Changes in the Sulfated Proteoglycans Synthesized by “Aging” Chondrocytes I. DISPERSED CULTURED CHONDROCYTES AND IN VIVO CARTILAGES*

Structural and chemical changes in the sulfated pro- teoglycans synthesized by “aging” cultures of dispersed chondrocytes were compared to those synthe- sized by freshly excised, intact cartilage explants in organ culture from chickens of various ages. In vitro, a with a decrease in the average size of the chondroitin sulfate chains; (ii) were to to regardless of size; (iii) maintained a relatively constant 6S/4S disaccharide ratio of for and (iv) a keratan remained attached to the substrate. The floating cells were concen- trated by centrifugation, washed with Eagle's minimal essential medium, resuspended in 0.25% trypsin for 10 min, and then plated at 5 X lo5 cells/100-mm dish as secondary subcultures. For further sub-culturing, the cells were washed with Eagle's minimal essential me- dium, treated for 3 to 5 min with a trypsin-EDTA solution, and plated at 5 X lo5 cells/100-mm dish. Fibroblast cultures were prepared from subdermal tissues of 11-day-old chick embryos and processed, as were the dispersed chondrocytes (19). Other experiments involved brief organ cultures of intact pieces of cartilage. Vertebral cartilages and entire cartilaginous heads from 11- day-old embryos, and tibial articular cartilages from 1- and 6-year-old chickens were dissected free of adhering connective tissues, cut into small l-mm2 pieces, and organ cultured. By Day 11, during chick embryo development, tibial epiphyseal heads are entirely cartilagi- nous. During further development, the cartilaginous model is replaced by bone, except at the site of the articular cartilage. These intact pieces of cartilage were incubated and analyzed in the same manner as that used for the dispersed chondrocytes. contents or-gan-cultured vertebral cartilages were determined by the fluorometric assay Hinegardner (20) using thymus DNA as a standard.

2 for reviews). One of the exceptions has been the delineation of changes in the structural and chemicophysical properties of the intercellular matrix in cartilages from young and old animals (3-7). More recently, the change in the synthetic behavior of chondrocytes under different culture conditions has been analyzed and proposed as a model for the aging of chondrocytes in vivo (8,9).
This report focuses on the question of whether the structural and chemical changes in the type IV-sulfated proteoglycan characteristic of cartilage (10-14) that occurs in vitro are qualitatively equivalent to those known to occur in vivo. This type of study is particularly feasible now, owing to the recent advances in knowledge of the molecular structure of sulfated proteoglycans (15). These molecules characteristic of cartilage consist of a core protein to which are covalently attached -100 chondroitin sulfate and 30 to 60 keratan sulfate chains with average M , = -20,000 and 5,000, respectively (16). One end of the core protein, which lacks glycosaminoglycan chains, possesses the property of interacting with hyaluronic acid. This allows large numbers of single proteoglycan monomers to form supramolecular aggregates within the cartilage matrix (17,18).
In this study, we have identified several alterations in the structural features of the sulfated proteoglycans synthesized by dispersed chondrocytes maintained in vitro for several weeks, and in cartilage explants isolated from animals of various ages and maintained in short term organ culture. Our data do not support the notion that the alterations observed in cultured chondrocytes are the same as those observed in cartilages of animals of increasing age. These differences that we and others have found with time in cultures of dispersed chondrocytes may reflect adaptations to the microenvironment in vitro rather than mirror aging in vivo.

Proteoglycans from Cultured Chondrocytes and in Vivo Cartilages
remained attached to the substrate. The floating cells were concentrated by centrifugation, washed with Eagle's minimal essential medium, resuspended in 0.25% trypsin for 10 min, and then plated at 5 X lo5 cells/100-mm dish as secondary subcultures. For further subculturing, the cells were washed with Eagle's minimal essential medium, treated for 3 to 5 min with a trypsin-EDTA solution, and plated at 5 X lo5 cells/100-mm dish. Fibroblast cultures were prepared from subdermal tissues of 11-day-old chick embryos and processed, as were the dispersed chondrocytes (19). Other experiments involved brief organ cultures of intact pieces of cartilage. Vertebral cartilages and entire cartilaginous heads from 11day-old embryos, and tibial articular cartilages from 1-and 6-year-old chickens were dissected free of adhering connective tissues, cut into small l-mm2 pieces, and organ cultured. By Day 11, during chick embryo development, tibial epiphyseal heads are entirely cartilaginous. During further development, the cartilaginous model is replaced by bone, except at the site of the articular cartilage. These intact pieces of cartilage were incubated and analyzed in the same manner as that used for the dispersed chondrocytes.
