Correlated metabolism of proteoglycans and hyaluronic acid in bovine cartilage organ cultures.

Proteoglycans exist in cartilage as complexes in which many proteoglycan molecules are bound to a central filament of hyaluronic acid. Many studies have investigated changes taking place in proteoglycan monomer structure during cartilage catabolism usually under the assumption that hyaluronic acid is a relatively inert metabolic component of the complex. In this paper we present organ culture data supporting a new hypothesis that the catabolism of proteoglycans and hyaluronic acid are coordinately regulated by chondrocytes. The data indicates that: 1) newly synthesized hyaluronate and proteoglycan maintain a nearly constant ratio, almost identical to that existing for the total chemical amounts of these two components in cartilage tissue; 2) these two components are catabolized with virtually identical kinetics; and 3) this catabolic relationship in vitro reflects the loss of hyaluronate and proteoglycans from native, undissociated aggregates as isolated from the tissue. We conclude that hyaluronate catabolism is an integral part of the overall mechanism of proteoglycan resorption in cartilage and that further understanding of this process may be key to the elucidation of the regulatory pathways for proteoglycan resorption in health and disease.

Proteoglycans exist in cartilage as complexes in which many proteoglycan molecules are bound to a central filament of hyaluronic acid. Many studies have investigated changes taking place in proteoglycan monomer structure during cartilage catabolism usually under the assumption that hyaluronic acid is a relatively inert metabolic component of the complex.
In this paper we present organ culture data supporting a new hypothesis that the catabolism of proteoglycans and hyaluronic acid are coordinately regulated by chondrocytes. The data indicates that: 1) newly synthesized hyaluronate and proteoglycan maintain a nearly constant ratio, almost identical to that existing for the total chemical amounts of these two components in cartilage tissue; 2) these two components are catabolized with virtually identical kinetics; and 3) this catabolic relationship in vitro reflects the loss of hyaluronate and proteoglycans from native, undissociated aggregates as isolated from the tissue. We conclude that hyaluronate catabolism is an integral part of the overall mechanism of proteoglycan resorption in cartilage and that further understanding of this process may be key to the elucidation of the regulatory pathways for proteoglycan resorption in health and disease.
Osteoarthritis is a major crippling disease that causes profound alterations in the integrity and biomechanical function of cartilage and bone. The precise biochemical events involved in the pathogenesis of this disease are not clearly understood, but it is generally agreed that an important early event in the disease is a loss of proteoglycans from the cartilage matrix (1). Maintenance of the appropriate structure and concentration of the proteoglycans is critical for healthy cartilage to resist compressive loads and shield the underlying bone (2). Thus, impairment of mechanisms controlling these processes would be expected to lead to loss of function and disease. A great deal of interest has focused on understanding mechanisms of proteoglycan catabolism. Proteoglycans exist in the cartilage matrix primarily as aggregates in which many proteoglycan monomers are attached to long strands of hyaluronic acid through a specialized binding region near the amino-terminal portion of the proteoglycan core protein, resulting in gigantic complexes with molecular weights on the order of 50 million. The aggregate structure is stabilized by a link protein which effectively "locks" the proteoglycan to the hyaluronic acid backbone (3,4).
* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The pioneering studies of Fell and Jubb (5) showed that proteoglycan resorption in cartilage is directly under the control of resident chondrocytes. Sandy et al. (6) subsequently showed that proteoglycans released from cultures of rabbit articular cartilage during the catabolic process had large hydrodynamic sizes but were unable to bind to hyaluronic acid. They proposed that the mechanism of release involved a hydrolytic cleavage above the hyaluronic acid binding region of the proteoglycan core protein. Later studies by other groups have been consistent with this hypothesis (7-9), although the issue has not been definitely settled. An alternate catabolic hypothesis is based on the finding that partially degraded link proteins accumulate in cartilage matrix with increasing age (10) and states that proteolytic cleavage of the link protein inactivates its ability to stabilize the proteoglycan-hyaluronate complex. Catabolism of the third component of the complex, hyaluronate, has not been investigated even though its size and integrity clearly contribute to the overall physical properties of the aggregate complex (11). In this paper we show that hyaluronate is a metabolically active component of the complex and is not, as has often been assumed, merely a stable anchor upon which proteoglycans are replaced during the normal course of proteoglycan catabolism.

