Background: LIM mineralization protein-1 (LMP-1), an intracellular protein, is thought to induce secretion of soluble factors that convey its osteoinductive activity. Although evidence suggests that LMP-1 may be a critical regulator of osteoblast differentiation in vitro and in vivo, little is known about its mechanism of action. The purpose of the present study was to identify candidates for the induced secreted factors and to describe the time sequence of histological changes during bone formation induced by LMP-1.
Methods: Human lung carcinoma (A549) cells were used to determine if LMP-1 overexpression would induce expression of bone morphogenetic proteins (BMPs) in vitro. Cultured A549 cells were infected with recombinant replication-deficient human type-5 adenovirus containing the LMP-1 or LacZ cDNA. Cells were subjected to immunohistochemical analysis after forty-eight hours. Finally, sixteen athymic rats received subcutaneous implants consisting of collagen disks loaded with human buffy-coat cells that were infected with one of the above two viruses. Rats were killed at intervals, and explants were studied with histological and immunohistochemical analyses.
Results: In vitro experiments with A549 cells showed that AdLMP-1-infected cells express elevated levels of BMP-2, BMP-4, BMP-6, BMP-7, and TGF-β1 (transforming growth factor-beta 1) protein. Human buffy-coat cells infected with AdLMP-1 also demonstrated increased levels of BMP-4 and BMP-7 protein seventy-two hours after ectopic implantation in athymic rats, confirming the in vitro hypothesis.
Conclusions: The osteoinductive properties of LMP-1 involve synthesis of several BMPs and the recruitment of host cells that differentiate and participate in direct membranous bone formation.
Clinical Relevance: Ex vivo gene therapy with the LMP-1 cDNA-induced secretion of multiple BMPs may provide an alternative to implantation of large doses of a single BMP to induce new bone formation.
Akihito Minamide, MD, PhD; Scott D. Boden, MD; Manjula Viggeswarapu, PhD; Gregory A. Hair, PhD; Colleen Oliver, DVM; Louisa Titus, PhD; Department of Orthopaedic Surgery, Emory Spine Center, Emory University School of Medicine, 2165 North Decatur Road, Decatur, GA 30033. E-mail address for S.D. Boden: email@example.com
Animal and in vitro studies have demonstrated a striking and consistent bone-forming effect with ex vivo gene transfer of the LIM mineralization protein-1 (LMP-1) cDNA with use of very low doses of adenoviral or plasmid vectors 1,2. However, little is known about the mechanism of action of LMP-1, how long the transduced cells survive, or which osteoinductive growth factors and cells participate in the induction of new bone and osteoblast differentiation 1-3. Furthermore, the mechanism of bone formation in vivo (endochondral or membranous) has not been determined. Understanding the mechanism of LMP-1 action would be helpful for optimal control of LMP-1-induced bone formation in the clinical setting and to increase the understanding of intracellular signaling pathways involved with osteoblast differentiation.
LMP-1 is a member of the heterogeneous LIM-domain family of proteins and is the first member to be directly associated with osteoblast differentiation 4-7. LMP-1 was identified in messenger ribonucleic acid (mRNA) from rat calvarial osteoblasts stimulated by glucocorticoid and later isolated from an osteosarcoma complementary deoxyribonucleic acid (cDNA) library 3. Unlike BMPs (bone morphogenetic proteins), which are extracellular proteins that act through cell surface receptors, LMP-1 is thought to be an intracellular signaling molecule that is directly involved in osteoblast differentiation 8-13. Thus, therapeutic use of LMP-1 requires gene transfer of its cDNA. On the basis of its association with bone development and the results of suppression and overexpression experiments, LMP is considered to induce secretion of soluble factors that convey its osteoinductive activity and to be a critical regulator of osteoblast differentiation and maturation in vitro and in vivo 3,14-16.
The purposes of the present study were (1) to identify candidates for the secreted osteoinductive factors induced by LMP-1, (2) to describe the histological sequence and type of bone formation induced by LMP-1, and (3) to determine how long the implanted cells overexpressing LMP-1 survive in vivo.
