Method for Inducing Bone Formation Using an Extract of Human Osteosarcoma Cell Line SAOS-2

This invention relates to a method for inducing bone formation in a mammal. More specifically, it relates to such a method in which a bone forming effective amount of an extract of human osteosarcoma cell line Saos-2 is administered to the subject. Also disclosed is the extract used in the method, and a process for making it.

The ability to induce bone formation and growth is an area of scientific inquiry which has seen a great deal of activity over the past 20-25 years. Initial observations by Urist et al, Science 150: 893-899 (1965), J. Dent. Res. 50: 1392-1406 (1971), that a substance residing in decalcified bone and now referred to as "bone morphogenic protein" ("BMP" hereafter) could induce ectopic cartilage and bone formation via endochonal ossification after implantation adjacent to skeletal muscles and away from preexisting bone have led to research to isolate and to characterize this material, as well as to identify sources for it. Efforts have also been directed to identifying and isolating other materials having the same or similar properties. For example, a material herein referred to as "BIA" for bone inducing agent has been shown to exist in some strains of transformed, cultured human epithelial and osteosarcoma cells. Exemplary of this are Anderson et al. Am. J. Path. 44:507-519 (1964) and J. Cell Biol. 33: 165-177 (FL transformed amnion cells); Anderson, et al, Fed; Proc; 27: 475 (1868) (HeLa cells); Wlodarski Exp; Cell Res; 57: 446-448 (1969) (WISH amnion cells); Wlodarski, et al Calcif. Tiss. Res. 7: 345-352 (1971); (neoplastic cells) Amitani et al., Gann 66: 327-329 (1975); Hanamura, et al., Clin. Orthop. & Rel. Res. 148: 274-280 (1980) (Dunn mouse osteosarcoma cells of BFO strain).

The patent literature in this field as it relates to BMP is growing. U.S. Pat. Nos. 4,294,753; 4,455,256; 4,526,909; 4,563,489; 4,596,574; 4,619,989; and 4,761,471, all issued to Urist, arise out of the work he has done on BMP. The most recent of these patents describes the isolation of BMP from frozen bone samples isolated from cadavers. Essentially, the bone was frozen, defatted, demineralized and freeze dried, after which a series of dissolving and dialyzing steps were carried out, ending in fractionation, dialysis, and SDS-PAGE analysis. The BMP is described as an acidic protein, with molecular weight of 17.5.+-.0.5 kd for the human variety, whereas bovine BMP has molecular weight of 18.5.+-.0.5 kd. The patent discusses how it is difficult to separate the bovine BMP from associated proteins, and how the associated proteins for both human and bovine BMP are necessary to optimize the effect of the protein on bone growth.

U.S. Patents have issued to others, as exemplified by U.S. Pat. Nos. 4,394,370; 4,472,840; 4,563,350; 4,642,120; and 4,968,590. For the most part, these patents deal with grafts for repairing osseous injuries. The '590 patent describes "pure mammalian osteogenic protein" and its uses. The protein, when glycosylated, has a molecular weight of 30 kd on SDS-PAGE, and is obtained from demineralized bone matrix. The bones are obtained from cattle. This protein may be identical to osteogenin, as described by Luyten et al, J. Biochem. 264: 13377-13380 (1989).

As the field has grown, it is convenient to refer to the class of proteins which have the specified effect as "BMPs". Thus, Wozney et al, Science 242: 1528-1534 (1989), have isolated genes for 3 BMPs, having molecular weights of 30, 18, and 16 kds. Recombinant BMPs have been produced via transfecting these genes into cultured monkey and hamster cells. Wang et al, Proc. Natl. Acad. Sci USA 87: 2220-2224 (1990), describe bone induction in animal when BMP-2A is implanted in GuHCl inactivated, decalcified bone matrix carriers. Sampath et al, J. Biol. Chem. 265: 13198-13205 (1990), describe "osteoinductive protein" (OP), also isolated from demineralized bovine bone. The OP is a 30 kd dimer, having 16+18 kd subunits. This protein also induces bone when implanted with a GuHCl inactivated bone matrix carrier. Bentz et al, J. Biol. Chem 264: 20805-20810 (1989) have described "osteoinductive factor" (OIF) from bovine bone, and characterize it as a 22-28 kd protein which must react with TGF-beta to have osteoinductive effect in vivo. Bessho et al, Biochem. Biophys. Res. Comm. 165: 595-601 (1989), also isolated an osteoinductive protein of 18 kd.

