Growth/differentiation factor 5 enhances chondrocyte maturation


  • Cynthia M. Coleman,

    1. Cartilage Biology and Orthopaedics Branch, National Institute of Arthritis, and Musculoskeletal and Skin Diseases, National Institutes of Health, Department of Health and Human Services, Bethesda, Maryland
    2. Department of Orthopaedic Surgery, Thomas Jefferson University, Philadelphia, Pennsylvania
    Search for more papers by this author
  • Rocky S. Tuan

    Corresponding author
    1. Cartilage Biology and Orthopaedics Branch, National Institute of Arthritis, and Musculoskeletal and Skin Diseases, National Institutes of Health, Department of Health and Human Services, Bethesda, Maryland
    2. Department of Orthopaedic Surgery, Thomas Jefferson University, Philadelphia, Pennsylvania
    • Cartilage Biology and Orthopaedics Branch, National Institute of Arthritis and Musculoskeletal and Skin Diseases, National Institutes of Health, Building 50, Room 1503, MSC 8022, Bethesda, MD 20892
    Search for more papers by this author

  • This article is a US Government work and, as such, is in the public domain in the United States of America.


Growth/differentiation factor 5 (GDF5) is required for limb mesenchymal cell condensation and joint formation during skeletogenesis. Here, we use a model consisting of long-term, high-density cultures of chick embryonic limb mesenchymal cells, which undergo the entire life history of chondrocyte development, to examine the effects of GDF5 overexpression on chondrocyte maturation. Exposure to GDF5 significantly enhanced chondrocyte hypertrophy and maturation, as determined by the presence of alkaline phosphatase activity, collagen type X protein production, and the presence of a sulfated proteoglycan-rich extracellular matrix. Histologic analysis also revealed an increase in cell volume and cellular encasement in larger lacunae in GDF5-treated cultures. Taken together, these results support a role for GDF5 in influencing chondrocyte maturation and the induction of hypertrophy in the late stages of embryonic cartilage development, and provide additional mechanistic insights into the role of GDF5 in skeletal development. Development Dynamics 228:208–217, 2003. Published 2003 Wiley-Liss, Inc.


Endochondral ossification, the process by which the bones of the limb are formed, involves the aggregation of mesenchymal cells to form cartilaginous tissue, which is subsequently replaced by mineralized bone (see reviews by DeLise et al., 2000a; Shum et al., 2003). Cartilage is characterized by the presence of an extracellular matrix (ECM) rich in collagen type II and sulfated proteoglycans secreted by chondrocytes. Further maturation of the chondrocyte involves cell cycle exit and hypertrophy, which includes cellular enlargement, a reduction in collagen type II secretion, an increase in collagen type X production, the presence of alkaline phosphatase, and the production of matrix vesicles. Terminal hypertrophic chondrocytes undergo apoptosis, which paves the way for vascular invasion and the transport of osteoblasts into the cartilaginous scaffold (Lewinson and Silbermann, 1992).

The continuous cartilaginous anlage present through the core of the limb bud early in development segments into individual skeletal elements through the formation of joints. The chick interzone forms from a population of predetermined cells, which flatten, become nonchondrogenic and form three layers: two layers of high cellular density surrounding one layer of low cellular density (Fell and Canti, 1934; Holder, 1977; Craig et al., 1987; Francis-West et al., 1999b). Subsequently, the joint interzone cavitates through a combination of altered ECM synthesis, apoptosis, and the coalescence of small fluid-filled cavities until the entire joint space is filled with synovial fluid (Edwards et al., 1994; Mori et al., 1995; Dowthwaite et al., 1998). In its final form, the joint consists of two superficial layers of cartilage and an internal fluid-filled space surrounded by the joint capsule.

Growth/differentiation factor 5 (GDF5), a member of the bone morphogenetic protein (BMP) family, is necessary for proper cellular condensation and joint formation during limb development. GDF5 mRNA expression localizes first to the condensing mesenchymal cells of the limb during endochondral ossification and then in the future joint spaces (Storm et al., 1994; Storm and Kingsley, 1996; Francis-West et al., 1999a). Its expression in the joint appears 24 hr before joint interzone formation and continues for a total of 2–3 days until all three layers of the interzone can be identified. Natural mutations in GDF5 are responsible for the brachypod (bp) phenotype in mice, resulting in alterations in skeletal patterning and development (Storm et al., 1994; Storm and Kingsley, 1996). The bp mouse has a normal axial skeleton and skull with severely malformed limbs. The long bones of the limb are reduced in length, and the autopod is flattened with multiple joint fusions and accessory bony elements (Gruneberg and Lee, 1973; Settle et al., 2003).