Materials and Methods
Phase 1: Detection of LMP-1-Induced Osteoinductive Factors in Vitro
The human LMP-1 cDNA with the human cytomegalovirus promoter was cloned into a transfer vector and subsequently was transferred into a recombinant replication-deficient (E1, E3 deleted) adenovirus as previously described 2. Human lung carcinoma cells (A549) are known for their high infectivity by human type-5 adenovirus. These cells were seeded at a density of 50,000 cells/cm 2 on two-well chamber slides (Nalge Nunc International, Naperville, Illinois) and were propagated in F12 Kaighn medium (Gibco BRL Life Technologies, Grand Island, New York), supplemented with 10% fetal bovine serum, and grown in a humidified 5% CO2 incubator at 37°C.
The A549 cells were infected for thirty minutes at 37°C at a multiplicity of infection of 10 pfu (plaque-forming units)/cell. Medium with 10% fetal bovine serum was added and the cells were grown for forty-eight hours at 37°C. The cells were infected with either AdLMP-1 (active LMP) or AdLacZ (Adβgal-adenoviral control) each driven by the human cytomegalovirus promoter 1-3. As an additional negative control, some cells were not infected with adenovirus (no-treatment control). After forty-eight hours, the cells on the chamber slides were fixed for two minutes in 50% acetone/50% methanol and then were subjected to immunohistochemical analysis with use of antibodies specific for LMP-1, BMP-2, BMP-4, BMP-6, BMP-7, TGF-β1 (transforming growth factor-beta 1), MyoD, and type-II collagen.
Phase 2: Histological Sequence of Bone Formation in Vivo
The experimental protocol was reviewed and approved by the Institutional Animal Care and Use Committee and the Human Investigation Committee. Rabbit or human peripheral blood (3 mL) was obtained by venipuncture, and the buffy-coat cells were isolated by centrifugation at 1200 times gravity for ten minutes. The cells were counted, and 1 × 10 6 cells were infected with adenovirus (AdLMP-1 or AdLacZ) at a multiplicity of infection of 4.0 pfu/cell for ten minutes at 37°C. After infection, the cells were resuspended in a final volume of 80 μL and applied to a 7 × 7 × 3-mm disk of bovine type-I collagen.
Sixteen athymic rats (Harlan, Indianapolis, Indiana) that were four to five weeks old were obtained and were housed in sterile conditions. The rats were anesthetized by inhalation of 1% to 2% isoflurane. Four 10-mm skin incisions were made on the chest of the athymic rats, pockets were developed by blunt dissection, and a collagen disk containing cells was implanted into each pocket. Implants consisted of a collagen disk loaded with buffy-coat cells infected with either AdLMP-1 (two per rat) or AdLacZ (two per rat). The skin was closed with resorbable suture. The animals were killed at one, three, five, seven, ten, fourteen, twenty-one, and twenty-eight days after implantation, and explants were studied with histological and immunohistochemical analyses.
The specimens were fixed for twenty-four hours in 10% neutral buffered formalin and then were prepared for undecalcified or decalcified sectioning. The specimens for the undecalcified sections were dehydrated through graded strengths of ethanol and embedded in paraffin. The specimens examined at twenty-one and twenty-eight days after implantation were decalcified with 10% ethylenediaminetetraacetic acid (EDTA) solution for three to five days. After decalcification, the specimens were dehydrated through graded strengths of ethanol, embedded in paraffin, and sectioned at a thickness of 5 μm on a microtome (Reichert Jung Heidelberg, Germany). The sections were then stained with hematoxylin and eosin, stained with Goldner trichrome, or subjected to immunohistochemical analysis with use of antibodies specific for BMP-4, BMP-7, CD-45, and type-I collagen.
Preparation of Primary Antibodies
The anti-LMP-1 antibody is an affinity-purified rabbit polyclonal antibody mapping within an internal region of human LMP-1, and it reacts with LMP-1 of rabbit and human origin. This antibody was used to identify LMP-1 protein at a dilution of 1:500 or 1:1000.