DNA contents of cultured, dispersed chondrocytes and intact organ-cultured vertebral cartilages were determined by the fluorometric assay of Hinegardner (20) using calf thymus DNA as a standard.
Analytical Procedures-Linear sucrose gradients were prepared by using a Buchler apparatus. Analytical Sepharose 2B and Sepharose 6B gel filtration columns (105 X 0.6 cm) were prepared and eluted with 0.5 M sodium acetate, pH 7.0. Average running time was 24 h. Effluent fractions were assayed for hexuronic acid with an automated carbazole procedure (21). For radioactivity measurements, 0.5-ml aliquots of samples were mixed with 1 volume of water and 10 ml of Aquasol. Samples were counted with a Beckman LS 3155 scintillation counter. Quenching was determined by the external standard method.
Proteoglycan Extraction and Characterization on Sucrose Gradients-Cultures were labeled by incubation for 6 h in 4 ml of fresh medium containing 10 to 30 pCi/ml of ["S]sulfate. Proteoglycan extraction was carried out essentially as described previously (10,11). Radioactive medium was removed from two 100-mm culture dishes and any floating cells removed by centrifugation. Proteoglycans in the medium were precipitated by the addition of ethanol and potassium acetate to final concentrations of 70 and l%, respectively, at 4°C. The precipitated material was collected by centrifugation (10,000 rpm X 15 min), resuspended in 4 ml of 4 M guanidine HC1/50 mM sodium acetate, pH 5.8, containing protease inhibitors, 0.1 M 6-aminohexanoic acid, 0.005 M benzamidine, and 0.05 M EDTA, and a few hours later, the material was precipitated again with ethanol/potassium acetate overnight at 4°C. This precipitation was repeated a 3rd time. The final precipitate was dissolved in Gdn-HC1 solution' by stirring overnight at 4°C.
The cells were disrupted using a Teflon-glass homogenizer in 1.0 to 1.5 ml of Gdn-HC1 solution. After stirring the suspension overnight at 4"C, the homogenate was clarified by centrifugation and the supernatant precipitated overnight with ethanol/potassium acetate. The precipitate was collected by centrifugation, washed two times with 70% ethanol, then dissolved with Gdn-HC1 solution by shaking overnight at 4OC. The extracts from cells and medium were clarified by centrifugation. The supernatants contained more than 90% of the total 35S-labeled material. Supernatants from cells and medium were mixed and the mixture was layered on a linear 5 to 20% sucrose gradient in 4 M guanidine HC1, 50 mM sodium acetate, pH 5.8 (10,11,22). The ionic strength of these gradients is equivalent to the dissociative conditions of Hascall and Sajdera (23), and thus results in the separation of proteoglycan monomers, as opposed to the aggregate molecule.
were labeled for 6 h with ["Slsulfate, washed five times with cold Organ-cultured, intact embryonic, and adult chicken cartilages Eagle's minimal essential medium, and then extracted in 10 volumes of Gdn-HC1 solution by shaking overnight at 4'C. The extracts were clarified by centrifugation and used directly for sucrose gradients. Gradients were centrifuged for 24 to 26 h in a Beckman SW 27.1 rotor at 27,000 rpm at 20°C. Fractions of 0.5 ml were collected from the bottom and counted in Aquasol for radioactivity as described above.
Isolation ofspecific Sulfated Proteoglycan-Specific sulfated pro- teoglycans from both the dispersed chondrocyte cultures and from the organ-cultured embryo vertebral and adult cartilages were isolated as follows. After labeling with [35]sulfate, medium and cells were separated. Medium was made 4 M in guanidine HCI by adding appropriate amounts of solid guanidine (-0.53 g/ml); concentrated sodium acetate buffer and protease inhibitors were also added to give the approximate final concentrations described above. Cell layers were extracted with 1.0 to 1.5 ml of Gdn-HC1 solution. Cells and medium were kept at 4OC overnight. After clarification by centrifugation, the supernatants from medium and cells were combined. The samples were made 5% glucose, then 1.0 to 1.5 ml was layered on a linear 7.5 to 20% sucrose gradient in 4 M guanidine, 50 mM sodium acetate, pH 5.8. Gradients were centrifuged for 24 h at 27,000 rpm; 31 fractions of 0.5 ml were collected from the bottom of the gradient.