Ectopic in vivo expression of GDF5 results in joint fusions and enhanced cellular differentiation. GDF5 has been shown to restrict the expression of joint specific markers, including its own expression (Storm and Kingsley, 1999), but is unable to induce the expression of joint markers, initiate ectopic joint formation or influence cellular apoptosis (Merino et al., 1999). Taken together, these data suggest that GDF5 does not specify the location of a joint or initiate joint cavitation. However, its presence does positively influence cellular proliferation and differentiation, thereby possibly regulating epiphyseal development (Francis-West et al., 1999a; Storm and Kingsley, 1999). In vivo misexpression of CDMP1, the human homologue of GDF5, in mice results in a larger zone of hypertrophic chondrocytes in the long bones of the limb, and reduces the size of the proliferative zone (Tsumaki et al., 1999)— data supporting the possible role for GDF5 in chondrocyte differentiation and maturation.

Based on these findings, we hypothesize that, during embryonic joint formation, GDF5 produced in the developing joint may act in part by regulating chondrocyte maturation. In this study, we have used high-density cultures of chick embryonic limb mesenchymal cells to recapitulate endochondral ossification in vitro, and the effects of retrovirally mediated, constitutive overexpression of GDF5 were analyzed.


GDF5-Misexpression Stimulated Chondrocyte Enlargement

Embryonic chick mesenchymal cell micromass cultures were electroporated with RCAS-GDF5 or empty RCAS vectors on the day of plating, and grown for 7, 14, or 21 days in vitro as described previously (Mello and Tuan, 1999). Morphology of the cultures was examined by hematoxylin and eosin staining of histologic sections. After 7 days in culture (Fig. 1A,B), both cultures consisted of newly formed cartilage nodules containing cells with the rounded phenotype of differentiated chondrocytes. In day 14 cultures (Fig. 1C,D), GDF5-expressing cultures contained distinct, mature chondrocytes with large, rounded cell bodies and large lacunae consistent with cellular hypertrophy. Cells in control cultures, while obviously also undergoing hypertrophy, were smaller in size with no extensive lacunae formation. The process of hypertrophy progressed in both control and GDF5-expressing cultures. After 21 days in culture (Fig. 1E,F), cells in GDF5-expressing cultures continued to enlarge and elaborate extensive ECM, resulting in fewer cells per microscopic field, compared with control cultures; large lacunae were also apparent. In comparison, day 21 control cultures contained round cells of smaller size with less-extensive ECM, resulting in a higher cell density. These morphologic differences were quantified in Figure 2. The average cross-sectional area of individual cell bodies (Fig. 2A) and the average number of cells per microscopic field in a defined region of the culture (Fig. 2B) were quantified. These results showed that misexpression resulted in a higher number of larger cells and a significant reduction in cell density (n > 300; *P < 0.0001).

Figure 1.

Effect of growth/differentiation factor 5 (GDF5) misexpression on cellular morphology in long-term micromass cultures of limb mesenchymal cells. Micromass cultures infected with GDF5 (A,C,E) or control RCAS constructs (B,D,F) were fixed, paraffin embedded, sectioned, and stained with hematoxylin and eosin after 7 (A,B), 14 (C,D), and 21 (E,F) days in culture. Cellular morphology appears similar in the two groups on day 7. On day 14, GDF5-infected cultures (C) contained large, rounded, mature chondrocytes. After 21 days in vitro, GDF5-infected micromass (E) contained fewer, yet larger, cells per unit area compared with control cultures (F). Scale bar = 10 μm in F (applies to A–F).

Figure 2.

Quantitation of growth/differentiation factor 5 (GDF5) -mediated effects on cellular morphology on the basis of cell area and density in long-term micromass cultures of limb mesenchymal cells. Cross-sectional area of cells in day 21 cultures (A) and the total number of cells per microscopic field in a defined region of day 14 and 21 cultures (B) were determined from hematoxylin and eosin–stained sections. The size distribution profile (A) indicates a difference between control and GDF5-treated cultures in that GDF5-treated cultures contained a higher number of larger cells. A statically significant difference (n > 300; *P < 0.0001) in the number of cells per field of view (B) was also observed between treatment groups.