Buffers and Experimental Conditions
The following buffers were used 1) Tris buffer (0.001 M Tris-HC1, 2 mM calcium chloride, pH 7.2); 2) papain digestion buffer (0.05 M sodium acetate, 5 mM EDTA, 0.5 mM dithiothreitol, pH 6.0); 3) 4 M guanidine HC1 buffer (4 M guanidine HCl, 0.05 M sodium acetate, pH Papain digestions were carried out for 6 h at 60 "C; enzyme concentration was 0.5 mg/ml and the reaction was terminated by addition of iodoacetamide to 5 mM. Streptomyces hyaluronidase digestions were done for 5 h at 60 "C and chondroitin ABC lyase digestions for 5 h at 37 "C. Hydroxyproline determinations were by the method of Woessner (12). For all the experiments described below, samples were stored at -20 "C until all the culturing and extraction procedures were carried out. The age of the animals used for the experiments described below ranged between 3 and 8 months (slaughterhouse estimate). Tissue Cultures and Extractwns-Articulm cartilage slices were prepared from the metacarpal-phalangeal joints of two calves as described previously (7). Approximately 1 g of tissue (wet weight) was placed in each of three batches which were cultured in 15 ml of Dulbecco's modified Eagle's medium with high glucose (DMEM) and 20% fetal calf serum (maintenance medium). Medium was changed daily and on days 8, 18, and 23, one batch was radiolabeled by incubation at 37 "C for 16 h in maintenance medium containing 60 pCi/ml [3H]glucosamine and 6 pCi/ml [36S]sulfate. The medium was then removed and the tissues extracted at 4 "C in 15 ml of the 4 M guanidine HC1 buffer for 24 h.
The extracted tissues were washed twice at 4 "C with 2 ml of Tris buffer to remove residual guanidine HCl. The washed tissue was incubated at 37 "C in 5 ml of Tris buffer containing 6600 units of type I11 collagenase (Advance Biofactures, New York). After digestion for 48 h, the tissues were dispersed by homogenization in a glass homogenizer. A remaining residue was incubated in 1 ml of Tris buffer with 850 units of collagenase overnight. This second digest was homogenized and pooled with the first. This procedure essentially dispersed all of the tissue; a small residue contained less than 5% of the total hydroxyproline in the tissue and was not analyzed further.
Biochemical Analysis of Tissue Extracts-The guanidine HCl extracts were dialyzed extensively against distilled water in the cold. An aliquot of each guanidine HCl and collagenase extract from days 8 and 23 (corresponding to approximately 200 mg wet weight of tissue) was dried and resuspended in papain digestion buffer and digested with enzyme as described above. The samples were then diluted 10-fold with 0.05 M sodium acetate, pH 6, applied to a DEAE-Sephacel column (3 ml) equilibrated in the same buffer, washed with 5 column volumes of buffer, and then eluted with a linear NaCl gradient (0-1.5 M) prepared in the same buffer. The yield of both radioisotopes in the combined breakthrough and gradient fractions was >go%.
Fractions were pooled as indicated by the bars in Fig. 1, A and B, and the peaks analyzed by selective enzyme digestions. Each pool was dried, resuspended in 500 pl of 0.1 M ammonium bicarbonate buffer, pH 8.5, and eluted on a PD-10 column of Sephadex G-25 in the same buffer to remove salts. Fractions eluting in the Vo were combined, portioned into various tubes for enzyme digestions, and dried. Recoveries were >go%. A set of tubes was analyzed by digestion with Streptomyces hyaluronidase as follows: 1) a control sample contained 100 pl of 0.05 M sodium acetate buffer, pH 5; 2) a digestion sample contained 100 p1 of this buffer and 10 units of Streptomyces hyaluronidase; and 3) a digestion sample contained the same additions as sample 2 plus 50 pg of a hyaluronate standard. The latter sample determined the elution position of authentic hyaluronic acid digestion products in the analytical Bio-Gel P-10 column (see below) and ensured that enzyme activity was not rate-limiting. The peaks eluting from DEAE with 0.8 M NaCl were also digested with chondroitin ABC lyase as follows: 1) a control sample contained 30 p l of 0.1 M Tris fluoride, pH 8; and 2) a digestion sample contained 30 pl of the same buffer plus 0.4 unit of chondroitin ABC lyase. Samples were chromatographed on a Bio-Gel P-10 column (0.7 X 100 cm) equilibrated in 0.1 M ammonium bicarbonate, pH 8.5, and eluted at a flow rate of approximately 2 ml/h. Yields were >SO%. Fractions containing the oligosaccharide digestion products were pooled as indicated by the bar in Fig. 2B, dried, resuspended in 0.25-0.5 ml of distilled water, and analyzed for hexuronate content by the method of Bitter and Muir (13). The Bio-Gel P-10 V, fractions for the untreated 0.8 M peaks (which contained >95% of the 3H radioactivity in chondroitin sulfate) were pooled, concentrated, and analyzed for hexuronate content (13).