The fist 17 fractions were mixed together when type IV-sulfated proteoglycans were isolated from cultured chondrocytes or organcultured cartilages; Fractions 10 through 20 were mixed when type I11 macromolecules were isolated (see Fig. 3 below). Pooled fractions were dialyzed against three changes of 0.5 M sodium acetate, pH 7.0, containing 20 mM sodium sulfate, and then against three changes of distilled water. Five milligrams of carrier bovine nasal proteoglycans (Al-Dl fraction) (23) were added to the labeled material prepared from two 100-mm dishes and the mixture was freeze-dried. Carrier proteoglycans are in vast excess in comparison to the labeled proteoglycans; consequently, the hexuronic acid analyses described below refer almost exclusively to the carrier proteoglycans.
Chromatographic Characterization of Sulfated Proteoglycans-Sepharose 2B and Sepharose 6B columns (105 X 0.6 cm) were prepared and eluted with 0.5 M sodium acetate, pH 7.0. Freeze-dried samples (0.5 mg) were each solubilized in 0.15 ml of 0.5 M sodium acetate buffer, pH 7.0, and applied to the appropriate column. Fractions of 0.52 ml were collected and aliquots of each (0.12 ml) were analyzed for hexuronic acid with an automated carbazole procedure (21) and then for radioactivity.
When assayed for ability to bind to hyaluronic acid, 0.5 mg of each freeze-dried sample was dissolved in 0.15 ml of 0.5 M sodium acetate buffer, pH 7.0. Then, 5 to 10 pg of hyaluronic acid were added and 12 to 16 h later, the mixture was applied to a Sepharose 2B column (12,17). Fractions of 0.52 ml were collected and analyzed as described above.2 Enzymatic Treatments-Portions of freeze-dried samples (0.5 mg) were digested with 100 pg of papain at 65°C for 4 to 5 h in 100 mM sodium acetate, pH 7.0, 5 mM EDTA, and 5 mM cysteine hydrochloride in a total volume of 0.2 ml. Samples were chromatographed on a Sepharose 6B column immediately or stored frozen until used (24). Fractions of 0.52 ml were collected and analyzed as described above.
Lyophilized samples were dissolved in 50 pl of Tris buffer, and chondroitinase ABC digestion (0.1 unit/0.5 mg of sample) was done as described (10,12,25). The digests were then spotted on Whatman No. 1 paper, and paper chromatograms developed for 20 to 22 h with 1-butanol/acetic acid/l N ammonia (2:3:1). The paper chromatograms were air-dried, 1.0-cm strips were cut, and each strip eluted in 0.5 ml of 10 mM HCI. Radioactivity contents in the eluted material were determined.
Keratan sulfate chains were analyzed as follows. Portions of the lyophilized samples were dissolved in Tris buffer and treated with chondroitinase ABC followed by alkaline borohydride mixture (26): 0.05 M sodium hydroxide, 1 M sodium borohydride for 24 h at 45OC. After neutralization with acetic acid, samples were applied to a Bio-Gel P-30 column (110 X 0.6 cm).
Electron Microscopy-Cultured cells were fiied in 3% gluteraldehyde in 0.25 M cacodilate buffer, pH 7.4, for 2 h at room temperature, and postfixed for 1 h in 1% osmium tetraoxide. Fixed cells were then dehydrated in a gradient series of ethanol and embedded in Epon 812. Ultrathin sections were stained with 0.5% uranyl acetate and 0.1% lead citrate. Sections were examined on a JEOL electron microscope.

RESULTS
Cultured Chondrocytes-Pure primary populations of chick embryo chondrocytes were obtained from vertebral cartilages, as described elsewhere (11). Briefly, floating chondrocytes which have accumulated matrix components on their peripheries were selectively harvested from the medium of primary cultures. The floating chondrocytes from 2-to 3-day-old Under these experimental conditions, only the cartilage-characteristic proteoglycans bound hyaluronic acid.
cultures were transferred to secondary cultures after treatment with proteolytic enzymes. As a consequence of this treatment, many chondrocytes become attached to the plastic substrate, whereas the remaining ones still float in the medium. The number of floating chondrocytes normally decreases with the age of the secondary culture: by Day 2 and 3, -50%, and by Day 4 and 5, -90% of the chondrocytes adhere to the culture dish-forming characteristic epithelioid colonies of polygonal cells. The chondrocytes were always subcultured just before reaching confluency.