Isolating protein from bone is an onerous task, requiring enormous amounts of raw materials for extremely small yields. The amount of processing involved cuts the available yield as well. Further, the source of the raw material (bone), must generally be animal rather than human. Given the diversity of individuals within an animal species, non-uniformity of protein from batch to batch is to be expected albeit being undesirable.

Given the drawbacks discussed supra, the use of cultured cells as a source of bone inducing agents and/or bone morphogenic proteins presents an attractive alternative. Cell masses can be grown up quickly, and when a cell line is used, one expects the protein to be uniform, since cell lines by definition are uniform.

The aforementioned epithelial cell lines, as indicated, are osteoinductive, but they are osteoinductive only when live cells are injected. The experiments showing this demonstrated efficacy using immunosuppressed animals. See Anderson, Connect. Tiss. Res. 24: 3-12 (1990); Anderson, et al., Am. J. Patho 44: 507-519 (1964); Wlodarski, et al., Calcif. Tiss. Res. 5: 70-79 (1970); Anderson et al, Fed. Proc. 27: 475 (1968). Implantation of live, foreign cells into a subject with a healthy or non-suppressed immune system will almost inevitably result in a strong, perhaps long term immunological response, so administration of live cells is not advisable. Also, when the cell line is tumorigenic, as most cell lines are, the ramifications of administering live cells include serious risks of transplanting neoplasia into the subject. Monitoring the site of implantation with removal at a "critical point" is not possible, due to the risk of malignant invasion by the implant. The immediate suggestion, i.e., to introduce dead epithelial cells killed by freezing repeatedly into the subject, has proven to be ineffective. Devitalized, freeze dried, FL, WISH, and HeLa cells, as well as their extracts, have not been effective as osteoinducing agents. Anderson et al, Clin. Orthop. Rel. Res. 119: 211-224 (1976).

It is rarely the case that one can obtain an osteoinductive cell line which has an effect when the cell line is devitalized. Workers out of the lab of Amitani and Takaoka, as per Amitani et al, Gann 66: 327-329 (1975); Takaoka, et al., Clin Orthop. Rel. Res. 164: 265-270 (1982); Takaoka, et al., Clin Orthop. Rel. Res. 144: 258-264 (1989), have reported positive results with an osteoinductive BFO strain of Dunn murine osteosarcoma, and human cell line H-OS-6. The aforementioned laboratory has not made these cell lines available to other researchers, so this work cannot be verified or compared to results with other cell lines. Also, as the art well knows, the odds against securing a second cell line with properties identical to a first one are astronomical. To be denied access to a cell line, even if one is aware of its lineage, renders it impossible for the artisan to repeat reported work. Even in the case where a researcher reports a detailed, careful protocol for how a cell line was produced, isolated or derived, repetition of this work hardly guarantees reproduction of the initial results. Hence, the existence of cell banks, depositories, and the general cordiality of researchers in sharing cell lines.

When treating humans, it is desirable to treat with material as close to native human material as possible. U.S. Pat. No. 5,035,901, to Anderson at al., describes that crude extracts of cell line Saos-2 induce bone formation, even in devitalized form. The results reported therein were surprising because (i) the prior publications in the field suggested that dead human cell lines were not osteoinductive, and (ii) comparison with many other cell lines showed that the properties of Saos-2 were not shared, either among other human cell lines or among other mammalian cell lines.

The work on Saos-2 has continued. In U.S. patent application Ser. No. 683,283, filed Apr. 10, 1991 extracts of Saos-2 not only possess osteoinductive effect but show activity superior to that possessed by the whole cells. This is surprising because the BIA from Saos-2 cell line has not yet been isolated or characterized. Extraction protocols are known to damage some molecules, and in the process of extraction, one inevitably loses a portion of the starting material. With this lack of certainty in mind, it was thus more than slightly surprising that an Saos- 2 extract, prepared as described infra, was effective in inducing bone formation in subject animals (i.e., rats) that were immunocompetent and have rejected less purified preparations of BIA.