The validity of the methodology described above for the quantitation of hyaluronate and chondroitin sulfate was tested by using an independent method of analysis for samples from the collagenase extract from the day 18 culture sample. The peaks of radioactivity eluting at 0.35 and 0.8 M NaCl from DEAE-Sephacel were separately pooled (see Fig. lB, bars). Bovine serum albumin (0.5 mg/ml) was added to each as carrier, and the samples were precipitated by adding ethanol to 80% (v/v). After 1 h at room temperature, precipitates were collected by centrifugation at 2000 rpm for 30 min and resuspended in 40 p1 of 0.1 M Tris fluoride, pH 8. Incubation of the samples with papain dispersed insoluble material. After inactivating the papain, 5 pl containing 0.4 unit of chondroitin ABC lyase in the same buffer were added to each sample. After digestion, samples were spotted on cellulose sheets, and thin layer chromatography was done as previously described (14). A mixture of standard chondroitin sulfate and hyaluronic acid disaccharides was spotted on each sheet and visualized by ultraviolet light after chromatography. Sample lanes were divided and cut into 5-mm horizontal strips. The cellulose was scraped off each strip with a razor blade and transferred into a scintillation vial. Disaccharides were eluted by addition of 1 ml of distilled water followed by gentle swirling for 1 h. Ten .ml of Hydrofluor scintillation solvent were added directly to the vials, and the 3H and 3sS radioactivities were determined. Recoveries were between 85-90%.
Catabolism of Glycosaminoglycans in the Total Tissue Pooled cartilage slices from two young calves were cultured in maintenance medium in batches for 6 days. One batch (approximately 1 g) was then radiolabeled by incubation for 19 h with 60 pCi/ml [3H] glucosamine and 15 pCi/ml [53]sulfate. The tissue was washed twice at 37 "C with 15 ml of DMEM to help remove unincorporated isotope and a portion of the tissue (approximately 100-150 mg wet weight) stored for later analysis (the experimental day 0 sample). Duplicate culture vials containing approximately 150-200 mg wet weight of the radiolabeled tissue were set up and maintained in 3 ml of maintenance medium with daily changes. After 26 days of culture, the tissue was extracted at 4 "C for 24 h with 4 M guanidine HCl buffer. The tissue residue was digested with collagenase as described above. After digestion, the clear collagenase extract was removed without homogenizing the tissue. The residual tissue was then digested with papain in 1.5 ml of buffer. The modified extraction protocol was used to determined whether collagenase digestion alone was sufficient to solubilize the hyaluronate and proteoglycan remaining in the tissue after guanidine HCl extraction. All extracts were analyzed by DEAE chromatography essentially as described above. In the case of the papain extracts, aliquots were removed for hydroxyproline analysis, the samples centrifuged, and the supernatants directly applied to DEAE-Sephacel columns. All the extracts from the day 0 and from one of the chase samples were analyzed by Streptomyces hyaluronidase digestion and Bio-Gel P-10 chromatography.

Catabolism of Glycosaminoglycans in the Proteoglycan Aggregate Fraction
Tissue Culture and Development of Extraction Procedures-Tissue slices were cultured in batches as described above. On day 6 (experimental day 0), approximately 1 g (wet weight) of tissue was radiolabeled by incubation for 20 h with 60 pCi/ml [3H]glucosamine and 15 pCi/ml [35S]sulfate. After washing and further incubation for 2 days to remove unincorporated isotope, two portions of the tissue (approximately 150-200 mg wet weight each) were removed, and 0.5 ml of Hank's balanced salt solution was added to each. Both samples were frozen by immersion in a Dry Ice-ethanol bath and stored (experimental day 2 sample). Another portion of the radiolabeled tissue (150-200 mg wet weight) was placed in a culture vial and maintained for another 18 days in 3 ml of maintenance medium with daily changes.
