Journal of Orthopaedic Research 9 4 8 5 4 9 4 Raven h e s s , Ltd., New York 0 1991 Orthopaedic Research Society
Localization of Type X Collagen in Canine Growth Plate and Adult Canine Articular Cartilage *James M. Gannon, *Gordon Walker, *Mark Fischer, *Randy Carpenter, *Roby C. Thompson, Jr., and *tTheodore R. Oegema, Jr. Departments of *Orthopaedic Surgery and TBiochemistry, University of Minnesota, Minneapolis, Minnesota, U.S.A.
Summary: Type X collagen was extracted from ends of canine growth plates by pepsin digestion after 4 M guanidine hydrochloride extraction, purified by stepwise salt precipitation (2.0 M NaCl in 0.5 M acetic acid), and chromatographed on a Bio-Gel A1.5 M column in 1.0 M CaCl,. Without reduction on sodium dodecyl sulfate (SDS) polyacrylamide gels, the preparation yielded a single, high-molecular-weight (mol wt) band; after reduction, a single band of relative mol wt 5.0 x lo4 was found. Polyclonal sera were raised against the purified collagen and used in the immunolocalization of canine type X collagen. As expected, indirect immunoperoxidase (IP) or indirect immunofluorescent staining with the polyclonal sera demonstrated that most of the immunoreactivity was localized in the zone of provisional calcification of the growth plate and in cartilage remnants in the metaphyseal region of the physis. A progressive decrease in staining toward the diaphysis of the fetal canine long bone was apparent as the trabecular structures were remodeled to bone. Unexpectedly, type X collagen was also detected in the zone of calcified, mature articular cartilage. It was concentrated in the pericellular matrix of the chondrocytes, appeared at or just above the tidemark, and was expressed immediately before mineralization. Identification of type X collagen in both the canine growth plate and the zone of calcified articular cartilage suggests that cells in the deep layer of cartilage and in the zone of calcified cartilage in the adult animal retain some characteristics of a growth plate and may be involved in regulation of mineralization at this critical interface. The expression of growth plate-like properties would allow the deep chondrocytes of mature articular cartilage to play a role in remodeling of the joint with age and in the pathogenesis of osteoarthritis. Key Words: Type X collagen-Zone of calcified cartilageImmunohistochemistry-Growth plate-Mineralization-Tidemark.
Much of the knowledge concerning type X collag e n has be en derived from th e chick model (7,11,12,21,23,35-39). The structure, biosynthesis, and location of type X collagen in chick cartilage have been demonstrated (7,11,12,17,21-24,27, 3540). Type X collagen, a unique product of hy-
pertrophic chondrocytes, is synthesized a s a procollagen composed of three identical pro a ( x ) chains of molecular weight (mol wt) of 59,000. The chick type X molecule is 138 nm long and consists of a triple helical domain flanked by a globular peptide on the amino terminus and a short noncollagenous piece on the carboxyl terminus. Type X collagen has been localized to zones of hypertrophic chondrocytes in chick cartilage by both biochemical and immunohistochemical tech-
Received June 5 , 1990; accpeted February 6, 1991. Address correspondence and reprint requests to Dr. Theodore R. Oegema, Jr., at Department of Orthopaedic Surgery, Box 310 UMHC, 420 Delaware St., SE, Minneapolis, MN 55455.
485
J . M . GANNON ET AL.
486
niques. Because of the unique localization of type X collagen some investigators have hypothesized that it provides a permissive matrix for calcification during endochondral bone formation or facilitates chondrocyte removal from the matrix during vascular invasion (21,37,39). Type X deposition appears to be programmed within chondrocytes, yet extracellular matrix components are involved in its gene expression (4,16,42). Type X collagen has also been found in chick fracture callus as well as in mature bovine tissue (7,19,28). Although less is known about mammalian type X, its presence has been shown in rabbit, bovine, and canine tissue (2,15,18,28,34,41). Rabbit type X collagen is apparently similar in organization to the chick molecule. In the bovine system, interchain disulfide bonds within the helical domain of the large pepsin-resistant a(x) chains help to distinguish type X from other collagen types. Monoclonal antibodies (MoAbs) against chick type X collagen have facilitated much of the research to date (11,24,37,40). It is unfortunate that antibodies to chick collagen do not cross-react with mammalian tissue. To study type X collagen distribution in the canine model, we developed a polyclonal sera against the molecule. Once characterized, the polyclonal sera was used to immunolocalize type X collagen in the canine growth plate and in adult canine articular cartilage. Immunoperoxidase (IP) and indirect immunofluorescent staining of the unfixed and decalcified fetal canine long bones demonstrated that most of the type X collagen is localized to the zone of provisional calcification of the growth plate and the metaphyseal region of the physis. Specific reactivity was also found in the zone of calcified adult articular cartilage using the immunohistochemical technique. MATERIALS AND METHODS
The purification of canine type X was modeled after that of Quarto and colleagues (33) for chicken type X. Cartilage was obtained from fetal dogs when calcification in the ribs and long bones was documented by roentgenograms. The cartilaginous epiphyses were harvested, washed extensively in phosphate-buffered saline (PBS), and homogenized. The homogenate was extracted for 24 h with two changes of 4 M guanidine hydrochloride at pH 5.8 in the presence of protease inhibitors (31). The residue was digested twice at 4°C for 18 h with 0.1% pepsin in 0.5 N acetic acid at pH 2.5. The pH of the
J Orthop Res, Vol. 9, No. 4, 1991
combined supernatants was raised to 8-9 with 2.0 M NaOH to inactivate the pepsin. The solution was clarified by centrifugation and dialyzed extensively at 4°C against 0.5 N acetic acid at pH 2.5. The collagens were fractionated by salt precipitation in 0.5 N acetic acid at 0.7, 1.2, and 2.0 M NaCl, pH 2.5 by dialysis. The 2.0 M NaCl precipitate was dissolved in a small volume of 0.5 N acetic acid and dialyzed exhaustively against 1.0 M CaC1, and 20 mM TrisHC1 at pH 7.4. The solution was heated at 42°C for 30 min, cooled quickly to room temperature, immediately chromatographed on a Bio-Gel Al.5 M column (1.1 x 175 cm), (Bio-Rad Laboratories, Richmond, CA, U.S.A.), equilibrated, and eluted at 2.4 ml/h in the same CaCl, buffer; 0.5 ml fractions were collected. Protein concentration was followed by absorbance at 220 nm. Aliquots from every third fraction of the peaks were dialyzed, lyophilized, and run on a 5% sodium dodecyl sulfate (SDS) polyacrylamide gel with or without reduction with dithiothreitol (DTT). In some cases, the electrophoretically separated proteins were transferred to nitrocellulose paper and Western blot analysis was performed (20). Type I1 and XI collagens were purified from fetal canine cartilaginous epiphyses at the time of type X purification as above. Type I1 was recovered as the precipitate in 0.5 N acetic acid at 0.7 M NaCl, and type XI was recovered as the 1.244 NaCl precipitate at 4°C (33). In each case, the pellet was dissolved in and desalted by extensive dialysis against 0.5 N acetic acid. The type I1 collagen was further purified as outlined by Miller and Rhodes (31), by dissolving the collagen in neutral salt buffer and precipitating it with increasing salt concentrations. The collagen was dialyzed against 1.O M NaC1,0.05 M Tris at pH 7.5 with protease inhibitors. The solution was cleared of insoluble material by centrifugation. The protein was then fractionally precipitated by dialysis against 1.8, 2.5, 3.5, and 4.0 M NaCl in 0.05 M Tris at pH 7.5. The desalted 4.044 precipitate was then run on SDS polyacrylamide gel electrophoresis (PAGE) to verify the purity of type I1 collagen. Collagen types I and 111were purified from fetal canine skin using essentially the same methods as described above (31). The collagens were extracted with pepsin and precipitated from 0.5 N acetic acid at 0.7 M NaC1. The weak acid was then exchanged for neutral salt by dialysis at 4°C. Type I was recovered from the neutral salt at 2.5 M NaCl and type I11 at 1.8 A4 NaCl.
