MICROSCOPY RESEARCH AND TECHNIQUE 43:156–173 (1998) Chondrogenic Potential of Skeletal Cell Populations: Selective Growth of Chondrocytes and Their Morphogenesis and Development In Vitro LOUIS C. GERSTENFELD,1* CYRIL D. TOMA,2 JONATHAN L. SCHAFFER,3 AND WILLIAM J. LANDIS4 1Musculoskeletal Research Laboratory, Boston University Medical Center, Boston, Massachusetts 02118 of Orthopedic Surgery, University of Vienna Medical School, Vienna, Austria 3Department of Orthopedic Surgery, Brigham and Woman’s Hospital and Harvard Medical School, Boston, Massachusetts 02115 4Department of Orthopedic Surgery, Children’s Hospital and Harvard Medical School, Boston, Massachusetts 02115 2Department KEY WORDS type X collagen; hypertrophic cartilage; bone formation ABSTRACT Most vertebrate embryonic and post-embryonic skeletal tissue formation occurs through the endochondral process in which cartilage serves a transitory role as the anlage for the bone structure. The differentiation of chondrocytes during this process in vivo is characterized by progressive morphological changes associated with the hypertrophy of these cells and is defined by biochemical changes that result in the mineralization of the extracellular matrix. The mechanisms, which, like those in vivo, promote both chondrogenesis in presumptive skeletal cell populations and endochondral progression of chondrogenic cells, may be examined in vitro. The work presented here describes mechanisms by which cells within presumptive skeletal cell populations become restricted to a chondrogenic lineage as studied within cell populations derived from 12-day-old chicken embryo calvarial tissue. It is found that a major factor associated with selection of chondrogenic cells is the elimination of growth within serum-containing medium. Chondrogenesis within these cell populations appears to be the result of permissive conditions which select for chondrogenic proliferation over osteogenic cell proliferation. Data suggest that chondrocyte cultures produce autocrine factors that promote their own survival or proliferation. The conditions for promoting cell growth, hypertrophy, and extracellular matrix mineralization of embryonic chicken chondrocytes in vitro include ascorbic acid supplementation and the presence of an organic phosphate source. The differentiation of hypertrophic chondrocytes in vitro is associated with a 10–15-fold increase in alkaline phosphatase enzyme activity and deposition of mineral within the extracellular matrix. Temporal studies of the biochemical changes coincident with development of hypertrophy in vitro demonstrate that proteoglycan synthesis decreases 4-fold whereas type X collagen synthesis increases 10-fold within the same period. Ultrastructural examination reveals cellular and extracellular morphology similar to that of hypertrophic cells in vivo with chondrocytes embedded in a well formed extracellular matrix of randomly distributed collagen fibrils and proteoglycan. Mineral deposition is seen in the interterritorial regions of the matrix between the cells and is apatitic in nature. These characteristics of chondrogenic growth and development are very similar in vivo and in vitro and they suggest that studies of chondrogenesis in vitro may provide a valuable model for the process in vivo. Microsc. Res. Tech. 43:156–173, 1998. r 1998 Wiley-Liss, Inc. INTRODUCTION Much of skeletal formation during vertebrate embryogenesis and postnatal growth occurs by a process referred to as the endochondral sequence. This is a developmentally regulated series of stages that proceeds in a highly coordinated temporal and spatial manner, in which there is sequential recruitment and differentiation of cells that form cartilage, vascular, and bone tissues. During the initial phases of this process, mesenchymal stem cells are recruited, and they condense and differentiate into chondrocytes to produce an initially unmineralized and avascular cartilage model of a developing bone. Subsequently, the chondrocytes within this model further differentiate and their elaborated matrix mineralizes to form calcified cartilage in specific areas. The mineralization proceeds and continr 1998 WILEY-LISS, INC. ues in the presence of vascularization of the tissue. This is initiated by means of blood vessels accompanied by undifferentiated mesenchymal cells entering cartilage regions containing hypertrophic chondrocytes. Chondrocytes within lacunae localized in these mineralized areas die and perivascular cells adjacent to cores of calcified cartilage secrete new matrix that results in bone deposition. The process is completed when metaphyseal remodeling occurs as the mineralized extracellular matrix of both bone and cartilage is ultimately Contract grant sponsor: NIH; Contract grant numbers: HD 22400, AR41452, N01-HD-2–3144; Contract grant sponsor: Max Kade Foundation. *Correspondence to: Dr. Louis C. Gerstenfeld, Musculoskeletal Research Laboratory, Boston University Medical Center, 715 Albany St., Housman 2, Boston, MA 02118. Received 16 July 1998; accepted in revised form 19 July 1998 CHONDROGENIC POTENTIAL OF SKELETAL CELLS resorbed and replaced with bone alone (Farnum and Wilsman, 1987; Ham and Cormack, 1979; Holtrop, 1972; Hunziker et al., 1984; Shapiro et al., 1977). In normal vertebrate development in vivo, an endochondral phase seldom occurs during the formation of the calvaria and the facial bones of the cranial skeleton. These tissues are generated by direct induction through a process known as membranous bone formation. On the other hand, cartilage tissue is present in some instances, including that of articular cartilage surfaces that form on the maxillary processes of the jaw by direct differentiation of ‘‘secondary’’ cartilage within the periosteum (Thorogood, 1979). In addition, specific experimental manipulation in vivo such as that causing calcium deficiency by growing chick embryos in a shell-less environment has been shown to give rise to cartilage formation within the calvaria (Jacenko and Tuan, 1986). Such changes to either cartilage or bone during vertebrate development would not be unanticipated considering that cell populations derived from embryonic calvaria from several species exhibit the potential to differentiate into one of several multiple cell lineages including those of osteoblasts (Aronow et al., 1990; Bellows and Aubin, 1989; Gerstenfeld et al., 1987), adipocytes (Bellows et al., 1994; Beresford et al., 1992; Kodama et al., 1982; Yamaguchi and Kahn, 1991), myoblasts (Grigoriadis et al., 1988; Yamaguchi and Kahn, 1991), and chondrocytes (Asahina et al., 1993; Bellows et al., 1989; Rifas et al., 1982; Toma et al., 1997). These results indicate that, although not normally observed, there is a subset of cells within the cell populations of the cranial bones that retains chondrogenic potential and can express a cartilage phenotype in vitro. The selection of cells having a cartilage phenotype in cultured calvarial cell populations has been shown to occur as a result of a variety of experimental conditions including treatment of the cultures with osteogenic protein (OP-1) (Asahina et al., 1993) or corticosteroids (dexamethasone) (Grigoriadis et al., 1988); growth of the cells on gelatin bone matrix (Terashima and Urist, 1975); or growth of the cells in the absence of serum (Cole et al., 1992; Rifas et al., 1982; Toma et al., 1997). These observations suggest that the epigenetic factors that regulate normal lineage progression of skeletal cells and presumably direct the changes so noted might be both identified and studied in vitro. Studies from several groups also indicate that both genetic factors (Gerstenfeld et al., 1989) and environmental factors (Eavey et al., 1988) alter chondrocyte phenotype and promote cell hypertrophy. With respect to the latter, chondrocyte hypertrophy can be stimulated in vitro by culturing cells in a three-dimensional (collagen or agarose) gel, selecting an appropriate cell plating density (Solursh and Meier, 1974), or supplementing the cells with appropriate nutrient and vitamin D3 metabolites (Gerstenfeld et al., 1990) or ascorbic acid (Gerstenfeld and Landis, 1991; Leboy et al., 1989). The identification of the required growth conditions has led to research studies that have examined the phenotypic developmental progression of chondrocyte hypertrophy and morphogenesis in vitro. The present report examines the mechanism(s) and contributory role of developmental versus environmental factors in the selection of cells to differentiate into 157 the chondroblastic lineage and to undergo further hypertrophic differentiation. For this purpose, 12-dayold embryonic chicken calvaria and vertebrae have been used as tissue sources for presumptive skeletal cells and committed growth chondrocytes, respectively. The results detail underlying mechanisms of cell type selection and