A procedure was developed for the extraction of intact aggregates from the cartilage. A day 2 sample was thawed, portioned into four samples, and each weighed. One was hydrolyzed in 6 N HCl to determine hydroxyproline content so that recoveries could be calculated on the basis of this index. The others were kept moist by adding collagenase buffer. Inhibitors were included in the collagenase buffers to inhibit the action of proteinases of the acidic, cysteine, and serine families. A portion of the tissue was sliced as thinly as possible with a razor blade and maintained at 4 "C in 0.5 ml of collagenase buffer (15). The other two portions were rapidly frozen in a small droplet of distilled water and sectioned into 30-pm slices in a cryostat (16). Frozen slices were recovered from the cold cryostat blade rapidly enough to prevent thawing of the tissue and transferred with forceps into a vial containing 0.5 ml of cold collagenase buffer. The manually sliced tissue and one of the cryostat-sectioned tissues were incubated at 37 "C for 5 h after addition of 0.5 ml of collagenase buffer containing 1000 units of the enzyme. The second tissue portion containing 30-pm slices was incubated under the same conditions without addition of the enzyme. Following incubation, guanidine HCI and ophenanthroline (to block metalloproteinase (collagenase) activity) were added to each sample to a final concentration of 0.5 M and 1 mM, respectively (associative extraction buffer). The tissues were then gently swirled at 4 "C. The following day the buffer was removed (associative extract) and the tissues suspended in 1 ml of collagenase buffer containing 1 mM o-phenanthroline. The tissues were gently homogenized, centrifuged, and the pellets extracted by swirling in 0.5 M NaOH at 4 "C overnight.
The extractability of %-labeled material was assessed. The manually sliced material was not effectively dispersed by homogenization following collagenase treatment, and the combined associative extract and homogenate yielded only 45% of the %-labeled material. The associative extracts from the collagenase-treated and untreated 30pm slices contained 69 and 31% of the "S label, respectively. Gentle homogenization of the residues extracted a further 17 and 46%, respectively. The total extraction was 86% for the collagenase-treated sample and 77% for the untreated sample. Separate experiments showed that the collagenase preparations did not degrade proteoglycan aggregate (aAl from the rat Swarm chondrosarcoma (17) generously supplied by Dr. J. Kimura, Rush Presbyterian St. Luke's Medical Center, Chicago) under experimental conditions comparable to those used here for treatment of the tissues, as determined by velocity sedimentation profiles in cesium sulfate gradients before and after exposure to the enzyme (data not shown).
Tissue Fractionation and Disaccharide Analysis-The remaining day 2 and 18 tissue samples were cryostat-sectioned and digested with collagenase. The collagenase-digested tissues were directly homogenized in the associative extraction buffer and the samples centrifuged. The supernatants (2-4 mg of proteoglycan in 1.5 ml) were layered on preformed cesium sulfate density gradients (0.15-0.5 M) and centrifuged at 20,000 rpm for 6 h as described before (18). The buffers used in the gradients contained the inhibitor mix described above including o-phenanthroline. Following centrifugation the gradients were fractionated by carefully introducing fine tubing to the bottom of the tube and pumping out at a flow rate of approximately 0.6 ml/min.  (19). The column was eluted at 1 ml/min, and 0.5-ml fractions were collected. Chromatographic recoveries were between 50-60%? One ml of 35% ethanol containing 0.5% Triton X-100 was added to each fraction and then mixed with 10 ml of Hydrofluor prior to scintillation counting.