TYPE X COLLAGEN IN CANINE TISSUE
Preparation of Anti-Type X Collagen Antiserum The polyclonal sera against canine type X collagen were produced in sheep. Preimmune sera were obtained from two I-year-old sheep. The animals were immunized subcutaneously (s.c.) with 100 pg purified canine type X in Freund's complete adjuvant and boosted with 100 pg canine type X in Freund's incomplete adjuvant at 1 and 5 months. The specificity of the antisera was assessed by enzymelinked immunosorbent assay (ELISA) against type I, 11, 111, and XI collagens. In a preliminary study under optimum reaction conditions, 96-well polystyrene flat-bottom plates (Immulon I plates, Dynatech Laboratories, Chantilly, VA, U.S.A.) were coated with varying concentrations of canine type X (25-1,000 ng/well in 0.1 N NaHCO,) at pH 9.6 at 37°C for 1 h. The plates were then washed further with 10 mM PBS (0.38 M NaCl), pH 8.0, and blocked with 0.01 N rabbit serum in PBS. They were washed again with PBSTween 20 and reacted with multiple dilutions of the sheep-derived antisera (150 to 1: 1,000). After addition of a horseradish peroxidase-conjugated antisheep antibody (Organon Teknika-Cappel, West Chester, PA, U.S.A.) at 1:1,000 for 1 h a t room temperature, the plates were developed using 0.03% H,02 and 67 mg/100 ml o-phenylene-diamine as substrates stopped with a final concentration of 0.63 N H,SO,, and the absorbance was read on an ELISA plate reader (Dynatech) at 490 nm. Trace amounts of cross-reacting antibodies to type 11, IX, and XI collagens were removed by immunoadsorption techniques. Type I1 reactivity was eliminated by gently shaking the antisera with cartilage powder. The cross-reactivity to type IX collagen was removed by adsorption of antisera to collagen coupled onto Sepharose 4B-200 (Pharmacia, Alameda, CA). Type XI, IX, and denatured IX collagens were coupled by a modification of the method of Axen and colleagues (I), as outlined by Timpl (43). Type IX in the 2.044 NaCl precipitate of the pepsin digest was separated from type X by chromatography on a Bio-Gel A1.5m (BioRad Labs., Richmond, CA) agarose column. Affinity columns were made against native collagen, as well as against reduced and denatured (DTT heatdenatured at 50°C for 30 min) type IX collagen. Immunohistochemistry Immunohistochemical staining was accomplished using both immunofluorescent and IP systems. All
487
bone-containing structures were decalcified in 0.3 M sodium EDTA (pH 7.0) for 5-14 days at 4°C. The EDTA solution was changed every other day. Before decalcification, specimens were cut to 1-mm slices with a Buehler isomet saw with a diamond wafering blade (Buehler, Ltd., Lake Bluff, IL, U.S.A.). For IP staining, a Vectastain ABC kit (Vector Laboratories, Burlingame, CA, U.S.A.) was used as outlined by the manufacturer. Formalin-fixed EDTA-decalcified specimens were dehydrated through graded ethanol and xylene and embedded in Paraplast at 59°C. Sections were cut at 6-7 pm and were covalently bonded to slides by the method outlined by Uhl (44). In this procedure, cleaned slides are heated at 70°C overnight in a 1% solution of 3-aminopropyltriethoxysilane(Aldnch Chemical, Milwaukee, WI, U.S.A.), pH 3.45, washed with distilled H,O for at least 2 h, and baked at 100°C for at least 8 h. Slides may be stored at this stage for s 6 months. Before use, the slides were activated with 10% glutaraldehyde in 0.1 M phosphate buffer (pH 7.0) for at least 30 min, rinsed in H,O, stabilized by mild oxidation in 0.1 M sodium metaperiodate for at least 15 min, and given a final rinse in 0.1 M phosphate buffer. This procedure is the most effective way to adhere sections to slides. The sections adhere even after harsh enzyme digestions. After the slides were deparaffinized and rehydrated, they were treated with 0.3% H,02 in 0.1 M sodium azide to quench endogenous peroxidase activity within the tissue (8). Enzymatic digestion of proteoglycans and other matrix molecules was necessary to unmask antigenic epitopes and was performed with a hyaluronidase and pepsin sequence or with Pronase (CalBiochem, La Jolla, CA, U.S.A.). In the first method, slides were treated with testicular hyaluronidase at 300-500 U/ml (Wyeth-Ayerst Laboratorie, Philadelphia, PA, U.S.A.) in PBS for 60 min at 2WC, then with 0.1% pepsin in 0.1 N acetic acid for 15 min. For the pronase digestion, a 0.1% solution in PBS was used for 12 min. The slides were washed with PBS after either digestion procedure. Nonspecific binding sites were blocked with dilute normal sera included in the Vectastain kit. Sections were incubated with varying dilutions (151:20) of purified sheep anticanine type X antisera for 60 min at 33"C, followed by PBS washes. Biotinylated antisheep secondary antibody was added for 30 min, and the slides were again washed with
J Orthop Res, Vol. 9, No. 4, 1991
J . h4. GANNON ET AL.
488
PBS. The sections were then reacted with the Vectastain ABC reagent. Localization of activity was followed by development with 0.05% 0-3,3-diaminobenzidine in 0.01% H202.In some cases, nickel chloride enhancement (9) and/or hematoxylin counterstaining were used. The sections were dehydrated in serial alcohol baths and coverslipped with Permount (Sigma Chemical, St. Louis, MO, U.S.A.). For both the indirect immunofluorescent and IP stains, control sections were incubated with preimmune sheep sera, PBS, or the second antibody. Similar techniques were used for both fetal and adult canine tissue specimens. Unfixed, decalcified frozen sections were used for the indirect immunofluorescent stains. An FITC-conjugated rabbit antisheep immunoglobulin (Organon Teknika-Cappel) was used to detect activity. The indirect immunofluorescent stained slides were viewed on a photomicroscope with a mercury vapor light source (Zeiss I11 RS, Carl Zeiss, Thornwood, NY, U.S.A.). RESULTS
As determined by electrophoretic mobility on SDS-PAGE, the 0 . 7 4 NaCl fraction contained mainly type I1 collagen, and the 1 . 2 4 fraction was greatly enriched in type XI. The 2.044 fraction contained collagen types IX and X. The type X was resolved from the type IX by taking advantage of
the greater thermal stability of type X chains (38). The sample was heated at 42°C for 30 min, then quickly cooled and chromatographed on a Bio-Gel A1.5-m column equilibrated with CaC1,. The results are shown in Fig. 1. The SDS-PAGE of peak A, as shown in Fig. 2, contained mainly type X collagen. Peaks B-D (Fig. 2) contained some domains of type IX, as well as some proteolytic fragments. The samples from peak A migrated on electrophoresis as a single highmolecular-weight (mol wt) band before reduction and as a band with mol wt -50,000 daltons after reduction (as compared with globular standards), with only traces of type IX collagen. The large change in migration after reduction by DTT substantiates the presence of disulfide bonds in canine type X collagen. Because some of the species present in peaks B-D (Fig. 2) were also reducible with DTT, the fractions may also contain disulfide bonds. We produced polyclonal sera against canine type X by immunizing sheep with the purified canine type X from pooled peak A after first obtaining preimmune sera samples. The preimmune sera showed essentially no endogenous anti-X activity by ELISA, whereas the postimmunization bleeds produced sera that bound reproducibly to type X collagen-coated plates and was competitively displaced by purified type X but not by type I, 11, 111, or XI collagens (data not shown). Some immunoreactivity was observed on Western blot analysis,
E =
2
0
FIG. 1. Elution profile from BioGel A 1.5-rn of the 2.044 NaCl acid salt fraction of fetal growth plate collagens. Peaks were pooled and analyzed by sodium dodecyl sulfate polyacrylarnide gel electrophoresis.
hl
N
0 0 C
m
:
I n 1
.n
a
0 40
60
100
80 sample
J Orthop Res, Vol. 9,No. 4, 1991
120
140
TYPE X COLLAGEN IN CANINE TISSUE
489
strated by significantly decreased Safranin-0 uptake in the growth plates. Immunohistochemistry was performed on unfixed, fast-frozen, and formalin-fixed tissue specimens. All samples were decalcified with EDTA. The unfixed specimens were used for immunofluorescent staining. Indirect immunofluorescent and IP stains yielded very similar results. The fixed and paraffin-embedded specimens yielded better tissue preservation, easier handling, and superior histology. Therefore, the IP-stained specimens were used for illustration. Most of the type X collagen was localized to the zone of provisional calcification in the growth plate and the metaphyseal region of the physis (Fig. 4A), but stain was easily detected at the level of the early hypertrophic chondrocytes. No staining was observed in control sections treated with preimmune
FIG. 2. Sodium dodecyl sulfate polyacrylarnide gel electrophoresis pattern obtained by pooling samples from peak A from the Bio-Gel A 1.5-rn column. The gel was stained with Coornassie blue R 250, 5 reduction with dithiothreitol (DTT). Peak A contained predominantly type X collagen with only traces of type IX. The presence of disulfide bonds is clearly evident after reduction with DTT.