For the present study, chondrocytes were grown for a total period of 6 weeks by means of six successive subcultures. After six subcultures (Fig. l), the great majority of the chondrocytes maintained their gross cytological characteristics: polygonal morphology, lack of overlapping, lack of motility, and a metachromatic extracellular matrix after staining with methylene blue (27). Only a small percentage (-1 to 2%) of "giant" cells and other fibroblast-like cells were observed. However, there are striking differences in the rough endoplasmic reticulum between early and late passaged chondrocytes (Fig. 2, A and  B ) . Cells subcultured two to three times display many prominent, parallel arrays of rough endoplasmic reticulum. By the 5th or 6th passage, these parallel arrays have invariably disappeared. Instead, the numerous cisternae appear to be randomly distributed and distended, containing a relatively homogeneously stained material.
[35SS/Sulfate Incorporation-Secondary cultures were prepared. At each time point, floating and substrate-bound chondrocytes were separated and pulsed independently with ["S]sulfate for 6 h. Table I shows the result of one typical experiment. The level of total ["S]sulfate incorporation/cell, as estimated by DNA content, was usually maximal in the early days of a given culture, dropping as that culture aged. In tertiary cultures, the level of incorporation was slightly lower than that of late secondary substrate-bound chondrocytes (Table I). Similar decreases in the level of incorporation were observed in subsequent subcultures, as shown in Table I. In all cases, -90% of the total "S-labeled macromolecules were digested by chondroitinase ABC (see Table I1 below).
Pattern of 35S-labeled Proteoglycans on Sucrose Gradients-The "S-labeled proteoglycans synthesized by chondrocytes can be separated into two size classes on a linear sucrose gradient in 4 M guanidine. The larger proteoglycan, operationally termed type IV monomer, is characteristic for cartilages and cultured chondrocytes. With the techniques used, the smaller type I 35SS-labeled proteoglycan fraction appears to be synthesized by many kinds of cells: fibroblasts, myotubes, spinal cord, etc. (see also, Ref. 14). Fig. 3A illustrates the patterns of sulfated proteoglycans synthesized by secondary floating chondrocytes as compared to secondary substrate-bound chondrocytes after a 3-day growth period in vitro. A profile for proteoglycans from organcultured, intact vertebral cartilages of the same embryonic age is shown for comparison. The figure demonstrates that the type IV monomers synthesized by both these dispersed cultured cell populations (i.e. floaters and substrate-bound chondrocytes) migrate on sucrose gradients more slowly than those synthesized by organ-cultured embryonic vertebral cartilages. Clearly, the monomers synthesized by cultured chondrocytes have smaller average molecular sizes than those synthesized by the intact vertebral cartilages. The figure shows also that the type IV monomer in floating chondrocytes has a larger molecular size than that of substrate-bound chondrocytes. In both cell populations, however, the overall proportion between type IV and type I sulfated proteoglycans is similar to that present in organ-cultured, intact vertebral cartilages.
The decrease in the molecular size of type IV monomers was more evident at later stages of chondrocyte growth in culture. Figure 3, B and C, shows results for chondrocytes grown for 6 and 10 days, respectively, in secondary cultures. The decrease of the molecular size of the monomers is related to the age of the culture. The decrease is faster in substratebound chondrocytes and the relative proportion between type IV and type I proteoglycan changes, with the proportion of type I greatly increased in 10-day-old substrate-bound chondrocytes.
Secondary chondrocytes were subcultured after trypsinization and grown for 6 days in tertiary cultures. The pattern of migration of the sulfated proteoglycans synthesized by these cells is shown in Fig. 30. There is no significant further decrease of the average molecular size of the type IV monomers synthesized in these tertiary as compared with the secondary cultures. On the other hand, the ratio between the modified type IV and type I sulfated proteoglycans is now greatly decreased.