Further purification has now been carried out, and it has been found that extracts of Saos-2 contain, at the least, "bone morphogenic protein 4", or "BMP4" as it is referred to hereafter, as well as transforming growth factor beta (TGF-.beta.). Surprisingly, these materials are present in the non-bone-inducing human osteosarcoma cell line designated U2-OS, but in U2-OS cells are insufficient to provoke the bone inducing effect achieved with Saos-2, indicating that the Saos-2 cell line harbors at least a third, unidentified factor which, in combination with the listed materials, leads to the surprising properties characteristic of Saos-2 extracts. These new extracts are the subject of the invention and are described in the following disclosure.

Cell line Saos-2, described by Fogh et al, in Fogh, ed. Human Tumor Cells In Vitro (Plenum, N.Y. 115-159) (1975) was received as a gift; however, it may also be obtained from public sources, such as the American Type Culture Collection Under Accession Number ATCC HTB 85. This is a non-restricted culture.

The Saos-2 cells were seeded at a density of approximately 3.times.10.sup.4 cells/cm.sup.2 in either glass prescription bottles or in 150 cm.sup.2 plastic "T" flasks. They were grown to confluence in approximately 1 week in Dulbecco's MEM, with 10% fetal calf serum, 2.2 mM L-glutamine added, penicillin (20 U/ml), amphotericin (50 ng/ml), and streptomycin (20 mg/ml) in 5% CO.sub.2 atmosphere at 37.degree. C. After reaching confluence, the cells were trypsinized, subdivided 1-4 or 1-5 and reseeded for further cell production. Approximately 2.0.times.10.sup.7 confluent cells were released from one glass flask at time of trypsinization or scraping, and viability, studied by trypan blue exclusion, was 95% or greater.

Confluent cultures of Saos-2 cells were gently rinsed twice with a sterile 0.9% sodium chloride solution to remove medium and stored at -70.degree. C. until needed. For construction of the pellets of devitalized cells or BIA extracts, the cells were thawed, scraped with a rubber policeman and 8.0.times.10.sup.7 cell aliquots were pelleted in 50 ml conical tubes by centrifugation at 2000 rpm for 10 minutes. The supernatant was discarded and the cells were defatted by resuspending the pellets in two washes of acetone. (Each wash was 5 times the volume of pellet). After acetone treatment, the cellular material was sedimented at 2000 rpm for 10 minutes and air dried at 4.degree. C. for one hour. The resultant pellets were then freeze-dried, weighed and stored at -20.degree. C. in a sterile, air tight container.

In order to facilitate retention of the pellets at the implantation site a collagen gel product (Vitrogen, Collagen Corporation, California) was mixed with the devitalized and defatted Saos-2 cell material or extracts thereof. First, SAOS-2 cells (80.times.10.sup.6 cells per implant) were dispersed in 300 .mu.l of 0.01N HCl (40 .mu.l acid per 10.times.10.sup.6 cells) for 30 minutes at room temperature. Then, 700 .mu.l of Vitrogen collagen gel, equal to the volume of the cell pellet and acid was added to the acidified pellet suspension. After further mixing, storage at room temperature for 3 hours and addition of 300 .mu.l of 0.1M pH7 sodium phosphate buffer (the same volume as acid) the Saos-2 collagen pellets were stored at -20.degree. C. In some experiments where pellets had been prepared with fewer than 80.times.10.sup.6 cells, the quantity of collagen added was increased in order to maintain a relatively constant total volume of 1.4 to 1.6 ml per pellet. The Saos-2 collagen pellets were frozen, lyophilized and stored at -20.degree. C. until implantation.

The freeze-dried cell pellets were used either "as is", or following extraction protocols. Examples 3 and 4, which follow, describe one extraction protocol using 4M guanidium hydrochloride (GuHCl), although similar results, not reported here, were also obtained using 6M urea, all other parameters being the same.

The defatted, freeze-dried Saos-2 cells were resuspended in 4M guanidium hydrochloride (Sigma), 0.1 g protein/10 ml of GuHCl (w/v) and continuously stirred at 4.degree. C. for 48 hours. After extraction, the residual particulates were sedimented at 3000 RPM for 10 min. and discarded. The supernatant was placed in dialysis tubing with a 6000 to 8000 MW cut-off and dialyzed exhaustively against 0.05 M, pH 7.4 phosphate-buffered saline (PBS) at 4.degree. C. for 24 to 48 hours. During dialysis, a precipitate formed that was resuspended and freeze-dried as described above. Before GuHCl extraction, each defatted and freeze-dried pellet derived from 100.times.10.sup.6 Saos-2 cells weighed approximately 100 mg (dry weight).