limitation for the development of chondrogenic and osteogenic lineages, define phenotypic changes that accompany hypertrophic differentiation, and describe chondrocyte and matrix morphogenesis during the hypertrophic progression of events. MATERIALS AND METHODS Cell Culture Skeletal cell populations were isolated by three sequential trypsin-collagenase treatments of 12-day-old chicken calvaria as previously described (Gerstenfeld et al., 1996). Only cells released from the third consecutive digest were used for experimental determinations. Cells were suspended in minimal essential medium (MEM) supplemented with 10% fetal bovine serum (FBS) (Sigma Chemical Co., St. Louis, MO) and plated at 1.25 x 105 cells/60 mm diameter and 5 x 105 cells/100 mm diameter culture dishes. After 24 hours, cells were washed once in phosphate buffered saline (PBS) with a subsequent change of medium to either MEM supplemented with 10% NU-Serum IV Culture Supplement (Collaborative Biomedical Products, Bedford, MA) (Wong and Tuan, 1993) or 10% FBS. Cultures were maintained in MEM/10% FBS or MEM/10% NUSerum for 2–3 weeks with medium changes every 2 days until cells reached confluence. At that time, the medium was changed to BGJb (Fitton-Jackson modification) supplemented with 10% FBS or 10% NU-Serum. After another 2 days, the medium was supplemented with 10 mM ␤-glycerophosphate (␤GPO4) and 2 days later this medium was changed into ‘‘mineralizing medium,’’ containing 10% FBS or 10% NUSerum, 10 mM ␤GPO4, and 12.5 µg/ml ascorbic acid. The time at which the cultures were introduced to the mineralizing medium was denoted as day 0. Thereafter, the medium was changed every 3 days until the experiments were terminated. All time points are referenced to day 0. In studies focusing on growth chondrocyte maturation and hypertrophic development, committed chondrocytes were prepared from vertebral bodies of 12-day-old embryonic chickens. In these experiments, the chondrocyte cultures were grown under conditions that were permissive for growth chondrocyte differentiation as previously described (Gerstenfeld and Landis, 1991). In other experiments, committed chondrocytes were prepared from the cephalic halves of 17-day embryonic chicken sterna and these cultures were grown under conditions identical to those used for the vertebral chondrocytes. In other investigations, conditioned medium supplementation experiments were carried out using medium from the sternal chondrocyte cultures in the following manner: Cephalic chondrocytes were selected and grown for at least 2 weeks in culture and, beginning at the end of the second week, medium was collected (this was termed ‘‘conditioned medium’’). Collected fresh from the sternal chondrocyte cultures through the duration of the experiments, this condi- 158 L.C. GERSTENFELD ET AL. tioned medium was mixed at a ratio of either one or three parts to either three or one part, respectively, of Dulbecco’s modified eagle’s medium (DME) (Sigma Chemical Co.) containing 10% FBS, 10 mM ␤GPO4, and 12.5 µg/ml ascorbic acid. Calvarial cells from 12- or 17-day-old embryonic chickens were then grown for 2 weeks in 1:3 or 3:1 DME:conditioned supplement. This medium was replenished with conditioned medium every 2 days over the time period. Bromodeoxyuridine Labeling BrdU labeling was as previously described (Gerstenfeld et al., 1996). Six days after their initiation, cell cultures prepared from 12-day-old embryonic chicken calvaria were pulsed for 3 hours with BrdU (Sigma Chemical Co.) at a final concentration of 20 mM. At the end of the pulse period, cell layers were fixed for 30 minutes at 4°C with 4% formaldehyde in 0.1 M phosphate buffer, pH 7.2. Cell layers were then reacted for 60 minutes with anti-BrdU antibody supplemented with nuclease (Amersham Inc., Arlington Heights, IL), diluted 1:10 in PBS buffer containing 0.05% bovine serum albumin (BSA) (w/v) and 0.5% Tween (v/v). Secondary antibody reactions followed with anti-mouse FTIC-conjugated IgG. Isolation and Analysis of mRNA Total RNA and DNA were isolated using Tri-ReagentTM, (Molecular Research Center, Cincinnati, OH) according to the manufacturer’s instructions. RNA was resolved on 1% agarose gels containing 2.2 M formaldehyde (Thomas, 1980) and 5 µg of total RNA based on OD260 were loaded per gel lane. Equal loading of RNA was verified by ethidium bromide staining of the gel before blotting onto Biotrans nylon membranes (ICN Biomedical Inc., Aurora, OH). Chicken cDNAs used for these studies were pro ␣1(I) collagen (Lehrach et al., 1979), pro ␣1(II) collagen (Young et al., 1984), osteocalcin (Neugebauer et al., 1995), osteopontin (Moore et al., 1991), and aggrecan (Sai et al., 1986). Type X collagen was detected using a 790 bp sequence beginning at the splice junction of exon 3 of a chicken genomic type X collagen clone (gcCOLX; L.C. Gerstenfeld, unpublished data). This sequence corresponds to the sequence published by LuValle et al. (1988). 32P-radiolabeled cDNA probes were synthesized by random priming and hybridization was carried out at 65°C in 50 mM PIPES, 100 mM NaCl, 50 mM Na phosphate, pH 7.0, 1 mM EDTA buffer containing 5% SDS (w/v) and 60 mg/ml single stranded salmon sperm DNA. Hybridization was carried out for 18 hours in a rotating hybridization oven (Robins Scientific, Sunnyvale, CA). Autoradiograms were quantitated using an LKB Ultrascan II laser scanning densitometer (LKB, Broma, Sweden) and values were normalized to 18S ribosomal RNA obtained by hybridization of each blot to a conserved nucleotide sequence probe of the 18S ribosomal subunit (Ambion Corp., Austin, TX). Analysis of Collagen and Proteoglycan Synthesis Collagen protein synthesis was analyzed by pulse labeling with 3H-proline. Cultures were labeled for 24 hours using 5 ml of 50 µCi/ml 3H-proline (110 mCi/ mmol) (Amersham Inc.) after 6 days of growth in mineralizing medium. The total cell layer-associated proteins and secreted medium proteins were processed in the presence of proteolytic inhibitors as described previously (Gerstenfeld et al., 1984). Fifty thousand cpm of medium-labeled proteins and 100,000 cpm of cell layer-associated proteins were applied per lane for each gel. Collagens were resolved by electrophoresis on 5–10% continuous gradient SDS-polyacrylamide gels (Laemmli, 1970) and fluorography was carried out (Bonner and Laskey, 1974). Quantitation of fluorographs was performed on an LKB Ultrascan II laser scanning densitometer (LKB, Broma, Sweden). For proteoglycan analysis, cell cultures were grown in 33 mm tissue culture wells. At day 6 after the change to complete medium, replicates (n ⫽ 6) were labeled with 50 µCi/ml Na2[35S]O4 (21 mCi/mmol) in 1.5 ml for 24 hours. Proteoglycans were precipitated overnight with 1.3% potassium acetate (v/v) in 95% ethanol. Samples were collected by centrifugation at 16,000g in a microcentrifuge for 30 minutes, resuspended, and precipitated a second time. Incorporation was determined by filtration of the precipitate onto glass filters followed by scintillation counting (Gerstenfeld et al., 1990). Statistical Analysis Biochemical assays were performed with a minimum of 6 replicates and were repeated with at least 2–3 separate preparations of cells. All values were normalized to DNA content. Data were tabulated as the mean ⫾ standard deviation (or standard error of the mean where appropriate) with a minimum sample size of 6. Statistical analysis consisted of ANOVA, Student’s t-test of significance, and the Tukey HSD test using Version 6.11 of SAS for Windows software package (SAS Institute Inc., Cary, NC). Values of p ⬍ 0.05 were considered significant. Mean ratios for cell count data were compared using an independent samples Student’s t-test (Bland and Altman, 1995). For all calculations, two-tailed values of p ⬍ 0.05 were considered statistically significant. Histological and Ultrastructural Analysis of Chondrocyte Cultures Alkaline phosphatase expression was examined by histochemical staining. Duplicate cultures were fixed with 2% paraformaldehyde in 0.1 M cacodylate buffer at pH 7.4 for 10 minutes, rinsed with fresh buffer, and stained for alkaline phosphatase activity using naphtol AS-MX phosphate and Fast Red TR salts (Sigma Chemical Co.). Matrix calcification was determined by hematoxylin/eosin and von Kossa (phosphate) staining. For ultrastructural analysis, samples (cells and matrix) were fixed and embedded directly in culture plates using either aqueous (glutaraldehyde-OsO4) or anhydrous (ethylene glycol) techniques (Landis et al., 1977). Specimens were sectioned (⬃80 nm) with diamond knives on a Reichert Ultracut S ultramicrotome (Leica, Deerfield, IL) and floated briefly on an aqueous or anhydrous (ethylene glycol) trough fluid (Landis et al., 1977). Some sections were stained with 8% uranyl acetate in absolute ethanol and with Sato’s lead. CHONDROGENIC POTENTIAL OF SKELETAL