"and rat extracellular matrix"
Proc. Nati. Acad. Sci. USA Vol. 80, pp. 6591-6595, November 1983 Cell Biology Homology of bone-inductive proteins from human, monkey, bovine, and rat extracellular matrix (cartilage/species specificity/alkaline phosphatase) T. K. SAMPATH AND A. H. REDDI Bone Cell Biology Section, Mineralized Tissue Research Branch, National Institute of Dental Research, National Institutes of Health, Bethesda, MD 20205 Communicated by Clifford Grobstein, August 12, 1983 ABSTRACT Allogeneic implantation of rat extracellular de- proteins of <50,000 daltons induces bone in vivo (11) and stim- mineralized diaphyseal bone matrix in subcutaneous sites induces ulates fibroblast proliferation in vitro (12). A recent study has a sequence of events resulting in the local differentiation of en- shown that these low molecular mass matrix fractions can in- dochondral bone. However, xenogenic subcutaneous implantation duce new bone formation in vivo when implanted intramus- of human, monkey, and bovine extracellular bone matrix into rat cularly (13). showed that bovine matrix had only a weak capacity to induce bone, There is a paucity of information regarding the species spec- whereas human and monkey matrix had none at all. This sug- ificity of matrix in bone induction (13, 14). In the present study gested that extracellular matrix-induced bone differentiation is we report that, although bone induction induced by intact ex- apparently species-specific. We recently reported that the ex- tracellular matrix is apparently species specific, the dissocia- traction of matrix with 4 M guanidine-HCI resulted in complete tively extracted and partially fractionated matrix components of removal of the ability to induce endochondral bone differentia- <50,000 daltons from human, monkey, and bovine bone matrix tion, with the biological activity of the matrix being again restored induce bone in rat after reconstitution with biologically inactive when the extracted active matrix components (<50,000 daltons) collagenous residue of rat. Our results reveal that there is a ho- were reconstituted with the inactive residue. To define the pos- mology in bone-inductive activity from human, monkey, bo- sible biochemical basis of species specificity, human, monkey, and vine, and rat extracellular matrix components. bovine extracellular bone matrices were extracted with 4 M guan- idine-HCI and the extracts were reconstituted with biologically in- MATERIALS AND METHODS active rat residue and bioassayed. The results were similar to those obtained with intact matrices and showed that total extracts of bo- Preparation of Demineralized Rat, Bovine, Monkey, and vine matrix had a weak capacity to induce bone, whereas cor- Human Bone Matrix. Dehydrated diaphyseal shafts of rat fe- responding extracts of human and monkey matrix did not induce murs and tibiae were pulverized in a CRC micromill (Techni bone. However, partial purification by gel filtration of 4 M guan- Laboratories, Vineland, NJ). Bone shafts were frozen in liquid idine HCI extracts from each species followed by reconstitution of nitrogen prior to and during pulverization to avoid possible heat the different fractions with inactive rat residue resulted in bone denaturation of matrix components. Pulverized bone particles induction by all species from fractions containing proteins of <50,000 were sieved to a discrete size of 74-420 ,Am. Matrix was de- daltons. These observations demonstrate that species specificity mineralized with 0.5 M HCl, extracted with water, ethanol, and of xenogenic extracellular bone matrix is due to immunogenic or ether, and prepared as described (4). Demineralization was ac- inhibitory components (or both) in the guanidine'HCI residue and complished either at 40C or at room temperature. Bovine de- solubilized extracellular matrix components of >50,000 daltons. mineralized bone matrix was prepared similarly and was made These results imply that there is homology in the bone inductive available by Dr. Joseph Nichols (Helitrex, Princeton, NJ). Dia- proteins from human, monkey, bovine, and rat extracellular bone physeal bones of monkey (Macaca fascicularis) were obtained matrices. from the Primate Research Center, National Institutes of Health. Human diaphyseal bone was a gift from M. Silbermann (Tech- There is a growing realization of the importance of extracellular nion, Haifa, Israel). Human and monkey demineralized bone matrix in adhesion, growth, and differentiation of cells (1-3). matrix was prepared as described above. Allogeneic implantation of rat demineralized extracellular dia- Dissociative Extraction. The demineralized bone matrix of physeal bone matrix in subcutaneous sites results in the local rat, bovine, monkey, and human was extracted (30 ml/g of ma- differentiation of endochondral bone. The sequential devel- trix) with constant stirring at 40C for 16 hr in 4 M Gdn-HCI/50 opmental cascade in bone induction includes: chemotaxis of mM Tris, pH 7.4, containing a mixture of protease inhibitors: mesenchymal cells, proliferation of progenitor cells, differ- 5 mM benzamidine-HCl/0. 1 M 6-aminohexanoic acid/0.5 mM entiation of cartilage and bone, and finally hematopoiesis (4- phenylmethylsulfonyl fluoride/5 mM N-ethylmaleimide (11). 8). The surface charge and geometry of the matrix are critical The extracts were centrifuged (40,000 X g, 30 min, 40C) and for bone induction (9, 10). We have recently reported that the the supernatants were dialyzed against water at 40C in Spec- extraction of matrix with 4 M guanidine-HCl (Gdn-HCl) re- trapor 3 tubing (Mr 3,500 cutoff) and lyophilized; the resi- sulted in the loss of the bone-inductive property. The loss of dues (insoluble demineralized bone matrix remaining after ex- the ability of matrix to induce endochondral bone differentia- traction) were washed three times in distilled water before tion could be restored by the reconstitution of soluble com- lyophilization. Each residue and each extract for each species ponents with inactive residue (11). Molecular sieve chroma- were bioassayed for their potential to induce endochondral bone tography ofthe Gdn-HCl extract showed that a fraction containing differentiation. Reconstitution. Portions of the extract from bovine, mon- The publication costs of this article were defrayed in part by page charge key, and human demineralized -bone matrices were reconsti- payment. This article must therefore be hereby marked "advertise- ment" in accordance with 18 U.S.C. §1734 solely to indicate this fact. Abbreviation: Gdn-HCl, guanidine hydrochloride. 6591 6592 Cell Biology: Sampath and Reddi Proc. Natl. Acad. Sci. USA 80 (1983) tuted with samples of extracted residues from their respective Table 1. Dissociative extraction and reconstitution of human, matrix or from rat matrix. Similarly, portions of the extract from monkey, bovine, and rat matrix for bone induction rat demineralized bone matrix were reconstituted with samples Alkaline Calcium from bovine, monkey, and human residue fractions. Reconsti- phosphatase, content, tution was accomplished by alcohol precipitation: 1 part (wt/wt) unit(s)/mg pg/mg Bone of lyophilized extract was dissolved in 2 ml of 4 M Gdn HCl; Group of protein of tissue histology 10 parts (wt/wt) of the insoluble demineralized bone matrix Rat matrix 2.56 ± 0.20 12.93 ± 1.67 +++ residue was added. The mixture was stirred for 2 hr at 40C. E (rat) + R (rat) 1.92 ± 0.52 7.43 ± 2.56 +++ Cold absolute ethanol (8.5 ml) was added to the mixture, which was then stirred for 30 min at 40C. After centrifugation (10,000 Bovine matrix 0.91 ± 0.12* 3.52 ± 0.84* + X g, 10 min, 40C), the supernatant was discarded. The recon- E (bovine) + R (rat) 1.05 ± 0.20* 4.90 + 1.43* + stituted matrix was washed three times with 85% ethanol in water and then lyophilized. Column fractions obtained by gel Monkey matrix 0.14 ± 0.02* 0.21 ± 0.02* - filtration were also reconstituted with residues in a similar fash- E (monkey) + R (rat) 0.32 ± 0.15* 0.75 ± 0.30* - ion. Gel Filtration. Aliquots of 4 M Gdn HC1 extracts obtained Human matrix 0.16 ± 0.04* 0.11 ± 0.06* - from demineralized rat, bovine, monkey, and human bone ma- E (human) + R (rat) 0.36 ± 0.44* 0.25 ± 0.29* - trix were applied (150 mg/2 ml) to two columns in tandem (2.6 Values are mean ± SEM of eight observations from four rats. E, x 100 cm) of Sepharose CL-6B, equilibrated in 4 M Gdn HCl/ Gdn HCl extract; R, Gdn HCl residue. 