AU2006226993B2 - Mechanisms of osteoinduction by LIM mineralization protein-1 (LMP-1) - Google Patents

Description

WO 2006/102417 PCT/US2006/010419 MECHANISMS OF OSTEOINDUCTION BY LIM MINERALIZATION PROTEIN-1 (LMP-1) BACKGROUND OF THE INVENTION 5 Many Americans are afflicted by low back pain, degenerative spinal disease, or bone fractures. These musculoskeletal problems are responsible for a major portion of the health care budget and are among the greatest causes of chronic disability and lost productivity in the United States. Orthopaedic surgical treatment of these problems frequently requires bone grafting to promote healing. Fusion of two or more bones with 10 cancellous bone graft may fail to heal in 25-45% of patients, and in even higher percentage of smokers and diabetic patients, co-morbidities which are more prevalent in the Veteran population. Use of osteoinductive proteins such as BMP-2 to induce bone formation in these patients is now possible. In 2002 the U.S. Food and Drug Administration approved rhBMP-2 for use as a bone graft substitute in interbody spine 15 fusions. Despite this regulatory milestone for BMP-2, this technology is not feasible for many patients with bone healing needs due to an unexpectedly high dose required in humans which has resulted in a very high cost (Boden SD, Zdeblick TA, Sandhu HS, and Heim SE. Spine 2000;25:376-81; Ackerman SJ, Mafilios MS, and Polly DW, Jr. Spine 2002;27:S94-S99). 20 A 15,000-fold higher concentration of BMP-2 is required to induce bone in humans (1.5 mg/mL) than in cell culture (100 ng/mL). Thus, without a dramatic improvement in BMP-2 responsiveness, healthcare economics may severely limit translation of one of the most seminal discoveries related to osteoblast differentiation in the last 50 years from helping large numbers of patients. 25 Consequently, a further understanding of the complex regulation of BMP-2 during osteoblast differentiation and the cellular responsiveness to such important bone forming proteins is critical so that their effect can be enhanced or their required dose limited to a more affordable quantity of protein especially in the most challenging orthopaedic healing environment - posterolateral lumbar spine fusion. 30 Several years ago a novel intracellular LIM domain protein critical to fetal and post-natal bone formation was identified (Boden SD, Liu Y, Hair GA et al. Endocrinology 1998;139:5125-34).

2 Termed LIM Mineralization Protein (LMP-1) it was the first LIM domain protein to be directly associated with osteoblast differentiation. Blocking LMP-l 5 expression prevents osteoblast differentiation in vitro, suggesting a critical functional role of this novel intracellular protein. Leukocytes expressing the LMP-l cDNA (via plasmid or adenoviral transduction) that are implanted into rabbits or athymic rats induce bone formation in bony and ectopic locations (Boden SD, Titus L, Hair G et al. Spine 1998;23:2486-92). The feasibility of LMP-I delivery by ex vivo gene therapy for 10 spine fusion and bone defect applications in rabbits and primates is currently being evaluated. LMP-l also has considerable potential as a local, regional, or systemic anabolic strategy for increasing bone density in patients with osteoporosis. However, before clinical applications can be seriously considered it will be critical to understand the mode of action of this protein. The present invention addresses this problem. 15 Osteoblasts are thought to differentiate from pluripotent mesenchymal stem cells, the maturation of which results in the secretion of an extracellular matrix which can mineralize and form bone. The regulation of this complex process involves a group of signaling glycoproteins known as bone morphogenetic proteins (BMPs), members of the transforming growth factor-beta (TGF-B) superfamily. Some BMPs are uniquely 20 capable of initiating the entire osteoblast differentiation cascade and BMP-2 is one of the most extensively studied. With Applicant's discovery that LMP-I can dramatically increase cellular responsiveness of mesenchymal stem cells (MSCs) to BMP-2 and mechanistic elucidation of various aspects of the signaling pathway of LMP-l, the present invention 25 provides combinatorial strategies including small molecules and peptide mimics, to overcome the dose-related translational barriers for BMP-2 therapeutics. Any discussion of documents, acts, materials, devices, articles or the like which has been included in the present specification is not to be taken as an admission that any or all of these matters form part of the prior art base or were common general 30 knowledge in the field relevant to the present invention as it existed before the priority date of each claim of this application.

2A SUMMARY OF THE INVENTION In a first aspect, the present invention provides a method for inducing deposition and maturation of bone in a subject having compromised bone conditions comprising a 5 co-therapeutic treatment regimen wherein at least one LMP agent and a subtherapeutically effective dose of at least one BMP agent are administered to a subject in need of such treatment. In a second aspect there is provided a method for decreasing the time required to form new bone in the presence of a BMP agent comprising co-administering at least 10 one LMP agent. In a third aspect, there is provided an osteogenic composition comprising at least one LMP agent; at least one BMP agent; and 15 optionally, at least one pharmaceutically acceptable carrier. In a fourth aspect there is provided an osteogenic composition comprising at least one Smurf binding agent; and at least one BMP agent; and 20 optionally, at least one pharmaceutically acceptable carrier. In a fifth aspect, there is provided an osteogenic composition comprising at least one phosphorylated Smad 1 competitive binding agent and at least one BMP agent. 25 In a sixth aspect there is provided an osteogenic composition comprising at least one phosphorylated Smad 5 competitive binding agent and at least one BMP agent.

2B In an seventh aspect there is provided a process for engineering bone tissue comprising combining at least one LMP agent and at least one BMP agent with a cell 5 selected from the group consisting of osteogenic cells, pluripotent stem cells, mesenchymal cells, and embryonic stem cells. In an eighth aspect there is provided an osteogenic composition comprising at least one phosphorylated Smad 4 competitive binding agent and 10 at least one BMP agent. In a ninth aspect there is provided a process for engineering bone tissue comprising combining at least one phosphorylated Smad 4 competitive binding agent and at least one BMP agent with a cell selected from the group consisting of osteogenic 15 cells, pluripotent stem cells, mesenchymal cells, and embryonic stem cells. In a tenth aspect, there is provided a method for inducing deposition and maturation of bone in a subject having compromised bone conditions comprising administering to the subject at least one JABI inhibiting agent and a therapeutically 20 effective dose of at least one BMP agent. Throughout this specification the word "comprise", or variations such as "comprises" or "comprising", will be understood to imply the inclusion of a stated element, integer or step, or group of elements, integers or steps, but not the exclusion of 25 any other element, integer or step, or group of elements, integers or steps. DETAILED DESCRIPTION The present invention derives from studies designed to elucidate the mechanism of LIM mineralization protein (LMP) action in modulating growth factor responsiveness in WO 2006/102417 PCT/US2006/010419 -3 cells, tissues and organisms. As a result of these studies, Applicant's have discovered that there is an unexpected synergistic result when an osteogenic composition comprising at least one LIM mineralization protein and at least one growth factor is administered to induce bone formation in a subject in need of bone repair, replacement or augmentation, 5 for example subject suffering from compromised bone conditions. Overview The studies disclosed herein demonstrate that LMP-1 enhances responsiveness to BMP-2 in MSCs. To elucidate the mechanism underlying this observation it is further 10 demonstrated that LMP- 1 interacts in vitro with an 85 kDa protein, identified as Smurfi, a key regulator of the degradation of BMP-2 signaling molecules, Smad1 and Smad5. It is also demonstrated here that endogenous Smurf1 and LMP-1 co-immunoprecipitate from cells, suggesting the physiological relevance of the interaction. The importance of the Smurfl/LMP-1 interaction is further documented by the fact that LMP-1 overexpression 15 increases levels of phosphorylated Smad1 (P-Smadl) in the nucleus and increases expression of BMP-2 regulated genes, expected outcomes of Smurfl/LMP-1 interaction. LMP- 1-induced inhibition of Smurf1 WW domain antibody binding to Smurf identified the WW domain as the region of Smurf1 that LMP-1 binds. Further, analysis of LMP-1 sequence has identified two potential WW domain interacting motifs within an 20 osteoinductive region of LMP-l. It is also demonstrated that LMP-1 increases BMPR1A levels in support of the hypothesis that LMP interrupts the Smurfl/Smad6 mediated degradation of the BMP receptor. LMP-1 is shown herein to interact with Jab 1, an adaptor protein which regulates degradation of the common Smad, Smad4 resulting in increased nuclear Smad4. 25 Furthermore, identified herein is the precise region of LMP which interacts with Smurfl. This discovery facilitates design of small compounds that mimic LMP's effects. The compounds include small proteins and peptides. In addition, the ability to use a single exposure dose of a recombinant TAT-LMP fusion protein is demonstrated, confirming that continuous LMP- 1 expression is not required for an effective therapeutic outcome and 30 opens the door for design of an LMP-mimic small compound. Also discovered is a novel interaction between LMP- 1 and Smurfi which represents a powerful control mechanism over BMP signaling and responsiveness. This WO 2006/102417 PCT/US2006/010419 -4 LMP-1 interaction occurs with the Smurf WW2 domain, is dependent on a specific PY motif in LMP-1, and can be mimicked by a small peptide containing only that motif. Further, LMP- 1 competitively binds to Smurfl, preventing ubiquitin-mediated proteasomal degradation of Smads, contributing to an enhanced cellular responsiveness to 5 BMP-2. These findings allow for the design of small molecule therapeutics that more efficiently control responsiveness of the BMP signaling pathway which would make clinical translation easier. Such small molecules would be more easily synthesized, stored, and delivered for clinical use to induce bone formation alone or with much lower doses of BMP-2 than are currently required in the clinical setting. 10 Thus, therapeutics that modulate the effects of LMP-1 have the potential to either replace BMP-2 as a strategy to induce bone formation or to serve as a method to enhance the efficacy of rhBMP-2, lowering the dose and cost of its use as an inducer of bone formation. Combination therapy 15 The invention relates to treatment of diseases using combination therapy. In particular, the novel LMP agents described herein may be used in conjunction with BMP agents. The present invention provides a method of inducing bone deposition by co administration of at least one LMP agent and a therapeutically effective dose of at least one BMP agent. It has been found that LMP agents are capable of accelerating bone 20 formation by enhancing the BMP agent's responsiveness. In the method of the invention, the LMP agent accomplishes this by affecting a BMP agent including but not limited to endogenous BMP protein, exogenous BMP protein, exogenous BMP protein fragment, and exogenous BMP protein variant fragment. The present invention may therefore be used to decrease the time required to form new bone in the presence of a BMP agent 25 comprising administering at least one LMP agent. As used herein LIM mineralization protein (LMP) "LMP" includes LMP- 1 and biologically active fragments thereof, LMPlt and biologically active fragments thereof, and LMP-3 and biologically active fragments thereof. More detailed descriptions, including sequences, can be found in US Patent 6,300,127, pending application USSN 30 10/951,236, and pending application USSN 09/959,578 filed by Boden et al., the entire teachings of which are incorporated herein by reference. LMP-2 is excluded as it is non- WO 2006/102417 PCT/US2006/010419 -5 osteogenic. Growth factors suitable in the invention include bone morphogenic proteins (BMP) including BMP-2. As used herein the term "LMP agent" includes a functional fragment of an LMP protein, a functional fragment of an LMP protein with a protein transduction domailn 5 (PTD) attached, an LMP protein with a PTD attached, an LMP protein without a PTD attached, a functional fragment of an LMP protein variant, an LMP protein variant with a PTD attached, an LMP protein variant without a PTD attached, an oligonucleotide sequence encoding any of the above, and an LMP gene. As used herein, the term "BMP agent" includes a functional fragment of a BMP 10 protein, a functional fragment of a BMP protein with a PTD attached, a BMP protein, a functional fragment of a BMP protein variant, a BMP protein variant, an endogenous BMP protein, exogenous BMP protein, an exogenous BMP protein fragment, an exogenous BMP protein variant fragment, an oligonucleotide sequence encoding any of the above, and a BMP gene. Particularly useful BMP is BMP-2, especially rhBMP-2. 15 Protein and peptide variants and derivatives Those skilled in the art will understand that one may make many molecules derived in sequence from the aforementioned LMP agents or BMP agents in which amino acids have been deleted ("deletion variants"), inserted ("addition variants"), or substituted ("substitution variants"). Molecules having such substitutions, additions, deletions, or any 20 combination thereof are termed individually or collectively "variant(s)." Such variants should, however, maintain at some level (including a reduced level) the relevant activity of the unmodified or "parent" molecule (e.g., an LMP variant possesses the ability to modulate BMP responsiveness or to bind Smurfl). Hereinafter, "parent molecule" refers to an unmodified molecule or a variant molecule lacking the particular variation under 25 discussion. There are two principal variables in the construction of amino acid sequence variant(s): the location of the mutation site and the nature of the mutation. In designing variant(s), the location of each mutation site and the nature of each mutation will depend on the biochemical characteristic(s) to be modified. Each mutation site can be modified individually or in series, e.g., by (1) deleting the target amino acid residue, (2) inserting 30 one or more amino acid residues adjacent to the located site or (3) substituting first with conservative amino acid choices and, depending upon the results achieved, then with more radical selections.

