesis might be achieved by loss-of-function studies. Inactivation of Bmp7 causes fused ribs and polydactyly of the hindlimbs,(14,15) whereas the loss of GDF-5 activity results in shortening of the appendicular skeleton in brachypod (bp) mice.(16) In these mice, cartilage is generally formed despite morphological changes to limited parts of the skeleton. Such relatively minor changes of the cartilage raise the possibility that other coexpressed BMP family members can compensate functionally for the absence of a protein in these mice. Our results also support this possibility. Expansion of cartilage in Col11a2-Bmp4 transgenic mice indicates that BMP-4 has a potent cartilage-forming effect. A study in Col11a2-GDF5 transgenic mice has shown that GDF-5 also has a potent cartilage-forming activity.(19) Therefore, the signals for formation of cartilage seem to be reinforced by multiple BMPs. Conversely, the absence of most cartilage in Col11a2-Nog transgenic mice clearly indicates the importance of BMPs for cartilage formation, because very little cartilage developed without BMPs. From these results, we speculate that multiple BMPs may be expressed in cartilage to ensure its formation during development, because of a highly important role of BMPs in cartilage formation.

Regarding bone formation in Col11a2-Nog transgenic mice, the calvarium appeared to be normal because the transgene was not expressed throughout the process of membranous ossification. On the other hand, the ribs and the long bones of the limbs were severely hypoplastic in these mice, with the limb bones being more severely affected (Fig. 1D). During endochondral ossification, Col11a2 promoter/enhancer sequences are reported to direct expression in proliferating chondrocytes and weakly in hypertrophic chondrocytes, but not in bone.(20) Therefore, hypoplasia of the limb bones and ribs in the transgenic mice was a consequence of hypoplastic primordial cartilage.

BMP-4 and GDF-5 have distinct roles in joint formation

The general skeletal phenotype of Col11a2-Bmp4 transgenic mice was very similar to that of transgenic mice expressing GDF5 under the control of the same Col11a2 promoter/enhancer sequences. Col11a2-Bmp4 and Col11a2-GDF5 transgenic mice both had a very similar chodrohysplasia-like skeletal phenotype with kyphosis and expansion of the cartilage (Fig. 1E).(19) These similar gross skeletal abnormalities confirmed that the pattern and level of transgene expression did not differ between Col11a2-Bmp4 and Col11a2-GDF5 transgenic mice. However, joint formation showed differences between these two types of mice. Col11a2-Bmp4 transgenic mice usually had well-formed joints, whereas Col11a2-GDF5 transgenic mice usually showed fusion of the joints. This difference may suggest the existence of unique signaling pathways in cartilage for each of these BMP family members. Various receptors for BMPs (BMPRs) have been identified so far, and it is known that the affinity for these receptors differs between BMPs.(32) Experiments using retroviral vectors to deliver activated BMPRs in chicks have demonstrated that BMPRIA and BMPRIB regulate distinct processes in the formation and differentiation of cartilage.(5) In addition, various Smads and other molecules may transduce intracellular signals from BMPs. It remains to be determined how each BMP activates specific signals and exerts its unique effect during skeletogenesis.

It has been reported that Noggin-deficient mice showed excess cartilage formation and impaired joint development due to BMP overactivity.(7) As shown in this study, activation of BMP-4 or GDF-5 also caused the expansion of cartilage, but joint formation was much less disturbed in Col11a2-Bmp4 transgenic mice (Fig. 2C) than in Col11a2-GDF5 transgenic mice (Fig. 2D). Therefore, the cartilage changes in Noggin knockout mice seem to arise from a combination of abnormalities caused by the enhanced signaling of several BMPs, including GDF-5 and BMP-4. Comparison of the phenotype of Noggin knockout mice with transgenic mice expressing each Bmp in cartilage may help to elucidate the mechanism by which multiple BMPs cooperate in the regulation of skeletal development. We speculate that the joint fusion seen in mice with inactivation of the Noggin gene may be caused by GDF-5 among the various BMPs expressed in cartilage.

Recently, it was proposed that Wnt-14 plays a critical role in the initiation of joint development and in the spacing of the joints.(33) Wnt-14 expressed in the developing joint interzone may induce expression of Gdf5, which could act on neighboring cartilage elements to prevent the induction of a new interzone. This action seems to be modulated by BMPs that are produced by cells surrounding cartilage elements.(34) BMP-2 and BMP-4 are expressed in perichondrium.(35) Our results support this mechanism for the spacing of the joints. Fused joints in Col11a2-GDF5 transgenic mice indicated that GDF-5 may inhibit joint formation, whereas well-formed joints in Col11a2-Bmp4 transgenic mice suggest that BMP-4 may promote joint formation. BMP-4 and GDF-5 may play opposing roles during the process of joint formation in the downstream of Wnt-14 signals.