CELLS 159 Fig. 1. Selection of chondrogenesis by growth of presumptive skeletal cell populations in serum-substituted medium. Left (top to bottom): Histochemical analysis of 12-day-old embryonic chicken calvarial cell cultures grown 6 days in medium containing either 10% fetal bovine serum (FBS) or 10% serum substitute (serum subt., SS). Cell layers were stained for either alkaline phosphatase (Apase) activity or proteoglycans based on alcian blue reactivity. Representative staining characteristics of culture dish regions are denoted. Right (top): Bar graph of alcian blue staining nodules in cultures grown in SS (solid bars) or 10% FBS (open bars). The number of alcian blue staining nodules was counted in 6 dishes and compared in 2 separate experiments. Error bars denote one standard deviation of these data. Right (middle): Analysis of proteoglycan synthesis in cultures grown 12 days in SS (solid bars) or 10% FBS (open bars). Cells were labeled on day 6 with 50 µCi/ml Na2[35S]O4 (21 mCi/mmol) in 1.5 ml for 24 hours and counted as noted in Materials and Methods. Error bars denote one standard deviation (n ⫽ 10) for two separate experiments. Right (bottom): Analysis of alkaline phosphatase enzyme activity (by p-nitrophenol, p-NP conversion) of cultures grown in SS (solid bars) or 10% FBS (open bars). Enzyme activity was measured on day 6. The results in all panels were determined from 12 wells in 2 separate experiments (exp 1 and exp 2). Error bars denote one standard deviation. Left panel magnification, 100x. Ultrastructural features of such sections were examined by conventional transmission microscopy at 60 or 80 kV in either a Philips EM 300 (Philips Electronic Instruments, Inc., Mahwah, NJ) or a JEOL 100C (JEOL USA, Peabody, MA) electron microscope. Mineral in unstained sections was studied by selected area 160 L.C. GERSTENFELD ET AL. Fig. 2. Comparison of chondrogenic and osteogenic expression in 12-day-old embryonic calvarial cultures grown under conditions permissive for chondrogenesis. A: RNAs were isolated from cells after 12 days of culture growth in mineralizing medium supplemented with either SS (serum substitute) or 10% FBS. Northern blot analysis was then carried out for bone and cartilage mRNAs, including aggrecan (AG), osteopontin (OPN), osteocalcin (OC), and ␣1(I), ␣1(II), and ␣1(X) collagen chains. The efficiency and uniformity of blot transfers are demonstrated by hybridization profiles to an 18S ribosomal probe and staining with ethidium bromide (EtBr). B: Collagen protein synthesis in cultures grown for 12 days in complete medium supplemented with either FBS or SS. The electrophoretic positions of 3H-proline-labeled ␣1(I ⫹ II), ␣2(I), pro ␣1(X), and ␣1(X) collagen chains isolated from either cell layers or medium are indicated. The horizontal line marking to the left of some blots indicates the position of 27S mRNA used as a molecular weight indicator. electron diffraction at 80 or 100 kV in the Philips instrument and by electron probe microanalysis at 25 kV in a STEM-modified JSM-50A scanning electron microscope (JEOL USA) (Landis and Glimcher, 1978, 1982). All microscopes were equipped with a liquid nitrogen anticontaminator to limit specimen artifacts. 35SO RESULTS Chondrogenic Selection Within Presumptive Skeletal Cell Populations Isolated From 12-Day-Old Embryonic Chicken Calvaria The growth of 12-day-old embryonic chicken calvarial cell populations in medium in which FBS was replaced with a serum substitute has been shown to promote chondrogenesis (Toma et al., 1997). As Figure 1 shows, histological staining of whole cultures with alcian blue identified the presence of chondrogenic cell nodules throughout the cell layers in such serum-substituted samples. The number of alcian blue staining nodules could be counted and compared between cultures grown in the absence or presence of medium containing FBS (Fig. 1, top right). These results demonstrate that the total number of chondrogenic nodules increased approximately 5–10-fold in the cultures grown in serum substitute. Chondrogenesis was biochemically determined by 4 labeling of sulfated proteoglycans (Fig. 1, middle right), and these results independently confirm an ⬃8–20-fold increase in the chondrogenic phenotype of the serum-substituted cultures. Differences in alkaline phosphatase enzyme activity were assessed in the two sets of cultures. At early time points, histochemical staining for alkaline phosphatase was seen only in the cells localized between the chondrocyte nodules in the cultures containing serum substitute (Fig. 1, middle left). In contrast, alkaline phosphatase staining was intense throughout the entire cell layer of the cultures containing FBS (Fig. 1, bottom left). Such histochemical results suggest that the enzyme activity would be much higher in the cultures grown in FBS. However, biochemical measurement of total enzyme activity was considerably higher in cultures grown in serum-substituted medium after 12 days (Fig. 1, bottom right). Histochemical staining after 12 days was also seen throughout the cell layers of cultures grown in either FBS or serum substitute (data not shown). Thus, since both hypertrophic chondrocytes and osteoblasts express alkaline phosphatase, this assay by itself cannot discriminate between osteogenic cells and hypertrophic chondrocytes. Moreover, as the chondrogenic cells progress toward hypertrophy, they CHONDROGENIC POTENTIAL OF SKELETAL CELLS will express very high levels of alkaline phosphatase (see below). The expression and hypertrophic progression of the chondrogenic phenotype were also examined by growing the 12-day-old embryonic chicken calvarial cultures under conditions that both permitted chondrogenesis and promoted mineralization. In these studies, a number of genes that are specific for either the chondrogenic or osteogenic cells were examined (Fig. 2A). A definitive increase in aggrecan and type II and X collagen gene expression (chondrogenesis markers) was seen in cultures grown in serum-substituted medium. However, as was previously demonstrated (Toma et al., 1997), the expression of type X collagen was observed only after longer periods of growth in serum-substituted medium that had been specifically supplemented with ascorbic acid and ␤GPO4 (see below). While chondrogenic cells were selectively increased when grown in serum substitute, there was only a slight diminishment in the osteogenic phenotype of the same cultures as determined from osteopontin and osteocalcin gene expression (osteogenesis markers), demonstrating that the cultures maintained their overall osteogenic differentiation. The relationship between levels of mRNA expression and new collagen synthesis was assessed in the cultures by pulse labeling with 3H-proline after 12 days of growth in mineralizing medium (Fig. 2B). Type I collagen synthesis was detected in cultures grown in both serum-substituted and FBS-containing medium as evidenced by the detection of the ␣2(I) collagen chain. In both cultures ␣2(I) was detected as a major secretory product based on its prevalence in the medium. An increased type II collagen expression was indicated by a greatly increased ␣1 to ␣2 ratio seen in serumsubstituted cultures. In comparing the cultures grown under the different conditions, proline labeled bands at the position of pro ␣1(X) and its processed form were detected only in the cultures grown in serum-substituted medium. Consistent with previous analysis (Gerstenfeld and Landis, 1991), type X collagen was shown to increase preferentially in its accumulation in the cell layers of the mineralizing culture. The mechanisms by which various growth conditions promote the selection of specific cell lineages within a population of presumptive skeletal cells may be examined by determining the total number of cells within the population that express osteogenic or chondrogenic lineage markers (Gerstenfeld et al., 1996; Toma et al., 1997). In the current study, this was carried out by immunolabeling the cell populations grown under the different culture conditions with markers selective for chondrogenic (type II collagen) and osteogenic (osteocyte SB-5) cells (Fig. 3). Analysis shows dense nodules of birefringent rounded cells cultured with serum substitute selectively labeled for type II collagen but few such chondrogenic cells in cultures grown under normal conditions in the absence of serum substitute (Fig. 3A, first and second rows of panels). In contrast, within cell populations grown under either culture condition, the terminal osteogenic marker SB-5 was present on cells throughout both cultures (Fig. 3A, third row of panels). In order to determine if there was selective proliferation of chondrogenic cells compared to osteogenic cells, proliferative cells were labeled by BrdU incorporation 161 (Fig. 3A, bottom panels). There was selective incorporation of BrdU