50 mM Tris, pH 7.0, and eluted with the same eluent at a flow * For difference from rat matrix, P < 0.01. rate of 15 ml/hr; 5-ml fractions were collected. The eluent was continuously monitored at 230 nm. Appropriate fractions were implanted (Fig. 1 A and B). Implantation of bovine matrix pooled (see Fig. 2), dialyzed against distilled water at 40C, and showed few islands of cartilage and bone histologically (Fig. lyophilized. Fractions were characterized by NaDodSO4/poly- 1C). Subcutaneous implantation of rat demineralized bone ma- acrylamide gel electrophoresis and bioassayed for endochondral trix (74-420 um) into allogeneic rats resulted in the local dif- bone differentiation activity after reconstitution with rat matrix ferentiation of bone. Bone formation is evident histologically, residue. as indicated by osteoblasts in apposition to implanted matrix Bioassay. Demineralized bone matrix of rat, bovine, mon- (Fig. LD). To rule out that acid demineralization inactivated the key, and human and the variously reconstituted bone matrix bone inductive activity, EDTA demineralization was also em- preparations were bioassayed for their ability to induce endo- ployed. The results were similar to those obtained with acid- chondral bone by subcutaneous implantation into ether anes- demineralized bone matrix for each species (data not shown). thesized male Long-Evans rats (28-35 day) at bilateral sites lo- Dissociative Extraction and Reconstitution. Dissociative ex- cated over the thorax. The day of implantation was designated traction of rat matrix with 4 M Gdn HCI resulted in a complete as day 0 of the experiment. On day 12, the subcutaneous but- ton-like plaques (implants) were dissected out and cleaned of loss of the ability of the residual matrix to induce endochondral adherent tissue. The implants were weighed and homogenized bone differentiation. Subcutaneous implantation of lyophilized in 2 ml of ice-cold 0.15 M NaCl/3 mM NaHCO3. Alkaline extract alone was also without effect (11). The biological activity phosphatase activity of the supernatants and calcium content of of rat matrix could be restored by reconstitution of the inactive the acid-soluble fraction of the sediment (by atomic absorption) residue with the extracted soluble components either by di- were determined as indices for bone formation as described (11). alysis (11) or by alcohol precipitation, as measured by the spe- The histological appearance of the implants was examined by cific activity of alkaline phosphatase and calcium content in day- fixing the implants in Bouin's fixative and embedding in JB4 12 plaques (10). However, an identical dissociative extraction plastic medium (Polyscience, Warrington, PA). One-microm- and reconstitution of bovine, monkey, and human matrix or the eter sections were cut and stained with toluidine blue. extracts of rat matrix when reconstituted with extracted resi- Polyacrylamide Slab Gel Electrophoresis. Dissociatively ex- dues from bovine, monkey, and human matrix did not induce tracted components of rat, bovine, monkey, and human matrix bone (data not shown). Extracellular components of bovine, were characterized by polyacrylamide slab gel electrophoresis monkey, and human bone matrix solubilized by 4 M Gdn HCl as described (11, 15). Gradient gels (5-20%) were used as the were reconstituted with rat matrix residue by alcohol precip- separating gel and a spacer gel of 3% acrylamide/2 M urea was itation and were bioassayed for endochondral bone differen- added to all of the gels. Electrophoresis was carried out in 50 tiation activity. The results were similar to those obtained with mM Tris glycine, pH 8.3/0.1% NaDodSO4 at 15 mA per slab intact tissues and revealed that extracts of bovine matrix had and 20°C for 12 hr. After electrophoresis, the gels were stained a weak capacity to induce bone and that extracts of monkey and with 0.25% Coomassie brilliant blue R 250 in 50% methanol/ human did not induce bone (Table 1). 