Enhanced Cartilage ECM Production in Cultures With Misexpressed GDF5

Alcian blue staining of sectioned micromass cultures revealed an abundant, sulfated ECM, characteristic of chondrocytes undergoing maturation (Fig. 3). Cultures expressing GDF5 for 7 days stained uniformly throughout the culture with Alcian blue (Fig. 3A). In comparison, day 7 control cultures consisted of nodules of differentiated chondrocytes interspersed with unstained, fibroblastic regions (Fig. 3B). After 14 days (Fig. 3C,D), cultures under both conditions stained intensely with Alcian blue. At day 21, GDF5-expressing cultures (Fig. 3E) exhibited a somewhat more intense staining of the cartilaginous matrix compared with control cultures (Fig. 3F). The difference in cell area and lacunae size noted with hematoxylin and eosin staining was also apparent in the Alcian blue–stained sections.

Figure 3.

Alcian blue staining of ECM in micromass cultures expressing growth/differentiation factor 5 (GDF5). Both GDF5-infected (A,C,E) and control (B,D,F) cultures stained positively with Alcian blue at days 7 (A,B), 14 (C,D), and 21 (E,F). B: Regions of control cultures at day 7 remained fibroblastic (arrow). A: Similar regions were not observed in GDF5-overexpressing cultures. Cultures stained with similar intensity at day 14. E: After 21 days in vitro, Alcian blue staining was enhanced in cultures exposed to misexpressed GDF5. Scale bar = 10 μm in F (applies to A–F).

Scanning electron microscopy was used to visualize the surface topography of the cultures at each time point (Fig. 4). Substantial ECM buildup was seen by day 7 (Fig. 4A,B), with the GDF5-expressing cultures showing more distinct cell-associated ECM, whereas control cells were more flattened with less accumulation of ECM (Fig. 4B). After 14 days (Fig. 4C,D), GDF5-expressing micromass cultures assumed a round shape, and a complex network of ECM and thick intercellular connections, more prominent than that in control cultures, was seen. At day 21, cells in GDF5-misexpressing cultures (Fig. 4E) were encased in a thick ECM that essentially filled all intercellular space, whereas control cultures (Fig. 4F) consisted of cells with a less pronounced coating of ECM, only slightly more substantial than that seen in day 14 cultures.

Figure 4.

Surface topography of day 21 growth/differentiation factor 5 (GDF5) -misexpressing and control micromass cultures as visualized by scanning electron microscopy (SEM). GDF5-infected (A,C,E) and control (B,D,F) cultures were examined by SEM at days 7 (A,B), 14 (C,D), and 21 (E,F). At day 7, cultures overexpressing GDF5 (A) show an increased amount of extracellular matrix (ECM) production, slight cell rounding, and a rougher topography as compared with the more flattened surface of day 7 control cultures (B). After 14 days in vitro, GDF5-expressing cells (C) were rounded and interconnected by a network of thick ECM compared with the cells in the control cultures, which appeared less organized and less embedded in ECM (D). Cultures expressing GDF5 for 21 days (E) displayed a thick, extensive, fibrous coating of ECM, whereas cells in control cultures (F) were still distinct and encased in less abundant ECM. Scale bars = 5.0 μm in A, 6.6 μm in B, 7.5 μm in C–E, 8.5 μm in F.

GDF5-Misexpression Enhances Chondrocyte Hypertrophy

Collagen type X protein production and the presence of alkaline phosphatase activity were used to assess chondrocyte hypertrophy in the long-term micromass cultures. mRNA and protein were extracted from GDF5-misexpressing and control micromass cultures after 7, 14, and 21 days in vitro. Reverse transcriptase-polymerase chain reaction (RT-PCR) analysis, using primers specific for chicken collagen type X, showed that the levels of collagen type X mRNA were enhanced in GDF5-expressing cultures at days 7, 14, and 21 (Fig. 5A,D; n = 3; *P < 0.0001). Immunoblotting also showed that collagen type X protein production was highly detectable in cultures infected with GDF5 compared with control cultures. Collagen type X protein level increased significantly in GDF5-misexpressing cultures from day 14 to day 21 (Fig. 5C,E; n = 3; *P < 0.0001).