WO 2006/102417 PCT/US2006/010419 -6 An amino acid sequence addition may include insertions of an amino- and/or carboxyl-terminal fusion ranging in length from one residue to one hundred or more residues, as well as internal intra-sequence insertions of single or multiple amino acid residues. Internal additions may range generally from about 1 to 20 amino acid residues, 5 preferably from about 1 to 10 amino acid residues, more preferably from about 1 to 5 amino acid residues, and most preferably from about 1 to 3 amino acid residues. An example of an amino- or a carboxy-terminus addition includes chimeric proteins comprising the amino-terminal or carboxy-terminal fusion of the parent molecules with all or part of a transduction peptide or other conjugate moiety. 10 Amino acid sequence deletions generally range from about 1 to 30 amino acid residues, preferably from about 1 to 20 amino acid residues, more preferably from about 1 to 10 amino acid residues and most preferably from about 1 to 5 contiguous residues. Amino-terminal, carboxy-terminal and internal intrasequence deletions are contemplated by the present invention. As used herein a "functional fragment" of a protein is any 15 fragment or portion of a protein which retains the characteristic of interest of the parent protein or peptide. As used herein "biologically active" means retaining that characteristic or property in question from the parent molecule. In one embodiment the protein or peptide may possess multiple activities such as would be provided by multiple binding sites. These binding sites or domains may be 20 identical or variable and may be in sequence or separated by non-binding site amino acids. In yet another embodiment, recombinant proteins, peptides or fusion proteins may be produced. In another embodiment the LMP agents or BMP agents of the present invention are conjugated to other proteins or peptides. Protein transduction domains (PTDs) and 25 attachment of these to proteins and peptides are contemplated. In one embodiment of the present invention, the PTD is the HIV-TAT protein. In one embodiment, a variant protein or peptide will preferably be substantially homologous to the amino acid of the parent molecule or a portion or a domain of the parent molecule from which it is derived. The term "substantially homologous" as used 30 herein means a degree of homology that is in excess of 80%, preferably in excess of 90%, more preferably in excess of 95% or most preferably even 99%. Homology is determined WO 2006/102417 PCT/US2006/010419 -7 relative to the smaller peptide or variant and is measured across that domain, site or fragment in the parent from which the variant or peptide is derived. The invention also comprises chemically modified derivatives of the parent molecule(s) in which the peptide is linked to a nonproteinaceous moiety (e.g., a polymer) 5 in order to modify its properties. These chemically modified molecules are referred to herein as "derivatives". Such derivatives may be prepared by one skilled in the art given the disclosures herein. Conjugates may be prepared using glycosylated, non-glycosylated or de-glycosylated parent molecule(s) and suitable chemical moieties. Typically non glycosylated molecules and water-soluble polymers will be used. Other derivatives 10 encompassed by the invention include post-translational modifications (e.g., N-linked or 0-linked carbohydrate chains, processing of N-terminal or C-terminal ends), attachment of chemical moieties to the amino acid backbone, and chemical modifications of N-linked or O-linked carbohydrate chains. The polypeptides may also be modified with a detectable label, such as an enzymatic, fluorescent, isotopic or affinity label to allow for detection 15 and isolation of the protein peptide. Water-soluble polymers are desirable because the protein or peptide to which each is attached will not precipitate in an aqueous environment, such as a physiological environment. Preferably, the polymer will be pharmaceutically acceptable for the preparation of a therapeutic product or composition. One skilled in the art will be able to 20 select the desired polymer based on such considerations as whether the polymer/protein conjugate will be used therapeutically and, if so, the therapeutic profile of the protein (e.g., duration of sustained release; resistance to proteolysis; effects, if any, on dosage; biological activity; ease of handling; degree or lack of antigenicity and other known effects of a water-soluble polymer on a therapeutic proteins). 25 Variants and/or derivatives may be screened to assess their physical properties in vitro and can be subsequently screened in vivo in the models described herein. It will be appreciated that such variant(s) will demonstrate similar properties to the unmodified molecule, but not necessarily all of the same properties and not necessarily to the same degree as the corresponding parent molecule. 30 Oligonucleotides Oligonucleotide sequences of the present invention include those polymeric nucleic acid sequences which would "code for" the protein or peptide of interest. Those of WO 2006/102417 PCT/US2006/010419 -8 ordinary skill in the art will appreciate the degeneracy of the genetic code and that variable codons may still produce the same protein on translation. As used herein the term "oligonucleotide" includes nucleic acid sequences which code for the proteins or peptides of the invention or their parent molecules, including but not limited to the LMP and BMP 5 agents and vectors encoding said agents as well as small interfering RNAs (siRNAs) designed to target the genes disclosed herein, especially those involved in BMP and LMP signalling pathways. Particular oligonucleotides of the present invention include siRNAs designed to LMP-1, Smurfl, Smurf2 and Jabl. Pharmaceutical compositions 10 The invention also provides for pharmaceutical compositions in the form of an osteogenic composition. As used herein an "osteogenic composition" is a composition comprising a therapeutically effective amount of at least one BMP agent combined with at least one other agent and optionally, at least one pharmaceutically acceptable diluent, carrier, solubilizer, emulsifier, preservative and/or adjuvant. In a preferred embodiment, 15 the osteogenic composition comprises a therapeutically effective amount of at least one LMP agent and at least one BMP agent and optionally, at least one pharmaceutically acceptable diluent, carrier, solubilizer, emulsifier, preservative and/or adjuvant. It is understood that the phrase "at least one" includes one and more than one; for example, two, three, four or more. 20 In another embodiment an osteogenic composition comprises at least one Smurf binding agent; and at least one BMP agent; and optionally, at least one pharmaceutically acceptable diluent, carrier, solubilizer, emulsifier, preservative and/or adjuvant. In the present invention, Smurf binding agents include LMP-1 protein, and LMP-1 protein fragments, variants or derivatives, and siRNA specific for Smurf mRNA. The Smurf is 25 selected from the group consisting of Smad Ubiquitin Regulatory Factor-1 (Smurfl) and Smurf 2. In yet another embodiment, the osteogenic composition comprises at least one phosphorylated Smad 1 competitive binding agent and at least one BMP agent. In the present invention, the phosphorylated Smad 1 competitive binding agent includes but is 30 not limited to an LMP-1 protein, and LMP-1 protein fragment, variant or derivative, and siRNA specific for phosphorylated Smad 1.