within the chondrogenic nodules, while in comparison few cells were labeled in the internodular regions composed primarily of osteogenic cells. These results show that the chondrogenic cells were more proliferative than the osteogenic cells. The percentage of cells that expressed either osteogenic or chondrogenic markers was determined from cell counts taken from random microscopic fields at identical magnifications and these results are summarized in Figure 3B. Under control growth conditions in medium supplemented with FBS, ⬃15% of the cells in the embryonic cultures expressed a chondrogenic phenotype while about ⬃40% of the cells in these cultures were positive for the expression of the osteogenic SB-5 marker. When these same cultures were grown in medium containing serum substitute, there was a ⬃3.5-fold increase in the cells that expressed the chondrogenic phenotype such that ⬃50% of the cells were now positive. This increase in the number of chondrogenic cells was associated with a slight decrease in the number of cells that expressed the osteocyte specific marker to ⬃35% of the cell population. This decrease was not statistically significant compared to the cell numbers determined from FBS cultures. Thus, growth of these cells in serum substitute did not change the overall percentage of cells that expressed an osteogenic phenotype based on their immunoreactivity for SB-5. The results presented in Figure 3 suggest that chondrocytes may produce autocrine factors that are selective for their own survival or further proliferation. Preliminary studies carried out to test this hypothesis are depicted in Figure 4. In these experiments, committed growth chondrocytes derived from the cephalic portions of 17-day-old embryonic chicken sterna were grown in culture and their medium was collected. This conditioned medium was added as a supplement to the medium of 12-day-old chicken calvarial cells grown in FBS. The 12-day-old embryonic calvarial cell populations were allowed to grow in the presence of different ratios of conditioned: FBS medium and the chondrogenic phenotype of the cultures was subsequently examined for expression of type X collagen within the cell layers. Figure 4 shows a comparison of the type X collagen labeling seen in these cultures supplemented in ratios of 1:3 or 3:1 and without conditioned medium. Increased labeling intensity and numbers of observable type X collagen-reactive chondrocyte nodules are found, dependent on the ratio of conditioned medium that was added. Immunostained cephalic chondrocyte cultures grown under conditions promoting chondrocyte hypertrophy were used as a positive control. Calvarial cell cultures from 17-day-old embryonic chickens were used as negative controls since previous studies demonstrated such cell populations lacked the ability to undergo chondrogenesis in response to growth in serum-substituted medium (Toma et al., 1997). This latter control was also used to address the question of whether there was adsorption of the type X collagen to the osteoblast cell layers because of type X presence in the conditioned medium isolated from the cephalic chondrocyte cultures. No immunofluorescent labeling with type X collagen antibody was apparent for the 17-day-old embryonic chicken calvarial cell populations supplemented with a 3:1 ratio of chondrocyte-conditioned medium. Type X collagen was Fig. 3 A. CHONDROGENIC POTENTIAL OF SKELETAL CELLS 163 Fig. 3. Chondrogenesis in presumptive skeletal cell populations. A: (page 162) Comparisons of chondrogenic and osteogenic phenotypic markers in 12-day-old chicken calvarial cell cultures grown in either SS (serum substitute) or 10% FBS. The osteocyte cell surface marker SB-5 and type II collagen were used as independent determinants for osteogenic or chondrogenic differentiation, respectively, for the cultured cells. Proliferative cells were labeled with BrdU. The photomicrographs depict cellular expression of the various markers carried out after 6 days of growth in complete medium. Top: Phase contrast micrographs of cells cultured with FBS or SS followed by matched immunochemical analysis with type II collagen (all panels, magnification, 100x). Intense staining of dense chondrogenic nodules may be noted. Immunochemical analysis for SB-5 expression shows numerous labeled cells in cultures grown with FBS or SS (magnification, 200x). Osteocyte processes can be clearly observed (magnification, 400x, inset, left panel). Similar numbers of SB-5 positive cells may be seen throughout the cultures. Bottom: Nuclear uptake of BrdU. Labeling is more pronounced within a chondrogenic nodule in a culture grown in SS than in an inter-nodular area of a culture supplemented with FBS (magnification, 400x). B: (above) A graphic depiction of the percentage of cells expressing chondrogenic and osteogenic phenotypes in calvarial cultures from 12-day-old chicken embryos. Quantitative data are from Toma et al. (1997). The percent of immunologically labeled cells for type II collagen (COL II) or SB-5 was calculated from 10 random micrographic fields such as represented in A (middle) for FBS or serum substitute (SS) cultures. Error bars denote one standard deviation of the measurements. Differences in type II collagen are significant at p ⫽ .005. found in cultures grown for 12 days; its presence can be attributed to induction of new chondrogenic differentiation. Finally, it is interesting to note the distinctive morphological location of the type X collagen that was seen within these cultures. Type X collagen labeled a dense fibrillar organization restricted to the chondrocyte nodules with a discernible pericellular distribution around individual cells within the 12-day calvarial cell cultures. A similar more heavily labeled distribution of the type X collagen was found in the cephalic chondrocyte cultures isolated from 17-day-old embryonic chickens. Labeling of these cultures was much more extensive compared to that of calvarial cells and was observed over the entire cell layer. This result was most likely caused by the higher levels of synthesis of the cephalic cultures. data. The progression of these cultures grown under different conditions was assessed qualitatively by histochemical analysis of individual culture wells (Fig. 5) and measured quantitatively (Table 1). Cultures grown in both DME and BGJb media containing ascorbic acid promote the endochondral sequence based on the observed increased alkaline phosphatase activity and mineral deposition that was seen in these cultures (Fig. 5A and Table 1). While the inclusion of organic phosphate (␤GPO4) increases mineral deposition in cultures grown in both DME and BGJb, it only enhanced alkaline phosphatase activity when included in BGJb. Overall proteoglycan deposition in the cultures did not appear to vary based on alcian blue staining (Fig. 5A). There was, however, a marked increase in both DNA content and total protein synthesis in cultures grown in BGJb medium. Results of an examination of the temporal sequence of endochondral progression are shown in Figure 5B. Ascorbic acid and ␤GPO4 introduced into the culture system at day 6 causes the cells to undergo rapid mineral deposition as determined from alkaline phosphatase and von Kossa staining. In previous studies of osteoblast (Gerstenfeld et al., 1987) and chondrocyte cultures (Gerstenfeld and Landis, 1991), ascorbic acid supplementation was shown to promote both collagen deposition and cellular differentiation. Figure 6 demonstrates that culture supplemen- Hypertrophic Progression of Growth Chondrocytes In Vitro Conditions can be described that promote the progression in vitro of the endochondral sequence of the chondrogenic program. Committed chondrocytes isolated from the vertebral columns (a tissue undergoing endochondral replacement in vivo (Shapiro, 1992)) of 12-day-old embryonic chickens were used in these studies. Figures 5 and 6 and Table 1 present relevant 164 L.C. GERSTENFELD ET AL. Fig. 4. Autocrine selection of chondrogenic growth of 12-day-old embryonic calvarial cell populations. Type X collagen matrix deposition in 12-day-old embryonic chicken calvarial cultures grown in the presence of conditioned medium from hypertrophic chondrocytes derived from 17-day-old embryonic chicken cephalic sterna. The identity and distribution of type X collagen extracellular matrix accumulation within chondrogenic nodules were determined by immunolocalization. Top: Calvarial cell cultures grown for 12 days in complete medium containing 10% FBS without conditioned medium (control) or with a 1:3 or 3:1 ratio of conditioned medium from cephalic sternal chondrocyte cultures. Bottom: Negative and positive controls, respectively, of 17-day-old chicken embryo calvarial cell populations treated with a 3:1 ratio of chondrocyte-conditioned medium (left) and cephalic sternal chondrocyte cultures after 