Figure 5.

Collagen type X expression in growth/differentiation factor 5 (GDF5) -expressing cultures assayed by reverse transcriptase-polymerase chain reaction (RT-PCR) and immunoblotting. GDF5-overexpressing and control cultures at days 7, 14, and 21 were analyzed for collagen type X mRNA and protein. Collagen type X mRNA expression (A) was enhanced in all GDF5 overexpressing cultures, as analyzed by RT-PCR, using glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as a control for mRNA loading (B). Expected size of RT-PCR products: collagen type X, 440 bp; GAPDH, 452 bp. C: Immunoblot of GDF5-infected or RCAS-infected micromass cultures showed that collagen type X protein was detected at high levels in GDF5-expressing cultures on days 14 and 21 with an increased level of expression at day 21. Molecular weight marker (85 kDa) is as indicated. D: Quantitation of collagen type X mRNA by RT-PCR. Densitometry revealed a statistically significant increase in collagen type X mRNA level in all GDF5-infected cultures compared with control cultures. These experiments were repeated twice. All intensities were measured in triplicate and normalized to those of GAPDH (*P < 0.0001). E: Quantitation of collagen type X protein levels. A significant enhancement of collagen type X protein expression is observed in day 14 and day 21 GDF5-treated cultures. The data are presented as a percentage of the control.

For alkaline phosphatase, no activity was detected in both GDF5-overexpressing and control cultures on day 7 (data not shown) or day 14 cultures (Fig. 6A,B). However, after 21 days in vitro, cultures infected with GDF5 virus (Fig. 6C) stained more significantly for alkaline phosphatase activity than control cultures (Fig. 6D).

Figure 6.

Effect of growth/differentiation factor 5 (GDF5) expression on alkaline phosphatase activity in micromass cultures. Sectioned micromass cultures were stained with nitroblue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate for visualization of alkaline phosphatase activity. For both GDF5-expressing cultures and control cultures, no staining was observed after 7 days (data not shown) or after 14 days (A,B) in vitro. C,D: After 21 days in culture, cells exposed to misexpressed GDF5 stained intensely for alkaline phosphatase activity (C), whereas control cultures showed minimal staining (D). Scale bar = 10 μm in D (applies to A–D).

Cellular apoptosis was also analyzed to assess the effect of GDF5 misexpression on chondrocyte hypertrophy. As shown in Figure 7, GDF5-misexpressing and control cultures were paraffin-embedded, sectioned, and assayed by TUNEL staining on days 7 (Fig. 7B,D), 14 (Fig. 7F,H), and 21 (Fig. 7J,L) to identify apoptotic nuclei, and then counterstained with Hoechst dye (Fig. 7A,C,E,G,I,K). Apoptotic nuclei were observable in all treatment groups at all time points. Quantitative analysis (Fig. 7M) revealed a general increase in cellular apoptosis over time; however, there was no significant difference between GDF5-misexpressing and control cultures.

Figure 7.

Effect of growth/differentiation factor 5 (GDF5) expression on apoptosis in micromass cultures as visualized by terminal deoxynucleotidyl transferase-mediated biotinylated UTP nick end labeling (TUNEL) staining. Cultures were stained with Hoechst dye (A,C,E,G,I,K) to visualize all nuclei and by means of TUNEL–fluorescein isothiocyanate to observe apoptotic nuclei (B,D,F,H,J,L). Control (C,D,G,H,K,L) and GDF5-overexpressing micromass (A,B,E,F,I,J) were assayed at days 7 (A–D), 14 (E–H), and 21 (I–L). There was no significant difference in the percentage of cells undergoing apoptosis (TUNEL-positive nuclei: Hoechst dye-stained nuclei) between GDF5-expressing and control cultures, as depicted in M. Scale bar = 50 μm in L (applies to A–L).


By using an in vitro system consisting of long-term high-density micromass cultures of chick embryonic limb mesenchymal cells, we have shown here that overexpression of GDF5 significantly enhanced chondrocyte differentiation and maturation. This culture system allows for the easy in vitro examination of the role of GDF5 in the later stages of chondrogenesis. Specifically, GDF5 misexpression in these cultures altered cellular morphology, including increased cell size and enlarged lacunae, enhanced cellular production of sulfated cartilaginous matrix, increased activity of alkaline phosphatase, and up-regulated expression of collagen type X. These data are consistent with a role for GDF5 as a positive regulator of growth cartilage maturation and hypertrophy during skeletal development.