10% acetic acid for 45 min and then were destained in 10% Fractionation of 4 M Gdn HCI-Solubilized Extracellular methanol/7.5% acetic acid. Protein was determined by the Components of Bone Matrix. A gel filtration analysis of 4 M method of Lowry et al. (16). Gdn&HCl-solubilized components of rat, bovine, monkey, and human on Sepharose CL-6B revealed similar column profiles, as shown in Fig. 2. Fractions assigned I-V were pooled, di- RESULTS alyzed against water at 4TC, and lyophilized. Lyophilized com- Influence of Rat, Bovine, Monkey, and Human Matrix. ponents of each fraction were reconstituted with 4 M Gdn HCI- Xenogenic implantation of bovine, monkey, and human matrix extracted rat residue and bioassayed. The results showed that in rat showed that bovine matrix had a weak capacity to induce only fraction IV of rat, bovine, monkey, and human restored bone and monkey and human matrix did not induce bone, as the activity to the rat matrix residue (Fig. 3). Histological ap- indicated by the specific activity of alkaline phosphatase and pearance of bone was evident (Fig. 1 A IV-D IV). The other calcium content in day-12 plaques (Table 1). Cartilage and bone fractions did not show biological activity. It is noteworthy that were absent histologically when human and monkey matrix was fraction III (apparent molecular mass, 50,000-100,000 daltons), Cell Biology: Sampath and Reddi Proc. Natl. Acad. Sci. USA 80 (1983) 6593 FIG. 1. Photomicrographs of histological sections of subcutaneous implants obtained from rats on day 12. (x 180.) (A) Human bone matrix: im- plants of demineralized human bone matrix did not induce bone formation. The implants consist of bone matrix (M), mesenchymal cells, and mul- tinucleate giant cells. Reconstitution of partially purified fraction IV of Gdn*HCl extract of human matrix with rat Gdn-HCI residue resulted in bone induction (A IV). Arrows indicate osteoblasts. (B) Monkey bone matrix: implants of demineralized monkey bone matrix did not induce bone formation. The implants consist of bone matrix (M), mesenchymal cells, and multinucleate giant cells. However, reconstitution of partially purified fraction IV of Gdn-HCI extract of monkey bone matrix with the rat Gdn*HCl residue resulted in bone induction (B IV). Arrows indicate osteoblasts. (C) Bovine bone matrix: implants of demineralized bovine bone matrix induced cartilage and traces of early bone formation (arrows). The chon- drocytes are in apposition to implanted matrix. Reconstitution of fraction IV of Gdn-HCl extract of bovine bone matrix with rat Gdn-HCI residue resulted in bone induction (C IV). It is noteworthy that in comparison to bovine matrix, the fraction IV of extract is more potent. Arrows indicate osteoblasts. (D) Rat bone matrix: implants of demineralized rat bone matrix induced new bone formation. Reconstitution offraction IV of Gdn-HCl extract of rat bone matrix with Gdn-HCl residue of rat resulted in bone induction (D IV). Arrows indicate osteoblasts. 6594 Cell Biology: Sampath and Reddi Proc. Natl. Acad. Sci. USA 80 (1983) Control A 4 20 3 15 2 10 0 1 ]fI ~ f~4 5 0. a) a) 0 4 co 101t bn 15 . co $.) 0 5 us E ._-' Control C 0 ta a.) 0 20S A: 4 To Mf 3 3 15 CID es4 -1:~ 4- 2 10 Q w 1 1 JnL m 5 -4 Control D 14 20 15 10 5 I II III IV V D Fraction 1.5 FIG. 3. Reconstitution of various fractions obtained on Sepharose CL-6B from 4 M Gdn-HCl extract of human (A), monkey (B), bovine (C), and rat (D) bone matrix. In each experiment the control represents the 1.0 F implantation of demineralized rat bone matrix. Among all of the spe- cies examined, only fraction IV has the bone-inductive activity, as quantitated by alkaline phosphatase (units/mg of protein), calcium 0.51- content (,ug/mg of tissue), and histology. tion. Further experiments revealed that the components that L 0 F - rI 60 80 R 100 I m 120 I T 140 I M . 160 I y 180 200 may be immunogenic or perhaps inhibitory to bone induction are present in the 4 M Gdn HCl residue or that there are pro- teins with an apparent molecular mass of >50,000 daltons in Fraction number the Gdn HCl extract (or both). An additional finding that de- serves comment is that the reconstitution of matrix compo- FIG. 2. Gel filtration of the 4 M Gdn HCI extract of human (A), nents for bioassay can be considerably shortened by the alcohol monkey (B), bovine (C), and rat (D) bone matrix on Sepharose CL-6B. The fractions (I-V) are pooled as indicated and bioassayed for bone in- precipitation technique as compared to the direct dialysis pro- duction by the reconstitution assay as described (see text). cedure reported earlier (11). This permits assay of several sam- ples concurrently. when reconstituted with the 4 M Gdn HCl residue, consis- The precise role of the collagenous matrix in bone induction tently elicited multinucleate giant cell