WO 2006/102417 PCT/US2006/010419 -9 In another embodiment, the osteogenic composition comprises at least one phosphorylated Smad 5 competitive binding agent and at least one BMP agent. In the present invention, the phosphorylated Smad 5 competitive binding agent includes but is not limited to an LMP- 1 protein, and LMP- 1 protein fragment, variant or derivative, and 5 siRNA specific for phosphorylated Smad 5. Also disclosed is an osteogenic composition comprising at least one phosphorylated Smad 4 competitive binding agent and at least one BMP agent. In the present invention, the phosphorylated Smad 4 competitive binding agent includes but is not limited to an LMP-1 protein, and LMP-1 protein fragment, variant or derivative, 10 siRNA specific for phosphorylated Smad 4. Pharmaceutically acceptable diluents, carriers, solubilizers, emulsifiers, preservatives and/or adjuvants are known to those skilled in the art. These include but are not limited to cells, vectors, gels, microspheres, macromolecules, biocompatible foams, biocompatible matrices, and implants. Compositions may also comprise incorporation of 15 any of the therapeutic molecules or agents into liposomes, microemulsions, micelles or vesicles for controlled delivery over an extended period of time. The term "therapeutically effective amount" means an amount which provides a therapeutic effect for a specified condition and route of administration. Whether an amount is therapeutically effective may be determined on a stand alone or combinatorial basis. Consequently, what might 20 represent a therapeutically effective amount of one agent may change when that agent is combined with a further agent. In a preferred embodiment of the invention, an LMP agent is administered in combination with a BMP agent, wherein the dose or amount of BMP agent is subtherapeutic as compared to conventional rhBMP-2 therapy alone. As a stand alone therapy, a BMP agent at subtherapeutic doses or amounts would not, therefore, be 25 therapeutically effective but when combined with the LMP agents of the present invention would represent a therapeutic dose or amount. Administration and Dose Depending on dosage form, the pharmaceutical compositions of the present invention may be administered in different ways, i.e., intrathecal injection, subcutaneous, 30 intravenous, intraperitoneal, intramuscular injection, in an implant or combinations thereof. The administration of the LMP agents of the present invention may occur before, after or simultaneously with the BMP agent and may be to a single targeted site or WO 2006/102417 PCT/US2006/010419 -10 separate sites. Sites for administration include, but are not limited to an intervertebral space, a facet joint, site of a bone fracture, bones of the mouth, chin and jaw, and an implant site. In yet another embodiment, the therapeutic methods of the invention further 5 comprises a co-therapeutic treatment regimen comprising administering a therapeutically effective amount of an LMP agent in combination with a therapeutically effective amount of a BMP agent to treat disease in a patient. As used herein a "co-therapeutic treatment regimen" means a treatment regimen wherein two agents are administered simultaneously, in either separate or combined formulations, or sequentially at different times separated by 10 minutes, hours or days, but in some way act together to provide the desired therapeutic response. Dosages of the LMP agent, BMP agent or compositions of the present invention may range from lnM to 200nM if delivered as a recombinant fusion protein with a PTD attached or from 0.1 to 100 MOI (multiplicity of infection, i.e. number of infectious viral 15 particles per cell) if delivered by an adenovirus or similar vector. In combination, it is understood that doses of one agent may be lowered when the dose of a co-administered agent is raised. For example, it is contemplated that on raising the dose of an LMP agent, the dose of the BMP agent administered may be lowered. It is also contemplated with synergistic compositions, that administration of one component of a combination will 20 mitigate the need for an equal dose of a second component as is demonstrated herein with the LMP agent synergistically increasing the responsiveness of cells to a BMP agent. In one embodiment, therapeutically effective dose of BMP agent is less than the currently acceptable therapeutically effective amount. Currently, 1.5mg/mL of bone formed is the therapeutic concentration of rhBMP-2 in primates in vivo with smaller doses 25 effective in cell culture and rodents. The therapeutically effective dose of said at least one BMP agent is at least 10-fold less than the dose required in conventional therapy. The dose required in conventional therapy can be 20 mg rhBMP-2 per site of 10 cc of bone formation. In other embodiments, the therapeutically effective dose of BMP is at least 20 fold, 50-fold, 100-fold, 1000-fold, 5000-fold or 10,000-fold less than the dose required in 30 conventional BMP therapy when administered in combination with an LMP agent.