Analysis of Medium Fractions-The medium collected during the 18 days of culturing was pooled, diluted 1.5-fold with 0.05 M sodium acetate, pH 6, and applied on a 15-ml column of DEAE-Sephacel equilibrated in the same buffer. The loaded column was washed with 30 ml of buffer and then eluted with 45 ml of the same buffer containing 1.5 M NaCl. CHAPS (0.5%) was added to the 1.5 M eluant which was then dialyzed first against 10 volumes of distilled water containing the detergent and then against papain buffer containing 0.5% CHAPS. The sample was digested with papain, concentrated by lyophilization, and dialyzed against 0.1 M ammonium bicarbonate, pH 8.5. A portion was dried, resuspended in 200 pl of 0.1 M Tris fluoride, pH 8, and 1 unit of chondroitin ABC lyase added. After incubation, the sample was treated and analyzed on Partisil 5 PAC as described above. The fraction that did not bind to the DEAE-Sephacel was dried by lyophilization, resuspended in 5 ml of distilled The Rpssibility of major selective loss of either hyaluronate or chondroitin sulfate in this experiment is very unlikely. The aggregate preparation contained 2.5-3% of its labeled glycosaminoglycan as hyaluronate and represented approximately 50% of the chondroitin sulfate in the guanidine HCl extract. It is assumed that this represents the percentage of aggregate in the total tissue chondroitin sulfate. Since hyaluronate was detected exclusively in the aggregate preparation, this represents a ratio of hyaluronate to chondroitin sulfate in the tissues of -1% which agrees well with the values determined independently in the other experiments (0.7-1.6%). water, and precipitated with 4 volumes of ethanol. The ethanol supernatant was dried, resuspended in 1 ml of 0.1 M ammonium bicarbonate, pH 8.5, and one-half chromatographed on a Bio-Gel P-10 column equilibrated in the same buffer. All the dialysis steps described above were carried out using Spectrapor 3500 tubing (nominal cut-off for peptide material = 3500 daltons) in order to prevent loss of intermediate to low molecular weight hyaluronate.

RESULTS
Quantitatwn of Glycosaminoglycans-In previous publications it was reported that bovine articular cartilage can be cultured under conditions that maintain a steady-state metabolism of constituent proteoglycans (7, 8). The work described in this paper first focused on: l) whether hyaluronate metabolism in cultured bovine articular cartilage also meets steady-state criteria in terms of both biosynthesis and maintenance of tissue concentrations over a period of 3-4 weeks; and 2) whether hyaluronate has a similar extractability as proteoglycans. Previous work showed that these cultures synthesize proteoglycans (7) and hyaluronate3 with linear kinetics over a 24-h period after changing the medium under the culture conditions used in this work.
Bovine articular cartilage was radiolabeled with [3H]glucosamine and [?3]sulfate after 8, 18, and 23 days of culture. Guanidine HCl and collagenase extracts of the tissues, which contained >95% of the incorporated 36S label, were treated with papain to release glycosaminoglycans. Fig. 1, A and B, show D E W elution profiles of these two extracts from the day 8 samples; the chromatographic profiles for days 18 and 23 showed the same peaks (data not shown). The guanidine HC1 extract contained two resolvable components eluting at 0.12 M NaCl (unsulfated) and 0.8 M (sulfated). Three labeled peaks were separated in the gradients of collagenase extracts: a sulfated peak eluting with 0.18 M NaC1, an unsulfated peak eluting with a salt concentration of approximately 0.35 M, and a sulfated peak in the region of 0.8 M.
The pooled peaks as shown by the bars in Fig. 1 were isolated from extracts of day 8 and 23 samples for digestion with Streptomyces hyaluronidase, which has absolute specificity for hyaluronic acid and produces a mixture of hexa-and tetrasaccharides in limit digests (20).  (Fig. lA, bar) contained 42 pg of hexuronatelmg of hydroxyproline. However, as mentioned above, the material was insensitive to Streptomyces hyaluronidase, indicating that this material did not contain hyaluronic acid.
The labeled hyaluronate-and chondroitin sulfate-containing peaks from the DEAE chromatography of the day 18 samples were analyzed by direct quantitation of disaccharides produced by chondroitin ABC lyase digestion using thin layer chromatography (Fig. 2, C and F ) . Sixty percent of the 3Hlabeled material in the hyaluronate-containing peak comigrated with ADi-HA standard, and 36 and 57% of the 3H material in the chondroitin sulfate-containing peak comigrated with ADi-6s and ADi-4S, respectively.
These results agree well with those obtained by analysis of the same peaks from day 8 and 23 samples by the Streptomyces hyaluronidase and the chondroitin ABC lyase procedure as described above.

FIG. 2. Analysis of pools I and I1 from DEAE-chromatography.