When the proteoglycans synthesized by chondrocytes subcultured three more times, i.e. grown for a total of 6 weeks in culture, were extracted and analyzed in the same manner, they appeared essentially identical with those synthesized by tertiary chondrocytes.
Interaction of Type IV Monomers with Hyaluronic Acid-Type IV proteoglycan monomers interact specifically with hyaluronic acid under the experimental conditions used. Neither type 111, 11, nor I obtained from normal dermal fibroblasts or muscle cells or chondrocytes treated with phorbol-12-myristate-13-acetate or 5"bromodeoxyuridine (10, 19) bind to hyaluronic acid under the conditions used here (data not shown). The ability of type IV monomers to interact with hyaluronic acid was tested. %-labeled type IV proteoglycan monomers from different cultures were collected after sepa-m ? p w m " r : : . . ration on sucrose gradients, and dialyzed and lyophilized in the presence of carrier bovine nasal cartilage proteoglycan. Carrier proteoglycans were used to facilitate the recovery of the limited amounts of material synthesized by cultured chondrocytes and to provide an internal standard for a direct comparison with the "S-labeled proteoglycans.
Portions of the lyophilized materials were dissolved in the appropriate buffer and then chromatographed on Sepharose 2B columns. The elution of the carrier was monitored by analysis for hexuronic acid, whereas the elution patterns of the experimental samples were monitored by radioactivity. As shown in Figs. 4, A , C, and E, type IV monomers from both vertebral cartilages and from cultured condrocytes (secondary or 6th passage) are included in the column. The vertebral cartilage monomers eluted earlier (KaV = 0.18) than the bovine nasal cartilage carrier (ICav = 0.21), whereas type IV monomers  shown in Fig. 4, B, D, and F. They demonstrate that the great majority of the 35S-labeled monomers synthesized by organcultured vertebral cartilages and by the cultured chondrocytes, as well as the carrier proteoglycan, eluted in the excluded volume, indicating that specific interactions with hyaluronic acid had occurred. Structural Characteristics of the Type IVMonomers-The decrease in the molecular weight of type IV monomers can be caused by several different structural changes in the macromolecules. One such change could be due to alteration of the average molecular size of the polysaccharide chains.
Portions of the 35S-labeled type IV monomers with carriers were digested with papain to release the polysaccharide chains. The digests were directly chromatographed on Sepharose 6B columns in 0.5 M sodium acetate to obtain an estimate of the average molecular weight of the released polysaccharide chains (24). The profiles obtained on Sepharose 6B refer almost exclusively to chondroitin sulfate chains, since 90 to 95% of sample radioactivity was digested by chondroitinase ABC (Table XI).
The relased chondroitin sulfate chains of organ-cultured vertebral cartilage and of the 3-day-old cultured chondrocytes (Fig. 5, A and B

P r o t e o g b n s from Cultured Chondrocytes and in Vivo Cartilages
the type IV monomers synthesized by cultured chondrocytes are shorter than those of the monomers synthesized by vertebral cartilages in organ culture. This structural change is even more evident in the type IV monomers from 10-day-old secondary chondrocytes (Fig. 5C), since their chains have a significantly lower molecular size (Kav = 0.56; M, = -15,500). Unexpectedly, the chondroitin sulfate chains synthesized by 6-wk-old chondrocytes appeared to have a slightly larger molecular size in this particular sample (Fig. 50). We then determined the relative ratios of the 6s and 4 s disaccharides in the isolated proteoglycans. The macromolecules were digested with chondroitinase ABC and digests were analyzed by paper chromatography. Table I1 shows that: (i) the 6S/4S disaccharide ratio remains practically unchanged throughout the entire series of chondroblast subcultures; (ii) the ratios are the same for the organ-cultured vertebral cartilage proteoglycans; and (iii) 90 to 95% of the radioactivity in any sample was digested by the chondroitinase ABC.
The constant presence of undigested material, which comprised 5 to 10% of the total sample radioactivity, gives a reasonably accurate estimation of the relative amount of keratan sulfate to chondroitin sulfate in the samples. We can therefore conclude that the type IV monomers synthesized at each chondrocyte subculture contained a constant ratio of keratan sulfate to chondroitin sulfate.  Embryonic and Adult Cartilages-Identical experiments were done for the "S-labeled proteoglycans synthesized by short term organ cultures of explants of embryonic tibial heads and articular cartilages directly isolated from chick embryos and adult chickens.