Anti-BMP-2, Anti-BMP-4, Anti-BMP-6, Anti-BMP-7, and Anti-TGF-β1 Antibodies
Polyclonal goat anti-BMP-2, anti-BMP-4, anti-BMP-6, anti-BMP-7, and anti-TGF-β1 antibodies (Santa Cruz Biotechnology, Santa Cruz, California) cross-react with mouse, rat, and human BMPs. The anti-BMP-2, anti-BMP-4, and anti-BMP-6 antibodies were raised against an epitope mapping at the amino terminus of BMP-2, BMP-4, and BMP-6 of human origin. The anti-BMP-7 antibody was an affinity-purified goat polyclonal antibody mapping within an internal region of human BMP-7. The anti-TGF-β1 antibody was an affinity-purified goat polyclonal antibody mapping at the carboxy terminus of the precursor form of human TGF-β1. These antibodies were used at a dilution of 1:100 and 1:500 or 1:1000.
A monoclonal mouse anti-human leukocyte common antigen (LCA), CD-45 antibody (purified IgG 1, kappa; DAKO, Carpinteria, California) consists of two antibodies, PD7/26 and 2B11, directed against different epitopes 17,18. The PD7/26 was derived from human peripheral blood lymphocytes maintained on T-cell growth factor. The 2B11 was derived from neoplastic cells isolated from T-cell lymphoma or leukemia. Both antibodies bind to lymphocytes and monocytes at the 94% to 96% range when tested by immunofluorescence. In the present study, this antibody was used at a dilution of 1:100 for the identification of human leukocytes.
A monoclonal anti-type-I-collagen antibody (mouse IgG1 isotype; Sigma Chemical, St. Louis, Missouri) was derived from the collagen type-I hybridoma produced by the fusion of mouse myeloma cells and splenocytes from BALB/c mice immunized with bovine skin type-I collagen. The antibody reacts with human, bovine, rabbit, deer, pig, and rat type-I collagen and was used at a dilution of 1:100.
A polyclonal rabbit anti-type-II-collagen antibody (Santa Cruz Biotechnology) was raised against an epitope corresponding to the amino terminus of the alpha-1 chain of human type-II collagen. The antibody reacts with type-II-collagen alpha-1 chain of mouse, rat, and human origin and was used at a dilution of 1:1000.
A polyclonal rabbit anti-MyoD antibody (Santa Cruz Biotechnology) was raised against an epitope corresponding to amino acids 1-318 representing full-length MyoD protein of mouse origin. The antibody reacts with MyoD (and not myogenin, Myf-5, or Myf-6) of mouse, rat, and human origin and was used at a dilution of 1:1000.
The staining procedure was performed with use of the labeled streptavidin-biotin method (LSAB method). A kit (Universal LSAB Kit, Peroxidase; DAKO) was used for immunostaining with antibodies against LMP-1, BMP-2, BMP-4, BMP-6, BMP-7, TGF-β1, CD-45, MyoD, type-I collagen, and type-II collagen. Appropriate biotinylated secondary antibodies were used, depending on the animal in which the primary antibody was raised. Endogenous peroxidase was blocked with methanol containing 0.3% hydrogen peroxide. Specimens were incubated with phosphate-buffered saline solution containing 5% normal rabbit or goat serum and 1% bovine serum albumin for fifteen minutes at room temperature to avoid nonspecific binding and then with the appropriate concentrations of primary antibodies at 4°C overnight in a humidified chamber. After washing with phosphate-buffered saline solution three times for five minutes, followed by incubation with biotinylated secondary antibody and streptavidin-biotin-peroxidase complex in a humidified chamber for ten minutes at room temperature, color was developed with use of 3,3′-diaminobenzidine tetrachloride (DAB; DAKO). Finally, the sections were counterstained with hematoxylin. To serve as negative controls, each primary antibody was incubated at room temperature for three hours with the corresponding blocking peptide (Santa Cruz Biotechnology) (1:40 dilution) prior to incubation with the specimens. In some experiments, primary antibody alone or secondary antibody alone was used as an additional negative control.