however, not only against type X but also against type IX both before and after reduction. Therefore, the antisera were affinity purified as described in the Materials and Methods section. The specificity of the resulting sheep affinity-purified anti-X antisera was verified by Western blot analysis. Semipurified type X collagen containing some type IX collagen, run on 5% SDS-PAGE and transblotted to nitrocellulose paper, produced a single immunoreactive band after exposure to the purified polyclonal sera, even after long development times (Fig. 3). The specificity of the adsorbed antisera was further confirmed by finding the staining pattern of type X collagen in the canine growth plate. Predigestion with proteases was necessary to unmask the type X collagen. This digestion resulted in substantial removal of proteoglycan, as demon-
FIG. 3. Western blot of semipurified types X and IX collagen after separation on 5% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). The bands were stained with purified anti-type X antisera at a dilution of 1 :200,which gives a single immunoreactive band. N, nonreduced; R, reduced.
J Orthop Res, Vol. 9, No. 4 , 1991
490
J . M . GANNON ET AL. B
A
FIG. 4. A: Canine growth plate section stained with sheep anti-canine type X collagen antisera and a peroxidase-conjugatedgoat anti-sheep IgG secondary antisera. Arrows indicate the most superior extent of type X expression. Arrowheads highlight staining in the metaphyseal cartilage remnants. 8: A control section stained by omitting the primary antibody, followed by the peroxidaseconjugated secondary antibody. No reactivity is evident.
sera or PBS with the antisera omitted (Fig. 4B). A progressive decrease in staining toward the diaphysis of the fetal canine long bone was apparent. The hypothesis that this decrease corresponds to a remodeling of the remaining cartilaginous anlage to bone is supported by a similar loss of Safranin-0 stain in undigested controls. When polyclonal sera dilutions were preincubated with known amounts of purified canine type X collagen before tissue exposure, the staining was lost (data not shown). Minor staining of blood vessels was noted, which probably represented crossreactivity of purified anti-type X antisera with type VIII collagen in the subendothelial space. Recent research has shown a high degree of homology between collagen types VIII and X (45). Otherwise, no reactivity was demonstrated with the sheep antiX polysera to tissues outside the skeletal system in
J Orthop Res, Vot. 9, NO.4,1991
the immature canine. Skin, kidney, lung, ligament, and tendon were examined in the tissue screen. Mature canine articular cartilage from 2- to 4-year old male dogs was examined next. The tissue sections were obtained from the distal femoral condyle, proximal humerus or patellar articular surface. Unexpectedly, immunohistochemical staining of the adult canine articular cartilage showed specific reactivity in the zone of calcified cartilage (ZCC) (Fig. 5A). Staining is localized directly adjacent to the chondrocytes in the pericellular matrix. No staining, or only very diffuse staining, was noted in the interterritorial matrix. Of additional interest, not all cells in the ZCC stained positively for type X. From 57 to 91% of the cells in a given field showed reactivity. The first area in which type X could be identified was around the chondrocytes located at or just above the tidemark (Fig. 6). This
TYPE X COLLAGEN IN CANINE TISSUE
491 B
A
FIG. 5. A: Sections of adult canine articular cartilage stained by the immunoperoxidase technique using sheep anti-canine type X collagen polysera. Specific reactivity is evident in the pericellular matrix around cells in the zone of calcified cartilage. Arrows indicate some sites of type X reactivity. 6: Control section of adult canine articular cartilage stained by omitting the primary antisera followed by peroxidase-conjugated secondary antisera. No specific staining is evident.
FIG. 6. Uncalcified-calcified cartilage interface of canine patella stained with the immunoperoxidase technique for type X collagen. Specific reactivity is evident around the cells at the tidemark (arrowheads). Section was counterstained with fast green to delineate between calcified and uncalcified tissues.
J Orthop Res, Val. 9, No. 4, 1991
492
J . M . GANNON ET AL.
FIG. 7. Higher power photograph of junction between the zone of calcified cartilage (ZCC) and the subchondral bone. Type X collagen staining can be seen to be remodeled away as the ZCC is replaced by mature bone.
distribution indicates that expression of type X precedes mineralization, as was observed in the growth plate. Thus, most of the staining is within the ZCC as delineated by the tidemark superiorly and the cement line inferiorly. Reactivity can be seen to be remodeled away as the ZCC is replaced with mature bone (Fig. 7). This reactivity was also competitively inhibited by preincubation of specific dilutions of anti-X polysera with known amounts of type X collagen. DISCUSSION Hyaline or articular cartilage contains collagen types 11, V, IX, and XI (13,30), occasionally type VI (13), and in special cases, type X. Our work with type X collagen in the canine evolved from our efforts to define an animal model ofjoint deterioration that reproduces changes observed in human osteoarthritis. Previous work in our laboratory (10) demonstrated that acute transarticular load to the adult canine patella-femoral joint causes reproducible swelling of the hyaline cartilage and alterations in the ZCC, including increased vascularity and duplication of the tidemark. Therefore, this research was undertaken to define the biology of mineralization of the ZCC and to compare these mechanisms with the biology of the growth plate. To compare the ZCC and the growth plate, we examined the
J Orthop Res, Val. 9, No. 4, 1991
expression of type X collagen, believed to be unique to these regions. The presence of type X collagen in mammalian tissue has been demonstrated in rabbit and bovine models (2,18,34); most of this research involved extraction of the molecule from growth plates or cultures of growth plate cartilage derived from young animals. Most recently in a preliminary report, the persistence of type X in adolescent and mature bovine ZCC was shown by chemical isolation (28). The immunohistochemical localization of canine type X collagen to the metaphyseal region of the physis is consistent with observations in the chick model. Both decalcification and proteoglycan digestion were necessary before immunoreactivity could be demonstrated. Type X has a peritrabecular distribution in the spongiosum, and the concentration progressively decreases toward the diaphysis of the fetal canine long bone, which is again consistent with the chick model. Work with the chick model has shown that type X collagen is present in cartilage matrix that contains hypertrophic chondrocytes. Although type X synthesis has been shown by cultured chondrocytes derived from permanent cartilaginous regions of the chick embryo sternum, little has been reported on the persistence of type X collagen in adult mammals. Therefore, we did not expect to see specific
TYPE X COLLAGEN IN CANINE TISSUE
staining in the adult ZCC. The identification of type X collagen in both the canine growth plate and the ZCC suggests that the ZCC in adult animals retains some characteristics of a growth plate. Whether type X collagen in adult articular cartilage is a growth plate remnant or is produced by hypertrophic chondrocytes in the adult remains to be determined. The function of type X collagen within the skeletal system remains unclear. Some investigators have suggested that it provides a permissive matrix for calcification and acts as a stimulus for chondrocyte removal. The expression of growth plate-like properties would allow the deep chondrocytes of mature articular cartilage to play a role in remodeling of the joint with age and in the pathogenesis of osteoarthritis. In summary, the presence of type X collagen in canine tissue has been demonstrated, and the distribution in the canine growth plate is consistent with observations in the chick model. The presence of type X collagen in permanent tissue, specifically the ZCC of adult canine articular cartilage, has also been shown and may indicate a similarity between the biology of mineralization of the growth plate and that of the ZCC of adult canine articular cartilage. Acknowledgment: This work was supported by NIH Grants No. AR39255 and AR07555 and in part by OREF Grant No. 89-492 from Bristol-Myers. We thank Thomas Schmid (Rush Presbyterian-St. Luke’s Medical Center, Chicago, IL, U.S.A.) for helpful advice and encouragement in the purification and characterization of type X collagen. We also thank Mark Oraskovich for help in developing the ELISA assay.