Small pellets (0.7.times.0.3 cm) of freeze-dried defatted cells or of reprecipitated GuHCl extract of Saos cells were implanted under metaphane anaesthesia beneath the latissimus dorsi muscles of Nu/Nu mice (obtained from Charles River Labs) and the skin incision was closed by stainless steel staples. At first, animals were sacrificed and implants were recovered after three weeks, but it was found with experience that only ten days to two weeks is required for full expression of cartilage, bone and bone marrow at the implantation site. Similar implants were prepared using the rat osteosarcoma cell line UMR-106 or the human osterosarcoma cell line U20S, where, in each case the cells were also prepared as indicated supra.

At the time of sacrifice, the implants were removed, bisected, and one half of each implant was fixed for 24 to 48 hours in 10% phosphate-buffered formalin, embedded in paraffin and sectioned and stained by conventional histologic methods without a decalcification step. Although the implants showed extensive calcification after only 10 days, it was possible to make sections of the implants with reasonable facility without decalcification. In some instances the ossified implants were fixed in 2.5% glutaraldehyde in cacodylate buffer, post-fixed in osmium tetroxide, embedded in Spurr low-viscosity resin and sectioned at 1 micron thickness with toluidine blue stain for light microscopy or at 40 nM with lead citrate and uranyl acetate stains for electron microscopy. Electron microscopy was carried out using a Zeiss EM10-A electron microscope operating at 80 KV.

The unfixed halves of selected implants were homogenized in N-butanol, and the resultant butanol extract was analyzed for alkaline phosphatase specific activity following Hsu et al, Biochem. Biophys. Acta 913: 329-334 (1987) as a quantitative measure of ossification.

The unique qualities of Saos-2 as an osteoinductive cell line were shown when comparative experiments were carried out, using various other osteosarcoma cell lines. Specifically, human osteosarcoma cell lines U2-OS, TE85 and MG63 were tested. The first two cell lines are available from the ATCC (HTB96, CRL 1543), while MG63 was a gift.

Following the same protocols for cell culture, preparation of pellets, and implants, the three cell lines listed above, plus UMR-106 rat osteosarcoma cells, and Saos-2 were tested in a side by side comparison. Six mice were tested using UMR-106 (8 pellets), 2 mice for MG63 (14 pellets), 3 mice for U2-OS (3 pellets), and thirteen mice (17 pellets) for Saos-2. Following cervical dislocation and analysis of the pellets when removed, all Saos-2 pellets showed bone, and some were very hard. The UMR-106 U2OS MG-63 and TE-85 pellets remained soft, showing no bone cartilage or marrow formation.

The failure of any osteosarcoma cell line other than Saos-2 to elicit bone formation led to the decision not to proceed with further extraction experiments on these cell lines.

An additional experiment was carried out to determine whether defatting, i.e., the use of acetone, was required for efficacy.

Either extraction with acetone or freeze drying will devitalize a cell, and in these experiments, cells were simply devitalized via freeze drying. Freeze drying, it will be recalled, only removes moisture from the cell sample.

Again, following the general protocols described surra, cells were grown up, pelleted, and implanted, the only difference being that in one set of experiments, no acetone extraction took place. These are the non-defatted cell lines.

In the following Table 2, the results for formation of bone, cartilage and marrow formation are presented summarizing the results of experiments on 3 animals. In this Table "+++" indicates much induction, "++" moderate induction, "+" a small amount, and "0" none whatsoever.

Further purification of the Saos-2 derived material was carried out. Specifically, 100-500 mg samples of Saos-2 pellets, prepared as described supra, were dissolved in 25 volumes of 4M GuHCl, again as described supra. These materials were then centrifuged at 25,000 rpm, to separate cellular debris therefrom. Following removal of the debris, the resulting supernatant was applied to a Pharmacia Sephacryl-200 column (5.times.100 cm), pre-equilibriated with 4M GuHCl. Fractions in the molecular weight range of 10-60 kd were removed, pooled, and dialyzed against phosphate buffer, using 10,000 kd cutoff dialysis tubing. The dialyzed retentates were then pooled, lyophilized, and prepared for implant. In these experiments, 500 mg samples of dried Saos-2 pellets yielded about 25-35 mg of proteinaceous material.