There is clear evidence implicating the importance of GDF5 in joint formation, particularly in the distal regions of the limb (Gruneberg and Lee, 1973; Storm et al., 1994). GDF5 has been hypothesized previously to regulate the shape of the epiphyseal regions by controlling proliferation (Gruneberg and Lee, 1973; Francis-West et al., 1999a) and accelerate the differentiation of chondrocytes to a hypertrophic state (Tsumaki et al., 1999). However, there is as yet no cellular mechanism for GDF5 in this skeletogenic process.

Our results showed that chondrocytes expressing GDF5 increased in cell size and were encased in large lacunae, a phenotype typical of mature, hypertrophic chondrocytes (Fig. 1). Complementing this morphology, there was a reduction in the number of cells in any given microscopic field of view due to the enlargement of cells as well as increased amount of a negatively charged ECM (Figs. 2–4). Therefore, the cellular phenotype resulting from exposure to GDF5 resembles that of a chondrocyte in the maturing growth cartilage and not that of a flattened cell in the developing joint interzone. Supporting our in vitro results, it has been shown that the in vivo misexpression of GDF5 will enhance cartilage development and differentiation, including larger cells and increased ECM production (Storm and Kingsley, 1999; Tsumaki et al., 1999, 2002). It is, therefore, likely that a principal action of GDF5 is to enhance chondrocyte differentiation and maturation in the cartilaginous regions adjacent to the joint.

Cultures misexpressing GDF5 also express, at a high level, markers typical of hypertrophic cartilage, i.e., collagen type X and alkaline phosphatase (Adams and Shapiro, 2002; Figs. 5, 6). Here, we demonstrate not only elevated collagen type X mRNA expression, but also increased translation to result in higher collagen type X protein production. In contrast, in control cultures, elevated collagen type X mRNA was not accompanied by increased levels of collagen type X protein. Possibly the presence of GDF5 is a permissive signal, allowing cells to progress to a hypertrophic state at an earlier point than those not exposed to GDF5. Gruneberg and Lee (1973) reported a decrease in long bone length and a delay in ossification in GDF5-deficient mice. The hypertrophic zone is also reduced in size in these mice (Nakamura et al., 1984). Storm and Kingsley (1999) also observed delayed ossification in bp limbs, which seem to eventually ossify by means of a pathway somewhat different from endochondral ossification. Each of these observations may be due to the absence of a hypertrophy-permissive signal by GDF5, resulting in a reduced number of chondrocytes entering hypertrophy, a reduced hypertrophic zone, and, therefore, a reduced length of the long bone.

Supporting this hypothesis, alkaline phosphatase activity was observed specifically in GDF5-treated cultures. High levels of alkaline phosphatase, particularly associated with matrix vesicles, have been shown previously to be a characteristic of the onset of cartilage calcification as a result of cartilage hypertrophy; the location of matrix vesicles is tightly associated with that of collagen type X (Ali et al., 1970; Osdoby and Caplan, 1981; Habuchi et al., 1985; Adams and Shapiro, 2002). Our findings, therefore, further suggest that during skeletal development, GDF5 acts in part on chondrocytes by promoting their differentiation, maturation, and hypertrophy, a sequence associated with the growth cartilage; this mode of action does not preclude GDF5 from also acting directly on the cells of the interzone. Such a role will also require GDF5 to act in a somewhat long-distance manner, i.e., the expression site within the interzone is some distance away from the growth plate where chondrocyte hypertrophy is taking place. It is noteworthy that the now well recognized parathyroid hormone–related protein (PTHrP)/Indian hedgehog (Ihh) crosstalk signaling pathway that regulates growth plate maturation also operates in a somewhat long-distance manner, i.e., the target cells are located some distance from the producer cells. Much research is currently devoted to discovering and deciphering the mechanistic roles of intermediary subcellular signaling molecules (Gli, Ptc, etc.; Jüppner, 2000; Kronenberg, 2003). Clearly, similar investigations are also needed to elucidate how GDF5 can act to modulate cartilage maturation and hypertrophy. Of interest, Wnt4, which is also expressed largely in the developing joint, accelerates chondrocyte differentiation and maturation. Like GDF5, the in vivo overexpression of Wnt4 results in an enlarged hypertrophic zone. It is possible, therefore, that Wnt4 and GDF5 share similar targets and have overlapping mechanisms of action. Additionally, Wnt4 expression has also been detected in the growth plate, suggesting it may also act in an immediately local manner (Hartmann and Tabin, 2000; Church et al., 2002). Finally, it remains possible that the hypertrophy inducing activity of GDF5 demonstrated in these cultures may not be important in embryonic skeletogenesis, but may play a role in the later stages of postnatal development, or in the pathogenesis of adult skeletal diseases. It is noteworthy, thus, that GDF5 has previously been localized to the articular surface of the adult joint, where it is thought to maintain surface integrity (Earlacher et al., 1998).