WO 2006/102417 PCT/US2006/010419 -11 Treatment outcomes The therapeutic agents and compositions of the present invention are useful in treating subjects having compromised bone conditions. The invention provides for a method of treating a bone disorder using a therapeutically effective amount of an LMP 5 agent in combination with a BMP agent. The bone disorder or compromised bone condition may be any disorder characterized by bone loss (osteopenia or osteolysis) or by bone damage or injury. Such bone conditions include but are not limited to broken bones, bone defects, bone transplant, bone grafts, bone cancer, joint replacements, joint repair, fusion, facet repair, bone degeneration, dental implants and repair, bone marrow deficits 10 and other conditions associated with bone and boney tissue. Examples of bone defects include but are not limited to a gap, deformation or a non-union fracture in a bone. Examples of bone degeneration include but are not limited to osteopenia or osteoporosis. In one embodiment, the bone defect is due to dwarfism. 15 The invention is especially useful for joint replacement or repair wherein the joint is vertebral, knee, hip, tarsal, phalangeal, elbow, ankle, sacroiliac or other articulating/non articulating joint. Bone tissue engineering 20 Furthermore, the engineering and use of cell- and animal-based models of bone disease in which the compounds of the invention may be useful are also described. The invention includes a process for engineering bone tissue comprising combining at least one LMP agent and at least one BMP agent with a cell selected from the group consisting of osteogenic cells, pluripotent stem cells, mesenchymal cells, and embryonic stem cells. 25 Also, disclosed is the engineered bone tissue produced by the above process. A method for inducing bone formation in a subject comprising administering the engineered bone tissue of the present invention is contemplated. Further, the invention includes a process for engineering bone tissue comprising combining at least one phosphorylated Smad 4 competitive binding agent and at least one 30 BMP agent with a cell selected from the group consisting of osteogenic cells, pluripotent stem cells, mesenchymal cells, and embryonic stem cells. Also included is engineered bone tissue produced by this process. In another aspect, the invention includes a method WO 2006/102417 PCT/US2006/010419 -12 for inducing bone formation in a subject comprising administering the engineered bone tissue as described in this paragraph. Also included in the invention is a method for inducing deposition and maturation of bone in a subject having compromised bone conditions comprising administering to the 5 subject at least one Jab1-inhibiting agent and a therapeutically effective dose of at least one BMP agent. Methods and Procedures Cell culture: Mesenchymal stem cells, (MSCs) at passage 2 are purchased from Cambrex Bio Sciences. Cells are grown at 37'C in 5% CO 2 in MSCBM media supplemented with 10 MSCGM Singlequots (Cambrex Bio Sciences), split at confluence, and plated at 3 x 104 cells/well in 6-well dishes at passage 4 in these studies. The next day treatments are applied in the presence of 50 uM L-Ascorbic Acid 2-Phosphate and 5 mM p-glycerol phosphate (Sigma-Aldrich). Medium is changed every 3-4 days with re-application of treatments where appropriate. Cells are transduced for 30 min with adenoviral constructs 15 in 0.5 ml serum free medium. rhBMP-2 will continue to be supplied as a gift from Wyeth (Genetics Institute) courtesy of John Wozney. Preparation of nuclear and cytoplasmic protein fractions: Cell pellets are suspended in buffer A (20 mM HEPES, pH 7.9, 10 mM KCl, 1 mM EGTA, 1 mM EDTA, 0.2 % Nonidet P-40, 10% Glycerol, 1 mM PMSF and lug/ml protease inhibitor mix (Sigma)), 20 incubated on ice for 10 min, and centrifuged. Supernatants (cytoplasmic fraction) are collected and nuclear pellets are suspended in high salt buffer B (buffer A plus 600 mM KCl, 20% glycerol), incubated on ice for 30 min and centrifuged. Supernatants are collected as the nuclear fraction. After protein determination, fractions are subjected to SDS-PAGE. 25 Measurement of phosphorylated Smad1 and Smad5 in the nucleus on overexpression of LMP-1: To show that increased levels of LMP-1 result in increased levels of phosphorylated Smad1 and Smad5, hMSCs are treated with the doses of LMP-1 that successfully synergize with BMP-2 to determine the timecourse of increased of phosphorylated Smad1 in the nucleus when the agents are applied alone or together. 30 Human MSCs are plated at 3 x 10 4 cells/well in 6-well plates, grown overnight, and treated with Ad5F35-LMP-1 (0, 5, 10 pfu/cell), BMP-2 (100 ng/ml) or both agents (control = Ad5F35-GFP). After 1, 2, 4, 8, 12, 24 and 48 hrs, cells are harvested and WO 2006/102417 PCT/US2006/010419 -13 nuclear proteins are analyzed by Western blot for Smad proteins. Both Smad1 and Smad5 are measured. Antibody to phosphorylated Smad1 is available and is used for Western analysis. Antibody to phosphorylated Smad5 is not available so an antibody that detects both phosphorylated and unphosphorylated Smad5 is used; and a phosphoserine antibody 5 on a separate blot is used to determine the phosphorylation state. Although there will be several molecules having phosphoserine, comparison of the two blots should allow determination of whether P-Smad5 also increases in the nucleus. Treating cells with the same adenoviral vector carrying GFP cDNA as a control is not expected to have an effect on nuclear levels of P-Smads. Increased P-Smad levels are expected to occur rapidly in 10 response to BMP-2, since that involves phosphorylation of existing Smad proteins to activate the intracellular signaling cascade. In contrast, LMP-1 cDNA must be transcribed and translated into an intracellular protein, a process requiring several hours. Thus a delay in increased nuclear levels of P-Smads is expected in response to LMP-1 as compared to the response to BMP-2. 15 SDS-PAGE and Westen blotting: SDS-PAGE is performed using 10% gels and transferred to nitrocellulose membrane. The membrane is blocked with milk protein, incubated with specific antibody, washed with Tris Buffered Saline containing 0.1% Tween 20 (TBST), incubated with anti-rabbit goat IgG-linked to horseradish peroxidase (NEN), and again washed with TBST. Chemiluminescent substrates are applied to the membrane and the 20 signal is detected by exposing the membrane to X-ray film 30 seconds. RNA extraction: RNA is isolated from cells grown in 6-well plates using RNeasy Mini Kits as specified by the manufacturer (Qiagen). Briefly, cells are harvested and disrupted in RLT buffer. The lysate is passed over QiaShredder columns, and the resulting eluate brought to 35% EtOH and passed over RNeasy columns to bind the RNA to the silica-gel 25 membrane. After washing the bound RNA with RW1 buffer and then RPE buffer, the RNA is eluted from the membrane with water. All RNA samples are DNase treated either using the Qiagen RNase-Free DNase Set during the RNeasy procedure or after final harvest of the RNA using the Ambion DNA-free Kit. After completion of the digestion, 5 1d of DNase Inactivation Buffer is added, the solution incubated for 2 minutes at RT, and 30 the samples centrifuged for 1 min in a microfuge. The RNA containing supernatant is removed and stored at -70 0 C. In addition to the above, if the RNA is being isolated from transfected cells, the initial RNA prep is digested for 1 h at 37*C with the restriction WO 2006/102417 PCT/US2006/010419 -14 enzyme RsaI to cleave any contaminating plasmid DNA, the RNA reisolated using an RNeasy kit, and DNase treated with the Ambion reagents. Each sample consists of RNA isolated from 2 wells of a 6-well plate and at least three samples are isolated for each treatment/time point. 5 Real-time Reverse Transcription-Polymerase Chain Reaction (PCR) of RNA: Two pg of total RNA is reverse transcribed in a 100I1 total volume containing 50 mM KCl, 10 mM Tris, pH 8.3, 5.5 mM MgCl 2 , 0.5 mM each dNTPs, 0.125 pM random hexamer, 40 units RNase Inhibitor, and 125 units MultiScribe (Applied Biosystems). In control samples the RNase inhibitor and MultiScribe are omitted. Samples are incubated for 10 min. at 25 0 C, 10 30 minutes at 48 0 C, and then 5 min. at 95 0 C to inactivate the enzyme. Real-time PCR is then performed on 5 pl of the resulting cDNA in a total volume of 25 pl containing 12.5 pl of 2X SYBR Green PCR Master Mix (Applied Biosystems), and 0.8 pM each primer. The PCR parameters used are 2 min. at 50'C, 10 min. at 95 0 C, and 45 cycles of 95 0 C for 15 sec. followed by 1 min. at 62 0 C. PCR is also performed as described on a 1/800 dilution 15 of the cDNA with 18S primers for normalization of the samples. Relative RNA levels were calculated using the A A Ct method (Applied Biosystems). The primers listed in Table 1 have been synthesized and successfully measured mRNA levels of gene expression in human MSCs. Table 1 20 RT PCR Primer/Probe Sets Primer Name Primer Sequence (5' to 3') SEQ ID NO Ad35LMP-1 TTCTGAGCTTCGATGTGTGTGA 1 Forward Ad35LMP-1 CATCATGGATTCCTTCAAGGTAGTG 2 Reverse Ad35LMP-1 Probe 6FAM-CATCGATGCTCAGCACCCAGTCACC-TAMRA 3 bmrSMAD-1 ACCCTGTCTGAGGAGCGTGTA 4 Forward hmrSMAD-1 ACCAAAGCGTCCACAGCTTT 5 Reverse hmrSMAD-5 CACCAAGATGTGTACCATTCGAA 6 Forward hmrSMAD-5 GAAGAGCCCATCTGAGTAAGGA 7 Reverse bmrSMAD-6 GGATCTGTCCGATTCTACATTGTCT 8 Forward hmrSMAD-6 TGTCCGGTGCTCCCAGTAC 9 Reverse hmrDLX5 Forward GGAGTTGGCCGCCTCTCTAG 10 hmrDLX5 Reverse TGGCGAGTTACACGCCATAG 11 hmrNoggin TGCCGAGCGAGATCAAAGG 12 WO 2006/102417 PCT/US2006/010419 -15 Forward hmrNoggin GTAGCGCGGCCAAAAGC 13 Reverse hBMP-2 Forward TCCAAGAGACATGTTAGGATAAGCA 14 hBMP-2 Reverse TCCACGTACAAAGGGTGTCTCTTAC 15 hmrSMURF-1 CCCAGAGACCTTAACAGTGTGAACT 16 Forward hmrSMURF-1 TTGAGTTGGCACTGGTGATTCA 17 Reverse hmrSMURF-2 TCTCGGTTGTGTTCGTCTTCTTT 18 Forward hmrSMURF-2 GCCTATTCGGTCTCTGGACTGAA 19 Reverse Osterix Forward TCAGACGCCCCGACCTT 20 Osterix Reverse ATTGGCAAGCAGTGGTCTAGAGA 21 siRNA treatment of cells: MSCs are transfected with Lipofectamine 2000 (Invitrogen) or Oligofectamine (Invitrogen) transfection reagent and either irrelevant siRNA or specific siRNA sequences (see Table 2). Silencing of the genes and specificity is confirmed by 5 real-time RT-PCR analysis of specific mRNA levels and Western analysis of protein levels. Table 2 siRNA sequences siRNA Sense Sequence (5'-3') SEQ ID NO: SMURF1 CCUUGCAAAGAAAGACUCtt 22 SMURF2 GGUGGUGGUUGAUGGAUCUtt 23 Jab -1 GCUCAGAGUAUCGAUGAAAtt 24 LMP-1 AGACCUUCUACUCCAAGAAtt 25 10 Mechanistic studies using Smurf1 siRNA 1. Interaction of LMP- 1 with Smurfl not Smurf2: To confirm the specificity of the Smurfi/LMP-1 interactivity, siRNA designed to selectively target Smurf1 mRNA are utilized. Conversely, selectively inhibiting Smurf2 levels and showing that LMP-1 15 interaction with its candidate binding protein is not affected may also be employed. To this end, human MSCs are plated at 3 x 104 cells/well in 6-well plates, grown overnight, and treated with an irrelevant siRNA or siRNA specific for Smurfl or Smurf2. Twenty-four hours after siRNA treatment, cells are treated with Ad5F35-LMP-1 (0, 5, 10 pfu/cell) or Ad5F35-GFP. Cells are harvested 2, 4, 8, 12, 24, or 48 hr later and both 20 cytoplasmic and nuclear proteins subjected to immunoprecipitation using LMP-1 antibody and Western analysis of co-localizing proteins using the WW domain antibody. In WO 2006/102417 PCT/US2006/010419 -16 addition cellular total RNA and protein fractions are harvested at the same timepoints and the levels of Smurfl and Smurf2 mRNA measured by real-time RT-PCR using primers specific for Smurfl or Smurf2. Reduction of protein levels by siRNA is confirmed by Western blots using newly acquired antibodies thought to be specific to Smurf1 and 5 Smurf2. Expected Results: Reduced Smurfi levels are expected to reduce the amount of the WW domain immunoreactive binding protein observed by Western analysis after immunoprecipitation using LMP- 1 antibody compared with controls that have normal levels of Smurfi. In contrast, reduced Smurf2 levels are expected to have no effect on the 10 amount of WW domain immunoreactive binding protein as we do not expect LMP-1 to bind Smurf2. These results would confirm that the interaction of LMP-1 is entirely with Smurfl. 2. Effects of decreasing Smurfl on phosphorvlated Smad1 and Smad5 -Nuclear studies: The hypothesis is that LMP-1 binding to Smurfl blocks Smurf1 from binding to P-Smadl 15 and P-Smad5, and, thus, reduces P-Smad proteasomal degradation. The expected overall effect is increased nuclear levels of P-Smadl and P-Smad5. While it is possible that LMP-1 may have several modes of action, it is believed the interaction with Smurfl is responsible for the osteoinductive properties of LMP-1. If the hypothesis is correct, decreasing Smurfl levels should produce the same effect as LMP-1 to increase bone 20 formation in vitro and responsiveness to BMP-2. To this end human MSCs are plated at 3 x 10 4 cells/well in 6-well plates, grown overnight, and treated with a control irrelevant siRNA or siRNA specific for Smurf1. BMP-2 (100 ng/ml) is applied to some cultures when the siRNA is removed. Cells are harvested 4, 8, 12, 24, 36, or 48 hr later and nuclear proteins are analyzed for the presence of P-Smads by 25 Western blot. The effectiveness of the siRNA applied to reduce the RNA and protein levels is monitored as described herein. Another group of MSCs plated as above are grown in differentiation medium for 21 days and stained with Alizarin Red to assess matrix mineralization. Expected Results: It is expected that there will be an increase in nuclear levels of P 30 Smadi and P-Smad5 in cells treated with siRNA to Smurfl as these proteins should not be targeted for proteasomal degradation in the absence of Smurf1. It is hypothesized that the preliminary results showing increased nuclear levels of Smads in response to WO 2006/102417 PCT/US2006/010419 -17 overexpression of LMP- 1 (see example 3), was caused by LMP- 1 blocking proteasomal degradation of the P-Smads. Thus, it is expected that a reduction in Smurfl levels to mimic LMP-1 blocking of Smurf1, will result in targeting P-Smads for degradation. It is also expected that an increased responsiveness to BMP-2 in the presence of Smurfl 5 siRNA as evidenced by increased levels of nuclear P-Smads and increased extracellular matrix mineralization will be observed. Consequently, siRNA to Smurfl would represent a therapeutic opportunity for matrix mineralization. 3. Effects of decreasing Smurf1 on phosphorylated Smadl and Smad5-Cytoplasmic studies: The hypothesis is that LMP-1 binding to Smurfl blocks Smurf1 from binding to 10 Smad1, reduces proteasomal degradation of Smad1, and increases the responsiveness of cells to activation of the BMP-2 pathway. The expected overall effect is increased cytoplasmic levels of Smad1. While it is possible that LMP l may have several modes of action, it is believe the interaction with Smurf1 is important for the osteoinductive properties of LMP-1. If the hypothesis is correct, decreasing Smurfl levels should 15 produce a similar effect as LMP- 1 in increasing responsiveness to BMP-2 as evidenced by the osteogenic response in vitro. To this end, human MSCs are plated at 3 x 10 4 cells/well in 6-well plates, grown overnight, and treated with a control irrelevant siRNA or siRNA specific for Smurfl. Cells are harvested 4, 8, 12, 24, 36, or 48 hr later, cytoplasmic proteins are analyzed for the presence of total Smadl and phospho-Smad1 by Western blot 20 and ELISA. The effectiveness of the siRNA to reduce the RNA and protein levels of Smurfi are monitored as described herein. Another group of MSCs plated as above are grown in differentiation medium for 21 days and stained with Alizarin Red to assess matrix mineralization. Expected Results: It is expected that treatment will result in increased cytoplasmic Smadl 25 in cells treated with siRNA to Smurf1 as this protein should not be targeted for proteasomal degradation in the absence of Smurfl. It has been shown, in other examples herein, that increased cytoplasmic levels of phospho-Smad1 in response to overexpression of LMP-1 and BMP-2 treatment. This result is believed to have been caused by LMP-1 blocking proteasomal degradation of Smad1. Thus, we expect reduction in Smurf1 levels 30 to mimic LMP-1 in blocking Smurf1 targeting of Smad1 for degradation. 4. Effects of decreasing Smurf1 expression on BMP-2 responsiveness: As stated above, it is expected that directly reducing the level of Smurfl will mimic the effect of LMP-1 WO 2006/102417 PCT/US2006/010419 -18 overexpression. To test this hypothesis human MSCs are plated at 3 x 104 cells/well in 6 well plates, grown overnight, and treated with a control irrelevant siRNA or siRNA specific for Smurfl. BMP-2 is applied to some cultures when the siRNA is removed. After 1, 2, 4, 8, 24, 36, 48, and 72 hrs cells are harvested for analysis of total RNA or secreted, 5 cytoplasmic and nuclear proteins including Dlx5, Smad6 and BMP-2 as previously described. Analysis by ELISA allows more accurate quantitation than can be achieved by Western blotting. Expected Results: While not wishing to be bound by theory, it is expected that expression of Dlx5, Smad6, and BMP-2 will be increased at both the protein and mRNA levels when 10 the Smurfl level is reduced using specific siRNA. Cells having reduced Smurfl levels also are expected to show an increased responsiveness to BMP-2 with regard to expression of these genes. An increased levels of expression of these genes is interpreted as a critical step for the synergistic response to BMP-2 observed in MSCs with reduced levels of Smurf1. 15 Biotin transfer Assay for detection of LMP-1 interacting proteins: Sulfo-SBED (Pierce), a trifunctional cross-linking agent, contains three functional groups (a photoactivatable aryl azide, a sulfonated N-hydroxy succinimide active ester with a cleavable disulfide group and a biotin moiety) and is widely used to identify interacting proteins (Neely KE, Hassan AH, Brown CE, Howe L, and Workman JL. Mol.Cell Biol. 2002;22:1615-25). LMP-1 is 20 labeled using this reagent, incubated as bait with nuclear proteins and cross-linked to interacting proteins by UV (365 nm). Proteins that physically interact with LMP- 1 retain the biotin group when suspended in SDS-PAGE reducing buffer. Biotin-containing target proteins are separated using neutravadin beads, detected by Western blotting with neutraviddin-HRP and the signal is developed with chemiluminescent substrate. 25 Corresponding protein bands are in-gel digested with trypsin. Tryptic peptides are recovered, concentrated and their mass profile is analysed by MALDI-TOF at the Emory University Microchemical Facility. LMP-1t: A 223aa osteoinductive truncated LMP-1 variant (missing LIM domains): Although LMP-1 is a LIM domain protein, it has been shown that a truncated 223aa 30 variant of LMP-1 which lacks the LIM domains still makes bone in vitro and in vivo (Liu Y, Hair GA, Boden SD, Viggeswarapu M, and Titus L., J.Bone Min.Res. 2002;17:406 14). If binding to Smurf1 and activation of the BMP-2 signaling pathway is critical for WO 2006/102417 PCT/US2006/010419 -19 LMP-1 action and induction of bone formation, then it would be expected that the truncated LMP-1 also binds Smurfl and mimics the downstream effects of the full length protein in the presence or absence of BMP-2. To this end truncated LMP-l (LMP-lt) fusion protein containing a Protein Transduction Domain (PTD) that readily enters cells 5 can be designed. The suitability of this PTD for use in these and subsequent experiments is validated by showing that the PTD-LMP- it fusion protein retains the same ability of overexpressed full length LMP- 1 to compete with WW domain antibody binding, to induce increased nuclear levels of P-Smadl and P-Smad5, and to increase expression of BMP/Smad regulated genes. Other data herein demonstrate the use of a full length TAT 10 LMP- 1 fusion protein to enter cells readily and induce bone formation. The PTD domain proposed for use in these studies been shown to be a more effective in protein transduction than the TAT protein transduction domain used in earlier studies (Mi Z, Mai J, Lu X, and Robbins PD., Mol.Ther. 2000; 2:339-47). Experimental Design: Initial studies will be required to determine a dose of PTD-LMP-lt 15 that enhances the effect of 100 ng/ml BMP-2 on mineralization of MSCs. Human MSCs are plated at 3 x 104 cells/well in 6-well plates, grown overnight, and treated with PTD LMP-lt (0.3-30 nM), BMP-2 (100 ng/ml) or both agents. Cells are grown in mineralizing medium as described in the Examples herein for overexpressed full length LMP-1 and the mineral stained with Alizarin Red. Two doses that synergize with BMP-2 are selected for 20 use in subsequent studies. Following dose selection, human MSCs are plated at 3 x 104 cells/well in 6-well plates, grown overnight, and treated with nothing or 2 doses of PTD-LMP-lt, BMP-1 (100 ng/ml) or both agents (control = PTD-pGal). After 1, 2, 4, 8, 12, 24, 36 and 48 hrs, cells are harvested and total RNA plus cytoplasmic and nuclear protein fractions are prepared. 25 The nuclear protein fraction from the untreated sample is analyzed for the ability of purified LMP-1 protein to compete with Smurfi antibody binding to Smurf1 on a Western blot. All protein samples are analyzed for the presence of phosphorylated Smadi and Smad5 by Western blot using previously described appropriate antibody. All RNA samples are analyzed by real-time RT-PCR for mRNA levels of Dlx5, Smad6, and BMP 30 2. Similarly secreted, cytoplasmic and nuclear proteins are analyzed by ELISA using commercially available antibodies to Dlx5, Smad 6, and BMP-2.