The materials described in Example 8 were then used in in vivo experiments. Twenty adult male Long Evans rats were divided into 4 groups of 5 rats each. Each rat was treated to create a bone injury. Specifically, a lateral incision was performed to expose the right femoral diaphysis, and a 5-hole, 1.5 mm AO plate was then placed on the anterior surface, using four cortical screws. The plates were then swung away from femurs, and a 4 mm central diaphysial defect was created, using a water cooled burr.

The four groups were then divided into a control (group 1), a group treated with pure bovine collagen (group 2), a group treated with an autograft obtained by morselizing resected femur segments (group 3), and a group treated with 10 mg of the extract of Example 8, combined with 10 mg of pure bovine collagen (Example 4). The implants for groups 2, 3 and 4 were packed into one half of a #4 gelatin capsule, slid over the distal cut end of femur. The plate was secured, and hemostatic closure performed. Animals were radiographed at four and eight weeks. After four weeks, two animals were sacrificed from each group at random, and the three remaining animals were sacrificed after eight weeks. Two sets of experiments were carried out. Coronal sections through the defect were decalcified, and stained with hematoxalin and epsin.

The results are summarized in Table 3, which follows. Control animals showed non-union at four or eight weeks, as the defects were filled with fibrous tissue and/or skeletal muscle, and there was only a trace of new bone at resecton edges, and subperiosteally. The second group (collagen only), also showed no healing at either four or eight weeks. Even at eight weeks, unresorbed residual collagen and fibrous tissue spanned the gap, with only a small quantity of new bone near cut ends. Group 3 animals (autograft) showed notable callous formation, but there was non-union in all four and eight week animals, with most of the grafted particles of bone being gradually resorbed over eight weeks.

Group 4, utilizing the combination of collagen and extract, showed early union, with cartilage containing calluses at four weeks, and complete union by lamellar bone, with a reformed medullary canal by eight weeks. None of the groups exhibited inflammation.

Given the osteoinductive properties of Saos-2 cell line extract, it was of interest to determine if the cell line expressed any known bone morphogenetic proteins. The analysis was based upon a study of mRNA in the cells. Specifically, total RNA was prepared from exponentially growing Saos-2 cells, by lysing samples in 0.5% SDS in 10 mM EDTA (ph 8.0), followed by acidification with 0.1 M Na acetate (pH 2.2). The lysate was then treated with H.sub.2 O, equilibriated with ethanol to extract total RNA. The extracted RNA (20 ug) was size fractionated using agarose gel electrophoresis, blotted onto nylon membranes, and then hybridized with .sup.32 P labelled cDNA probes for known osteoinducers "BMP2", "BMP3", and "BMP4", using standard Northern blotting. Similar experiments were carried out using cDNA for TGF-.beta.. The results revealed that both BMP-4 and TGF-.beta. mRNA were present in the Saos-2 cells, but no RNA for BMP-2 and BMP-3 was found.

An experiment similar to that carried out on Saos-2 was performed on cell line U2-OS. The expression pattern was identical, i.e., both BMP4 and TGF-.beta. mRNA were found in U2-OS, but neither of BMP-2 or BMP-3 were found. In both of Examples 10 and 11, the tests with anti-TGF-.beta. antibody showed a stronger reaction with two high molecular weight proteins (65 kd, 72 kd), than with any material of the known molecular weight of TGF-.beta., i.e., about 12.5 kd. This suggests that some intermediate form of TGF-.beta. is present, although this is by no means certain.

The foregoing examples also demonstrate that bone induction was obtained when extracts of devitalized Saos-2 cell line were applied to the site of injury. Results were obtained not only with extracts obtained via treatment with "weak denaturing agents", i.e., those materials known to permit proteins to refold from whatever conformation is possessed by the denatured protein to the active form, but also from extracts which have been further treated, i.e., subjected to purification on an absorbing or separating gel. Specifically, those fractions in the molecular weight range of from about 10 kd to about 60 kd are effective. "Weak denaturing agents", as the term is used herein, is identical to that used Builder, e.g., U.S. Pat. No. 4,511,502, the disclosure of which is incorporated by reference. In addition to the GuHCl and urea solutions described herein, other solvents possessing the recited properties will be known to the skilled artisan, and need not be repeated here.