12 days of growth in mineralizing medium. All photomicrographs: magnification, 400x. tation with ascorbic acid promotes chondrocyte hypertrophic differentiation as demonstrated by the induction of high levels of type X collagen mRNA expression and protein synthesis. Figure 7 summarizes the overall temporal relationships between proteoglycan and collagen synthesis to mineral deposition and relates these to aggrecan and types II and X collagen mRNA expression. Here, it is apparent that proteoglycan synthesis and message levels fall rapidly after a few days of culture time while collagen content increases and then drops and mineral continuously increases. The rise and fall in collagen synthesis is mirrored by a preceding type II collagen message expression. Type X expression gradually rises throughout the culture period. The morphogenesis of the chondrocytes and their matrices under culture growth conditions permissive for endochondral progression is shown in Figures 8–11. Corresponding to the changes in gene expression and biochemistry described above, there are changes in the structure of the chondrocytes and their extracellular regions that likewise reflect a progression through the endochondral sequence. The temporal development of the cultures morphologically shows increasing numbers of cells having rounded shapes and residence within lacunae (Fig. 8). Changes in cell size, shape, distribution, and cytoplasmic content in the cultures are apparent (Figs. 8–11) with larger chondrocytes being associated with not only more mature cultures (grown for longer periods) but also the relatively deeper regions of a culture dish where mineralization is becoming prominent (Figs. 8, 10). Representative calvarial chondrocytes from 12-day-old embryonic chickens and grown for 17–18 days in culture are generally marked by typical intracellular organelles, including a single nucleus, extensive endoplasmic reticulum, vesicles, and vacuoles (Fig. 9). Many chondrocytes, particularly the larger ones, are observed within a highly staining pericellular region characterized by the thin fibrils of CHONDROGENIC POTENTIAL OF SKELETAL CELLS 165 Fig. 5. Histochemical analysis of proteoglycan, alkaline phosphatase, and mineral deposition in cultures of vertebral chondrocytes grown under different conditions. A: All cultures were grown for 15 days in BGJb or DME with or without ␤-glycerophosphate (␤-GPO4). On day 15, separate 33 mm wells were analyzed using histochemical staining to assess proteoglycan deposition (alcian blue), mineral deposition (von Kossa silver nitrate), or alkaline phosphatase (Apase, fast red). B: Time course over 20 days of mineral accumulation and alkaline phosphatase activity in 33 mm culture wells containing vertebral chondrocytes grown in BGJb medium supplemented with ␤-GPO4. Images are of the full 33 mm wells. type II collagen and small condensations suggestive of proteoglycans (aggrecan) (Fig. 11). Mineral in the form of dense, irregularly shaped deposits is exclusively extracellular in location and is found overlying the collagen (Fig. 10), frequently following individual fibrils and in clear association with them (Fig. 11). There are many cytoplasmic processes that invade the extracellular regions of the cultures (Fig. 11) and these appear to terminate in small vesicles (Fig. 11). Consistent association of such vesicles with mineral is equivocal. With time, the cultures mineralize more completely, and the numerous mineral deposits can be identified structurally as apatitic on the basis of their electron diffraction character (data not presented; see Gerstenfeld and Landis, 1991). by which mesenchymal stem cells are restricted to the osteogenic vs. the chondrogenic lineage was examined in skeletal cell populations isolated from embryonic chicken calvaria. The use of cell populations from calvaria is particularly informative since this tissue undergoes direct bone induction in the absence of an endochondral stage. Thus, the appearance of chondrocytes in these cell populations arises through the induction and differentiation of the mesenchymal stem cells to chondrocytes. The data presented and reviewed here indicate that the developmental restriction of skeletal cells of the calvaria was not a result of positive selection for osteogenic differentiation but a negative selection against the progressive growth of chondrogenic cells in the absence of either a permissive or inductive environment. This conclusion is based on the findings that the 12-day-old embryonic calvarial cell populations selectively proliferated to chondrogenic cells but were little changed in their percentage of osteogenic cells when they were induced to undergo chondrogenesis. DISCUSSION The data presented here focus on the growth conditions in vitro that either induce chondrogenesis from presumptive or uncommitted cells or promote chondrocyte maturation and terminal differentiation toward a hypertrophic phenotype. The underlying mechanism(s) 166 L.C. GERSTENFELD ET AL. Fig. 6. Effect of ascorbic acid on expression of types II and X collagen. A: Northern blot analysis of type II and type X collagen mRNA expression in 12-day-old embryonic chicken cultures of vertebral chondrocytes grown 15 days in culture, in the absence (⫺) or presence (⫹) of ascorbic acid. Five micrograms of total RNA were applied to each lane. Horizontal lines to the left of the blots indicate the position of 27S mRNA as a molecular weight marker. B: Collagen synthesis in ascorbic acid treated (⫹/⫺) cultures. Intact procollagen and collagen ␣ chains for type II and type X of either cell layer or medium were separately extracted as described in Materials and Methods. The electrophoretic positions of various 3H-proline-labeled collagen chains of interest are indicated. TABLE 1. Culture conditions and their effects on culture parameters* Mineralization1 DNA content (µg/33-mm well) Apase activity (nM p-nitrophenol/µg DNA/30 min) Total protein synthesis (3H-leucine/µg DNA)2 Percent collagen synthesis3 DME DME ⫹ Asc DME ⫹ Asc ⫹ ␤GPO4 BGJb ⫹ Asc BGJb ⫹ Asc ⫹ ␤GPO4 ⫺ 5.8 ⫾ 0.9 15.0 ⫾ 4.0 1.07 ⫻ 105 34 ⫾ 8 ⫹ 6.2 ⫾ 0.5 140.0 ⫾ 30.0 ND 48 ⫾ 12 ⫹⫹ 7.4 ⫾ 0.6 162.0 ⫾ 26.5 ND 53 ⫾ 14 ⫹⫹⫹ 10.6 ⫾ 0.6 148.0 ⫾ 25.0 6.8 ⫻ 104 58 ⫾ 14 ⫹⫹⫹⫹ 13.7 ⫾ 0.7 403.0 ⫾ 46.0 4.1 ⫻ 104 62 ⫾ 15 *Reprinted with permission from Gerstenfeld, L.C., and Landis, W.J. (1991) Gene expression and extracellular matrix ultrastructure of a mineralizing chondrocyte cell culture system. J. Cell Biol., 112:501–513. 1Gross extracellular matrix calcification was assessed qualitatively by von Kossa staining. 2Total protein synthesis was assessed after 24 hours of labeling of day 15 cultures and represents the addition of both medium and cell layer nondialyzable TCA precipitable counts per minute. 3Percent collagen synthesis was determined by collagenase treatment of day 15 cultures after 24 hours 3H-proline labeling and samples represent the addition of both medium and cell layer-associated nondialyzable counts per minute. ND ⫽ Not determined; Asc ⫽ ascorbic acid; ␤GPO4 ⫽ ␤-glycerophosphate. It is interesting to speculate that the balance between chondrogenesis and osteogenesis may be controlled through epigenetic signals that promote presumptive skeletal cell populations to differentiate independently toward either the osteogenic or chondrogenic lineage. Thus, while intrinsic genetic regulation controls the gross anatomic patterning of the skeletal organs, micropatterning and growth control probably reside in a wide variety of extrinsic mediators that may be related to structural functions of the organs or their physiological role in mineral homeostasis. Studies of secondary cartilage formation of the articulating surfaces of the vertebrate maxillary process have shown that during embryogenesis the absence of movement of this joint inhibited secondary growth cartilage formation (Thorogood, 1979). Correlative to this finding is the report that bone repair in rigidly fixed fractures of minimized micro-motion results in less cartilage formation and direct induction of new bone formation (Probst and Spiegel, 1997). In addition, systemic factors such as calcium stasis (Jacenko and Tuan, 1986), levels of 1,25(OH)2D3 (Gawande and Tuan, 1990), or the presence of corticosteroids such as dexamethasone (Bellows et al., 1989) have all been shown to affect the balance of osteogenic versus chondrogenic differentiation. Finally, the presence of local morphogens such as OP-1 (Asahina et al., 1993), transforming growth factor-␤ (TGF␤) (Derynck, 1994), and basic fibroblast growth factor CHONDROGENIC POTENTIAL OF SKELETAL CELLS 167 Fig. 7. Comparison of chondrocyte mineral deposition, collagen synthesis, proteoglycan synthesis, and gene expression over a 20-day growth period in vitro. Top: Total collagen synthesis (l), proteoglycan synthesis (●), and mineral deposition (o) were determined from triplicate 33 mm culture dishes of duplicate experiments with 12-day-old embryonic chicken vertebral chondrocytes. Bottom: Comparison of type II collagen (l), type X collagen (o), and aggrecan (●) mRNA gene expression over 21 days. The left axis of the plot pertains to quantitative levels of aggrecan (Ag) and the right axis to quantitative levels of either type II or type X collagen. Error bars denote total range of experimental variation. 