Mechanism for regulation of cartilage formation by BMPs

Cartilage is composed of chondrocytes embedded in an abundant extracellular matrix. Histological examination of transgenic mice showed that cartilage formation occurred through production of matrix and an increase in the number of chondrocytes. First, we analyzed the expression of the Col2a1 and Sox9 genes to examine how BMPs controlled the production of cartilage matrix components. Sox9 encodes a transcriptional factor that regulates the expression of cartilage-specific collagen genes, including Col2a1 and Col11a2.(25,29,30) In Col11a2-Nog transgenic mice (in which BMP activity might be depressed), expression of Sox9 was decreased, probably resulting in a decrease of Col2a1 expression. This observation is consistent with the effects of Noggin on the limbs of chicks when delivered with a retroviral vector.(36) On the other hand, overactivation of BMP-4 in Col11a2-Bmp4 transgenic mice might cause an increase of Sox9 expression, leading to high levels of Col2a1 expression. It has been reported that inactivation of Sox9 results in the abolition of cartilage formation, because there is no cartilage in teratomas derived from Sox9−/− embryonic stem cells.(37) The severe cartilage hypoplasia in Col11a2-Nog transgenic mice seems to be a similar result to the effect of inactivating Sox9 in embryonic stem cells. These observations suggest that BMPs may control cartilage formation by regulating the expression of the Sox9 gene. In this context, it has also been reported that application of BMP-2 to chick limbs(38) and mesenchymal cells(39) results in the upregulation of Sox9.

Next, we examined the mechanism of chondrocyte proliferation. In the early stage of chondrocyte development, two forms of type II procollagen are generated by alternative splicing of exon 2.(31) The longer form (type IIA) containing the exon 2 sequence is predominantly expressed by immature chondroprogenitor cells.(26) Expression of type IIA collagen mRNA was greater in the cells around Col11a2-Bmp4 transgenic cartilage than in cells in wild-type cartilage (Figs. 3G and 3I). An increase of type IIA mRNA expression is also observed in the cartilage of transgenic mice with overexpression of GDF5.(19) In these mice, the activation of BMP signaling in cartilage might enhance the commitment of mesenchymal cells to the chondrocytic lineage, contributing to expansion of the primordial cartilage.

BMP signaling and chondrocyte differentiation

A striking feature in the cartilage of Col11a2-Bmp4 transgenic mice was the increased thickness of the hypertrophic zone accompanied by a reduced thickness of the proliferating chondrocyte zones. Similar findings were also observed in the cartilage of Col11a2-Gdf5 transgenic mice.(19) Endochondral bone formation is initiated when chondrocytes in the center of the primordial cartilage proliferate and differentiate into hypertrophic chondrocytes. Hypertrophic chondrocytes are eventually replaced by osteoblasts in the process of bone formation. This change radiates outward with formation of the growth plates at both ends of the primordial cartilage and these events are represented histologically by zones of proliferative and hypertrophic chondrocytes.(40) The reduction in the height of the proliferating zone in Col11a2-Bmp4 transgenic mice seems to be caused by accelerated differentiation into hypertrophic chondrocytes. It is conceivable that enhanced differentiation of these cells into hypertrophic chondrocytes caused the increased height of the zone of hypertrophy in the transgenic mice. In situ hybridization with Col10a1, a marker for hypertrophic chondrocytes, showed enlargement of the hypertrophic zone. Conversely, Col11a2-Nog transgenic mice had hypoplastic cartilage that lacked signals for Col10a1 (Fig. 4H). In addition, type IIA collagen mRNA, a marker of immature chondroprogenitor cells, was expressed by most of the chondrocytes in primordial cartilage (Fig. 3H), suggesting that Noggin may inhibit differentiation and maintain chondrocytes in an immature state. Taken together with the absence of mature hypertrophic chondrocytes in Col11a2-Nog transgenic mice, the increase of hypertrophic cells in Col11a2-Bmp4 (Fig. 4C) and Col11a2-Gdf5 transgenic mice(19) indicates that activation of BMP-4 or GDF-5 results in the acceleration of chondrocyte differentiation.

Role of BMPs in cartilage development

By overexpressing Nog in chondrocytes, we created mice that lacked cartilage, thus showing that BMP signaling is essential for cartilage development. In addition, we characterized the influence of BMP-4 on cartilage by assessing the response to its overactivation in mouse chondrocytes. Activation of BMP-4 resulted in cartilage expansion and promoted chondrocyte differentiation, as did activation of GDF-5. These observations indicate that the signals for cartilage production and chondrocyte differentiation are reinforced by multiple BMPs. In addition, the lack of cartilage in Col11a2-Nog transgenic mice suggests that the loss of BMP signals could not be compensated by other growth factors such as fibroblast growth factors (FGFs) or hepatocyte growth factor (HGF), despite possible interactions of the intracellular signaling pathways for these factors.(41) Comparison of skeletal differences between Col11a2-Bmp4 and Col11a2-GDF5 transgenic mice suggested the possible existence of unique signaling pathways in the cartilage for each BMP family member. These unique intracellular signaling pathways are subjects for further investigation.

Acknowledgements

We are grateful to Richard M. Harland for the mouse Noggin cDNA plasmid and Y. Yamada, L. Sandell, and K.S.E. Cheah for probes. We thank K. Itoh, K. Takaoka, A. Nifuji, and F.P. Luyten for advice. Financial support was partly provided by the Japan Spina Bifida and Hydrocephalus Research Foundation and the Japanese Ministry of Education (13470309).

Ancillary