A and E represent the analysis of pool I (Fig. 1) by Streptomyces hyaluronidase digestion and Bio-Gel P-10 elution of degradation products from day 8 samples. D shows the analysis of pool I1 (Fig. 1) by Streptomyces hyaluronidase digestion, and E shows the analysis of the same pool by chondroitin ABC lyase digestion; both panels represent Bio-Gel P-10 analysis of degradation products for day 8 samples. C and F represent analyses of pools I and 11, respectively, by chondroitin ABC lyase digestion followed by separation and quantitation of disaccharides by thin layer chromatography for day 18 samples. The arrows represent the migration positions of standard disaccharides. undigested material eluted in the Vo of the column, while treatment with the enzyme resulted in a 50% shift of the tritiated material to two included peaks. Chromatography of a digest of a separate sample which contained both tritiated material and a hyaluronate standard showed that the two tritiated peaks coeluted with the authentic hyaluronate digestion products. The finding that the same amount of 3H-labeled material (50%) was susceptible to Streptomyces hyaluronidase in the presence of a large excess of hyaluronic acid (70% of the total hexuronate) showed that the amount of enzyme used was not rate-limiting for digestion of the radiolabeled materials. No other peak in the DEAE profiles ( Fig. 1) contained any material susceptible to Streptomyces hyaluronidase. Analysis of day 23 samples gave the same results (see Fig. 3).
The sulfated peaks eluting at 0.8 M NaCl from DEAE chromatography of both extracts (Fig. 1) were identified as predominantly chondroitin sulfate by their susceptibility to chondroitin ABC lyase (>95%) and their resistance to Streptomyces hyaluronidase (Fig. 2, D and E, see also results of disaccharide analyses below). The level of biosynthesis of hyaluronate was determined from the quantitation of the 3H radioactivity in the hyaluronate oligosaccharide peaks (Fig.  2B) and the biosynthesis of chondroitin sulfate (used to quantitate proteoglycan synthesis) determined from the amount of 3H and 35S radioactivity in the 0.8 M peaks (Fig.  20, Control). The total chemical amounts of these components in the tissue were assessed by hexuronate analyses of the same fractions. The 0.12 M peak from the guanidine HC1 Articular Cartilage- Fig. 3, A and B Fig. 3C also shows that there were no significant amounts of unlabeled hyaluronate in the guanidine HC1 extracts.
Catabolism of Hyaluronate-The fact that the amount of hyaluronate remained constant while it was actively synthesized suggests that hyaluronate is actively catabolized. The similar ratios of hyaluronate and proteoglycan in the matrix (1.4-1.8%) and in rates of synthesis (1.1-1.6%) suggests that the metabolism of these two components is correlated. Direct evidence for this was sought in an experiment in which articular cartilage was labeled with [3H]glucosamine and [36S] sulfate as precursors and then chased for a period of 26 days. Labeled hyaluronate and proteoglycan were quantitated in the tissues immediately after biosynthesis and at the end of the experiment by DEAE chromatography and Bio-Gel P-10 analysis. Fig. 4A shows that both labeled components were lost with time from one of the cultures, indicating an active catabolism. The estimated time required for 50% of the radiolabeled components to be released from the tissue is shown   Fig. 4B also shows an independent method for estimating the rate of catabolism of chondroitin sulfate by measuring the amounts of 35S activity released into the culture medium each day. This method bypasses the purification steps since >90% of the [35S]sulfate is incorporated into proteoglycans in these tissues (see Fig. 1). The calculated average tlh of 29 days for replicate cultures is very close to the average for replicates obtained with 3H label (27 days).
In this experiment, the guanidine HCl extraction was followed by collagenase digestion without subsequent homogenization of the tissue (see "Materials and Methods"). Fig. 4, C, D, and E, shows that even though the collagenase digest solubilized approximately one-half of the collagen of the tissue, it did not remove significant amounts of the remaining proteoglycans or of the hyaluronate. The association of hyaluronate with the tissue after collagenase treatment persisted in samples chased for 26 days (not shown). Even though more than 50% of the labeled hyaluronate had been lost from the tissue, the extractability of the molecules left behind was not appreciably altered.
Catabolism of Hyaluronate in Aggregates-Hyaluronate is found in the pericellular environment in isolated chondrocyte cultures (21). Even though this hyaluronate pool has not been shown directly in intact cartilage, its possible existence or the existence of any other hyaluronate pools apart from the pool involved in proteoglycan aggregation could make a significant contribution to the overall kinetics of hyaluronate metabolism in cartilage cultures. For this reason, an experiment was designed to follow the rate of catabolism of hyaluronate in a native proteoglycan aggregate fraction.