168 L.C. GERSTENFELD ET AL. Fig. 8. A light micrograph of a 12-day-old embryonic chicken calvarial chondrocyte culture grown for 18 days in mineralizing medium. The culture has been sectioned to show its full thickness profile. The tissue was fixed in glutaraldehyde and osmium tetroxide and stained with toluidine blue. The interface between the culture dish and the cultured cells and matrix is marked by the relatively straight line along the bottom of the micrograph. The culture consists of numerous cells, each having a generally round shape and residing in single lacunae. The size, shape, and disposition of the cells are similar to those features of growth plate chondrocytes observed in situ. The extracellular matrices vary in their intensity of staining as do some of the cells. Pericellular staining marks the lacunae of certain cells (arrowheads) and, in the deeper portions of the culture, nearer the culture dish bottom, there is deposition of considerable mineral (arrows). Magnification, 1,200x; bar ⫽ 10 µm. (FGF) (Rousseau et al., 1994) are also suggested to influence the balance between chondrogenesis and osteogenesis for presumptive skeletal cell populations. The apparent local control in the balance of chondrogenic and osteogenic differentiation has led to the concept that these processes are coupled in some manner (Gerstenfeld and Shapiro, 1996). For example, promotion of chondrogenesis by an absence of serum in growth medium is consistent with studies showing that either embryonic calvarial cell cultures or undifferentiated limb bud cells in serum-free conditions become chondrogenic rather than osteogenic (Cole et al., 1992; Rifas et al., 1982; Toma et al., 1997; Wong and Tuan, 1995). Unlike many other cell types, chondrocytes appear to be specifically adapted to grow in serum-free conditions in vitro (Ballock and Reddi, 1994; Bruckner et al., 1989; Quarto et al., 1992). These data might also suggest that serum contains factors that either inhibit chondrogenesis or promote selective growth of osteogenic cells. Chondrocytes themselves may produce factors that promote their own differentiation or growth. This conclusion is supported by studies that have shown that chondrogenesis can occur when cells are grown at high plating densities or in micromass cultures, even in the presence of serum supplementation (Osdoby and Caplan, 1979). Other studies demonstrate that chondrocytes cultured at high densities produce autocrine factors that enable them to survive at low cell densities in the absence of serum (Ishizaki et al., 1994), and it is interesting to speculate that growth in serum substitute reported here and elsewhere (Toma et al., 1997; Wong and Tuan, 1995) may have either promoted chondrogenesis or maintained critical numbers of chondrogenic cells that allow their survival and expansion in culture. Further work reports that calvarial cell populations isolated from 14-day avian embryos and grown in vitro produce factors that inhibit limb bud mesenchymal cells from undergoing chondrogenesis (Wong and Tuan, 1995). Moreover, the limb bud mesenchymal cells produce factors that stimulate selected calvarial cell populations to undergo chondrogenesis (Wong and Tuan, 1995). The results in the present study provide yet additional evidence that chondrocytes produce factors that promote chondrogenic differentiation. All the above results lead to the conclusion that, under normal growth conditions in vivo, osteogenic differentiation predominates as a result of regulated selection against progressive chondrogenic differentiation, which itself is probably under the control primarily of autocrine and paracrine factors produced by the local precursor cell population. Considerable data has accumulated pertaining to the culture conditions that promote growth chondrocyte differentiation. From the research summarized here and that of several other groups (Gerstenfeld and Landis 1991; Leboy et al., 1989; Sandell and Daniel, 1988), one of the most critical factors leading to maintenance of the chondrocyte phenotype, progressive culture growth, and hypertrophic differentiation of these cells is the inclusion of ascorbic acid in the growth media. The role of ascorbic acid as an essential cofactor in collagen synthesis is well documented (Lyons and Schwartz, 1984), and the biochemical data presented here show that ascorbic acid increased both total collagen synthesis and matrix accumulation. Ultrastructural analysis further demonstrates that over the period of chondrocyte growth in vitro, these cells elaborated an extensive extracellular matrix that retained morphological features comparable to those seen in vivo, including increasing collagen content with time. Recent studies of both chondrocytes and osteoblast growth in vitro describe the crucial role the extracellular matrix (ECM) plays in stimulating cellular differentiation (Franceschi et al., 1994; Gerstenfeld et al., 1988; Gerstenfeld and Landis, 1991). These reports and the present data suggest that the differentiation of specialized connective tissue cells not only is controlled by extrinsic soluble regulators but also is dependent on some form of feedback regulatory process related to the assembly and deposition of their matrices. While the exact mechanism(s) by which the ECM may exert its effects on cellular differentiation in this manner are incompletely understood, considerable information has accumulated to support the concept that both the maintenance of normal cell morphology and specific receptor interactions with the ECM are quite CHONDROGENIC POTENTIAL OF SKELETAL CELLS 169 Fig. 9. An electron micrograph of a calvarial chondrocyte culture grown for 17 days in mineralizing medium. The region contains a number of cells typically having a single nucleus and nucleolus as well as cytoplasmic organelles marked by numerous vesicles and vacuoles and extensive endoplasmic reticulum. The cells reside within lacuna (L) defined by relatively electron transparent pericellular spaces (S). The extracellular matrix beyond the pericellular regions of the tissue is characterized principally by collagen fibrils (C). Mineral (M) appears as electron dense, small punctate and larger irregularly shaped extracellular deposits. This specimen is torn (T) in places as a result of sectioning through the mineral. Magnification, 2,850x; bar ⫽ 5.0 µm. important in promoting normal skeletal cell differentiation (Franceschi, 1992). While the data presented here define growth conditions that promote growth chondrocyte differentiation, they also summarize several fundamental features that characterize synthetic changes associated with hypertrophic cartilage maturation and matrix formation. One of the most striking is an inverse relationship between the levels of collagen and proteoglycan synthesis in comparison to mineral deposition. Thus, increasing collagen synthesis may be correlated with mineral accumulation while proteoglycan synthesis levels decrease. Such a relationship is consistent with earlier work (Gerstenfeld and Landis, 1991) and with the suggestions that proteoglycans inhibit mineral nucleation (Buckwalter et al., 1987) while the development of a collagenous matrix is a prerequisite to mineral deposition (Glimcher, 1992). It is also apparent that the synthetic levels in collagen and proteoglycan observed here are primarily controlled at a pre-translational level as reflected by comparisons in collagen and aggrecan mRNA. Further, while overall collagen synthesis is increasing, it is largely a result of increasing levels of type X collagen expression. The reciprocal relationship found here in the expression of types II and X collagen is comparable to the reported temporal patterns of expression for these genes observed during the endochondral differentiation of appendicular growth cartilage in vivo (Schmid and Linsenmayer, 1985). Both the cells and extracellular matrices within the cultures described here develop and retain a number of ultrastructural features that have their close correlates with the biochemical results presented above as well as the structural characteristics of chondrocytes and matrix in vivo. These include, for example, the presence of high levels of alkaline phosphatase as mineralization occurs and progresses in culture. Microscopically, alkaline phosphatase and other enzymes may be identified by specific immunoreactivity of culture sections or inferred