Hyaluronan Receptor-directed Assembly of Chondrocyte Pericellular Matrix

Initial assembly of extracellular matrix occurs within a zone immediately adjacent to the chondrocyte cell surface termed the cell-associated or periceUular matrix. Assembly within the periceUular matrix compartment requires specific cell-matrix interactions to occur, that are mediated via membrane receptors. The focus of this study is to elucidate the mechanisms of assembly and retention of the cartilage pericellular matrix proteoglycan aggregates important for matrix organization. Assembly of newly synthesized chondrocyte pericellular matrices was inhibited by the addition to hyaluronan hexasaccharides, competitive inhibitors of the binding of hyaluronan to its cell surface receptor. Fully assembled chondrocyte pericellular matrices were displaced using hyaluronan hexasaccharides as well. When exogenous hyaluronan was added to matrixfree chondrocytes in combination with aggrecan, a pericellular matrix equivalent in size to an endogenous matrix formed within 30 min of incubation. Addition of hyaluronan and aggrecan to glutaraldehyde-fixed chondrocytes resulted in matrix assembly comparable to live chondrocytes. These matrices could be inhibited from assembling by the addition of excess hyaluronan hexasaccharides or displaced once assembled by subsequent incubation with hyaluronan hexasaccharides. The results indicate that the aggrecanrich chondrocyte pericellular matrix is not only on a scaffolding of hyaluronan, but actually anchored to the cell surface via the interaction between hyaluronan and hyaluronan receptors. T am initial assembly of the chondrocyte extracellular matrix most likely occurs near the cell surface within a zone termed the cell-associated or pericellular matrix. How this assembly is regulated by the chondrocytes and how the matrix itself is anchored to the cell surface is largely unknown. Hyaluronan, a high molecular weight matrix polysaccharide, has a central role in the organization of the extracellular matrix of cartilage as the backbone of the cartilage proteoglycan aggregate (Hascall and Heinegard, 1974; Christner et al., 1977). In the pericellular matrix of chondrocytes where it is an essential matrix component, hyaiuronan may not only serve a structural role, but other functions as well. Chondrocytes in culture exhibit large pericellular matrices extending from the plasma membrane which can be readily visualized by a particle exclusion assay (Clarris and Fraser, 1968; Goldberg and Toole, 1984; Knudson and Toole, 1985). Small particles when added to low density cultures settle onto the culture dish, and are excluded from a distinct zone surrounding the chondrocytes. This zone or "coat" defines the chondrocyte pericellular matrix. A major component of this matrix is the large, chondroitin sulfate-rich aggregating proteoglycan, termed '~ggrecan" (Doege et al., 1991). Treatment of chondrocytes with a small amount of Streptomyces hyaluronidase, which specifically degrades hyaluronan (Ohya and Kaneko, 1970; Harrison et al., 1986), removes this pericellular matrix (Goldberg and Toole, 1984; Knudson and Toole, 1985; McCarthy and Toole, 1989) suggesting that the majority of the aggrecan is in the form of aggregates with hyaluronan. The focus of this study is the mechanism of retention of these proteoglycan aggregates in order to elucidate the interactions important to establish and maintain chondrocyte pericellular matrix organization. Many cells, including chondrocytes (Knudson and Toole, 1987; McCarthy and Toole, 1989; Toole et al., 1989; Yu et al., 1992), have specific cell surface binding proteins, or receptors, for hyaluronan. Hyaluronan receptors present on chondrocytes have properties similar to hyaluronan receptors reported on other cell types such as SV-40 transformed 3T3, BHK, and human bladder carcinoma cells (Underhill and Toole, 1979; Underhill, 1989; Nemec et ai., 1987). These hyaluronan receptors are a family of non-integrin, hydrophobic membrane proteins, termed the hyaladherins (Toole, 1991). Studies have also suggested that hyaluronan receptors are related or identical to the CD44 family of lymphocyte homing receptors (Aruffo et al., 1990; Culty et al., 1990). The hyaiuronan receptors are grouped together as a family based on their similar physical and functional properties which include: (a) a high binding affinity for hyaluronan (Kd ------10 -9 M); (b) a high degree of specificity for hyaluronan (when receptor is assayed on non-extracted, intact membranes); (c) binding affinity for hyaluronan that increases 9 The Rockefeller University Press, 0021-9525/93/02/825/10 $2.00 The Journal of Cell Biology, Volume 120, Number 3, February 1993 825-834 825 on July 3, 2017 jcb.rress.org D ow nladed fom with increase in ionic strength; (d) binding that is stable to mild fixation of the receptor with glutaraldehyde; and lastly (e) binding that can be competed by hyaluronan oligosaccharides with a minimum size of six monosaccharides. These and other properties help to distinguish this family of hyaluronan binding proteins/receptors from other hyaluronan binding proteins including aggrecan and link protein. One important property shared by all the CD44-1ike cell surface hyaluronan receptors is the minimum size of hyaluronan oligosaccharide required to effectively compete for the binding of native hyaluronan to its receptor, which is a hyaluronan hexasaccharide (HA0 ~. This property is helpful in differentiating the specific binding of hyaluronan to the receptor from the aggregation of hyaluronan with other matrix macromolecules, in particular aggrecan and link protein, which require a minimum sequence of 10-12 monosaccharides for competition (Hascall and Heinegard, 1974; Christner et al., 1977; Hardingham et al., 1992). Incubation of cells with HA~ has been shown to displace pre-bound 3H-hyaluronan from cell surfaces (Nemec et al., 1987), to compete with binding when added together with 3H-hyaluronan to intact ceils (Underhill and Toole, 1979) or isolated cell membranes (Underhill et al,, 1983), and in this study to inhibit assembly of newly synthesized chondrocyte pericellular matrices. A large proportion of endogenous hyaluronan can be displaced from the cell surface by exogenous hyaluronan or HA~; proteoglycans are also displaced along with the hyaluronan (Knudson and Toole, 1987; Knudson, C. B,, L. J. Coombs, and K. E. Kuettner, 1991. Ortho Trans. 15:467-468; Hua, Q., C. B. Knudson, and W. Knudson, 1992. Trans. Orthop. Res. Soc. 17:29). As shown in this paper, HA~ can displace fully assembled chondrocyte pericellular matrices. This study characterizes the roles of hyaluronan, aggrecan, and cell surface hyaluronan receptors in the assembly, organization, and maintenance of the chondrocyte pericellular matrix. Two cell types are used as model systems for study; embryonic chick tibial chondrocytes (Kim and Conrad, 1977; Knudson and Toole, 1985, 1987) and adult rat chondrocytes, derived from the well-characterized Swarm rat chondrosarcoma (Kimura et al., 1979; Sun et al., 1986). The results indicate that the aggrecan-rich chondrocyte pericellular matrix is anchored to the cell surface via the interaction between hyaluronan and hyaluronan receptors, Therefore, the hyaluronan receptors may direct the assembly of the chondrocyte pericellular matrix. Materials and Methods All reagents, unless specified otherwise, were purchased from Sigma Chemical Co. (St. Louis, MO). Tissue culture samples were from Gibco BRL (Grand Island, NY). Chondrocytes and Culture Conditions Embryonic chick chondrocytes were released from stage 38 tibiae (Hamburger and Hamilton, 1951), zone II cartilage segments (Kim and Conrad, 1977), by trypsin (type II)/collagenase (type II; Wbrthington, Freehold, NJ) treatment (Knudson and Toole, 1985). Rat chondrosareoma chondrocytes from long-term continuous cell cultures were obtained from Dr. James H. 1. Abbreviations used in this paper: CMF-PBS, Calciumand magnesiumfree PBS; HAs, hyaluronan hexasaecharides; polyhema, poly-2-hydroxyethylmethacrylate. Kimura (Henry Ford Hospital, Detroit, MI). The two chondrocyte types were cultured in DME (4.5 g/l glucose) containing 10% FBS (Hyclone, Logan, UT), 1% penicillin/streptomycin solution, 50 #g/ml ascorbic acid and 2 mM glntamine, at 37~ in 5% CO2/95% air. Chondrocytes were also grown in suspension cultures over poly-2-hydroxyethylmcthacrylate coating (polyhema; Aldrich, Milwaukee, WI) in complete medium. To coat the surface, 35-mm tissue culture dishes containing 2 ml of 0.06% polyhcma in ethanol were incubated to dryness at 37~ Particle Exclusion Assay The particle exclusion assay followed a protocol described previously for cells in monolayer culture (Knudson and Toole, 1985). The culture medium was removed and replaced with 750 #l of a suspension of formaldehydefixed RBCs (108/ml) in PBS/0.1% BSA. The particles settled in 10 rain, and the cells were then observed by phase-contrast microscopy. To visualize the pericellular matrix assembled by chondrocytes in suspension culture, cells were transferred to 6-well flat-bottom tissue culture plates and spun at 500 g for 15 rain in an Omnifuge microtiter plate holder before the addition of particles. Morphometric Analysis of Matrix to Cell Area A matrix/cell ratio, defined as the ratio of the area delimited by the perimeter of the pericellular matrix to the area delimited by the plasma membrane, was determined by tracing the matrix and cell perimeters on a digitizing tablet using Sigma Scan software (Jandel Scientific, Corte Madera, CA; Knudson and Toole, 1985). The matrix/cell ratio was ,x,l.0 if no detectable periccllular matrix was present. The values given represent the mean matrix/cell ratio and the 95 % confidence range from the mean. Experimental groups were compared with regard to control by a t test. Enzymatic Removal of PericeUular Matrix For studies on assembly of newly synthesized pericellular matrix, the endogenous pericellular matrices were removed by adding 20 ~1 of a 100 U/ml solution of Streptomyces hyaluronidase (type IX) directly to 24-h chondrocyte cultures, containing complete culture medium with 10% FBS, and incubating the cells for 1 h at 37~ followed by HBSS washes. After this treatment, the ceils were "matrix-free" by the particle exclusion assay. Following Streptomyces hyaluronidase treatment, some cells were fixed in 1% glutaraldehyde (EMS, Fort Washington, PA) in PBS, for 5 rain, and then rinsed with PBS/I% BSA. Matrix Assembly and Displacement Studies The influence of HA6 or high molecular weight hyaluronan on matrix assembly was studied. HAs can compete for the binding of hyaluronan to its receptor but cannot support proteoglycan aggregate formation. Chondrocytes were made matrix-free with Streptomyces hyaluronidase. 1Ceassembly of the endogenous pericellular matrix was monitored in complete medium or medium containing hyaluronan (250/~g/mi), or HA6 (25-125 #g/ml). The ability of exogenous hyaluronan and HAS to displace the endogenous matrix made by chondrocytes in suspension culture over polyhema, or chondrocytes grown as attached ceils in monolayer, was also determined. Cells grown under these culture conditions for 2 to 48 h were incubated for 2 to 3 h more in the presence or absence of hyaluronan or HA6 before the particle exclusion assay. Preparation and Addition of Exogenous Macromolecules To test the ability of chondrocytes to assemble a pericellular matrix with exogenous macromolecules, hyaluronan and aggrecan were added to matrix-free live or glutaraldchydc-fixed cells. Initially, aggrecan derived from several sources was tested; monomer preparations from the Swarm rat chondrosarcoma were gifts from Drs. J. Kimura and 1". Giant (Henry Ford Hospital, Detroit, MI and Rush-Presby~rian-St. Luke's Medical Center, Chicago, IL, respectively), and, proteoglycan monomers derived from adult bovine articular or nasal cartilages were gifts from Dr. E. Thonar (RushPresbyterian-St. Luke's Medical Center). Subsequently, proteoglycan was extracted from rat chondrosarcoma tumor homogenate according to Faltz et al. (1979) and