REFERENCES 1. Axen R, Porath J, Ernback S: Chemical coupling of peptides to polysaccharides by means of cyanogen halides. Nature 214~1302-1304, 1967 2. Ayad S, Kwan APL, Grant ME: Partial characterization of type X collagen from bovine growth plate cartilage. FEBS Lett 220:181-186, 1987 3. Bates GP, Schor SL, Grant ME: A comparison of the effects of different substrata on chondrocyte morphology and the synthesis of collagen types IX and X. In Vitro Cell Dev Biol 23:374-380, 1987 4. Brown CC, Balian G : Effect of heparin on synthesis of short chain collagen by chondrocytes and smooth muscle cells. J Cell Biol 105:1007-1012, 1987 5. Bullough PG, Yawitz PS, Tafra L, Boskey AL: Topographical variations in the morphology and biochemistry of adult canine tibia1 plateau articular cartilage. J Orthop Res 3: 1-16, 1985
493
6 . Castagnola P, Dozin B, Moro G, Cancedda R: Changes in the expression of collagen genes show two stages in chondrocyte differentiation in vitro. J Cell Biol 106:461467, 1988 7. Castagnola P, Torella G, Cancedda R: Type X collagen synthesis by cultured chondrocytes derived from the permanent cartilaginous region of chick embryo sternum. Dev Biol 123:332-337, 1987 8. Chin Yang LI, Ziesman SC, Lazcano-Villereal 0: Use of azide and hydrogen peroxide as an inhibitor for endogenous peroxidase in the IP method. J Hisrochem Cytochem 35: 1457-1460, 1987 9. De Jong ASH, Van Kessel-Van Vark M, Raap AK: Sensitivity of various visualization methods for peroxidase and alkaline phosphatase activity in immunoenzyme histochemistry. Histochem J 17:1119-1130, 1985 10. Donohue JM, Buss D, Oegema TR, Thompson RC: The effects of indirect blunt trauma on adult canine articular cartilage. J Bone Joint Surg [Am] 65:948-957, 1983 11. Eavy RD, Schmid TM, Linsenmayer TF: Development of the chick columella: Immunohistochemical studies with anticollagen monoclonal antibodies. Int J Pediatr Otorhinolaryngol 13:99-105, 1987 12. Eavy RD, Schmid TM, Linsenmayer TF: Intrinsic and extrinsic controls of the hypertrophic program of chondrocytes in the avian columella. Dev Biol 1265742, 1988 13. Eyre DR, Wu J, Apone S: A growing family of collagens in articular cartilage: Identification of 5 genetically distinct types. J Rheumatol 14(suppl 14):25-27, 1987 14. Eyre DR, Apon S, Wu JJ, Ericsson LH, Walsch KA: Collagen type IX: Evidence for covalent linkages to type I1 collagen in cartilage. FEBS Lett 220:337-341, 1987 15. Gannon JM, Fischer M, Hofmeister F , Carpenter R, Oegema TR Jr, Thompson RC Jr: Localization of type X collagen in fetal canine growth plate and adult canine articular cartilage. Trans Orthop Res Soc 1459, 1989 16. Gibson GJ, Kielty CM, Gamer C, Schor SL, Grant ME: Identification and partial characterization of three lowmolecular-weight collagenous polypeptides synthesized by chondrocytes cultured within collagen gels in the absence and in the presence of fibronectin. Biochem J 211:417426, 1983 17. Gibson GJ, Flint MH: Type X collagen synthesis by chick sternal cartilage and its relationship to endochondral development. J Cell Biol 101:277-284, 1985 18. Grant WT, Sussman MD, Balian G: A disulfide-bonded short chain collagen synthesized by degenerative and calcifying zones of bonive growth plate cartilage. J Biol Chem 260:379&3803, 1985 19. Grant WT, Wang GJ, Balian G: Type X collagen synthesis during endochondral ossification in fracture repair. J Biol Chem 262:9844-9849, 1987 20. Habuchi H, Conrad HE, Glaser JH: Coordinate regulation of collagen and alkaline phosphatase levels in chick embryo chondrocytes. J Biol Chem 260: 13029-13034, 1985 21. Jimenez SA, Yankowski R, Reginato AM: Quantitative analysis of type X collagen biosynthesis by embryonic chick sternal cartilage. Biochem J 233:357-367, 1986 22. Kielty CM, Holmes DJS, Schor SL, Grant ME: Embryonic chick cartilage collagens. FEBS Lett 169:179-184, 1984 23. Kielty CM, Kwan APL, Holmes DF, Schor SL, Grant ME: Type X collagen: A product of hypertrophic chondrocytes. Biochem J 27545-554, 1985 24. Kwan APL, Freemont AJ, Grant ME: Immunoperoxidase localization of type X collagen in chick tibiae. Biosci Rep 6: 155-162, 1986 25. Kwan APL, Sear CHJ, Grant ME: Identification of disul-
J Orthop Res, Vol. 9,N O . 4 , 1991
494
J . M . GANNON ET AL.
phide-bonded type X procollagen polypeptides in embryonic chick chondrocyte cultures. FEBS Lett 206:267-272, 1986 26. Laemmli UK: Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:68& 685, 1970 27. Linsenmayer TF, Gibney E, Schmid TM: Segmental appearance of type X collagen in the developing avian notochord. Dev Biol 113:467-473, 1986 28. Love11 TP, Eyre DR: Unique biochemical characteristics of the calcified zone of articular cartilage. Trans Orthop Res Soc 13511, 1988 29. Manduca P, Castagnola P, Cancedda R: Dimethyl sulfoxide interferes with in vitro differentiation of chick embryo endochondral chondrocytes. Dev Biol 125:234-236, 1986 30. Mayne R, Irwin MH: Collagen types in cartilage. In: Kuettner K, Schleyerbach R, Hascall VC, ed. Articular Cartilage Biochemistry. New York: Raven Press. 1986:23-38 31. Miller EJ, Rhodes K: Preparation and characterization of the different types of collagen. Methods Enzyrnol82:33-64,1982 32. Ninomiya Y, Gordon M, van der Rest M, Schmid TM, Linsenmayer TF, Olsen BR: The developmentally regulated type X collagen gene contains a long open reading frame without introns. J Biol Chem 2615041-5050, 1985 33. Quarto N, Cancedda R, Descali-Cancedda F: Purification and characterization of the low molecular mass (type X) collagen from chick embryo tibial cartilage. Eur J Biochem 147:397-400, 1985 34. Remington MC, Bashey RI, Brighton CT, Jimenez SA: Biosynthesis of a low molecular weight collagen by rabbit growth plate cartilage organ cultures. Cot1 Re1 Res 3:271278, 1983 35. Schmid TM, Conrad HE: Metabolism of low molecular weight collagen by chondrocytes obtained from histologi-
J Orthop Res, Vol. 9, No. 4, 1991
cally distinct zones of the chick embryo tibiotarsus. J Biol Chem 257:12451-12457, 1987 36. Schmid TM, Linsenmayer TF: Developmental acquisition of type X collagen in the embryonic chick tibiotarsus. Dev Biol 107~373-381,1985 37. Schmid TM, Linsenmayer TF: Immunohistochemical localization of short chain cartilage collagen (type X) in avian tissues. J Cell Biol 100:49&605, 1985 38. Schmid TM, Linsenmayer TF: Type X collagen. In: Mayne R, Burgeson RE, eds. Structure and Function of Collagen Types. Orlando: Academic Press. 1987:223-259 39. Solursh M, Jensen KL, Reiter RS, Schmid TM, Linsenmayer TF: Environmental regulation of type X collagen production by cultures of limb mesenchyme, mesectodern and sternal chondrocytes. Dev Biol 117:90-101, 1986 40. Summers TA, Irwin MH, Mayne R, Balian G: Monoclonal antibodies to type X collagen. J Biol Chem 263:581-587, 1988 41. Sussman MD, Ogle RC, Balian G: Biosynthesis and processing of collagens in different cartilaginous tissues. J Orthop Res 2:134-142, 1984 42. Thomas JT, Grant ME: Cartilage proteoglycan aggregate and fibronectin can modulate the expression of type X collagen by embryonic chick chondrocytes in collagen gels. Biosci Rep 8:163-171, 1988 43. Timpl R: Antibodies to collagens and procollagens. Methods Enzymol. 82: Academic Press, New York, 1982,pp 472-498 44. Uhl GR, ed: In Situ Hybridization in Brain. New York: Plenum Press. 1986:266-267 45. Yamaguchi N, Benya PD, van der Rest M, Ninomiya Y: The cloning sequence of al(VII1) collagen cDNAs demonstrate that type VIII collagen is a short chain collagen and contains triple-helical and carboxyl-terminal non-triple-helical domains similar to those of type X collagen. J Biol Chem 264:16022-16029, 1989
Localization of Type X Collagen in Canine Growth Plate and Adult Canine Articular Cartilage *James M. Gannon, *Gordon Walker, *Mark Fischer, *Randy Carpenter, *Roby C. Thompson, Jr., and *tTheodore R. Oegema, Jr. Departments of *Orthopaedic Surgery and TBiochemistry, University of Minnesota, Minneapolis, Minnesota, U.S.A.