The absorbing or separating gel, as used herein, means any gel to which a protein containing molecule will adhere to when a solution passing through the gel contains the protein. "Sephacryl" was used in the experiments, however, any of the standard gels used in protein separation such as Sepharose, etc., may be used in its stead.

In experiments not reported herein, molecules of a larger molecular weight were also tested, but failed to induce ectopic bone formation in animals, using the same model reported herein. Thus, the smaller molecular weight fraction of the Saos-2 extract is clearly the active portion.

The analyses of osteoinductive factor mRNA in the cell extracts is significant. The results show that both of Saos-2 and U2-OS express BMP-4 and TGF-.beta.; however, it is also shown, supra, that U2-OS does not induce ectopic bone formation. Thus, as of yet at least one unidentified factor is believed to contribute to the osteoinductive properties of Saos-2. This factor does not appear to be present in the other osteosarcoma cell lines tested, as these do not show osteoinductive properties.

Thus, the invention is a method for inducing bone formation in a mammal, such as a human, by administering or treating the subject mammal in need of bone formation or healing with an amount of an extract of devitalized, freeze dried Saos-2 cells sufficient to promote bone healing or bone formation, wherein said extract is prepared by contacting freeze dried Saos-2 cells with a solvent containing a weak denaturing agent to form a supernatant and a sediment, separating the supernatant therefrom and treating the supernatant by contacting it to a separating gel, and removing 10-60 kd fractions from the gel. The 10-60 kd fractions contain the bone inducing activity and may be further treated by, e.g., dialyzing against a salt solution with a low molecular weight cut-off point. A "low molecular weight cut-off" is e.g., 6000-8000 kd. It will be understood that the specific dialysis step described herein, and other methods for concentrating the active material, available to the skilled artisan, including other dialysis protocols are optional. Following dialysis, the 10-60 kd "retentate" may be further treated to form a precipitate, pellet, solution, lyophilisate and so forth. This precipitate may be pelleted and freeze dried, and administered to the subject mammal in that form.

Also described herein is the precipitate used in the treatment of the subject immunocompetent animal, prepared as described surra as are the freeze dried, devitalized Saos-2 cells per se. The extract, i.e., the precipitate, can be further treated to form a pellet, or can be combined with a pharmaceutically acceptable carrier to form a composition useful in inducing bone formation in a mammal.

As with the devitalized Saos-2 cells, the extract described herein can be used in the treatment of mammalian and veterinary diseases and neoplasms. An example of the latter is the use of the extract in the repair of large fractures or displacements, such as in equine and related species. Such fractures admit to treatment via splinting and/or bone replacement, using the extract described herein, either alone or in combination with biodegradable materials or pharmaceutical carriers. Similar efficacy is envisioned for, e.g., repair of deviated limbs in canine and other related species, as well as spine fusion.

The extract may be useful for systemic as well as local treatment to augment bone growth, to prevent bone loss in metabolic diseases such as osteoporosis, and to speed fracture healing and bone repair and bone replacement via systemic administration. In short, pathological conditions which result in general or specific bone loss, such as osteoporosis can be treated using the materials described herein. Other conditions which may be usefully treated include the replacement of bone removed for tumor surgery, for maxillofacial repair, to speed rigid fusion of vertebral bodies in spinal surgery ("slipped discs," scoliosis, etc), bone grafting, to speed healing of traumatic fractures, to augment bonding of resected bone surfaces to porous, biocompatible prostheses, to effect repair of non-uniform fractures, and so forth. Any of the conditions which the art speaks of being treatable by BMP or BMPs may be treated with the subject extract.

The materials described herein are preferably combined with collagen to form material suitable for implant or other in situ administration. The collagen may be bovine collagen, or collagen derived from some other species, preferably one identical to the subject being treated, or from an evolutionarily related species. Autologous collagen is also preferred, as is collagen that has been treated in some way to render it non-immunogenic.

It will be understood that the specification and examples are illustrative but not limitative of the present invention and that other embodiments within the spirit and scope of the invention will suggest themselves to those skilled in the art.