isolated by dissociative cesium chloride equilibrium centrifugation in 4.0 M guauldine HC1 with protease inhibitors at a starting density of 1.5 g/ml (Hascall vtal., 1976) for 50 h at 100,000 g at 10~ The Journal of Cell Biology, Volume 120, 1993 826 on July 3, 2017 jcb.rress.org D ow nladed fom The bottom 1/4th of the gradient, density >1.6 g/ml, was collected and recentrifuged. The bottom 1/4th of the second gradient was collected (DID1 fractions), dialyzed, and lyophilized. The D1DI fractions were incubated with Streptomyces hyaluronidase to degrade small concentrations of hyaluronan found within these preparations. The proteoglycans were recovered by another dissociative equilibrium centrifugation. HA6 were prepared by digestion of hyaluronan (grade I) with testicular hyaluronidase (type I-S) with 172 USP/NFU hyaluronidase/mg hyaluronan (Knudson et al., 1984; Kimura et al., 1979; Hascall and Heinegard, 1974). The hyaluronan oligosaccharides were separated on a 2.5 x 118 cm column of Bit-gel P30 in 0.5 M pyridinium acetate (Solursh et al., 1980; Knudson and Knudson, 1991). In parallel experiments, the glycosaminoglycan chondroitin sulfate was added to chondrocytes cultures. Chondroitin sulfate (grade III) was pretreated with Streptomyces hyaluronidase, boiled, precipitated in 1.3% potassium acetate in 95% ethanol. Chondroitin sulfate hexasaccharides were prepared by testicular hyaluronidase digestion, purified by Bit-gel P-30 chromatography as above. Additions to matrix-free chondrocytes were DME containing either: (a) no additional components; (b) 2.0 mg/ml proteoglycan monomer; (c) 12 ~g/ml hyaluronan; (d) 2.0 mg/ml proteoglycan monomer plus 12 /~g/ml hyaluronan (determined empirically as optimal, data not shown); or (e) 2.0 mg/ml proteoglycan monomer, 12 t~g/ml hyaluronan, and 100/~g/ml HA6. The hyaluronic acid binding region of D1D1 proteoglycan was separated from the chondroitin sulfate-rich fragment by dissociative cesium chloride gradient centrifugation as above, after clostripain digestion (Caputo et al., 1980). Reduction/alkylation of proteoglycan was by treatment with 20 mM DTT followed by addition of iodoacetamide (Hardingham et al., 1976). The products of chondroitinase ABC digestion (Hascall and Heinegard, 1974), were separated by Sephacryl S-300 chromatography in PBS. The proteoglycan core protein, containing chondroitin-4-sulfate stubs, was detected using 9-A-2/E9 antibody (ICN Biomedicals, Inc., Irvine, CA). Residual chondroitinase activity was detected in later eluting fractions by 3H-hyaluronan degradation assays (Knudson and Toole, 1987). Modified proteoglycan was lyophilized, dissolved in DME at 10 ng/ml, and added to matrix-free chondrocytes at the same concentration as the native proteoglycan. The ability of aggrecan to associate with chondrocytes in monolayer culture was tested by adding 3H-aggrecan to matrix-free chondrocytes or to chondrocytes with an endogenous pericellular matrix. The 3H-aggrecan (Byers et al., 1987) was added at 1 mg/ml (62,000 cpm/mg) serum-free DME per 105 cells for 2 h at 37~ Cells were then rinsed three times with serum-free DME; the radioactivity was released from the culture dishes with 1% SDS. Proteoglycan Biosynthesis in the Presence of Hyaluronan Hexasaccharides Matrix-free chondrocytes were incubated with 25 #Ci/ml of [35S]sulfuric acid (43 Ci/mg; ICN Biomedicals) in the presence or absence of 250 t~g/ml HA6. The media fractions were collected at 24 and 48 h and analyzed by descending chromatography (Knudson et al., 1984). Chondroitinase ABCsensitive radioactivity at the origin of the strip represented incorporation into chondroitin sulfate proteoglycan. Samples were also run on a Sepharose CL-2B column in 0.5 M sodium acetate, 0.1% Triton X-100, pH 6.8, with and without pre-aggregation with hyaluronan for 24 h (Sandy et al., 1989). The percentage of macromolecular [35S]sulfate-labeled material eluting at the Vo of the column represents the capacity of the radiolabeled proteoglycan to form stable aggregates with hyaluronan.