Finally, we have also analyzed the effect of GDF5 on cellular apoptosis, in view of the finding that the induction of apoptosis is believed to lead to joint cavitation. Although there is some indication of increased apoptosis in GDF5-misexpressing cultures, the differences were statistically insignificant (Fig. 7). This finding is in agreement with pervious in vivo studies (Merino et al., 1999). It should be noted that apoptosis is also a part of the terminal phase of chondrocyte hypertrophy, contributing to vascular and osteoblastic invasion of the cartilaginous scaffold (Lewinson and Silbermann, 1992; reviewed in Adams and Shapiro, 2002). Therefore, GDF5 promotes chondrocyte hypertrophy, but not the terminal stage of apoptosis.

Taken together, our observations suggest that a principal action of GDF5 on limb mesenchymal cells is to promote their chondrogenic differentiation and to enhance and/or permit their maturation into hypertrophic chondrocytes. These findings, thus, are more consistent with the role of GDF5 establishing the formation and inducing the maturation of cartilage elements adjacent to the articulating joint. We thus propose that GDF5 action on joint formation is an indirect yet enabling one, such that abnormal joint formation in GDF5 mutants may in fact be the result of incomplete establishment of growth cartilage units at specific sites of the cartilage anlage of the developing limb.


Embryonic Limb Mesenchymal Cell Micromass Cultures

The protocol described here is a modification of the methods described by Ahrens et al. (1977) and DeLise and Tuan (2000). Fertilized White Leghorn chicken eggs from Charles River SPAFAS (Preston, CT) were incubated at 37°C in a humidified incubator until Hamburger–Hamilton stage 23–24. Limb buds were dissected and digested in trypsin (Sigma, St. Louis, MO) and type I collagenase (Worthington, Lakewood, NJ) for 1 hr at 37°C. Dissociated cells were collected by centrifugation and adjusted to 40 × 106 cells/ml. Cells were then transfected by electroporation with RCAS-GDF5 cDNA, kindly provided by Dr. Philippa Francis-West (Francis-West et al., 1999a). A 400-μl aliquot of cells was placed in a 0.4-cm gap width electroporation cuvette (Invitrogen, Carlsbad, CA) along with 10 μg of RCAS-GDF5 or empty RCAS plasmid as a control. Cells were electroporated by using the following parameters: 380 V, 250 μF, ∞ Ω by using an Invitrogen electroporator. Electroporation leaves 50% of the cells viable, resulting in an approximate plating density of 20 × 106 cells/ml (DeLise and Tuan, 2000, 2002). Twelve-microliter drops were plated on Corning tissue culture plates (Corning, Corning, NY) and incubated at 37°C in a humidified incubator with 5% CO2 for 2 hr, after which they were flooded with medium.

Cultures were maintained in Ham's F-12 medium (Bio-Whittaker, Walkersville, MD), 10% fetal bovine serum (Atlanta Biologicals, Norcross, GA), 1% penicillin-streptomycin (GIBCO, Rockville, MD), and 0.2% embryo extract (Sigma) for the first 48 hr in culture. After 2 days, cultures were incubated with a 1:1 mixture of Ham's F-12 and DMEM (Bio-Whittaker, Walkersville, MD) containing 1.1 mM CaCl2, 1% glucose (Sigma), 1% penicillin–streptomycin, 10% fetal bovine serum, 2.5 mM β-glycerophosphate (Sigma), 0.3 mg/ml L-glutamine (Sigma), and 25 mg/L ascorbic acid (Sigma), as described previously (Mello and Tuan, 1999; DeLise et al., 2000b). Animal use was approved by the Institutional Animal Care and Use Committees of Thomas Jefferson University and the National Institutes of Health.