Summary: Type X collagen was extracted from ends of canine growth plates by pepsin digestion after 4 M guanidine hydrochloride extraction, purified by stepwise salt precipitation (2.0 M NaCl in 0.5 M acetic acid), and chromatographed on a Bio-Gel A1.5 M column in 1.0 M CaCl,. Without reduction on sodium dodecyl sulfate (SDS) polyacrylamide gels, the preparation yielded a single, high-molecular-weight (mol wt) band; after reduction, a single band of relative mol wt 5.0 x lo4 was found. Polyclonal sera were raised against the purified collagen and used in the immunolocalization of canine type X collagen. As expected, indirect immunoperoxidase (IP) or indirect immunofluorescent staining with the polyclonal sera demonstrated that most of the immunoreactivity was localized in the zone of provisional calcification of the growth plate and in cartilage remnants in the metaphyseal region of the physis. A progressive decrease in staining toward the diaphysis of the fetal canine long bone was apparent as the trabecular structures were remodeled to bone. Unexpectedly, type X collagen was also detected in the zone of calcified, mature articular cartilage. It was concentrated in the pericellular matrix of the chondrocytes, appeared at or just above the tidemark, and was expressed immediately before mineralization. Identification of type X collagen in both the canine growth plate and the zone of calcified articular cartilage suggests that cells in the deep layer of cartilage and in the zone of calcified cartilage in the adult animal retain some characteristics of a growth plate and may be involved in regulation of mineralization at this critical interface. The expression of growth plate-like properties would allow the deep chondrocytes of mature articular cartilage to play a role in remodeling of the joint with age and in the pathogenesis of osteoarthritis. Key Words: Type X collagen-Zone of calcified cartilageImmunohistochemistry-Growth plate-Mineralization-Tidemark.
Much of the knowledge concerning type X collag e n has be en derived from th e chick model (7,11,12,21,23,35-39). The structure, biosynthesis, and location of type X collagen in chick cartilage have been demonstrated (7,11,12,17,21-24,27, 3540). Type X collagen, a unique product of hy-
pertrophic chondrocytes, is synthesized a s a procollagen composed of three identical pro a ( x ) chains of molecular weight (mol wt) of 59,000. The chick type X molecule is 138 nm long and consists of a triple helical domain flanked by a globular peptide on the amino terminus and a short noncollagenous piece on the carboxyl terminus. Type X collagen has been localized to zones of hypertrophic chondrocytes in chick cartilage by both biochemical and immunohistochemical tech-
Received June 5 , 1990; accpeted February 6, 1991. Address correspondence and reprint requests to Dr. Theodore R. Oegema, Jr., at Department of Orthopaedic Surgery, Box 310 UMHC, 420 Delaware St., SE, Minneapolis, MN 55455.
485
J . M . GANNON ET AL.
486
niques. Because of the unique localization of type X collagen some investigators have hypothesized that it provides a permissive matrix for calcification during endochondral bone formation or facilitates chondrocyte removal from the matrix during vascular invasion (21,37,39). Type X deposition appears to be programmed within chondrocytes, yet extracellular matrix components are involved in its gene expression (4,16,42). Type X collagen has also been found in chick fracture callus as well as in mature bovine tissue (7,19,28). Although less is known about mammalian type X, its presence has been shown in rabbit, bovine, and canine tissue (2,15,18,28,34,41). Rabbit type X collagen is apparently similar in organization to the chick molecule. In the bovine system, interchain disulfide bonds within the helical domain of the large pepsin-resistant a(x) chains help to distinguish type X from other collagen types. Monoclonal antibodies (MoAbs) against chick type X collagen have facilitated much of the research to date (11,24,37,40). It is unfortunate that antibodies to chick collagen do not cross-react with mammalian tissue. To study type X collagen distribution in the canine model, we developed a polyclonal sera against the molecule. Once characterized, the polyclonal sera was used to immunolocalize type X collagen in the canine growth plate and in adult canine articular cartilage. Immunoperoxidase (IP) and indirect immunofluorescent staining of the unfixed and decalcified fetal canine long bones demonstrated that most of the type X collagen is localized to the zone of provisional calcification of the growth plate and the metaphyseal region of the physis. Specific reactivity was also found in the zone of calcified adult articular cartilage using the immunohistochemical technique. MATERIALS AND METHODS
The purification of canine type X was modeled after that of Quarto and colleagues (33) for chicken type X. Cartilage was obtained from fetal dogs when calcification in the ribs and long bones was documented by roentgenograms. The cartilaginous epiphyses were harvested, washed extensively in phosphate-buffered saline (PBS), and homogenized. The homogenate was extracted for 24 h with two changes of 4 M guanidine hydrochloride at pH 5.8 in the presence of protease inhibitors (31). The residue was digested twice at 4°C for 18 h with 0.1% pepsin in 0.5 N acetic acid at pH 2.5. The pH of the
J Orthop Res, Vol. 9, No. 4, 1991
combined supernatants was raised to 8-9 with 2.0 M NaOH to inactivate the pepsin. The solution was clarified by centrifugation and dialyzed extensively at 4°C against 0.5 N acetic acid at pH 2.5. The collagens were fractionated by salt precipitation in 0.5 N acetic acid at 0.7, 1.2, and 2.0 M NaCl, pH 2.5 by dialysis. The 2.0 M NaCl precipitate was dissolved in a small volume of 0.5 N acetic acid and dialyzed exhaustively against 1.0 M CaC1, and 20 mM TrisHC1 at pH 7.4. The solution was heated at 42°C for 30 min, cooled quickly to room temperature, immediately chromatographed on a Bio-Gel Al.5 M column (1.1 x 175 cm), (Bio-Rad Laboratories, Richmond, CA, U.S.A.), equilibrated, and eluted at 2.4 ml/h in the same CaCl, buffer; 0.5 ml fractions were collected. Protein concentration was followed by absorbance at 220 nm. Aliquots from every third fraction of the peaks were dialyzed, lyophilized, and run on a 5% sodium dodecyl sulfate (SDS) polyacrylamide gel with or without reduction with dithiothreitol (DTT). In some cases, the electrophoretically separated proteins were transferred to nitrocellulose paper and Western blot analysis was performed (20). Type I1 and XI collagens were purified from fetal canine cartilaginous epiphyses at the time of type X purification as above. Type I1 was recovered as the precipitate in 0.5 N acetic acid at 0.7 M NaCl, and type XI was recovered as the 1.244 NaCl precipitate at 4°C (33). In each case, the pellet was dissolved in and desalted by extensive dialysis against 0.5 N acetic acid. The type I1 collagen was further purified as outlined by Miller and Rhodes (31), by dissolving the collagen in neutral salt buffer and precipitating it with increasing salt concentrations. The collagen was dialyzed against 1.O M NaC1,0.05 M Tris at pH 7.5 with protease inhibitors. The solution was cleared of insoluble material by centrifugation. The protein was then fractionally precipitated by dialysis against 1.8, 2.5, 3.5, and 4.0 M NaCl in 0.05 M Tris at pH 7.5. The desalted 4.044 precipitate was then run on SDS polyacrylamide gel electrophoresis (PAGE) to verify the purity of type I1 collagen. Collagen types I and 111were purified from fetal canine skin using essentially the same methods as described above (31). The collagens were extracted with pepsin and precipitated from 0.5 N acetic acid at 0.7 M NaC1. The weak acid was then exchanged for neutral salt by dialysis at 4°C. Type I was recovered from the neutral salt at 2.5 M NaCl and type I11 at 1.8 A4 NaCl.