WO 2006/102417 PCT/US2006/010419 -20 Expected Results: It is expected that LMP-lt protein will enhance the osteoinductive responsiveness of MSCs to BMP-2 as seen with overexpression of the full length LMP-1 protein. It is also expected that truncated LMP-1 will bind to Smurf1 and prevent WW domain antibody binding. In addition, it is expected that there will be increased nuclear 5 Smad levels and activation of Smad-regulated genes to exactly mimic the outcome observed with overexpression of full length LMP-1 in the presence or absence of BMP-2. This result would support the hypothesis that the LIM domains are not required for LMP-1 to induce bone formation or for LMP-1 to enhance responsiveness to BMP-2. Smurf1 -WW domain and LMP- 1 interaction assay: Nuclear proteins are separated by 10 SDS-PAGE and blotted onto nitrocellulose membrane. Protein blots are blocked with 5% milk protein and pre-incubated with purified LMP-1 protein (10 uM) or TBST buffer. Blots are incubated with Smurf antibody at 1:5000 dilution (Rabbit antibody raised to WW-domain peptide). After washes, blots are incubated with HRP-labeled Anti-rabbit second antibody. The washed blots are then incubated with ECL substrate solution and the 15 membranes are exposed to X-ray film for signal detection. Protein A-based immunoprecipitation assay: Protein A-agarose beads are incubated with LMP-1 antibody, washed 3 times, incubated with nuclear proteins, and washed again to remove unbound protein. Bound proteins are eluted by 2 washes in 0.1 M citric acid, pH 2.7. The eluates are neutralized with Tris base and concentrated by centricon tubes 20 (Ambicon) prior to SDS-PAGE and Western blotting. Pulse labeling of Smads: Cells are incubated for 1 h in the presence of ( 35 S)Methionine, washed extensively, incubated 30 min with Ad5F35LMP-1 (control = Ad5F35GFP, or nothing), rinsed and returned to normal medium in the presence of BMP-2 (100 ng/ml). 1, 2, 3, 4 h after transduction, cells are harvested, nuclear fractions prepared, and 25 radiolabeled P-Smadl or P-Smad5 immunoprecipitated as above. Immunoprecipitates are subjected to SDS-PAGE, visualized and quantitated by autoradiography. Modified pulse chase assay for nuclear and cytoplasmic analyses of proteins. Nuclear analysis: It has been shown herein that phosphorylated Smad1 is increased in the nucleus within 4 hrs in response to LMP- 1 overexpression and is further increased at 30 8 hours. Possible explanations include: 1) increased gene expression of Smad1; 2) increased phosphorylation of the cytoplasmic pool of unphosphorylated Smad1 with subsequent translocation to the nucleus; and 3) reduced degradation of Smad1. The short WO 2006/102417 PCT/US2006/010419 -21 time required for the P-Smad increase to occur does not rule out increased transcription and translation of Smad1. However, the 4 hr time frame makes it less likely that LMP-1 can be transcribed and translated and also cause increased subsequent transcription and translation of Smad1. The hypothesis for the increased nuclear level of P-Smadl in 5 response to LMP overexpression is that LMP- 1 binds Smurf1, preventing Smurfl from targeting P-Smad1 for proteasomal degradation. Similar accumulation of P-Smad5 is expected but has not yet been measured. The increased levels of P-Smadl and P-Smad5 are expected to result in activation of BMP signaling, a requirement of which is that they localize in the nucleus in order to alter gene expression. 10 To this end, a modified pulse-chase experiment using 35 S-Methionine to label newly synthesized Smad proteins and determine their rate of degradation in the presence or absence of overexpressed LMP-1 can be performed. This approach allows one to distinguish between Smad synthesis in response to LMP-1 from P-Smad accumulation in response to LMP-1. 15 For the pulse-chase experiment, human MSCs are plated at 3 x 10 4 cells/well in 6 well plates and grown overnight. 3 5 S-Methionine ( 35 S-Express, NEN) are applied for 1 hr to pulse label the pool of newly synthesized proteins. Upon removal of the radiolabel and subsequent washing, cells are incubated 30 min with Ad5F35-LMP-1 (0, 5, 10 pfu/cell), Ad5F35-GFP, or nothing. BMP-2 (100 ng/ml) is applied to some cultures at the end of the 20 transduction incubation to ensure phosphorylation of the Smad1. One, 2, 3, or 4 hr after transduction, cells are harvested and nuclear protein fractions prepared. Radiolabeled P Smad1 or P-Smad5 are immunoprecipitated from the nuclear fraction using specific antibodies and analyzed by SDS-PAGE and fluorography. Expected Results: Without wishing to be bound by theory, it is expected that Smad1 and 25 Smad5 proteins will be synthesized during the labeling period and that treatment of cells with BMP-2 will result in rapid phosphorylation of some of the newly synthesized molecules. The labeled P-Smadl and P-Smad5 would then bind Smad4 and move into the nucleus. LMP-1 and Smurf1 are predominately found in the nucleus. As overexpressed LMP-1 protein is translated it is expected to move into the nucleus and bind Smurfi. If 30 the hypothesis is correct, this event should block subsequent degradation of P-Smadl and P-Smad5 and result in a reduced rate of degradation compared with P-Smads in cells not overexpressing LMP-1.

WO 2006/102417 PCT/US2006/010419 -22 Cytoplasmic analysis: Human MSCs are plated at 3 x 10 4 cells/well in 6-well plates, grown overnight and incubated 30 min with Ad5F35-LMP-1 (0, 5, 10 pfu/cell), Ad5F35-GFP, or nothing. To determine stability of unphosphorylated Smad1, 35

S

Methionine ( 35 S-Express, NEN) is applied for 30 min on the next day to pulse label the 5 pool of newly synthesized proteins and the cells subsequently washed. Cells are incubated in chase medium containing cold methionine with or without cycloheximide (10 ug/mL) for 2-3 hrs during which cells are harvested at various time points. Cells are harvested in the presence of protease inhibitors and cytoplasmic protein fractions prepared. Radiolabeled total Smad1 are immunoprecipitated from the cytoplasmic fraction 10 using specific antibody and analyzed by SDS-PAGE and auto-fluorography. To measure stability of phospho-Smad1, similar experiments are performed in control MSCs or MSCs overexpressing LMP-1, as above. Cells are labeled with 32 P-orthophosphate for 2 hr in the presence of the nuclear transport inhibitor, leptomycin B (25uM), to assure that only decay of the labeled phospho-Smad, not the net amount of decay and transport into the nucleus is 15 measured. 32 P is removed and cells are treated for 1 hr with BMP-2 (100 ng/ml). Cytoplasmic fractions are prepared in the presence of protease and phosphatase inhibitors (to prevent loss of the 32 P label). Radiolabeled 32 P-Smadl are immunoprecipitated from cytoplasmic fractions and analyzed as above. Expected Results: Without wishing to be bound by theory, in the absence of BMP-2, it is 20 expect that unphosphorylated Smadl protein will be synthesized and incorporate 3S- Methionine during the labeling period. As overexpressed LMP-1 protein is translated it is expected to bind cytoplasmic Smurf1. If the hypothesis is correct, this event should block subsequent degradation of Smadl and result in a reduced rate of degradation of Smadl compared with cells not overexpressing LMP-1. It is expected that this block to be true of 25 both the unphosphorylated and phosphorylated form of Smad. To determine the relative susceptibility of unphosphorylated Smad1 and phospho-Smadl to Smurf1-mediated proteasomal degradation,comparisons of the kinetic curve for the decay of unphosphorylated Smadl ( 3 5 S-Methionine labeled) with that of phospho-Smadl ( 32