principally through the observation of electron dense condensates appearing within extracellular vesicles. In this paper, the many such dense vesicles are present in the extracellular regions of calvarial chondrocytes as cultures of these cells grow and mineralize. Their detection follows the progressive changes in alkaline phosphatase measured with nitrophenol. A second correlate is reflected in the overall chondrocyte culture thickness that is elaborated and increases with time. The changes observed can be attributed to a significant degree of hydration of the culture matrix through water retention mediated by proteoglycans. Evidence for these macromolecules is clear in micro- 170 L.C. GERSTENFELD ET AL. Fig. 10. Microscopic view of the hypertrophic state in vitro. The lower portions of a culture nearer to the culture dish bottom may be marked by calvarial cells (C) undergoing hypertrophy and having sparse cytoplasmic elements. The extracellular matrix at these loca- tions contains many mineral deposits (M) overlying or otherwise in apparent association with collagen fibrils of narrow diameter. This micrograph is taken from a culture grown for 17 days. Magnification, 4,400x; bar ⫽ 5.0 µm. graphs taken from regions spanning wide portions of the culture volume, a result that is consistent with the known presence, location, and distribution of proteoglycans in vivo. In addition, the ultrastructural studies here reveal numerous type II collagen fibrils comprising the culture matrices and whose appearance corresponds to their biochemical detection in the same culture samples as well as their documented presence in vivo. Perhaps the most important observation made in this work is that these culture systems will mineralize and they do so in a reproducible manner. As such, these systems are distinct from others that mineralize unpredictably or not at all or elaborate mineral that is abnormal. In the present cultures, the mineral is apatitic, having a chemical nature identical by electron diffraction means to that found in vivo. The mineral is also discretely deposited and is associated principally with the many type II collagen fibrils comprising the matrices. In these aspects, the chondrocyte cultures again follow the manner of mineralization that has been reported extensively, based on many studies of growth cartilage in vivo. In summary, the data reviewed here indicate that studies in vitro have led to considerable understanding of mechanisms that regulate chondrocyte differentiation, morphogenesis, extracellular matrix formation, and mineralization. These data suggest that chondrogenesis occurs in three stages: the differentiation of chondrogenic cells within a population of presumptive skeletal cells; the maintenance and expansion of these chondrocytes; and chondrocyte terminal maturation. The progressive differentiation of chondrocytes in vitro demonstrates that the cells develop in a manner comparable to that found in vivo. Chondrocyte maturation is characterized by an increasing expression of type X collagen and alkaline phosphatase and a decreasing expression of type II collagen and aggrecan. These changes are accompanied by correlative morphogenetic modulations including chondrocyte progression to a hypertrophic phenotype and the development of a ma- CHONDROGENIC POTENTIAL OF SKELETAL CELLS 171 Fig. 11. Electron micrograph of the chondrocyte-matrix interface. An enlargement of a region from a 17-day-old cell culture shows cytoplasmic organelles, including mitochondria (arrows), the Golgi apparatus (G), dense vacuoles (DV), endoplasmic reticulum (ER), and other elements. The cell appears to be secreting numbers of thin collagen fibrils adjacent to the cell envelope and its infoldings. The narrow diameter and faint periodicity (arrowheads) of the fibrils are consistent with structural features of type II collagen. Cell processes (P) sectioned into a variety of profiles, possible vesicles (V) containing alkaline phosphatase, highly dense condensations of putative proteoglycans (PG), and mineral deposits (M) may be observed in the extracellular matrix. The mineral is located over the collagen network in the matrix and some deposits follow individual collagen fibrils (double arrowheads). The cellular and matrix characteristics illustrated here are similar to those described for cartilage in vivo. Magnification, 25,600x; bar ⫽ 1.0 µm. trix characterized ultimately by extensive apatitic mineral deposition corresponding to that observed in vivo. ated figures. Other antibodies were obtained from the Developmental Studies Hybridoma Bank, which is maintained by the Department of Pharmacology and Molecular Sciences at Johns Hopkins University School of Medicine, Baltimore, MD, and the Department of Biology at the University of Iowa, Des Moines, IA, under contract NO1-HD-2–3144 from the National Institutes of Health. ACKNOWLEDGMENTS This investigation was supported in part by NIH grant HD22400 (L.C.G.), NIH AR41452 (W.J.L.), and the Max Kade Foundation (C.D.T.). The authors express thanks to Ms. Karen J. Hodgens (Children’s Hospital, Boston, MA) for superb technical assistance and manuscript preparation and to Drs. Tom Linsenmayer (Tufts University School of Medicine, Boston, MA), Scott Bruder (Osiris Corp., Baltimore, MD), and Arnold Caplan (Case Western Reserve University, Cleveland, OH) for generously providing type X collagen and osteoblast lineage-specific antibodies, respectively, which were used in these studies. The authors are also indebted to Mr. Christopher George (Boston University Medical Center) for assistance with the computer gener- REFERENCES Aronow, M.A., Gerstenfeld, L.C., Owen, T.A., Tassinari, M.S., Stein, G.S., and Lian, J.B. (1990) Factors that promote progressive development of the osteoblast phenotype in cultured fetal rat calvaria cells. J. Cell. Physiol., 143:213–221. Asahina, I., Sampath, T.K., Nishimura, I., and Hauschka, P.V. (1993) Human osteogenic protein-1 induces both chondroblastic and osteoblastic differentiation of osteoprogenitor cells derived from newborn rat calvaria. J. Cell Biol., 123:921–933. Ballock, R.T., and Reddi, A.H. (1994) Thyroxine is the serum factor 172 L.C. GERSTENFELD ET AL. that regulates morphogenesis of columnar cartilage from isolated chondrocytes in chemically defined medium. J. Cell Biol., 126:1311– 1318. Bellows, C.G., and Aubin, J.E. (1989) Determination of numbers of osteoprogenitors present in isolated fetal rat calvaria cells in vitro. Dev. Biol., 133:8–13. Bellows, C.G., Heersche, J.N.M., and Aubin, J.E. (1989) Effects of dexamethasone on expression and maintenance of cartilage in serum-containing cultures of calvaria cells. Cell Tissue Res., 256:145– 151. Bellows, C.G., Wang, Y.H., Heersche, J.N.M., and Aubin, J.E. (1994) 1,25-dihydroxyvitamin D3 stimulates adipocyte differentiation in cultures of fetal rat calvaria cells: Comparison with the effects of dexamethasone. Endocrinology, 134:2221–2229. Beresford, J.N., Bennett, J.H., Devlin, C., Leboy, P.S., and Owen, M.E. (1992) Evidence for an inverse relationship between the differentiation of adipocytic and osteogenic cells in rat marrow stromal cell cultures. J. Cell Sci., 102:341–351. Bland, J.M., and Altman, D.G. (1995) Multiple significance tests: The Bonferroni method. Br. Med. J., 310:170. Bonner, W.M., and Laskey, R.A. (1974) A film detection method for tritium-labeled proteins and nucleic acids in polyacrylamide gels. Eur. J. Biochem., 46:83–88. Bruckner, P., Horler, I., Mendler, M., Houze, Y., Winterhalter, K.H., Eich-Bender, S.G., and Spycher, M.A. (1989) Induction and prevention of chondrocyte hypertrophy in culture. J. Cell Biol., 109:2537– 2545. Buckwalter, J.A., Rosenberg, L.C., and Ungar, R. (1987) Changes in proteoglycan aggregates during cartilage mineralization. Calcif. Tissue Int., 41:228–236. Cole, A.A., Luchene, L.J., Linsenmayer, T.F., and Schmid, T.M. (1992) The influence of bone and marrow on cartilage hypertrophy and degradation during 30-day serum-free culture of the embryonic chick tibia. Dev. Dyn., 193:277–285. Derynck, R. (1994) Transforming growth factor-␤. In: Cytokine Handbook. 