Journnl of Orrhopaedic Research 10276-284 Raven Press, Ltd., New York 0 1992 Orthopaedic Research Society
Influence of Continuous Infusion of Interleukin- la on the Core Protein and the Core Protein Fragments of the Small Proteoglycan Decorin in Cartilage P. Witsch-Prehm, *A. Karbowski, B. Ober, and H. Kresse Institute of Physiological Chemistry and Puthobiochemistq; und *Orthopedic University Hospital, Wniversio of Miinster, Munster, Germuny
Summary: Decorin, a collagen-binding small proteoglycan, is considered to have a specific function in the organization or stability of the collagen network. Therefore, alteration of its molecular properties may be of pathophysiological relevance during the development of cartilage damage. It is shown here that normal cartilage from rabbit knee-joint contains glycosaminoglycan chainbearing core protein fragments of 39, 23, and 18 kDa, each one amounting to - 5 4 % of the intact decorin core protein. Continuous infusion of human recombinant interleukin-la for 14 days (200 ngiday) into a knee-joint led in condylar cartilage to a reduction in the amount of intact core protein from 2 pg/mg wet tissue to about 1.1 pg/mg. The increase in its quantity found after infusion of heat-inactivated interleukin- 1 was not statistically significant. The concentration of all three core protein fragments became reduced to a similar extent as the intact core protein under the influence of the cytokine, and additional fragments were not found. Surprisingly, there was a much smaller response to interleukin-l-treatment in patellar cartilage. Key Words: Small proteoglycan II-Decorin-Interleukinl-Core protein-Proteolysis.
or small proteoglycan I1 is composed of a core protein of 36.5 kDa, a single glycosaminoglycan chain and either two or three asparagine-bound oligosaccharides (10,18). A third small proteoglycan, tentatively designated proteoglycan- 100, according to the size of the core protein ( 5 ) , has been detected in human cartilage, too (P. Witsch-Prehm and H. Kresse, unpublished observations). In fetal cartilage, the major small proteoglycan to be synthesized was biglycan. In the adult, however, biglycan synthesis was barely detectable relative to decorin production (2 1j . The glycosaminoglycan composition of these small proteoglycans may vary with age and with the tissue of origin (30). The functions of biglycan and of proteoglycan100 are not yet known. However, there are some clear indications of the function of decorin. It binds
Articular cartilage contains aggregating and nonaggregating proteoglycans. The large aggregating cartilage proteoglycan, aggrecan, confers upon the tissue the property of compressive resistance to applied load (for a review, see refs. 13,15,16). Two types of small chondroitinidermatan sulfate proteoglycans have additionally been identified in cartilage (27j, the sequence of their core proteins has been elucidated by molecular cloning (9,18). Biglycan or small proteoglycan I is characterized by a core protein of 38 kDa and the presence of either one or two glycosaminoglycan chains (6,9). Decorin Received October 5 , 1990; accepted July 24, 1991. Address correspondence and reprint requests to P. WitschPrehm at Institute of Physiological Chemistry and Pathobiochemistry, University of Miinster, Waldeyerstr. 15, D-4400 Munster, Germany.