Cartilaginous tibial heads from a 11-day-old chick embryo and articular cartilages from 1-and 6-year-old chickens were dissected and labeled immediately in vitro for 6 h with [35S]sulfate. Labeled proteoglycans were extracted and analyzed on linear sucrose gradients. The patterns obtained, shown in Fig. 6, demonstrate that: (i) the proteoglycans synthesized by embryonic tibial cartilages have nearly the same molecular sizes as those observed in embryonic vertebral cartilages; (ii) the type IV monomers synthesized by the 6year-old chicken articular cartilages have much smaller sizes than those synthesized by the embryonic cartilage; and (iii) the overall proportion between the type IV and type I classes of 35S-labeled proteoglycans is greatly different in embryonic as compared to adult cartilages.
Type IV monomers were collected from the sucrose gradients, dialyzed, and lyophilized in the presence of carrier bovine nasal proteoglycans. The isolated monomers were then tested for the ability to bind to hyaluronic acid and for glycosaminoglycan chain sizes. Fig. 7 shows that type IV monomers synthesized by both embryonic cartilages and adult chicken articular cartilages are able to bind to hyaluronic acid, and, therefore, the decrease of the molecular size observed for labeled adult chicken cartilage monomers (Kav = 0.31) compared with embryonic cartilage (ICav = 0.16) does not affect their characteristic property to bind to hyaluronic acid.
Since almost 40% of the polysaccharides present in the 35Slabeled type IV monomers of the 6-year-old articular cartilages are keratan sulfate (see Fig. 9), the pattern reported in Fig. 8B contains a mixture of -60% chondroitin sulfate and 40% keratan sulfate chains.
Aliquots of the isolated type IV monomers were then digested with chondroitinase ABC, and the digests were resolved on chromatographic paper. Table I11 shows that 95% of the type IV monomer radioactivity from embryonic tibial cartilaginous heads was digested by the enzyme, and that the relative proportion of 6 s and 4s disaccharides is similar to .
. that found in embryonic vertebral cartilage monomers. On the contrary, only 77 and 60% of the monomer radioactivity, from 1-and 6-year-old articular cartilages, respectively, were digested by chondroitinase ABC, of which -35% migrated as 6 s disaccharide and 65% as 4 s disaccharide.These findings demonstrate that the amount of keratan sulfate increases from 5 to 10% in proteoglycans synthesized by embryonic cartilage to 20 to 40% in proteoglycans synthesized by adult articular cartilages.
Aliquots of the isolated type IV monomers were treated with chondroitinase ABC and then with alkaline borohydride to /?-eliminate the keratan sulfate chains. The samples were then applied to a P-30 column (28). About 10% of the total radioactivity from embryonic cartilage monomers eluted as a peak (ICav = 0.32) in the column, whereas the remaining 90% eluted with the total column volume (Fig. 9). Conversely, -45% of the total radioactivity from 6-year-old articular cartilage monomers eluted as a peak (ICav = 0.40) in the column.  ( The percentages of radioactivity included in the column and eluted with the total column volume give accurate estimations of the keratan sulfate and the chondroitin sulfate, respectively, in the isolated monomers (28-29). These data confirm that the relative proportion of keratan sulfate is much higher in adult articular cartilage proteoglycans than in embryonic cartilage proteoglycans.

DISCUSSION
Modifications of proteoglycan structure have been reported to occur in cartilages from different species as they age, although the relationship of such modifications to the issue of aging is still problematical (3,5-7,30). For example, Hjertquist and Wasteson (30) reported a shortening of the chondroitin sulfate chains in articular cartilages from normal, older human individuals, a change that also occurs precociously in individuals affected by osteoarthritis . Inerot et al. (7) reported a progressively increased keratan sulfate content and a decreased chondroitin sulfate content in normal hip articular cartilage proteoglycans isolated from dogs of increasing age; however, they did not observe a concomitant change in the average molecular size of the chondroitin sulfate chains. These and other experimental observations demonstrate that the structural modification of proteoglycans associated with aging cartilages in vivo may vary modestly from species to species. Changes similar to those described by these investigators, which were based on chemical analysis of proteoglycans, also occur in chick cartilages as described above. Moreover, by using freshly excised cartilage for short term organ culture, we have in this paper shown that the changes in proteoglycan structures such as decreased monomer size with time, reflect changes in the biosynthesis of proteoglycans. Thus, the changes in chondroitin sulfate chain size (30) and in the relative amounts of keratan sulfate (7), which characterize proteoglycans extracted from the cartilage matrix of aging animals, may be due primarily to changes in their biosynthesis (Figs. 8 and 9). Clearly, we cannot conclusively rule out the alternative, although very unlikely possibility, that the structural changes in the newly synthesized proteoglycans occur after the proteoglycans are released into the extracellular environment.