TYPE X COLLAGEN IN CANINE TISSUE
Preparation of Anti-Type X Collagen Antiserum The polyclonal sera against canine type X collagen were produced in sheep. Preimmune sera were obtained from two I-year-old sheep. The animals were immunized subcutaneously (s.c.) with 100 pg purified canine type X in Freund's complete adjuvant and boosted with 100 pg canine type X in Freund's incomplete adjuvant at 1 and 5 months. The specificity of the antisera was assessed by enzymelinked immunosorbent assay (ELISA) against type I, 11, 111, and XI collagens. In a preliminary study under optimum reaction conditions, 96-well polystyrene flat-bottom plates (Immulon I plates, Dynatech Laboratories, Chantilly, VA, U.S.A.) were coated with varying concentrations of canine type X (25-1,000 ng/well in 0.1 N NaHCO,) at pH 9.6 at 37°C for 1 h. The plates were then washed further with 10 mM PBS (0.38 M NaCl), pH 8.0, and blocked with 0.01 N rabbit serum in PBS. They were washed again with PBSTween 20 and reacted with multiple dilutions of the sheep-derived antisera (150 to 1: 1,000). After addition of a horseradish peroxidase-conjugated antisheep antibody (Organon Teknika-Cappel, West Chester, PA, U.S.A.) at 1:1,000 for 1 h a t room temperature, the plates were developed using 0.03% H,02 and 67 mg/100 ml o-phenylene-diamine as substrates stopped with a final concentration of 0.63 N H,SO,, and the absorbance was read on an ELISA plate reader (Dynatech) at 490 nm. Trace amounts of cross-reacting antibodies to type 11, IX, and XI collagens were removed by immunoadsorption techniques. Type I1 reactivity was eliminated by gently shaking the antisera with cartilage powder. The cross-reactivity to type IX collagen was removed by adsorption of antisera to collagen coupled onto Sepharose 4B-200 (Pharmacia, Alameda, CA). Type XI, IX, and denatured IX collagens were coupled by a modification of the method of Axen and colleagues (I), as outlined by Timpl (43). Type IX in the 2.044 NaCl precipitate of the pepsin digest was separated from type X by chromatography on a Bio-Gel A1.5m (BioRad Labs., Richmond, CA) agarose column. Affinity columns were made against native collagen, as well as against reduced and denatured (DTT heatdenatured at 50°C for 30 min) type IX collagen. Immunohistochemistry Immunohistochemical staining was accomplished using both immunofluorescent and IP systems. All
487
bone-containing structures were decalcified in 0.3 M sodium EDTA (pH 7.0) for 5-14 days at 4°C. The EDTA solution was changed every other day. Before decalcification, specimens were cut to 1-mm slices with a Buehler isomet saw with a diamond wafering blade (Buehler, Ltd., Lake Bluff, IL, U.S.A.). For IP staining, a Vectastain ABC kit (Vector Laboratories, Burlingame, CA, U.S.A.) was used as outlined by the manufacturer. Formalin-fixed EDTA-decalcified specimens were dehydrated through graded ethanol and xylene and embedded in Paraplast at 59°C. Sections were cut at 6-7 pm and were covalently bonded to slides by the method outlined by Uhl (44). In this procedure, cleaned slides are heated at 70°C overnight in a 1% solution of 3-aminopropyltriethoxysilane(Aldnch Chemical, Milwaukee, WI, U.S.A.), pH 3.45, washed with distilled H,O for at least 2 h, and baked at 100°C for at least 8 h. Slides may be stored at this stage for s 6 months. Before use, the slides were activated with 10% glutaraldehyde in 0.1 M phosphate buffer (pH 7.0) for at least 30 min, rinsed in H,O, stabilized by mild oxidation in 0.1 M sodium metaperiodate for at least 15 min, and given a final rinse in 0.1 M phosphate buffer. This procedure is the most effective way to adhere sections to slides. The sections adhere even after harsh enzyme digestions. After the slides were deparaffinized and rehydrated, they were treated with 0.3% H,02 in 0.1 M sodium azide to quench endogenous peroxidase activity within the tissue (8). Enzymatic digestion of proteoglycans and other matrix molecules was necessary to unmask antigenic epitopes and was performed with a hyaluronidase and pepsin sequence or with Pronase (CalBiochem, La Jolla, CA, U.S.A.). In the first method, slides were treated with testicular hyaluronidase at 300-500 U/ml (Wyeth-Ayerst Laboratorie, Philadelphia, PA, U.S.A.) in PBS for 60 min at 2WC, then with 0.1% pepsin in 0.1 N acetic acid for 15 min. For the pronase digestion, a 0.1% solution in PBS was used for 12 min. The slides were washed with PBS after either digestion procedure. Nonspecific binding sites were blocked with dilute normal sera included in the Vectastain kit. Sections were incubated with varying dilutions (151:20) of purified sheep anticanine type X antisera for 60 min at 33"C, followed by PBS washes. Biotinylated antisheep secondary antibody was added for 30 min, and the slides were again washed with
J Orthop Res, Vol. 9, No. 4, 1991
J . h4. GANNON ET AL.
488
PBS. The sections were then reacted with the Vectastain ABC reagent. Localization of activity was followed by development with 0.05% 0-3,3-diaminobenzidine in 0.01% H202.In some cases, nickel chloride enhancement (9) and/or hematoxylin counterstaining were used. The sections were dehydrated in serial alcohol baths and coverslipped with Permount (Sigma Chemical, St. Louis, MO, U.S.A.). For both the indirect immunofluorescent and IP stains, control sections were incubated with preimmune sheep sera, PBS, or the second antibody. Similar techniques were used for both fetal and adult canine tissue specimens. Unfixed, decalcified frozen sections were used for the indirect immunofluorescent stains. An FITC-conjugated rabbit antisheep immunoglobulin (Organon Teknika-Cappel) was used to detect activity. The indirect immunofluorescent stained slides were viewed on a photomicroscope with a mercury vapor light source (Zeiss I11 RS, Carl Zeiss, Thornwood, NY, U.S.A.). RESULTS
As determined by electrophoretic mobility on SDS-PAGE, the 0 . 7 4 NaCl fraction contained mainly type I1 collagen, and the 1 . 2 4 fraction was greatly enriched in type XI. The 2.044 fraction contained collagen types IX and X. The type X was resolved from the type IX by taking advantage of
the greater thermal stability of type X chains (38). The sample was heated at 42°C for 30 min, then quickly cooled and chromatographed on a Bio-Gel A1.5-m column equilibrated with CaC1,. The results are shown in Fig. 1. The SDS-PAGE of peak A, as shown in Fig. 2, contained mainly type X collagen. Peaks B-D (Fig. 2) contained some domains of type IX, as well as some proteolytic fragments. The samples from peak A migrated on electrophoresis as a single highmolecular-weight (mol wt) band before reduction and as a band with mol wt -50,000 daltons after reduction (as compared with globular standards), with only traces of type IX collagen. The large change in migration after reduction by DTT substantiates the presence of disulfide bonds in canine type X collagen. Because some of the species present in peaks B-D (Fig. 2) were also reducible with DTT, the fractions may also contain disulfide bonds. We produced polyclonal sera against canine type X by immunizing sheep with the purified canine type X from pooled peak A after first obtaining preimmune sera samples. The preimmune sera showed essentially no endogenous anti-X activity by ELISA, whereas the postimmunization bleeds produced sera that bound reproducibly to type X collagen-coated plates and was competitively displaced by purified type X but not by type I, 11, 111, or XI collagens (data not shown). Some immunoreactivity was observed on Western blot analysis,
E =
2
0
FIG. 1. Elution profile from BioGel A 1.5-rn of the 2.044 NaCl acid salt fraction of fetal growth plate collagens. Peaks were pooled and analyzed by sodium dodecyl sulfate polyacrylarnide gel electrophoresis.
hl
N
0 0 C
m
:
I n 1
.n
a
0 40
60
100
80 sample
J Orthop Res, Vol. 9,No. 4, 1991
120
140
TYPE X COLLAGEN IN CANINE TISSUE
489
strated by significantly decreased Safranin-0 uptake in the growth plates. Immunohistochemistry was performed on unfixed, fast-frozen, and formalin-fixed tissue specimens. All samples were decalcified with EDTA. The unfixed specimens were used for immunofluorescent staining. Indirect immunofluorescent and IP stains yielded very similar results. The fixed and paraffin-embedded specimens yielded better tissue preservation, easier handling, and superior histology. Therefore, the IP-stained specimens were used for illustration. Most of the type X collagen was localized to the zone of provisional calcification in the growth plate and the metaphyseal region of the physis (Fig. 4A), but stain was easily detected at the level of the early hypertrophic chondrocytes. No staining was observed in control sections treated with preimmune
FIG. 2. Sodium dodecyl sulfate polyacrylarnide gel electrophoresis pattern obtained by pooling samples from peak A from the Bio-Gel A 1.5-rn column. The gel was stained with Coornassie blue R 250, 5 reduction with dithiothreitol (DTT). Peak A contained predominantly type X collagen with only traces of type IX. The presence of disulfide bonds is clearly evident after reduction with DTT.