Phase 1: Detection of LMP-1-Induced Osteoinductive Factors in Vitro
The A549 cells infected with AdLMP-1 showed strong intracellular staining for LMP-1 protein ( Fig. 1 ). Strong staining for BMP-2, BMP-4, and BMP-7 was observed in the AdLMP-1-treated cells, especially in the cytoplasm ( Fig. 2 ). The cells treated with AdLMP-1 also stained positive with anti-BMP-6 and anti-TGF-β1 antibodies ( Fig. 3 ); however, the reactions were somewhat less intense than those seen with the other BMPs. Neither the Adβgal-infected cells nor the untreated controls had any specific reaction for LMP-1, any of the BMPs, or TGF-β1. A blocking peptide for each antibody confirmed that the reaction was specific. There was no specific reaction with the anti-type-II-collagen or anti-MyoD antibodies.
Phase 2: Histological Sequence of Bone Formation in Vivo
Immunolocalization of CD-45 in leukocytes: At one and three days after implantation, cells stained with anti-CD-45 antibody were abundantly present in buffy-coat preparations within both the AdLMP-1 (active) and Adβgal (control) treated implants ( Fig. 4 ). The number of cells staining with the specific anti-human-CD-45 reaction decreased after day 3, especially in the center of the implants. Positive staining still was observed in the periphery of the implant at five days, but ten days after implantation there were few cells staining with anti-CD-45. The pattern of decreased staining was the same in active and control implants.
Immunolocalization of BMPs: Immunohistochemical analysis of the AdLMP-1 treated implants three and five days after implantation revealed strong BMP-4 ( Fig. 5 ) and BMP-7 ( Fig. 6 ) staining within cells on the collagen fibers. There was no specific staining for BMP-4 or BMP-7 in cells on the Adβgal (control) implants. Moreover, the strong staining with anti-BMP-4 and anti-BMP-7 antibodies was also seen in the AdLMP-1 implants at each time-point beyond ten days. Strong staining for BMP-4 and BMP-7 was observed in two temporal phases; the first phase of staining was seen in a limited number of buffy-coat cells in the early days (three and five days after implantation), and the second was seen after day 10, in osteoblast-like cells surrounded by matrix that most likely were responding cells rather than transplanted buffy-coat cells ( Fig. 7 ).
Immunolocalization of type-I collagen: Strong staining for anti-type-I-collagen antibody was observed in the AdLMP-1 implants seven, ten, fourteen, twenty-one, and twenty-eight days after implantation. At the early time-points, the specific reaction was seen adjacent to osteoblast-like cells and on the periphery of the cells themselves. There was minimal staining for type-I collagen in the control implants treated with Adβgal.
Hematoxylin and Eosin and Goldner Trichrome Staining
Because the results were the same regardless of whether rabbit or human buffy coat cells were used, the following description and corresponding illustrations will be for the human donor cells only. At one and three days after implantation, the Ad-LMP implants had increased numbers of cells at their edges ( Fig. 8 ), whereas fewer cells were seen at the periphery of the Adβgal controls at the same time-point. These observations suggest that host cells migrated into the implants with cells expressing LMP-1 ( Figs. 9-A and 9-B ). These cells were a mixture of monocytes, polymorphonuclear leukocytes, and histocyte-appearing cells.