Structural modifications occurring with time, such as the reduction of the average monomer weight and the shortening of the chondroitin sulfate chains, are common to sulfated proteoglycan monomers synthesized by in vitro chondrocytes, as well as to those synthesized by freshly excised cartilage.
However, there exist two main qualitative structural alterations; namely, the decrease of the chondroitin sulfate/keratan sulfate ratio and the decrease of the 6S/4S disaccharide ratio, which are expressed only in adult and aging cartilages in vivo; these alterations are not observed in "old" or frequently subcultured chick chondrocytes. Consequently, although cultured chondrocytes are useful for many types of studies relating to chondrogenesis, they, nevertheless, may be a poor general model for the study of aging as it occurs in the animal. Alternatively, cultured chondrocytes may prove to be useful for the study of specific aspects of aging chondrocytes; for example, the mechanism of shortening the chondroitin sulfate chains. In this context, it will be important to compare the kinds of sulfated proteoglycans synthesized in vitro by chondrocytes derived from embryos with those derived from old animals.
Under the culture conditions used in this study, a reasonably stable cartilage phenotype can be maintained for many cell doublings, at least in terms of cell morphology and the continued synthesis of a modified type IV sulfated proteoglycan. Alternatively, simply by adding embryo extract, 5"bromodeoxyuridine, fibronectin, or phorbol-12-myristate-13 acetate, cultured chondroblasts can be induced to dedifferentiate rapidly (31). Such dedifferentiated chondrocytes promptly cease synthesizing their type IV monomer, inhibit their synthesis of type I1 collagen chains, and initiate the synthesis of different forms of type I collagen chains (19,(32)(33)(34)(35). Similarly, Benya et al. (9) and von der Mark et al. (36) have shown that the types of collagen chains may vary as normal chondrocytes are subcultured. At present, it is not possible to relate precisely which of these changes are uniquely associated with dedifferentiation in vitro and which are characteristic for aging in vivo. It is not even clear whether: (i) the modest numbers of aging and dying chondrocytes that appear as part of the normal developmental program of embryonic vertebral and limb bud cartilages, or (ii) the changes displayed by the aging and dying hypertrophying chondrocytes at the epiphyseal plates of young and growing animals undergo the same changes that occur in aged animals.
The systematic qualitative changes in the sulfated proteoglycans, or even in collagen chains of aging cells in culture do not fulfii the notion of the accumulation of "faulty molecules" predicted by Orgel's aging hypothesis (37). On the other hand, the accumulation of systematically altered complex molecules, such as sulfated proteoglycans, which require sequential enzymatic additions, may be more prone to systematic changes in entire regulatory pathways than in the straightforward translation of a simple protein.
Pending more detailed comparisons of the same terminally differentiated cells in vitro and in vivo, claims regarding aging or "senescence" (8) in the former must be interpreted with considerable caution. Possible causes for these differences could be the very different microenvironments of the chondrocytes in vitro and in vivo. One very conspicuous difference is the higher rate of replication in cultured chondrocytes as compared to chondrocytes in vivo. The chondrocytes in culture double in number approximately every 48 h. For the most part, the definitive chondrocytes in vivo rarely replicate. The constantly dying chondrocytes in vivo are replaced by the generation of new cells in the perichondrium (38,39). I n vitro, it is primarily the upper surface of the chondrocyte that is covered with mat* the undersurface adhering to the plastic substrate generally has very little matrix3 (27). This contrasts with the chondrocyte in vivo which is totally surrounded by a sizeable mass of matrix. That some of these factors may be involved in the differences in the activity of chondrocytes growing as replicating monolayers and their activity in older animals is suggested by the results obtained when sizeable pieces of intact vertebral columns are organ-cultured for periods up to 10 days (see following paper, Ref. 40).