however, not only against type X but also against type IX both before and after reduction. Therefore, the antisera were affinity purified as described in the Materials and Methods section. The specificity of the resulting sheep affinity-purified anti-X antisera was verified by Western blot analysis. Semipurified type X collagen containing some type IX collagen, run on 5% SDS-PAGE and transblotted to nitrocellulose paper, produced a single immunoreactive band after exposure to the purified polyclonal sera, even after long development times (Fig. 3). The specificity of the adsorbed antisera was further confirmed by finding the staining pattern of type X collagen in the canine growth plate. Predigestion with proteases was necessary to unmask the type X collagen. This digestion resulted in substantial removal of proteoglycan, as demon-
FIG. 3. Western blot of semipurified types X and IX collagen after separation on 5% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). The bands were stained with purified anti-type X antisera at a dilution of 1 :200,which gives a single immunoreactive band. N, nonreduced; R, reduced.
J Orthop Res, Vol. 9, No. 4 , 1991
490
J . M . GANNON ET AL. B
A
FIG. 4. A: Canine growth plate section stained with sheep anti-canine type X collagen antisera and a peroxidase-conjugatedgoat anti-sheep IgG secondary antisera. Arrows indicate the most superior extent of type X expression. Arrowheads highlight staining in the metaphyseal cartilage remnants. 8: A control section stained by omitting the primary antibody, followed by the peroxidaseconjugated secondary antibody. No reactivity is evident.
sera or PBS with the antisera omitted (Fig. 4B). A progressive decrease in staining toward the diaphysis of the fetal canine long bone was apparent. The hypothesis that this decrease corresponds to a remodeling of the remaining cartilaginous anlage to bone is supported by a similar loss of Safranin-0 stain in undigested controls. When polyclonal sera dilutions were preincubated with known amounts of purified canine type X collagen before tissue exposure, the staining was lost (data not shown). Minor staining of blood vessels was noted, which probably represented crossreactivity of purified anti-type X antisera with type VIII collagen in the subendothelial space. Recent research has shown a high degree of homology between collagen types VIII and X (45). Otherwise, no reactivity was demonstrated with the sheep antiX polysera to tissues outside the skeletal system in
J Orthop Res, Vot. 9, NO.4,1991
the immature canine. Skin, kidney, lung, ligament, and tendon were examined in the tissue screen. Mature canine articular cartilage from 2- to 4-year old male dogs was examined next. The tissue sections were obtained from the distal femoral condyle, proximal humerus or patellar articular surface. Unexpectedly, immunohistochemical staining of the adult canine articular cartilage showed specific reactivity in the zone of calcified cartilage (ZCC) (Fig. 5A). Staining is localized directly adjacent to the chondrocytes in the pericellular matrix. No staining, or only very diffuse staining, was noted in the interterritorial matrix. Of additional interest, not all cells in the ZCC stained positively for type X. From 57 to 91% of the cells in a given field showed reactivity. The first area in which type X could be identified was around the chondrocytes located at or just above the tidemark (Fig. 6). This
TYPE X COLLAGEN IN CANINE TISSUE
491 B
A
FIG. 5. A: Sections of adult canine articular cartilage stained by the immunoperoxidase technique using sheep anti-canine type X collagen polysera. Specific reactivity is evident in the pericellular matrix around cells in the zone of calcified cartilage. Arrows indicate some sites of type X reactivity. 6: Control section of adult canine articular cartilage stained by omitting the primary antisera followed by peroxidase-conjugated secondary antisera. No specific staining is evident.
FIG. 6. Uncalcified-calcified cartilage interface of canine patella stained with the immunoperoxidase technique for type X collagen. Specific reactivity is evident around the cells at the tidemark (arrowheads). Section was counterstained with fast green to delineate between calcified and uncalcified tissues.
J Orthop Res, Val. 9, No. 4, 1991
492
J . M . GANNON ET AL.
FIG. 7. Higher power photograph of junction between the zone of calcified cartilage (ZCC) and the subchondral bone. Type X collagen staining can be seen to be remodeled away as the ZCC is replaced by mature bone.
distribution indicates that expression of type X precedes mineralization, as was observed in the growth plate. Thus, most of the staining is within the ZCC as delineated by the tidemark superiorly and the cement line inferiorly. Reactivity can be seen to be remodeled away as the ZCC is replaced with mature bone (Fig. 7). This reactivity was also competitively inhibited by preincubation of specific dilutions of anti-X polysera with known amounts of type X collagen. DISCUSSION Hyaline or articular cartilage contains collagen types 11, V, IX, and XI (13,30), occasionally type VI (13), and in special cases, type X. Our work with type X collagen in the canine evolved from our efforts to define an animal model ofjoint deterioration that reproduces changes observed in human osteoarthritis. Previous work in our laboratory (10) demonstrated that acute transarticular load to the adult canine patella-femoral joint causes reproducible swelling of the hyaline cartilage and alterations in the ZCC, including increased vascularity and duplication of the tidemark. Therefore, this research was undertaken to define the biology of mineralization of the ZCC and to compare these mechanisms with the biology of the growth plate. To compare the ZCC and the growth plate, we examined the
J Orthop Res, Val. 9, No. 4, 1991
expression of type X collagen, believed to be unique to these regions. The presence of type X collagen in mammalian tissue has been demonstrated in rabbit and bovine models (2,18,34); most of this research involved extraction of the molecule from growth plates or cultures of growth plate cartilage derived from young animals. Most recently in a preliminary report, the persistence of type X in adolescent and mature bovine ZCC was shown by chemical isolation (28). The immunohistochemical localization of canine type X collagen to the metaphyseal region of the physis is consistent with observations in the chick model. Both decalcification and proteoglycan digestion were necessary before immunoreactivity could be demonstrated. Type X has a peritrabecular distribution in the spongiosum, and the concentration progressively decreases toward the diaphysis of the fetal canine long bone, which is again consistent with the chick model. Work with the chick model has shown that type X collagen is present in cartilage matrix that contains hypertrophic chondrocytes. Although type X synthesis has been shown by cultured chondrocytes derived from permanent cartilaginous regions of the chick embryo sternum, little has been reported on the persistence of type X collagen in adult mammals. Therefore, we did not expect to see specific
TYPE X COLLAGEN IN CANINE TISSUE
staining in the adult ZCC. The identification of type X collagen in both the canine growth plate and the ZCC suggests that the ZCC in adult animals retains some characteristics of a growth plate. Whether type X collagen in adult articular cartilage is a growth plate remnant or is produced by hypertrophic chondrocytes in the adult remains to be determined. The function of type X collagen within the skeletal system remains unclear. Some investigators have suggested that it provides a permissive matrix for calcification and acts as a stimulus for chondrocyte removal. The expression of growth plate-like properties would allow the deep chondrocytes of mature articular cartilage to play a role in remodeling of the joint with age and in the pathogenesis of osteoarthritis. In summary, the presence of type X collagen in canine tissue has been demonstrated, and the distribution in the canine growth plate is consistent with observations in the chick model. The presence of type X collagen in permanent tissue, specifically the ZCC of adult canine articular cartilage, has also been shown and may indicate a similarity between the biology of mineralization of the growth plate and that of the ZCC of adult canine articular cartilage. Acknowledgment: This work was supported by NIH Grants No. AR39255 and AR07555 and in part by OREF Grant No. 89-492 from Bristol-Myers. We thank Thomas Schmid (Rush Presbyterian-St. Luke’s Medical Center, Chicago, IL, U.S.A.) for helpful advice and encouragement in the purification and characterization of type X collagen. We also thank Mark Oraskovich for help in developing the ELISA assay.