There were fewer buffy-coat cells associated with the collagen fibers over time, and the number of cells surviving in the center of the Adβgal treated implants was diminished by five days after implantation ( Fig. 10 ). Seven days after implantation, new bone matrix was visible adjacent to osteoblast-like cells between collagen fibers at the periphery of the AdLMP-1 implants ( Fig. 11-A ). There was rapid mineralization of the surrounding matrix without classic osteoid seams and without any specific orientation. The lack of organized bone orientation was not surprising given the fact that these were subcutaneous implants that were not substantially loaded. More abundant osteoblast-like cells were observed in the AdLMP-1 implants ten days after implantation and were growing into the voids between the collagen fibers. By fourteen days after implantation, osteoblast-like cells occupied the central region of the AdLMP-1 implants. In contrast, fibroblast-like cells were filling the voids of the collagen in the Adβgal-treated implants. Twenty-one days after implantation, new bone matrix was mineralized and was forming in most or all of the central regions of the AdLMP-1 implants. Mature new bone had formed in the spaces located in the most central regions of the AdLMP-1 implants twenty-eight days after implantation. Osteoblasts were seen covering surfaces of osteoid and newly formed bone, while osteoclasts could be seen remodeling the primary woven bone ( Fig. 11-B ). Hematopoietic marrow tissue was also seen forming within the bone ( Fig. 11-C ). In the Adβgal-treated controls, the implanted collagen was mostly resorbed by day 28 and was replaced with fibrous tissue.
LMP-1 is a novel LIM-domain protein associated with early osteoblast differentiation. LMP-1 transcripts are first detectable in mesenchymal cells adjacent to the hypertrophic cartilage cells in developing embryonic long bones just before osteoblasts appear at the center of the cartilage anlage 3. The LMP-1 protein is a member of the heterogeneous family of LIM-domain proteins, many of which are involved with growth and differentiation in a variety of cell types. However, the precise mechanisms of action of LIM-domain proteins remain poorly understood 4-7. Although LMP-1 is an LIM-domain protein, we have recently shown that the LIM domains themselves are not necessary for osteoblast differentiation 19. LMP-1 is thought to be a potent intracellular signaling molecule that is capable, at very low doses, of inducing osteoblast differentiation in vitro and de novo bone formation in vivo, yet its mechanism of action remains unknown 3. We have hypothesized that LMP-1 may directly or indirectly result in the synthesis and secretion of one or more BMPs and possibly of ancillary proteins that enhance the activity of BMPs 3.
Four important results have emerged from this series of experiments concerning the mechanism of action of LMP-1. There is now compelling evidence from two separate experimental systems that LMP-1 induces the expression of several BMPs. The evidence is most compelling for BMP-4 and BMP-7, which can be detected as early as forty-eight hours after insertion of the LMP-1 cDNA in vitro and seventy-two hours after insertion in vivo. In vivo studies showed that most of the implanted buffy-coat cells expressing LMP-1 survived for less than a week in vivo, but there was indirect evidence of an influx of host cells that differentiated into bone-forming cells. Lastly, LMP-1 appears to induce membranous bone formation without a clear cartilage interphase, which is common with many of the BMPs.
The present study also showed that cells treated with AdLMP-1 produced LMP-1, BMP-2, and to a lesser extent BMP-6 and TGF-β1 protein. However, these results were tested only in the in vitro system and have not been evaluated in vivo, to our knowledge. BMP-4 and BMP-7 remain two strong candidates for secreted osteoinductive factors induced by LMP-1. We have performed preliminary antisense oligonucleotide experiments that suggested that BMP-4 and BMP-7 were necessary for the osteoinductive effects of LMP-1 to transfer to other cells (unpublished data), but those experiments did not demonstrate whether LMP-1 induced synthesis of these BMPs.
The importance of the A549 experiments in the present report was that they showed that the BMPs were not induced by the adenovirus itself and were not expressed in the untreated cells. The A549 experiments also showed that two proteins not related to osteoblast differentiation (type-II collagen and MyoD) were not induced by LMP-1. A549 lung carcinoma cells were chosen rather than osteoblasts because the A549 cells had no basal expression of BMPs. If we had used osteoblasts in our experiments, we would not have been able to make as direct a link between LMP expression and BMP induction. In osteoblasts, any nonspecific initiation of osteoblast differentiation would ultimately result in BMP expression, and the link to LMP expression would have been less clear. Finally, the in vivo experiments in human buffy-coat cells confirmed these observations in cells and an environment in which bone was actually forming to ensure that the observations were true in a physiologic bone-formation setting.