2 76
277
INTERLEUKIN-la EFFECT ON DECORIN
with its core protein (35,40) to types I and I1 collagen (39), and can be found orthogonally arrayed at the d band of collagen fibrils (34). This binding may be functionally equivalent to the incorporation of collagen type TX proteoglycan into mixed fibrils of types 11, IX, and XI collagens in cartilage (22). Therefore, both decorin and collagen type IX proteoglycan may be important in the regulation of fibrillogenesis and the organization of the collagen network in vivo. Cartilage destruction is a common feature of rheumatoid arthritis and osteoarthritis. There is ample evidence that interleukin- lkatabolin is a key mediator by which synoviocytes as well as chondrocytes enhance the production of collagen- and proteoglycan-degrading proteases (17,23-26,29,42). Interleukin-1 has also been shown to inhibit the expression of aggrecan and cartilage-specific collagens (3,11,38). In the present study, special attention is given on the effect of interleukin-1 on the content of the collagen-binding small proteoglycan, decorin, and its core protein fragments in articular cartilage. Furthermore, since interleukin- l was cleared very rapidly (half-life, 0.38-0.48 h) upon injection into the synovial cavity (7), the cytokine was continuously infused at a dose of 200 ng/day into the rabbit knee joint (8). A period of 14 days of treatment was chosen because of the development of an arthritis with features of both an acute and chronic inflammation during that time although the time course of the development of the alterations had not been systematically studied. MATERIALS AND METHODS Human recombinant interleukin-la (IL-1) was obtained from Biogen, Geneva, Switzerland; [35S]methionine (sp. radioactivity 5 1 TBqimmol) and sodium ["Slsulfate (carrier-free) from Amersham-Buchler, Braunschweig, Germany; DEAETrisacryl from Pharmacia LKB GmbH, Freiburg, Germany; chondroitin ABC lyase (EC 4.2.2.4) from Seikagaku Kogyo, Tokyo, Japan; nitrocellulose membranes BA 83 (0.2 pm) from Schleicher and Schuell, Dassel, Germany; and antirabbit IgG goat antibodies, conjugated with peroxidase from BioRad, Munich, Germany. Decorin was prepared from the secretions of cultured human skin fibroblasts exactly as described (14). A rabbit antiserum against its core protein was obtained (10) and affinity purified (41) as described
previously. Briefly, the antiserum was first passed through a column of Sepharose 4B covalently linked with decorin core protein. Bound antibodies were desorbed with 7 M urea, dialyzed against 10 mM sodium phosphate buffer, pH 7.4, containing 0.15 M NaC1, and applied to a second affinity matrix containing a neoglycoprotein made from bovine serum albumin and unsaturated hexeneuronosylGalNAc-4i6-sulfate disaccharides. Antibodies not retarded by this column were specific for epitopes on the decorin core protein. Cross-reactivity with the structurally related biglycan core protein (9) was not observed. Immune staining of sections of human cartilage indicated that the antiserum did not react with the chondroitin ABC lyase-digested large proteoglycan, aggrecan (41). Animal Preparation The implantation of Alzet osmotic minipumps into the right knee joint of 10-week-old white New Zealand rabbits has been described previously (8). The animals received 200 ng/day of active or heatinactivated (XO'C, 20 min) human recombinant interleukin-la (12 pliday) over a period of 14 days. The cytokine had been dissolved in 5 mM sodium phosphate buffer, pH 7.5, containing 0.15 M NaCl and 5% rabbit serum to avoid unspecific adsorption. In all cases, 2-3 days after surgery the rabbits used the right leg normally again. At the end of the infusion period, the animals were killed with embutramide. Cartilage samples were removed from medial and lateral condyles, from the patella, and the tibia of both the treated and untreated knee joints and stored frozen at -70°C for up to 6 months. Care was exercised to exclude subchondral bone from cartilage specimens. Proteoglycan and Core Protein Preparation Five milligrams of wet tissue were carefully minced and extracted for 16 h at 4°C under constant mixing with 500 pl of 50 mM sodium acetate buffer, pH 6.0, containing 4 M guanidine hydrochloride, 0.1 M 6-aminohexanoic acid, 10 mM EDTA, 10 mM N-ethylmaleimide and 5 mM benzamidine hydrochloride. After centrifugation (10 min, 10,000 g) the supernatant was dialyzed against 0.15 M NaCl in buffer A (20 mM Tris/HCl buffer, pH 7.4, containing 0.1% Triton X-100 and the protease inhibitors mentioned above). The solution was then loaded onto a 1 ml column of DEAE-Trisacryl equilibrated