Histology of Micromass Cultures

After 7, 14, and 21 days in culture, micromass cultures were fixed in 4% paraformaldehyde (FD NeuroTechnologies, Baltimore, MD), dehydrated, embedded in paraffin (Fisher, Pittsburgh, PA), and sectioned at 8 μm thickness. Sections were stained with hematoxylin (FD NeuroTechnologies) and eosin (Sigma) to visualize culture morphology or Alcian blue to visualize sulfated proteoglycan secretion to the ECM (pH 1.0, Rowley Biochemical, Danvers, MA).

Assays for Cellular Hypertrophy

Cell size and number.

The average cellular area of each cell was calculated by using the best fitting elliptical outline drawn around the cell membrane using IP Lab version 3.3.5 software (n > 300 for each treatment group). Cell number was determined by averaging the number of cells in 3 to 9 fields of view at ×40 magnification.


Terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) was used to detect cells undergoing apoptosis. The Fluorescent In Situ Cell Death Detection Kit (Roche Molecular Biochemicals, Mannheim, Germany) was used according to the manufacturer's protocol to identify apoptotic nuclei. Sections were also stained with Hoechst 33342 (Molecular Probes, Eugene, OR) to visualize all nuclei. The number of TUNEL-positive cells per field of view was determined by examining three to five fields of view at ×40 magnification and expressed as a percentage of total number of cells estimated by Hoechst staining.


Cellular protein was extracted with a Tris-buffered saline (TBS) containing 1 μM CaCl2, 0.02% Triton-X and 0.02% NP40 for 3 hr at 4°C. Aliquots of cellular protein (30 μg each) were separated on a 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis gel (Bio-Rad, Hercules, CA), standardized with a Kaleidoscope marker (Bio-Rad), and electroblotted to a 0.22 μm nitrocellulose membrane (Bio-Rad). After blocking in 3% bovine serum albumin (Sigma) in TBS-containing 0.05% Tween-20 (TBS-T), the blots were incubated with a chicken collagen type X monoclonal antibody (X-AC9 from Developmental Studies Hybridoma Bank, Iowa City, IA) at a 1:250 dilution, then probed with an alkaline phosphatase–conjugated secondary antibody diluted 1:2,000 in TBS-T. Blots were developed by using 5-bromo- 4-cholor-3-indolyl phosphate (BCIP) and nitroblue tetrazolium (NBT; Zymed, San Francisco, CA).


RT-PCR was performed with Invitrogen One-Step RT-PCR kit (Invitrogen) following the manufacturer's protocol with an annealing temperature of 58°C for 35 cycles to yield a 440 bp product. The primers used for collagen type X are as follows: forward primer 5′ ATT GCC AGG GAT GAA GGG ACA TAG 3′ and reverse primer 5′ AGG TAT TCC TGA AGG TCC TCT TGG 3′ (Daumer et al., manuscript in preparation). RT-PCR products were analyzed electrophoretically and ethidium bromide staining intensities were determined densitometrically.

Scanning electron microscopy.

Cells were cultured on Thermanox coverslips (EM Sciences, Ft. Washington, PA) for 7, 14, or 21 days. Cultures were fixed in 2.5% glutaraldehyde (EM Sciences), dehydrated in a graded series of alcohol, and air-dried. Samples were sputter-coated (MED 010, Balzers Union) with 4 nm of gold and examined by using a Hitachi S-4500 scanning electron microscope equipped with a cold cathode field emission gun at an accelerating voltage of 10 kV. Photomicrographs were digitally recorded (Quartz Imaging Corporation, Vancouver, Canada).

Alkaline phosphatase.

Enzyme histochemistry was carried out on paraffin sections by using BCIP and NBT as substrates and counterstained with eosin as described by Stott and Chuong (1997) and Daumer et al. (manuscript in preparation).


All data were analyzed for statistical significance (P < 0.05) by using Fisher's protected least significant difference test for relative percent.


We thank Dr. Philippa Francis-West for supplying the RCAS-GDF5. We also thank Dr. David J. Hall and Kathleen Daumer for their advice during these experiments and Dr. Patrice Laquerriere for scanning electron microscopy. Initial stages of this work were supported in part by a grant from the National Institutes of Health.