We recognize that LMP-1 may induce other proteins, including other BMPs or possibly helper proteins that facilitate the action/activity of very small amounts of BMPs as seen in physiologic bone-healing situations. This phenomenon would not be surprising given the high potency of small doses of LMP-1 and the difficulty of observing its induction of individual BMP proteins by less sensitive techniques such as Western blotting.
The use of buffy-coat cells from ordinary venous blood for ex vivo gene therapy is a relatively new concept 2. One relevant question raised has been how long the buffy-coat cells transfected with LMP-1 cDNA survive in vivo and enhance the synthesis, secretion, and activity of BMPs. To attempt to answer this question in the present study, we examined the CD-45 antigen, which is well known as a marker of white blood cells 17,18. The number of cells specifically reacting with the anti-CD-45 primary antibody decreased progressively and was minimal by ten days following implantation. The loss of anti-CD-45 staining, the dropout of cells in the center of the implant by seven days, and the centripetal pattern of bone formation all suggested that the transplanted cells, including those expressing LMP-1 cDNA, may not survive long. This observation suggests, but does not confirm, the notion that LMP-expressing cells may participate only indirectly in the bone formation process through induction of secreted factors that subsequently recruit host progenitor cells and modulate their differentiation into mature osteoblasts. LMP-1 seems to start a cascade of events, including the secretion of several osteoinductive proteins (BMPs). Therefore, we believe that the expression of LMP-1 does not need to occur in very many cells or need to persist for very long in vivo.
Our studies demonstrated the histological healing sequence of bone induced by ex vivo gene transfer of LMP-1 cDNA to peripheral blood buffy-coat cells implanted in an ectopic location. This work has begun to answer some of the questions about the mechanism of bone formation with LMP-1 at the macroscopic level. A better understanding of the mechanism of action of LMP-1 will facilitate its translation to the clinical setting and improve the understanding of intracellular signaling pathways involved in LMP action.
Investigation performed at the Department of Orthopaedic Surgery, Emory Spine Center, Emory University School of Medicine, Decatur, and the Atlanta Veterans Affairs Medical Center, Atlanta, Georgia
In support of their research or preparation of this manuscript, one or more of the authors received grants or outside funding from Medtronic Sofamor Danek, Atlanta Research and Education Foundation, Veterans Affairs Medical Center Merit Award, Veterans Affairs Research Enhancement Award Program, and the ERC Program of the National Science Foundation Award EEC-9731643. In addition, one or more of the authors received payments or other benefits or a commitment or agreement to provide such benefits from a commercial entity (licensing agreement with Medtronic Sofamor Danek). No commercial entity paid or directed, or agreed to pay or direct, any benefits to any research fund, foundation, educational institution, or other charitable or nonprofit organization with which the authors are affiliated or associated.
1. Boden SD, Titus L, Hair G, Liu Y, Viggeswarapu M, Nanes MS, Baranowski C. Lumbar spine fusion by local gene therapy with a cDNA encoding a novel osteoinductive protein (LMP-1). Spine
, 1998;23: 2486-92.
2. Viggeswarapu M, Boden SD, Liu Y, Hair GA, Louis-Ugbo J, Murakami H, Kim HS, Mayr MT, Hutton WC, Titus L. Adenoviral delivery of LIM mineralization protein-1 induces new-bone formation in vitro and in vivo. J Bone Joint Surg Am
, 2001;83: 364-76.
3. Boden SD, Liu Y, Hair GA, Helms JA, Hu D, Racine M, Nanes MS, Titus L. LMP-1, a LIM-domain protein, mediates BMP-6 effects on bone formation. Endocrinology
, 1998;139: 5125-34.
4. Kong Y, Flick MJ, Kudla AJ, Konieczny SF. Muscle LIM protein promotes myogenesis by enhancing the activity of MyoD. Mol Cell Biol
, 1997;17: 4750-60.