J Orthop ReAj Vol. 10, N o . 2 , 1992
2 78
P. WITSCH-PREHM ET AL.
with the same buffer. The column was developed with 3 ml of buffer A containing 0.15 M NaCl, 3 ml of buffer A containing 0.3 M NaC1, and 2 ml of buffer A containing 1.O M NaCl. The last fraction was dialyzed against water, brought to dryness in a Speed Vac concentrator (Bachofen, Reutlingen, Germany) and incubated with 10 mU chondroitin ABC lyase and protease inhibitors or with buffer alone for 2 h at 37°C (28). The samples were dried again, and salts were removed by washing with methanol. Decorin Core Protein Analysis Chondroitin ABC lyase-treated and undigested proteoglycans were subjected to SDSipolyacrylamide gel electrophoresis at a total acrylamide concentration of 12.5% (w/v) in the separation gel (19). Decorin core protein from human skin fibroblasts (0.14-1.1 pg corresponding to 0.2-1.58 nmol of hexuronic acid in the intact proteoglycan [10,18]) was applied simultaneously to each gel for calibration purposes. Electrophoresis was followed by Western blotting (36). The nitrocellulose membrane was saturated with 3% bovine serum albumin in buffer B (20 mM Tris/HCl, pH 7.5, containing 0.15 M NaCl) for 16 h at 4°C before it was treated for 2 h at 2022°C with a 1:SOO dilution of decorin core protein antibodies in buffer Bi0.196 bovine serum albumin. The blot was washed five times, 5 min each, with buffer B and was then exposed for 2 h at 20-22°C to a 1 :1,000 dilution (with buffer BiO. 1% bovine serum albumin) of the peroxidase-conjugated second antibody. After five washing steps with buffer B, two reference lanes were cut off, and the core proteins were visualized by incubation for up to 30 min at ambient temperature with 3 mM 4-chloro-1-naphtol in buffer B containing 0.003% H,O, and 17% (v/v) methanol. The R, values of the upper and lower margins of the decorin core protein bands and of the bands of lower molecular weight fragments were determined, and corresponding areas from the unstained sample lanes of the blot were cut out, transferred to Eppendorf tubes and developed with a soluble chromogen by adding 500 pl of 1 mM 2,2'azino-bis(3-ethyl-benzthiazoline-6-sulfonicacid) in 50 mM sodium citrate buffer, pH 4.0, containing 0.003% H,O,. The reaction was stopped by adding 2 mM NaN,. A 250 pl aliquot of the solution was then transferred to a microtiter plate, and the absorbance at 410 nm was recorded by a Dynatech MR 710 microplate reader.
J Orthop Res, Voi. 10, N o . 2,1992
0.4
0.2
0.5
1.o
Amount of decorin core protein ( p g ) FIG. 1. Quantification of decorin core protein after Western blotting. The indicated amounts of decorin core protein prepared from the secretions of human fibroblasts were subjected to polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate followed by Western blotting as described in the Materials and Methods section.
From the calibration curve shown in Fig. 1, it is evident that there is a linear relationship between color development and an amount of decorin core protein between 0.1 and 1.2 kg. For the analysis of decorin core protein from cartilage, the values fell into the linear range when the extracts of 0.25 mg of tissue were applied. For an analysis of core protein fragments extracts from 2.5 mg of tissue were required. It had been ascertained that there was quantitative transfer of proteins with M, values
Influence of Continuous Infusion of Interleukin- la on the Core Protein and the Core Protein Fragments of the Small Proteoglycan Decorin in Cartilage P. Witsch-Prehm, *A. Karbowski, B. Ober, and H. Kresse Institute of Physiological Chemistry and Puthobiochemistq; und *Orthopedic University Hospital, Wniversio of Miinster, Munster, Germuny
Summary: Decorin, a collagen-binding small proteoglycan, is considered to have a specific function in the organization or stability of the collagen network. Therefore, alteration of its molecular properties may be of pathophysiological relevance during the development of cartilage damage. It is shown here that normal cartilage from rabbit knee-joint contains glycosaminoglycan chainbearing core protein fragments of 39, 23, and 18 kDa, each one amounting to - 5 4 % of the intact decorin core protein. Continuous infusion of human recombinant interleukin-la for 14 days (200 ngiday) into a knee-joint led in condylar cartilage to a reduction in the amount of intact core protein from 2 pg/mg wet tissue to about 1.1 pg/mg. The increase in its quantity found after infusion of heat-inactivated interleukin- 1 was not statistically significant. The concentration of all three core protein fragments became reduced to a similar extent as the intact core protein under the influence of the cytokine, and additional fragments were not found. Surprisingly, there was a much smaller response to interleukin-l-treatment in patellar cartilage. Key Words: Small proteoglycan II-Decorin-Interleukinl-Core protein-Proteolysis.