formation. Electropho- is not yet clear. Other experiments (9, 10) reveal that the ge- resis of Sepharose CL-6B fraction IV of rat, bovine, monkey, ometry of the collagenous residue has a profound influence on and human showed proteins of <50,000 daltons. bone induction. The collagenous matrix may provide a suitable substratum for anchorage-dependent proliferation and differ- DISCUSSION entiation of cells (17-19). The present study further illustrates this and shows, in addition, that reconstitution with xenogenic The experiments described address the species specificity of Gdn HCl-extracted residue is ineffective, despite the fact that bone matrix-induced local bone differentiation. The salient the major component present in all species is type I collagen. finding of this study is that the bone-inductive potential of di- Furthermore, as demonstrated by the xenogenic reconstitution verse mammals appears to be homologous. It is likely that there experiments, the apparent species specificity may be due, in is a family of related proteins that are involved in bone induc- part, to the collagenous residue. Cell Biology: Sampath and Reddi Proc. Natl. Acad. Sci. USA 80 (1983) 6595 The sequential multistep cascade of events elicited in re- 3. Reddi, A. H. (1983) in Biochemistry of Extracellular Matrix, eds. sponse to extracellular bone matrix is reminiscent of the changes Piez, K. A. & Reddi, A. H. (Elsevier, New York), in press. observed in fracture healing in long bones and in amphibian 4. Reddi, A. H. & Huggins, C. B. (1972) Proc. Nati. Acad. Sci. USA limb regeneration (3). It is conceivable that extracellular matrix 69, 1601-1605. 5. Reddi, A. H. & Huggins, C. B. (1975) Proc. Natl. Acad. Sci. USA of bone may play a role in specifying morphogenetic infor- 72, 2212-2216. mation locally at the site of fracture and during normal bone 6. Reddi, A. H. & Anderson, W. A. (1976)J. Cell Biol. 69, 557-572. remodeling. In this regard extracellular bone matrix in the solid 7. Reddi, A. H. (1981) Collagen Rel. Res. 1, 209-226. state may function as an affinity matrix to locally govern the 8. Urist, M. R. (1965) Science 150, 893-899. function of chemotactic, mitogenic, and differentiation regu- 9. Reddi, A. H. & Huggins, C. B. (1973) Proc. Soc. Exp. Biol. Med. lating proteins (3, 10-12) in skeletal tissues, which are endowed 143, 634-637. 10. Reddi, A. H. (1976) in Biochemistry of Collagen, eds. Ramachan- with enormous potential for regenerative growth, repair, and dran, G. N. & Reddi, A. H. (Plenum, New York), pp. 449-478. remodeling. 11. Sampath, T. K. & Reddi, A. H. (1981) Proc. Nati. Acad. Sdi. USA In conclusion, the results described indicate that there is ho- 78, 7599-7603. mology among the osteoinductive proteins from diverse species 12. Sampath, T. K., DeSimone, D. P. & Reddi, A. H. (1982) Exp. Cell of mammals. The apparent species specificity of matrix-in- Res. 142, 460-464. duced bone formation is due to species-specific immunogens or 13. Urist, M. R., DeLange, R. & Finerman, G. A. M. (1983) Science 220, 680-686. inhibitors (or both) present in the Gdn HCl-extracted collage- 14. Huggins, C. B., Wiseman, S. & Reddi, A. H. (1970)J. Exp. Med. nous residue and the Gdn HCI-solubilized extracellular bone 132, 1250-1258. matrix components of >50,000 daltons. The principles learned 15. Laemmli, U. K. (1970) Nature (London) 227, 680-685. from the present investigation can be applied to the rational 16. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. design of matrix materials for optimal predictable bone induc- (1951)]. Biol. Chem. 193, 265-275. tion to correct acquired and congenital craniofacial anomalies 17. Stoker, M., O'Neill, C., Berryman, S. & Waxman, V. (1968) Int. J. Cancer 3, 683-693. and skeletal defects (20). 18. Folkman, J. & Greenspan, H. P. (1975) Biochim. Biophys. Acta 1. Grobstein, C. (1975) in Extracellular Matrix Influences on Gene 417, 211-236. Expresson, eds. Slavkin, H. & Greulich, R (Academic, New York), 19. Gospodarowicz, D., Greenburg, G. & Birdwell, C. R. (1978) pp. 9-16. Cancer Res. 38, 4155-4171. 2. Hay, E. D. (1977) in International Cell Biology, eds. Brinkley, B. 20. Glowaki, J., Kaban, L. B., Murray, J. E., Folkman, J. & Mulli- R. & Porter, K. R. (Rockefeller Univ. Press, New York), pp. 50- ken, J. B. (1981) Lancet i, 959-963. 57.