5. Sadler I, Crawford AW, Michelsen JW, Beckerle MC. Zyxin and cCRP: two interactive LIM domain proteins associated with the cytoskeleton. J Cell Biol , 1992;119: 1573-87.
6. Salgia R, Li JL, Lo SH, Brunkhorst B, Kansas GS, Sobhany ES, Sun Y, Pisick E, Hallek M, Ernst T, Tantravahi R, Chen LB, Griffin JD. Molecular cloning of human paxillin, a focal adhesion protein phosphorylated by P210BCR/ABL. J Biol Chem , 1995;270: 5039-47.
7. Way JC, Chalfie M. mec-3, a homeobox-containing gene that specifies differentiation of the touch receptor neurons in C. elegans. Cell
, 1988;54: 5-16.
8. Boden SD, Schimandle JH, Hutton WC, Chen MI. 1995 Volvo Award in basic sciences. The use of an osteoinductive growth factor for lumbar spinal fusion. Part I: Biology of spinal fusion. Spine
, 1995;20: 2626-32.
9. Cook SD, Wolfe MW, Salkeld SL, Rueger DC. Effect of recombinant human osteogenic protein-1 on healing of segmental defects in non-human primates. J Bone Joint Surg Am , 1995;77: 734-50.
10. Schimandle JH, Boden SD, Hutton WC. Experimental spinal fusion with recombinant human bone morphogenetic protein-2. Spine , 1995;20: 1326-37.
11. Spector JA, Luchs JS, Mehrara BJ, Greenwald JA, Smith LP, Longaker MT. Expression of bone morphogenetic proteins during membranous bone healing. Plast Reconstr Surg , 2001;107: 124-34.
12. Suzawa M, Takeuchi Y, Fukumoto S, Kato S, Ueno N, Miyazono K, Matsumoto T, Fujita T. Extracellular matrix-associated bone morphogenetic proteins are essential for differentiation of murine osteoblastic cells in vitro. Endocrinology , 1999;140: 2125-33.
13. Wozney JM, Rosen V, Celeste AJ, Mitsock LM, Whitters MJ, Kriz RW, Hewick RM, Wang EA. Novel regulators of bone formation: molecular clones and activities. Science
, 1988;242: 1528-34.
14. Boden SD, McCuaig K, Hair G, Racine M, Titus L, Wozney JM, Nanes MS. Differential effects and glucocorticoid potentiation of bone morphogenetic protein action during rat osteoblast differentiation in vitro. Endocrinology
, 1996;137: 3401-7.
15. Knutsen R, Honda Y, Strong DD, Sampath TK, Baylink DJ, Mohan S. Regulation of insulin-like growth factor system components by osteogenic protein-1 in human bone cells. Endocrinology , 1995;136: 857-65.
16. Yeh LC, Adamo ML, Duan C, Lee JC. Osteogenic protein-1 regulates insulin-like growth factor-I (IGF-I), IGF-II, and IGF-binding protein-5 (IGFBP-5) gene expression in fetal rat calvaria cells by different mechanisms. J Cell Physiol
, 1998;175: 78-88.
17. Kurtin PJ, Pinkus GS. Leukocyte common antigen-a diagnostic discriminant between hematopoietic and nonhematopoietic neoplasms in paraffin sections using monoclonal antibodies: correlation with immunologic studies and ultrastructural localization. Hum Pathol
, 1985;16: 353-65.
18. Pulido R, Cebrian M, Acevedo A, de Landazuri MO, Sanchez-Madrid F. Comparative biochemical and tissue distribution study of four distinct CD45 antigen specificities. J Immunol
, 1988;140: 3851-7.
Copyright 2003 by The Journal of Bone and Joint Surgery, Incorporated
19. Liu Y, Hair GA, Boden SD, Viggeswarapu M, Titus L. Overexpressed LIM mineralization proteins do not require LIM domains to induce bone. J Bone Miner Res
, 2002;17: 406-14.