or small proteoglycan I1 is composed of a core protein of 36.5 kDa, a single glycosaminoglycan chain and either two or three asparagine-bound oligosaccharides (10,18). A third small proteoglycan, tentatively designated proteoglycan- 100, according to the size of the core protein ( 5 ) , has been detected in human cartilage, too (P. Witsch-Prehm and H. Kresse, unpublished observations). In fetal cartilage, the major small proteoglycan to be synthesized was biglycan. In the adult, however, biglycan synthesis was barely detectable relative to decorin production (2 1j . The glycosaminoglycan composition of these small proteoglycans may vary with age and with the tissue of origin (30). The functions of biglycan and of proteoglycan100 are not yet known. However, there are some clear indications of the function of decorin. It binds
Articular cartilage contains aggregating and nonaggregating proteoglycans. The large aggregating cartilage proteoglycan, aggrecan, confers upon the tissue the property of compressive resistance to applied load (for a review, see refs. 13,15,16). Two types of small chondroitinidermatan sulfate proteoglycans have additionally been identified in cartilage (27j, the sequence of their core proteins has been elucidated by molecular cloning (9,18). Biglycan or small proteoglycan I is characterized by a core protein of 38 kDa and the presence of either one or two glycosaminoglycan chains (6,9). Decorin Received October 5 , 1990; accepted July 24, 1991. Address correspondence and reprint requests to P. WitschPrehm at Institute of Physiological Chemistry and Pathobiochemistry, University of Miinster, Waldeyerstr. 15, D-4400 Munster, Germany.
2 76
277
INTERLEUKIN-la EFFECT ON DECORIN
with its core protein (35,40) to types I and I1 collagen (39), and can be found orthogonally arrayed at the d band of collagen fibrils (34). This binding may be functionally equivalent to the incorporation of collagen type TX proteoglycan into mixed fibrils of types 11, IX, and XI collagens in cartilage (22). Therefore, both decorin and collagen type IX proteoglycan may be important in the regulation of fibrillogenesis and the organization of the collagen network in vivo. Cartilage destruction is a common feature of rheumatoid arthritis and osteoarthritis. There is ample evidence that interleukin- lkatabolin is a key mediator by which synoviocytes as well as chondrocytes enhance the production of collagen- and proteoglycan-degrading proteases (17,23-26,29,42). Interleukin-1 has also been shown to inhibit the expression of aggrecan and cartilage-specific collagens (3,11,38). In the present study, special attention is given on the effect of interleukin-1 on the content of the collagen-binding small proteoglycan, decorin, and its core protein fragments in articular cartilage. Furthermore, since interleukin- l was cleared very rapidly (half-life, 0.38-0.48 h) upon injection into the synovial cavity (7), the cytokine was continuously infused at a dose of 200 ng/day into the rabbit knee joint (8). A period of 14 days of treatment was chosen because of the development of an arthritis with features of both an acute and chronic inflammation during that time although the time course of the development of the alterations had not been systematically studied. MATERIALS AND METHODS Human recombinant interleukin-la (IL-1) was obtained from Biogen, Geneva, Switzerland; [35S]methionine (sp. radioactivity 5 1 TBqimmol) and sodium ["Slsulfate (carrier-free) from Amersham-Buchler, Braunschweig, Germany; DEAETrisacryl from Pharmacia LKB GmbH, Freiburg, Germany; chondroitin ABC lyase (EC 4.2.2.4) from Seikagaku Kogyo, Tokyo, Japan; nitrocellulose membranes BA 83 (0.2 pm) from Schleicher and Schuell, Dassel, Germany; and antirabbit IgG goat antibodies, conjugated with peroxidase from BioRad, Munich, Germany. Decorin was prepared from the secretions of cultured human skin fibroblasts exactly as described (14). A rabbit antiserum against its core protein was obtained (10) and affinity purified (41) as described
previously. Briefly, the antiserum was first passed through a column of Sepharose 4B covalently linked with decorin core protein. Bound antibodies were desorbed with 7 M urea, dialyzed against 10 mM sodium phosphate buffer, pH 7.4, containing 0.15 M NaC1, and applied to a second affinity matrix containing a neoglycoprotein made from bovine serum albumin and unsaturated hexeneuronosylGalNAc-4i6-sulfate disaccharides. Antibodies not retarded by this column were specific for epitopes on the decorin core protein. Cross-reactivity with the structurally related biglycan core protein (9) was not observed. Immune staining of sections of human cartilage indicated that the antiserum did not react with the chondroitin ABC lyase-digested large proteoglycan, aggrecan (41). Animal Preparation The implantation of Alzet osmotic minipumps into the right knee joint of 10-week-old white New Zealand rabbits has been described previously (8). The animals received 200 ng/day of active or heatinactivated (XO'C, 20 min) human recombinant interleukin-la (12 pliday) over a period of 14 days. The cytokine had been dissolved in 5 mM sodium phosphate buffer, pH 7.5, containing 0.15 M NaCl and 5% rabbit serum to avoid unspecific adsorption. In all cases, 2-3 days after surgery the rabbits used the right leg normally again. At the end of the infusion period, the animals were killed with embutramide. Cartilage samples were removed from medial and lateral condyles, from the patella, and the tibia of both the treated and untreated knee joints and stored frozen at -70°C for up to 6 months. Care was exercised to exclude subchondral bone from cartilage specimens. Proteoglycan and Core Protein Preparation Five milligrams of wet tissue were carefully minced and extracted for 16 h at 4°C under constant mixing with 500 pl of 50 mM sodium acetate buffer, pH 6.0, containing 4 M guanidine hydrochloride, 0.1 M 6-aminohexanoic acid, 10 mM EDTA, 10 mM N-ethylmaleimide and 5 mM benzamidine hydrochloride. After centrifugation (10 min, 10,000 g) the supernatant was dialyzed against 0.15 M NaCl in buffer A (20 mM Tris/HCl buffer, pH 7.4, containing 0.1% Triton X-100 and the protease inhibitors mentioned above). The solution was then loaded onto a 1 ml column of DEAE-Trisacryl equilibrated