P

orthophosphate labeled) can be made. It is expected that LMP-1 will impact primarily the 30 unphosphorylated form. Metabolic Pulse-Chase Analysis for half life determination: To determine metabolic half life of Smad1, MSCs are transfected with Ad5F35-LMP-l or AD5F35-GFP and grown for WO 2006/102417 PCT/US2006/010419 -23 24 hr prior to metabolic labeling. Cells are washed in methionine-free medium and incubated for 15 min to deplete endogenous methionine. Cells are then incubated with the pulse-labeling medium containing [ 35 S]methionine (190 pCi/ml) for 30 min. After washing with chase medium (containing 150 mg/L unlabeled methionine). Cells are incubated in 5 chase medium with or without cycloheximide (10 ug/mL) for 2-3 hr. Cells are lysed in buffer containing protease inhibitors (Sigma). Smad1 is immunoprecipitated from cytoplasmic fractions using specific antibody and analyzed by SDS-PAGE and auto fluorography. To determine the half life of phospho-Smad1, cells are transfected with Ad5F35 10 LMP-1 or AD5F35-GFP and grown for 24hr prior to metabolic labeling. For in vivo

[

32 P]orthophosphate labeling, cells are pre-incubated with phosphate-free media for 1 h and exposed to 1 mCi/ml [ 32 P]orthophosphate for 2 h at 37 0 C. Leptomycin B (1Ong/ml) or Ratjadone (10ng/ml) (Cal Biochem) is incubated with cells for 30 min to inhibit nuclear translocation of the labeled Smadl. Cells are then treated with BMP-2 (100 ng/ml) for 1 15 hr, lysed in buffer containing protease and phosphatase inhibitors, and the cytoplasmic fractions subjected to immunoprecipitation with phospho-Smad1 specific antibody. Immunoprecipitates are visualized by SDS-PAGE followed by auto-fluorography. Enzyme Linked Sorbent Assay (ELSA): Purified Smurfl is coated to individual wells of an Immulon 1B plate and the remaining surface blocked. Incubation of Smad1 and varying 20 concentrations of competing ligand (LMP-1) and/or vice versa are performed overnight at 4*C. Using appropriate primary and enzyme linked-secondary antibodies, optical density is monitored at specific wavelengths using the BioLumin 960 microtiter plate reader or the SpectraMax M2 microtiter plate reader. The assay can be adapted several ways to suit binding partner proteins. When biotin-labeled Smad1 is assayed, streptavidin-alkaline 25 phsophatase is used as the secondary reagent. After determining maximum binding between LMP-1/Smurfl and Smad1/Smurfl, mutual competition curves at various concentrations will provide data for Scatchard plot analysis to obtain binding affinity, dissociation constant and number of binding sites. Ubiquitin assay for determining extent of ubiquitination of Smadl/5 and/or LMP-1: The 30 ELSA assay described above is modified for the ubiquitin assay as follows: Microtiter plates are coated with Smadl/5 or LMP-1 antibody and the remaining active sites are blocked by 1% BSA. Smad1/5 or LMP-1 are then captured by incubating nuclear proteins WO 2006/102417 PCT/US2006/010419 -24 in wells. Using specific enzyme/fluorescent-linked Ubiquitin antibody and the appropriate substrate solution, the extent of ubiquitination can be assayed. The same assay can be adapted to study inhibitory effect of LMP-1 derived peptides on ubiquitination of Smads using an in vitro assay system with purified and commercially available ubiquitination 5 assay reagents (Boston Biochemicals). Transfection of MSCs for in vivo Bone Induction: MSCs are grown to confluence, incubated with agents being tested, trypsinized, washed 2x with PBS, suspended to 10 20M/mL DMEM and 100 uL of the suspension applied to a sterile disc (2 x 5 mm) of bovine collagen. Implants are surgically placed subcutaneously on the chest of 4-6 wk 10 athymic rats (rnu~/rnu~). The animals are euthanized after 4 weeks; the explants removed, fixed in 70% ethanol, and analyzed by radiography and undecalcified histology. Mechanistic Investigations of Ubiquitination 1. Ubiquitination of Smad1 and Smad5. It is hypothesized that the consequence of LMP-1 binding to Smurf1 is a reduction in the 15 number of ubiquitinated Smad1 and Smad5 proteins. To this end, LMP-1 is overexpressed in hMSCs, Smad1 or Smad5 are captured on wells coated with the specific Smad antibody, and the level of ubiquitination of the Smads is quantitated by ELISA. Human MSCs are plated at 3 x 10 4 cells/well in 6-well plates, grown overnight and treated with Ad5F35-LMP-1 (0, 5, 10 pfu/cell), or Ad5F35GFP control plasmid. After 4, 8, 12, 20 24, 48, and 72 hr, cells are harvested and cytoplasmic and nuclear protein fractions prepared. The cell fractions are incubated with antibody to Smad1 or Smad5 in a 96 well plate coated with the antibody. After washing to remove nuclear proteins not associated with Smad1 or Smad5, fluorescence tagged-ubiquitin antibody is applied and the fluorescence is detected using a Biolumin 960 microtiter plate reader. Use of the 25 fluorescent tagged antibody increases sensitivity and allows quantitation of the low level of ubiquitin expected to be present in the sample. To validate the ELISA results, another aliquot of the cell fractions undergoes immunoprecipitation using anti-Smad1 or anti Smad5 antibody and the precipitated proteins are analyzed by Western blotting using antibody to ubiquitin as well as antibody to Smadl or Smad5. 30 Expected Results: It is expected that the Western analysis will show a smear with both the ubiquitin and Smad1 or Smad5 antibodies representing Smads conjugated with ubiquitin chains of various lengths. The Smad1 antibody is expected to capture only Smad1 and the WO 2006/102417 PCT/US2006/010419 -25 Smad5 antibody is expected to capture only Smad5. Free ubiquitin is removed during the ELISA washes. Additional ubiquitinated proteins are not expected to be observed. Compared with untreated or GFP treated cells, it is expected that lower levels of Smad ubiquitination will occur when LMP-1 is overexpressed due to interruption of Smurfl 5 function. It should be noted that this method will not distinguish whether decreased ubiquitination is due to variation in the number of ubiquitin subunits linked to Smads or fewer Smad molecules that are conjugated to ubiquitin chains. Examination of Western blots performed for Smad1 or Smad5 and ubiquitin may resolve this issue. 2. Effects of LMP-1 siRNA treatment on ubiquitination. 10 Upon demonstration that forced expression of LMP- 1 can alter the level of ubiquitinated Smadl and Smad5, reduction of endogenous levels of LMP-1 using siRNA would be expected to increase the rate of degradation of those Smad proteins. Other examples herein have shown that siRNA targeting LMP- 1 are effective at reducing LMP-1 levels. Using these siRNA, studies investigating the relationship between LMP-l 15 and ubiquitination of Smads can be performed. To show that application of siRNA to reduce the endogenous LMP-1 levels results in increased levels of ubiquitinated Smadl/5 human MSCs are plated at 3 x 10 4 cells/well in 6-well plates, grown overnight, and treated with a control irrelevant siRNA or specific LMP-1 siRNA. Cells are harvested 4, 8, 12, 24, 48, or 72 hr later and both cytoplasmic and nuclear fractions prepared. The cell 20 fractions are incubated with antibody to Smadl or Smad5 in a 96 well plate coated with the antibody. After washing to remove nuclear proteins not associated with Smadl or Smad5, fluorescence tagged-ubiquitin antibody are applied and the fluorescence detected as in other examples herein. This approach allows quantitation of the low level of ubiquitin that is expected to be present in the sample. 25 Expected Results: It is expected that reduced levels of endogenous LMP-1 will increase ubiquitination of both Smad1 and Smad5. We would interpret this as being due to greater availability of Smurf1 molecules for Smad binding. 3. Ubiquitination of Smadl by Smurfl It is believed that LMP-1 prevents Smurfl interaction with Smadl, resulting in decreased 30 ubiquitination of Smad1. Reduced ubiquitination is therefore expected to result in reduced proteasomal degradation of Smad1 protein. To demonstrate that the interaction of LMP-1 WO 2006/102417 PCT/US2006/010419 -26 with Smurf1 inhibits ubiquitination of Smadl, in vitro ubiquitination assays are performed. To this end, biotin labeled Smad1 is prepared by the TNT-coupled reticulocyte lysate system (Promega) and impurities removed by capturing Smadi using neutravidin-linked 5 resin. Smad1 is mixed with a mixture of purified El and E2 ligases and incubated with the Smad ubiquitin E3 ligase, Smurf1, in the presence or absence of recombinant LMP- 1 protein. The reaction mixture also contains ubiquitin and the creatine kinase-ATP generating system. The reaction mixture is analyzed by SDS-PAGE and Western blots using specific antibody to ubiquitin or Strepavidin-HRP (to detect Biotin-Smadi). 10 Expected Results: It is expected that there will exist a baseline ubiquitination of Smad1 by Smurfl under the conditions of the reaction. Further, it is expect that addition of LMP-1 will inhibit this reaction. This will identify Smurfl as an E3 ligase with which LMP- 1 interacts to reduce Smurf1 -induced ubiquitination of Smad1. 4. Effects of Smurf1 siRNA treatment on ubiquitination. 15 It is possible that LMP-1 binding to Smurfl leads to its own ubiquitination and subsequent proteasomal degradation. However, there are examples of proteins that bind E3 ligases without being degraded (Murillas R, Simms KS, Hatakeyama S, Weissman AM, and Kuehn MR. J.Biol.Chem. 2002; 277:2897-907). Our empirical data that relatively small amounts of LMP- 1 alone profoundly activate the BMP-2 pathway suggest that LMP-1 20 itself may not be targeted for proteasomal degradation by Smurfl, but rather occupy the Smadl/5 binding site, preventing Smadl/5 targeting and degradation. To determine whether reducing the levels of Smurfl results in reduced LMP-1 ubiquitination and increased levels of endogenous LMP-1, human MSCs are plated at 3 x 104 cells/well in 6 well plates, grown overnight, and Smurfi or control siRNA applied. Cells are harvested 4, 25 8, 12, 24, 48, or 72 hr later and both cytoplasmic and nuclear fractions prepared. The cell fractions are incubated with antibody to LMP-1 in a 96 well plate coated with the antibody. After washing to remove nuclear proteins not associated with LMP-1, fluorescence-tagged ubiquitin antibody is applied and fluorescence detected as above. This approach allows quantitation of the level of ubiquitin in the sample. In other wells, 30 captured LMP-1 levels is quantitated by applying fluorescence tagged-LMP-1 antibody. To validate that the ELISA measurements are specific to ubiquitinated LMP-1, another aliquot of the cell fractions undergoes immunoprecipitation using anti-LMP-1 antibody WO 2006/102417 PCT/US2006/010419 -27 and the precipitated proteins are analyzed by Western blotting using antibody to ubiquitin as well as antibody to LMP-1. Expected Results: It is not believed that LMP-1 undergoes ubiquitination by Smurfl. Thus, it is expected that decreased levels of Smurfl will not change the level of 5 ubiquitinated LMP-1 compared with cells with normal levels of Smurf1. Further it is predicted that reduced levels of Smurfl will have no effect on the level of LMP-1 protein within cells. In concluding that Smurf1 does not target LMP-1 for proteasomal degradation, it is suggested that LMP-1 binds Smurfl at the site that also binds Smad1 or Smad5, blocking Smad ubiquitination and targeting for proteasomal degradation. 10 Therefore, this interruption of Smurf1 function would be a primary mechanism by which LMP-1 enhances signaling of the BMP-2 pathway. 5. Ubiquitination reaction in vitro: Purified Smad1 (100 ng) is buffer-exchanged to ubiquitination buffer (50 uM Tris-HC1 pH 7.8, 5 mM MgCl 2 , 0.5 mM dithiothreitol (DTT), 2 mM NaF, and 3 pM okadaic acid). 15 Smad1 is then combined with a mixture of purified El and E2 enzymes and incubated with Smurf1 (E3 ligase) in the presence or absence of recombinant LMP-1 or LMP-2 protein. The reaction mixture also contains 2 mM ATP, ubiquitin (150 pM), ubiquitin aldehyde (5 pLM), and creatine kinase-ATP generating system (Boston Biochem). The ubiquitin aldehyde is included to prevent hydrolysis of polyubiquitin chains. The reaction 20 mixture (40 gL) is incubated 4 hr at 37'C. Aliquots at various time points are taken for SDS-PAGE and western blotting using specific antibody for Smad1 and/or ubiquitin. Mechanistic studies of the BMP-2 receptor (BMPR1A) 1. Smurf1/Smad6 complexes One hypothesis by which LMP-1 enhances cellular responsiveness to BMP-2 is by 25 increasing the BMPR1A (ALK3) level in the plasma membrane. Smurf1 has been shown to regulate the proteasomal degradation of BMP-2 receptors through its interaction with I Smads (Murakami, G., Watabe, T., Takaoka, K., Miyazono, K., and Imamura, T., Mol. Biol. Cell. 2003, 14:2809-2817, 2003. It is suggested that I-Smad/Smurfl complexes form in the nucleus and translocate to the plasma membrane where I-Smad binding to the 30 receptor occurs. Once bound to the receptor complex, Smurfl ubiquitinates the I-Smad and BMPR1A receptor, targeting them for proteasomal degradation (Ebisawa, T., Fukuchi, M., Murakami, G., Chiba, T., Tanaka, K., Imamura, T., and Miyazono, K., J. Biol. Chem.