REFERENCES 1. Axen R, Porath J, Ernback S: Chemical coupling of peptides to polysaccharides by means of cyanogen halides. Nature 214~1302-1304, 1967 2. Ayad S, Kwan APL, Grant ME: Partial characterization of type X collagen from bovine growth plate cartilage. FEBS Lett 220:181-186, 1987 3. Bates GP, Schor SL, Grant ME: A comparison of the effects of different substrata on chondrocyte morphology and the synthesis of collagen types IX and X. In Vitro Cell Dev Biol 23:374-380, 1987 4. Brown CC, Balian G : Effect of heparin on synthesis of short chain collagen by chondrocytes and smooth muscle cells. J Cell Biol 105:1007-1012, 1987 5. Bullough PG, Yawitz PS, Tafra L, Boskey AL: Topographical variations in the morphology and biochemistry of adult canine tibia1 plateau articular cartilage. J Orthop Res 3: 1-16, 1985
493
6 . Castagnola P, Dozin B, Moro G, Cancedda R: Changes in the expression of collagen genes show two stages in chondrocyte differentiation in vitro. J Cell Biol 106:461467, 1988 7. Castagnola P, Torella G, Cancedda R: Type X collagen synthesis by cultured chondrocytes derived from the permanent cartilaginous region of chick embryo sternum. Dev Biol 123:332-337, 1987 8. Chin Yang LI, Ziesman SC, Lazcano-Villereal 0: Use of azide and hydrogen peroxide as an inhibitor for endogenous peroxidase in the IP method. J Hisrochem Cytochem 35: 1457-1460, 1987 9. De Jong ASH, Van Kessel-Van Vark M, Raap AK: Sensitivity of various visualization methods for peroxidase and alkaline phosphatase activity in immunoenzyme histochemistry. Histochem J 17:1119-1130, 1985 10. Donohue JM, Buss D, Oegema TR, Thompson RC: The effects of indirect blunt trauma on adult canine articular cartilage. J Bone Joint Surg [Am] 65:948-957, 1983 11. Eavy RD, Schmid TM, Linsenmayer TF: Development of the chick columella: Immunohistochemical studies with anticollagen monoclonal antibodies. Int J Pediatr Otorhinolaryngol 13:99-105, 1987 12. Eavy RD, Schmid TM, Linsenmayer TF: Intrinsic and extrinsic controls of the hypertrophic program of chondrocytes in the avian columella. Dev Biol 1265742, 1988 13. Eyre DR, Wu J, Apone S: A growing family of collagens in articular cartilage: Identification of 5 genetically distinct types. J Rheumatol 14(suppl 14):25-27, 1987 14. Eyre DR, Apon S, Wu JJ, Ericsson LH, Walsch KA: Collagen type IX: Evidence for covalent linkages to type I1 collagen in cartilage. FEBS Lett 220:337-341, 1987 15. Gannon JM, Fischer M, Hofmeister F , Carpenter R, Oegema TR Jr, Thompson RC Jr: Localization of type X collagen in fetal canine growth plate and adult canine articular cartilage. Trans Orthop Res Soc 1459, 1989 16. Gibson GJ, Kielty CM, Gamer C, Schor SL, Grant ME: Identification and partial characterization of three lowmolecular-weight collagenous polypeptides synthesized by chondrocytes cultured within collagen gels in the absence and in the presence of fibronectin. Biochem J 211:417426, 1983 17. Gibson GJ, Flint MH: Type X collagen synthesis by chick sternal cartilage and its relationship to endochondral development. J Cell Biol 101:277-284, 1985 18. Grant WT, Sussman MD, Balian G: A disulfide-bonded short chain collagen synthesized by degenerative and calcifying zones of bonive growth plate cartilage. J Biol Chem 260:379&3803, 1985 19. Grant WT, Wang GJ, Balian G: Type X collagen synthesis during endochondral ossification in fracture repair. J Biol Chem 262:9844-9849, 1987 20. Habuchi H, Conrad HE, Glaser JH: Coordinate regulation of collagen and alkaline phosphatase levels in chick embryo chondrocytes. J Biol Chem 260: 13029-13034, 1985 21. Jimenez SA, Yankowski R, Reginato AM: Quantitative analysis of type X collagen biosynthesis by embryonic chick sternal cartilage. Biochem J 233:357-367, 1986 22. Kielty CM, Holmes DJS, Schor SL, Grant ME: Embryonic chick cartilage collagens. FEBS Lett 169:179-184, 1984 23. Kielty CM, Kwan APL, Holmes DF, Schor SL, Grant ME: Type X collagen: A product of hypertrophic chondrocytes. Biochem J 27545-554, 1985 24. Kwan APL, Freemont AJ, Grant ME: Immunoperoxidase localization of type X collagen in chick tibiae. Biosci Rep 6: 155-162, 1986 25. Kwan APL, Sear CHJ, Grant ME: Identification of disul-
J Orthop Res, Vol. 9,N O . 4 , 1991
494
J . M . GANNON ET AL.
phide-bonded type X procollagen polypeptides in embryonic chick chondrocyte cultures. FEBS Lett 206:267-272, 1986 26. Laemmli UK: Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:68& 685, 1970 27. Linsenmayer TF, Gibney E, Schmid TM: Segmental appearance of type X collagen in the developing avian notochord. Dev Biol 113:467-473, 1986 28. Love11 TP, Eyre DR: Unique biochemical characteristics of the calcified zone of articular cartilage. Trans Orthop Res Soc 13511, 1988 29. Manduca P, Castagnola P, Cancedda R: Dimethyl sulfoxide interferes with in vitro differentiation of chick embryo endochondral chondrocytes. Dev Biol 125:234-236, 1986 30. Mayne R, Irwin MH: Collagen types in cartilage. In: Kuettner K, Schleyerbach R, Hascall VC, ed. Articular Cartilage Biochemistry. New York: Raven Press. 1986:23-38 31. Miller EJ, Rhodes K: Preparation and characterization of the different types of collagen. Methods Enzyrnol82:33-64,1982 32. Ninomiya Y, Gordon M, van der Rest M, Schmid TM, Linsenmayer TF, Olsen BR: The developmentally regulated type X collagen gene contains a long open reading frame without introns. J Biol Chem 2615041-5050, 1985 33. Quarto N, Cancedda R, Descali-Cancedda F: Purification and characterization of the low molecular mass (type X) collagen from chick embryo tibial cartilage. Eur J Biochem 147:397-400, 1985 34. Remington MC, Bashey RI, Brighton CT, Jimenez SA: Biosynthesis of a low molecular weight collagen by rabbit growth plate cartilage organ cultures. Cot1 Re1 Res 3:271278, 1983 35. Schmid TM, Conrad HE: Metabolism of low molecular weight collagen by chondrocytes obtained from histologi-
J Orthop Res, Vol. 9, No. 4, 1991
cally distinct zones of the chick embryo tibiotarsus. J Biol Chem 257:12451-12457, 1987 36. Schmid TM, Linsenmayer TF: Developmental acquisition of type X collagen in the embryonic chick tibiotarsus. Dev Biol 107~373-381,1985 37. Schmid TM, Linsenmayer TF: Immunohistochemical localization of short chain cartilage collagen (type X) in avian tissues. J Cell Biol 100:49&605, 1985 38. Schmid TM, Linsenmayer TF: Type X collagen. In: Mayne R, Burgeson RE, eds. Structure and Function of Collagen Types. Orlando: Academic Press. 1987:223-259 39. Solursh M, Jensen KL, Reiter RS, Schmid TM, Linsenmayer TF: Environmental regulation of type X collagen production by cultures of limb mesenchyme, mesectodern and sternal chondrocytes. Dev Biol 117:90-101, 1986 40. Summers TA, Irwin MH, Mayne R, Balian G: Monoclonal antibodies to type X collagen. J Biol Chem 263:581-587, 1988 41. Sussman MD, Ogle RC, Balian G: Biosynthesis and processing of collagens in different cartilaginous tissues. J Orthop Res 2:134-142, 1984 42. Thomas JT, Grant ME: Cartilage proteoglycan aggregate and fibronectin can modulate the expression of type X collagen by embryonic chick chondrocytes in collagen gels. Biosci Rep 8:163-171, 1988 43. Timpl R: Antibodies to collagens and procollagens. Methods Enzymol. 82: Academic Press, New York, 1982,pp 472-498 44. Uhl GR, ed: In Situ Hybridization in Brain. New York: Plenum Press. 1986:266-267 45. Yamaguchi N, Benya PD, van der Rest M, Ninomiya Y: The cloning sequence of al(VII1) collagen cDNAs demonstrate that type VIII collagen is a short chain collagen and contains triple-helical and carboxyl-terminal non-triple-helical domains similar to those of type X collagen. J Biol Chem 264:16022-16029, 1989