J Orthop ReAj Vol. 10, N o . 2 , 1992
2 78
P. WITSCH-PREHM ET AL.
with the same buffer. The column was developed with 3 ml of buffer A containing 0.15 M NaCl, 3 ml of buffer A containing 0.3 M NaC1, and 2 ml of buffer A containing 1.O M NaCl. The last fraction was dialyzed against water, brought to dryness in a Speed Vac concentrator (Bachofen, Reutlingen, Germany) and incubated with 10 mU chondroitin ABC lyase and protease inhibitors or with buffer alone for 2 h at 37°C (28). The samples were dried again, and salts were removed by washing with methanol. Decorin Core Protein Analysis Chondroitin ABC lyase-treated and undigested proteoglycans were subjected to SDSipolyacrylamide gel electrophoresis at a total acrylamide concentration of 12.5% (w/v) in the separation gel (19). Decorin core protein from human skin fibroblasts (0.14-1.1 pg corresponding to 0.2-1.58 nmol of hexuronic acid in the intact proteoglycan [10,18]) was applied simultaneously to each gel for calibration purposes. Electrophoresis was followed by Western blotting (36). The nitrocellulose membrane was saturated with 3% bovine serum albumin in buffer B (20 mM Tris/HCl, pH 7.5, containing 0.15 M NaCl) for 16 h at 4°C before it was treated for 2 h at 2022°C with a 1:SOO dilution of decorin core protein antibodies in buffer Bi0.196 bovine serum albumin. The blot was washed five times, 5 min each, with buffer B and was then exposed for 2 h at 20-22°C to a 1 :1,000 dilution (with buffer BiO. 1% bovine serum albumin) of the peroxidase-conjugated second antibody. After five washing steps with buffer B, two reference lanes were cut off, and the core proteins were visualized by incubation for up to 30 min at ambient temperature with 3 mM 4-chloro-1-naphtol in buffer B containing 0.003% H,O, and 17% (v/v) methanol. The R, values of the upper and lower margins of the decorin core protein bands and of the bands of lower molecular weight fragments were determined, and corresponding areas from the unstained sample lanes of the blot were cut out, transferred to Eppendorf tubes and developed with a soluble chromogen by adding 500 pl of 1 mM 2,2'azino-bis(3-ethyl-benzthiazoline-6-sulfonicacid) in 50 mM sodium citrate buffer, pH 4.0, containing 0.003% H,O,. The reaction was stopped by adding 2 mM NaN,. A 250 pl aliquot of the solution was then transferred to a microtiter plate, and the absorbance at 410 nm was recorded by a Dynatech MR 710 microplate reader.
J Orthop Res, Voi. 10, N o . 2,1992
0.4
0.2
0.5
1.o
Amount of decorin core protein ( p g ) FIG. 1. Quantification of decorin core protein after Western blotting. The indicated amounts of decorin core protein prepared from the secretions of human fibroblasts were subjected to polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate followed by Western blotting as described in the Materials and Methods section.
From the calibration curve shown in Fig. 1, it is evident that there is a linear relationship between color development and an amount of decorin core protein between 0.1 and 1.2 kg. For the analysis of decorin core protein from cartilage, the values fell into the linear range when the extracts of 0.25 mg of tissue were applied. For an analysis of core protein fragments extracts from 2.5 mg of tissue were required. It had been ascertained that there was quantitative transfer of proteins with M, values