WO 2006/102417 PCT/US2006/010419 -28 2001, 276:12477-12480; Izzi, L. and Attisano, L., Oncogene. 2004, 23:2071-2078). To test the hypothesis that LMP-1/Smurfl interaction prevents formation of Smurfl/I-Smad (Smad6) complexes, recombinant Smurf1, LMP-1 and biotin-Smad6 are purified and tested in an ELSA binding competition assay. The assay is described fully herein. 5 Briefly, wells of a 96-well plate are coated with Smurf1 protein and preincubated in the presence or absence of different amounts of LMP-1. After washing, biotin-Smad6 is incubated with the Smurfl complexes and binding detected using strepavadin-alkaline phosphatase. Absorbance is monitored at 405 mu using the SpectraMax M2 microplate reader. 10 Expected Results: It is expected that, in the absence of LMP-1, Smurfl/Smad6 complexes will be detected. In the presence of increasing concentrations of LMP-1, decreasing amounts of Smurfl/Smad6 complex are expected to be formed until complete inhibition of Smad6 binding is achieved. 2. Smurf- 1 mediated export of Smad6 from the nucleus. 15 In the absence of receptor activation, I-Smads have been shown to primarily reside in the nucleus, but activation of the receptor results in translocation of I-Smads to the cytoplasm (Itoh, F., Asao, H., Sugamura, K., Heldin, C. H., ten Dijke, P., and Itoh, S., EMBO J. 2001, 20:4132-4142; Nakayama, T., Gardner, H., Berg, L. K., and Christian, J. L., Genes Cells, 1998, 3:387-394). It has been suggested that the export of Smad6 from the nucleus 20 is facilitated by Smurf1 (Izzi, L. and Attisano, L., Oncogene 2004, 23:2071-2078); Suzuki, C., Murakami, G., Fukuchi, M., Shimanuki, T., Shikauchi, Y., Imamura, T., and Miyazono, K., J. Biol. Chem., 2002, 277:39919-39925). Since the hypothesis is that LMP 1 binds Smurfl in the same site as Smurf1 binds Smads, it is expected that the consequence of LMP-1/Smurfl interaction will be that less Smad6 moves to the 25 cytoplasmic compartment as a result of BMP-2 activation of its receptor. To this end, Flag-Smad6 is overexpressed to show that LMP-1 blocks the BMP-2-mediated translocation of Smad6 into the cytoplasm. Overexpressed Flag-Smad6 is used rather than endogenous Smad6 because of the specificity of the Flag antibody and cross-reactivity of Smad6 antibodies with other Smads. 30 Experimental Design: Human MSCs are plated at 3 x 104 cells/well in 6-well plates, grown overnight, and transfected with 3 ug empty vector or plasmid containing Flag Smad6. After the transfection, cells are allowed to recover for 1 hr prior to transduction WO 2006/102417 PCT/US2006/010419 -29 with Ad5F35-LMP-1 or Ad5F35-GFP (0, 5, 10 pfu/cell). Twenty-four hours after the transduction is complete, BMP-2 (100 ng/mL) is applied to the cells. Cells are harvested 1, 2, 4, 8, and 12 hrs after BMP-2 treatment is initiated and cytoplasmic and nuclear fractions prepared. Fractions are subjected to SDS-PAGE and Western analysis using 5 Flag specific antibody. Expected Results: An increase in Smad6 in the cytoplasm upon BMP-2 receptor activation is expected. However, in cells overexpressing LMP-1 this increase is not expected. These results would suggest LMP-1/Smurfl complexes form and interrupt the formation of Smurfl/Smad6 complexes that move Smad6 into the cytoplasm. It is 10 believed that disrupting movement of Smad6 from the nucleus to the cytoplasm is a critical first step that is required for LMP-1 disruption of Smurfl/Smad6-mediated proteasomal degradation of BMPR1A. This finding would also be important in elucidation of the overall mechanism of LMP-l enhancement of responsiveness of MSCs to BMP-2. If LMP-1 blocks Smurfl 15 from interacting with Smad6, it would seem that the free Smad6 would still be available in the cytoplasm to oppose the other effects of LMP-1 on BMP action. However, it is expected that Smad6 is not exported from the nucleus to the cytoplasm in the presence of LMP- 1. Thus, it is important to demonstrate that Smad6 levels do not increase in the cytoplasm. 20 3. LMP-1 effects on the amount of BMPR1A in the plasma membrane of MSCs. It is believed that the consequence of reducing Smurfl/Smad6 interaction will be an increase in the level of BMPR1A in the plasma membrane. Results disclosed herein suggest that BMP-2 increases BMPR1A in cells overexpressing LMP-1 more than in control cells. This finding would represent the sum of BMP-2 action to increase receptor 25 number and the action of LMP-1 to reduce proteasomal degradation of the receptor. This experiment allows for the elucidation of the effect of LMP-l alone on this increase in receptor. Increasing receptor number is an extremely powerful mechanism for increasing the responsiveness of MSCs to BMP-2. Experimental Design: Human MSCs are plated at 3 x 10 4 cells/well in 6-well plates, 30 grown overnight, and treated with Ad5F35-LMP-1 or Ad5F35-GFP (0, 5, 10 pfu/cell). After 1, 2, 4, 8, 12, 24 and 48 hrs, cells are harvested and plasma membrane enriched WO 2006/102417 PCT/US2006/010419 -30 fractions prepared. Fractions are subjected to SDS-PAGE and Western analysis using BMPR1A specific antibody. Expected Results: Increased levels of BMPR1A in the plasma membranes of MSCs overexpressing LMP-1 are expected. This increase would likely be due to decreased 5 proteasomal degradation of the receptor by Smurfl/Smad6 complexes in the presence of increased LMP-1. Consequently, this finding would represent a second significant mechanism by which LMP-l enhances the osteoinductive efficacy of BMP in MSCs. 4. Effects of LMP-1 on ubiquitinated BMPR1A. It is hypothesized that the receptor increase shown above would be due to decreased 10 ubiquitination and proteasomal degradation because of reduced interaction with the Smurfl/Smad6 complex. To test that hypothesis, the levels of ubiquitinated BMPR1A are measured. Past experience suggests that detection of ubiquitinated receptors by ELISA will be difficult because receptor proteins do not adhere to plastic consistently. Thus, Western blot analysis is used to determine changes in the level of ubiquitination. 15 Experimental Design: Human MSCs are plated at 3 x 104 cells/well in 6-well plates, grown overnight and treated with Ad5F35-LMP-1 (0, 5, 10 pfu/cell), or Ad5F35GFP control plasmid. After 4, 8, 12, 24, and 48 hr, cells are harvested and plasma membrane enriched fractions prepared. BMPR1A are immunoprecipitated from the cell fractions using beads coated with specific antibody. The beads are washed and the precipitated 20 proteins are analyzed by Western blotting using antibody to ubiquitin as well as antibody to BMPR1A. Expected Results: Western analysis is expected to show a smear with both the ubiquitin and BMPR1A antibodies representing BMPR1A conjugated with ubiquitin chains of various lengths. Additional ubiquitinated proteins are not expected. Compared with 25 untreated or GFP treated cells, lower levels of BMPR1A ubiquitination are expected when LMP-1 is overexpressed due to interruption of Smurfl/Smad6 function. It should be noted that this method may not quantitatively distinguish whether decreased ubiquitination is due to variation in the number of ubiquitin subunits linked to BMPR1A or fewer BMPR1A molecules that are conjugated to ubiquitin chains. Nevertheless, a comparison 30 between treatments can be made, reflecting different amounts of ubiquitinated BMPR1A among various treatments.