Collagenase treatment of 30-pm slices, the most effective method found for the extraction of aggregates from the tissue, was used for the analyses described below. Each digest was directly homogenized in the associative extraction solvent (see "Materials and Methods"), and the clarified samples layered on preformed cesium sulfate density gradients. Kimata et al. (17) previously showed that following centrifugation in these gradients for 6 h, aggregates are recovered in the bottom one-fifth of the gradient, while monomers are well separated and recovered in the upper two-thirds of the gradient. Free hyaluronic acid and link protein are recovered in the upper third of the gradient. It was also shown that separation was independent of initial concentration of proteoglycans between 0.4-8 mg/ml. Fig. 5, A and C, shows the centrifugation profiles for the day 2 and 20 tissue extracts from tissue that was labeled on day 0 and then chased (see "Materials and Methods"). The extraction of 35S-labeled material from the tissue in each case was approximately 60% with 50-60% of this material recovered in the aggregate fraction, tubes 1-6. This is in close agreement with the results of Manicourt et al. (15) using similar procedures to extract native aggregates from rabbit and dog articular cartilage. Thus, the aggregate fraction con-' For estimating catabolic rates of the 3H-labeled components, the 3H-labeled hyaluronate or chondroitin sulfate per pg of hydroxyproline for the day 0 samples was set to 100% and then the time required for 50% loss calculated from the actual loss that occurred after 26 days assuming a first-order rate. The turnover rate of the 3sS-labeled component was calculated by adding the radioactivity released each day into the culture medium (starting 2 days after radiolabeling) and the amount remaining in the tissue after 26 days; the percent remaining in the matrix was then calculated and the tu determined by extrapolation of the curve to 50% loss (see Fig. 4). tains about 35% of the total tissue proteoglycan.
Fractions 1-6 of the cesium sulfate gradients (Fig. 5, A and C) were pooled and subjected to CsCl density gradient centrifugation. Because the aggregates have high density they will equilibrate in the bottom of the CsCl gradients, whereas any particulate membrane or debris which might have sedimented to the bottom of the Cs2S04 gradients will float in the CsCl gradients. The aggregate fractions in the bottom of the CsCl gradients (Fig. 5, B and D ) were recovered, digested with papain and then chondroitin ABC lyase, and analyzed for disaccharides on an HPLC Partisil 5 PAC column (19). Fig.  6 shows representative elution profiles from this column. There was good separation of all the disaccharides including resolution of the hyaluronate and chondroitin disaccharides. Fig. 7A shows that both radiolabeled chondroitin sulfate and hyaluronate are lost from the aggregate fraction during the chase. Fig. 7B shows the calculated t,,+ for [3H]chondroitin sulfate and [3H]hyaluronate: 17 and 20 days, respectively. The former value corresponds closely with that obtained for the total release of 35S-labeled material from the tissue (tIh = 18 days, B ) . The percent of [3H]hyaluronate of the total 3Hlabeled glycosaminoglycans in aggregate was 2.5-3%. Analysis of the rest of the fractions from the velocity gradients showed that they contained chondroitin sulfate, but not hyaluronate. Fig. 6B shows a representative profile, and Table I shows the disaccharide analyses for the various samples.
Analysis for Hyaluronate in the Medium-Attempts were made to detect hyaluronate degradation products in the medium fractions. The media collected from the 18 days of chase were pooled and fractionated on a DEAE-Sephacel column. The bound fraction which would be expected to contain any large to relatively small molecular weight hyaluronate contained only chondroitin sulfate (Table I). Experiments with standard oligosaccharides produced by Streptomyces hyaluronidase digestion of hyaluronate showed that the tetrasaccharide is not bound by the column, while the hexasaccharide is weakly bound (eluted by 0.1 M NaC1). Thus, the unbound fraction from the DEAE chromatography was concentrated  (Fig. 5 8 , bar). B represents the chromatography of the disaccharides from the material in the velocity gradients, fractions 7-16 for the day 2 sample (Fig. 5A Figs. 5 and 6. B shows the total loss of %S activity from these cultures. The arrows indicate the ts calculated for the %Sand 3H-labeled components; the latter were calculated from analysis of A . and precipitated with 3 volumes of ethanol. The hexa-and tetrasaccharides would be soluble in this procedure. The ethanol supernatant was analyzed by Bio-Gel P-10 chromatography. All of the ethanol-soluble material was recovered in the Vt of this column. Thus, there was no evidence for the presence of hyaluronate oligosaccharides in the medium.