2nd ed. A.W. Thomson, ed. Academic Press, London and San Diego, pp. 319–342. Eavey, R.D., Schmid, T.M., and Linsenmayer, T.F. (1988) Intrinsic and extrinsic controls of the hypertrophic program of chondrocytes in the avian columella. Dev. Biol., 126:57–62. Farnum, C.E., and Wilsman, N.J. (1987) Morphologic stages of the terminal hypertrophic chondrocyte of growth plate cartilage. Anat. Rec., 219:221–232. Franceschi, R.T. (1992) The role of ascorbic acid in mesenchymal differentiation. Nutr. Rev., 50:65–70. Franceschi, R.T., Iyer, B.S., and Cui, Y. (1994) Effects of ascorbic acid on collagen matrix formation and osteoblast differentiation in murine MC3T3-E1 cells. J. Bone Miner. Res., 9:843–854. Gawande, S.R., and Tuan, R.S. (1990) Action of 1,25(OH)2 vitamin D3 on calcium metabolism and skeletal development of the developing chick embryo. Trans. Ortho. Res. Soc. 15:401. Gerstenfeld, L.C. and Shapiro, F.D. (1996) Expression of bone-specific genes by hypertrophic chondrocytes: Implications of the complex functions of the hypertrophic chondrocyte during endochondral bone development. J. Cell. Biochem., 62:1–9. Gerstenfeld, L.C., Crawford, D.R., Boedtker, H., and Doty, P. (1984) Expression of type I and III collagen genes during differentiation of embryonic chicken myoblasts in culture. Mol. Cell. Biol., 4:1483– 1492. Gerstenfeld, L.C., Chipman, S.D., Glowacki, J., and Lian, J.B. (1987) Expression of differentiated function by mineralizing cultures of chicken osteoblasts. Dev. Biol., 122:49–60. Gerstenfeld, L.C., Chipman, S.D., Kelly, C.M., Hodgens, K.J., Lee, D.D., and Landis, W.J. (1988) Collagen expression, ultrastructural assembly, and mineralization in cultures of chicken embryo osteoblasts. J. Cell Biol., 106:979–989. Gerstenfeld, L.C., Finer, M.H., and Boedtker, H. (1989) Quantitative analysis of collagen expression in embryonic chick chondrocytes having different developmental fates. J. Biol. Chem., 264:5112– 5120. Gerstenfeld, L.C., Kelley, C.M., Von Deck, M., and Lian, J.B. (1990) Effect of 1,25-dihydroxyvitamin D3 on induction of chondrocyte maturation in culture: Extracellular matrix gene expression and morphology. Endocrinology, 126:1599–1609. Gerstenfeld, L.C., Zurakowski, D., Schaffer, J.L., Nichols, D.P., Toma, C.D., Broess, M., Bruder, S.P., and Caplan, A.I. (1996) Variable hormone responsiveness of osteoblast populations isolated at different stages of embryogenesis and its relationship to the osteogenic lineage. Endocrinology, 137:3957–3968. Glimcher, M.J. (1992) The nature of the mineral component of bone and the mechanism of calcification. In: Disorders of Bone and Mineral Metabolism. Coe, F.L. and Favus, M.J., eds. Raven Press, Ltd., New York, pp. 265–286. Grigoriadis, A.E., Heersche, J.N.M., and Aubin, J.E. (1988) Differentiation of muscle, fat, cartilage, and bone from progenitor cells present in a bone-derived clonal cell population: Effect of dexamethasone. J. Cell Biol., 106:2139–2151. Ham, A.W., and Cormack, D.H. (1979) Histology. 8th ed. J.B. Lippincott Company, Philadelphia and Toronto, pp. 377–462. Holtrop, M.E. (1972) The ultrastructure of the epiphyseal plate. II: The hypertrophic chondrocyte. Calcif. Tissue Res., 9:140–151. Hunziker, E.B., Herrmann, W., Schenk, R.K., Mueller, M., and Moor, H. (1984) Cartilage ultrastructure after high pressure freezing, freeze substitution, and low temperature embedding. I: Chondrocyte ultrastructure—implications for the theories of mineralization and vascular invasion. J. Cell Biol., 98:267–276. Ishizaki, Y., Burne, J.F., and Raff, M.C. (1994) Autocrine signals enable chondrocytes to survive in culture. J. Cell Biol., 126:1069– 1077. Jacenko O., and Tuan, R.S. (1986) Calcium deficiency induces expression of cartilage-like phenotype in chick embryonic calvaria. Dev. Biol., 115:215–232. Kodama, H.A., Amagai, Y., Koyama, H., and Kasai, S. (1982) Hormonal responsiveness of a preadipose cell line derived from newborn mouse calvaria. J. Cell. Physiol., 112:83–88. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227:680–685. Landis, W.J., and Glimcher, M.J. (1978) Electron diffraction and electron probe microanalysis of the mineral phase of bone tissue prepared by anhydrous techniques. J. Ultrastruct. Res., 63:188– 223. Landis, W.J., and Glimcher, M.J. (1982) Electron optical and analytical observations of rat growth plate cartilage prepared by ultracryomicrotomy: The failure to detect a mineral phase in matrix vesicles and the identification of heterodispersed particles as the initial solid phase of calcium phosphate deposited in the extracellular matrix. J. Ultrastruct. Res., 78:227–268. Landis, W.J., Paine, M.C., and Glimcher, M.J. (1977) Electron microscopic observations of bone tissue prepared anhydrously in organic solvents. J. Ultrastruct. Res., 59:1–30. Leboy, P.S., Vaias, L., Uschmann, B., Golub, E., Adams, S.L., and Pacifici, M. (1989) Ascorbic acid induces alkaline phosphatase, type X collagen, and calcium deposition in cultured chick chondrocytes. J. Biol. Chem., 264:17281–17286. Lehrach, H., Frischauf, A.M., Hanahan, D., Wozney, J., Fuller, F., and Boedtker, H. (1979) Construction and characterization of pro ␣1 collagen complementary deoxyribonucleic acid clones. Biochem., 18:3146–3152. LuValle, P., Ninomiya, Y., Rosenblum, N.D., and Olsen, B.R. (1988) The type X collagen gene. Intron sequences split the 5’-untranslated region and separate the coding regions for the non-collagenous amino-terminal and triple-helical domains. J. Biol. Chem., 263: 18378–18385. Lyons, B.L., and Schwarz, R.I. (1984) Ascorbate stimulation of PAT cells causes an increase in transcription rates and a decrease in degradation rates of procollagen mRNA. Nucleic Acids Res., 12:2569– 2579. Moore, M.A., Gotoh, Y., Rafidi, K., and Gerstenfeld, L.C. (1991) Characterization of a cDNA for chicken osteopontin: Expression during bone development, osteoblast differentiation, and tissue distribution. Biochemistry, 30:2501–2508. Neugebauer, B.M., Moore, M.A., Broess, M., Gerstenfeld, L.C., and Hauschka, P.V. (1995) Characterization of structural sequences in the chicken osteocalcin gene: Expression of osteocalcin by maturing osteoblasts and by hypertrophic chondrocytes in vitro. J. Bone Miner. Res., 10:157–163. Osdoby, P., and Caplan, A.I. (1979) Osteogenesis in cultures of limb mesenchymal cells. Dev. Biol., 73:84–102. Probst, A., and Spiegel, H.U. (1997) Cellular mechanisms of bone repair. J. Invest. Surg., 10:77–86. Quarto, R., Campanile, G., Cancedda, R., and Dozin, B. (1992) Thyroid hormone, insulin, and glucocorticoids are sufficient to support chondrocyte differentiation to hypertrophy: A serum-free analysis. J. Cell Biol., 119:989–995. Rifas, L., Uitto, J., Memoli, V.A., Kuettner, K.E., Henry, R.W., and Peck, W.A. (1982) Selective emergence of differentiated chondrocytes during serum-free culture of cells derived from fetal rat calvaria. J. Cell Biol., 92:493–504. Rousseau, F., Bonaventure, J., Legeai-Mallet, L., Pelet, A., Rozet, J.M., Maroteaux, P., Le Merrer, M., and Munnich, A. (1994) Mutations in the gene encoding fibroblast growth factor receptor-3 in achondroplasia. Nature, 371:252–254. CHONDROGENIC POTENTIAL OF SKELETAL CELLS Sai, S., Tanaka, T., Kosher, R.A., and Tanzer, M.L. (1986) Cloning and sequence analysis of a partial cDNA for chicken cartilage proteoglycan core protein. Proc. Natl. Acad. Sci., U.S.A., 83:5081–5085. Sandell, L.J., and Daniel, J.C. (1988) Effects of ascorbic acid on collagen mRNA levels in short term chondrocyte cultures. Connect. Tissue Res., 17:11–22. Schmid T.M., and Linsenmayer, T.F. (1985) Developmental acquisition of type X collagen in the embryonic chick tibiotarsus. Dev. Biol., 107:373–381. Shapiro, F. (1992) Vertebral development of the chick embryo during days 3–19 of incubation. J. Morphol., 213:317–333. Shapiro, F., Holtrop, M.E., and Glimcher, M.J. (1977) Organization and cellular biology of the perichondrial ossification groove of Ranvier: A morphological study in rabbits. J. Bone Joint Surg., 59A:703–723. Solursh, M., and Meier, S. (1974) Effects of cell density on the expression of differentiation by chick embryo chondrocytes. J. Exp. Zool., 187:311–322. Terashima, Y., and Urist, M.R. (1975) Differentiation of cartilage from calvarial bone under the influence of bone matrix gelatin in vitro. Clin. Orthop. Rel. Res., 113:168–177. Thomas, PS. (1980) Hybridization of denatured RNA and small DNA 173 fragments transferred to nitrocellulose. Proc. Natl. Acad. Sci., U.S.A., 77:5201–5205. Thorogood, P. (1979) In vitro studies on skeletogenic potential of membrane bone periosteal cells. J. Embryol. Exp. Morphol., 54:185– 207. Toma, C., Schaffer, J., Meazzini, M.C., Zurakowski, D., Nah, H.D., and Gerstenfeld, L.C. (1997) Developmental restriction of embryonic mesenchymal stem cells as characterized by their in vitro potential for differentiation. J. Bone Miner. Res., 12:2024–2039. Wong, M., and Tuan, R.S. (1993) Nuserum, a synthetic serum replacement, supports chondrogenesis of embryonic chick limb bud mesenchymal cells in micromass culture. In Vitro Cell. Dev. Biol., Animal, 29A:917–922. Wong, M., and Tuan, R.S. (1995) Interactive cellular modulation of chondrogenic differentiation in vitro by subpopulations of chick embryonic calvarial cells. Dev. Biol., 167:130–147. Yamaguchi, A., and Kahn, A.J. (1991) Clonal osteogenic cell lines express myogenic and adipocytic developmental potential. Calcif. Tissue Int., 49:221–225. Young, M.F., Vogeli, G., Nunez, A.M., Fernandez, M.P., Sullivan, M., and Sobel, M.E. (1984) Isolation of cDNA and genomic DNA clones encoding type II collagen. Nucleic Acids Res., 12:4207–4228.