DISCUSSION
The results of this study support the novel hypothesis that the metabolism of both hyaluronate and proteoglycans in cartilage is highly coordinated. The data presented support this conclusion in that: 1) newly synthesized hyaluronate and proteoglycans maintain a nearly constant ratio that is very similar to that existing between the total chemical amounts of these two components in the steady-state culture system; The values of percent of hyaluronate disaccharide are in brackets to indicate that there were no distinct peaks in this area of the elution profile (refer to Fig. 6).
bThe daily conditioned media from days 2-18 of culture were pooled and fractionated on DEAE-Sephacel, and the bound peak was dialyzed and treated with chondroitin AEX lyase as indicated under "Materials and Methods." Disaccharides were analyzed by the HPLC procedure ("Materials and Methods," Ref. 2) the catabolism of these two components occurs with virtually identical kinetics; and 3) this catabolic relationship holds for the loss of both hyaluronate and proteoglycans from native, undissociated aggregates isolated from intact tissue. This latter finding argues that the kinetic interrelationship seen in whole tissue is not the fortuitous result of the coexistence of various hyaluronate pools turning over with different kinetics. Furthermore, we have investigated general protein catabolism by chondrocytes in this system and found that the t% of this pool is much faster than for hyaluronate, chondroitin sulfate, or the core protein of proteoglycan.' This indicates that the rate of release of the aggregate components is not simply a reflection of the general rate of protein catabolism in the tissue.
These new data suggest that it is important to introduce the problem of hyaluronate processing into any hypothesis dealing with the mechanism of proteoglycan catabolism. In previous work we found that the release of proteoglycans from the matrix is a highly targeted process in which proteoglycan monomers destined for removal from the matrix are excised without leaving partially truncated proteoglycans behind (7,8). The majority (>go%) of the molecules released into the culture medium are of large hydrodynamic size and are comparable to, or slightly smaller than, newly synthesized molecules (8); <IO% of the proteoglycans released into the medium are completely degraded, presumably by lysosomes, as monitored by the generation of free [3sS]sulfate from catabolized proteoglycan.
It cannot be ascertained at this point which is the decisive or primary event that releases a proteoglycan monomer from its complex association in the tissue, but it is clear that the major portion of the protein core of the proteoglycan escapes extensive processing. Hyaluronate is lost from the tissue with similar kinetics as for proteoglycans, but no detectable large hyaluronic acid molecules or hyaluronic acid oligosaccharide fragments were detected by the methodology used. The fate of the catabolized hyaluronate will be the focus of a separate investigation.
A number of cell types including liver endothelial cells (221, 3T3 fibroblasts (23), and embryonic myocardial cells (24) have been shown to metabolize hyaluronate by receptor- mediated endocytosis followed by extensive intracellular degradation. The possibility that cartilage proteoglycan aggregate catabolism involves a similar process is considered in the following hypothesis. Proteinase(s) activated close to or on the surface of the cell may cleave proteoglycan monomers preferentially in the hyaluronic acid binding region and in the link protein, releasing large fragments from the aggregate complex while the hyaluronate backbone is internalized for intracellular degradation. Dingle and Dingle (25) originally suggested that proteoglycan catabolism occurs in close vicinity to the cell and provided indirect evidence for this by showing that proteoglycans introduced into small cavities in cultured cartilage remain unhydrolyzed while adjacent cartilage undergoes extensive resorption. In addition, the observation of Saklatvala and Bird (26) that cytochalasin B inhibits the rapid proteoglycan losses induced by interleukin 1 suggests that an intact cytoskeleton is vital to the catabolic process and thus implies a cell surface involvement. However, the possible existence of extracellular neutral hyaluronidases, yet unidentified, which contribute to the overall process of hyaluronate resorption cannot be ruled out.
Whatever the exact mechanisms involved in the catabolism of hyaluronic acid, this study clearly shows that this process is an integral part of the mechanism of proteoglycan aggregate metabolism. This might reflect a highly regulated process by which whole aggregates become catabolized, thereby insuring that aggregate structures containing molecules damaged by proteolysis or by wear and tear of the tissue do not accumulate in the matrix.