WO 2006/102417 PCT/US2006/010419 -31 If ubiquitinated BMPR1A is not detectable it is likely to be caused by rapid degradation of the ubiquitinated receptor. In that case, the experiment would be performed in the presence of the proteasome inhibitor, lactacystin, to accumulate ubiquitinated BMPR1A for detection in Western blots. 5 This method may not be sensitive enough to detect the low levels of ubiquitinated BMPR1A that are present, but it is felt that ELISAs are likely to be inconsistent using hydrophobic receptor proteins. Therefore, if there are not enough ubiquitinated receptors, Smurfi and Smad6 can be overexpressed in the presence or absence of LMP-1 and the experiment can be repeated. Alternatively, if there are too few receptors to detect by 10 Western analysis, Flag-BMPR1A can be overexpressed the experiment repeated using Flag antibody to immunoprecipitate proteins for Western blot analysis. Mechanistic investigations of LMP- 1 and Jab 1 1. Effect on BMP-2 responsiveness on interrupting Jab 1-mediated proteasomal degradation of Smad4 15 One hypothesis is that LMP-1 enhances responsiveness of MSCs to BMP-2 by binding to Jab 1 and preventing Jab 1-induced proteasomal degradation of Smad4. Jab 1 is an adapter protein that targets Smad4 for ubiquitination and regulates the rate of Smad4 degradation, an important role in controlling the cytoplasmic levels of Smad4. Smad4 is required for translocation of phospho-Smadl/5 into the nucleus and has been shown to be critical in 20 regulating the responsiveness of cells to BMP-2 (Hata, A., Lagna, G., Massague, J., and Hemmati-Brivanlou, A., Genes Dev. 1998, 12:186-197; Moren, A., Hellman, U., Inada, Y., Imamura, T., Heldin, C. H., and Moustakas, A., J. Biol. Chem. 2003, 278:33571 33582). Other studies herein have demonstrated that LMP- 1 overexpression increases 25 Smad4 levels. Thus, reduction of endogenous levels of LMP-1 using siRNA would be expected to decrease Smad4. It is hypothesized that this decrease would be caused by increased proteasomal degradation of Smad4 and that it would not occur in the presence of the proteasomal inhibitor, lactacystin. More specific evidence that reducing LMP-1 enhances proteasomal degradation of Smad4 requires demonstration of increased 30 ubiquitination of Smad4 protein when LMP-1 levels are reduced. To this end, studies are performed that demonstrate that reduction of endogenous LMP-1 by siRNA increases proteasomal degradation of Smad4.

WO 2006/102417 PCT/US2006/010419 -32 Experimental Design: Human MSCs are plated at 3 x 104 cells/well in 6-well plates, grown overnight, and treated with a control irrelevant siRNA or specific LMP-1 siRNA in the presence or absence of lactacystin. Cells are harvested 4, 8, 12, 24, 48, or 72 hr later and total RNA or cytoplasmic protein fractions prepared. An aliquot of the protein 5 fraction is analyzed by Western blot using antibodies to Smad4. Another aliquot of cytoplasmic proteins is incubated with antibody to Smad4 in a 96 well plate coated with the antibody. After washing to remove proteins not associated with Smad4, fluorescence tagged-ubiquitin antibody is applied and the fluorescence detected using a SpectraMax M2 combined fluorescence and absorbance microplate reader. This approach allows 10 quantitation of the low level of ubiquitin that is expected to be present in the sample. Expected Results: An increase in ubiquitination would likely be due to greater availability of functional Jab 1 molecules for Smad4 binding. Taken together these findings would suggest that LMP-1 could block proteasomal degradation of Smad4. A less likely alternative is that the LMP-1/Jab 1 interaction results in increased sumoylation of Smad4 15 which is an alternative mechanism of enhanced stability of Smad4 (Lee, P. S., Chang, C., Liu, D., and Derynck, R., J. Biol. Chem. 2003, 278:27853-27863). Another possibility is that the interaction of Jab 1 with Smad7, a TGF-B inhibitory Smad, could augment the BMP-2 responsiveness portion of these experiments by decreasing availability of TGF-B phospho-R-Smads (Kim, B. C., Lee, H. J., Park, S. H., Lee, S. R., Karpova, T. S., 20 McNally, J. G., Felici, A., Lee, D. K., and Kim, S. J., Mol. Cell Biol. 2004, 24:2251 2262). 2. Investigation of the requirement of Jabi for LMP-1 induced effects. Studies herein have demonstrated a direct interaction between Jab 1 and LMP- 1 and it is well known that ectopic expression of Jab1 in certain cell lines can decrease Smad4 25 steady-state levels (Wan, M., Cao, X., Wu, Y., Bai, S., Wu, L., Shi, X., Wang, N., and Cao, X., EMBO Rep. 2002, 3:171-176). Experimental Design: To demonstrate that Jab 1 is required for LMP-1-induced effects on Smad4, human MSCs are plated at 3 x 10 4 cells/well in 6-well plates, grown overnight, and treated with a control irrelevant siRNA or specific Jab 1 siRNA in the presence or 30 absence of rhBMP-2 (100 ng/ml). Cells are harvested 4, 8, 12, 24, 48, or 72 hr later and total RNA or cytoplasmic and nuclear protein fractions prepared. Cell fractions are analyzed by Western blots using specific antibodies to Smad4 and phospho-Smad1. The WO 2006/102417 PCT/US2006/010419 -33 efficacy of Jab 1 -specific siRNA to lower Jab 1 levels are assessed by real time RT-PCR and Western analysis using Jab-1 specific antibody. Expected Results: It is expected that reducing endogenous Jabl will increase cytoplasmic Smad4 levels as has been shown herein with overexpression of LMP-1. Further, in cells 5 treated with rhBMP-2, it is expected that reducing Jab 1 will increase nuclear phospho Smad1, as a result of the increased pool of cytoplasmic Smad4 available to translocate activated R-Smads to the nucleus. These results would indicate that Jab 1 regulates responsiveness of BMP-2 in MSCs by regulating Smad4 proteasomal degradation. This would complement work by others demonstrating this relationship in transformed cell 10 lines or cancer cells (Wan, M., Cao, X., Wu, Y., Bai, S., Wu, L., Shi, X., Wang, N., and Cao, X., EMBO Rep., 2002 3:171-176). 3. Effect of LMP-1 on Jab l-mediated decrease in Smad4 If decreasing endogenous Jab 1 as above increases Smad4, mimicking the effect of LMP, it would be expected that overexpression of Jabl would decrease cytoplasmic Smad4. 15 Further, if LMP- 1 exerts its effects on Smad4 levels by decreasing its Jab l-induced degradation, then overexpressed LMP- 1 should reverse the effect of overexpressing Jab . Experimental Desi n: To test this hypothesis, human MSCs are plated at 3 x 104 cells/well in 6-well plates, grown overnight, and transfected with plasmid containing JabI or empty vector. After a 1 hour recovery period cells are transduced with Ad5F35-LMP-1 20 or Ad5F35-GFP (0, 5, 10, 25, 50, 100 pfu/cell). After 1, 2, 4, 8, 12, 24 and 48 hr, cells are harvested and cytoplasmic fractions prepared. Cytoplasmic fractions are analyzed for Smad4 and Jab1 by Western blot using specific Smad4 or Jab1 antibody. Expected Results: It is expected that increasing Jab1 levels will result in decreased cytoplasmic Smad4 in the absence of overexpressed LMP-1. If the hypothesis that LMP-1 25 increases Smad4 through its interaction with Jab1 is correct, then overexpression of LMP 1 should overcome the effect of Jabl overexpression. These results would confirm that LMP-1 blocks Jab1-induced proteasomal degradation of Smad4. The LMP-1/Jabl interaction may either prevent Jab 1 binding to Smad4 or block the ability of Jab 1 to serve as an adapter protein that increases targeting of Smad4 for ubiquitination and subsequent 30 proteasomal degradation. This question could be resolved by competition binding studies. In either case, however, the overall effect of a LMP-1/Jabl interaction would be to increase Smad4 levels.

WO 2006/102417 PCT/US2006/010419 -34 If LMP-1 overexpression fails to overcome the effect of Jab 1 overexpression on Smad4 levels there may be insufficient LMP-1 levels to overcome the high level of Jab1 expression. To resolve this issue, the experiment would be repeated with higher doses of Ad5F35-LMP-1. While it has been demonstrated in other studies herein that the doses of 5 Ad5F35-LMP-1 that are applied synergize with BMP-2 in MSCs, the doses to overcome the effect of Jab 1 overexpression have not been tested. If toxicity is observed as a result of "double overexpression," a recombinant LMP-1 fusion protein (TAT-LMP-1) could be used instead of Ad5F35-LMP-1 as demonstrated in other studies herein. 4. Enhancement of physiologically relevant markers of BMP-2 responsiveness 10 To investigate the combined effect on BMP-2 responsiveness, a series of experiments are performed to demonstrate the effect of LMP-1 overexpression. First, a Smad-responsive luciferase reporter construct (9xGCCG) is first used to determine the effect of LMP-1 overexpression on luciferase activity. Next, expression of the BMP-2 regulated gene, Dlx5 is measured. Other studies disclosed herein suggest Dlx5 is important for the 15 synergistic effects of LMP-1 and BMP-2 to induce the osteoblast phenotype in MSCs. Experimental Design: Human MSCs are plated at 3 x 104 cells/well in 6-well plates, grown overnight transfected with 3ug 9xGCCG/luciferase reporter construct and treated with Ad5F35-LMP-1 (0, 5, 10 pfu/cell), (control = Ad5F35GFP), BMP-2 (100 ng/mL), or both. After 4, 8, 12, 24, 48, and 72 hr, cells are harvested and luciferase activity 20 determined (Promega). A second set of experiments is performed without transfection of luciferase reporter in which RNA is harvested at the same time points for detection of Dlx5 expression using real time RT-PCR. Expected Results: LMP-1 overexpression is expected to greatly enhance the reporter 25 construct activity in response to BMP-2. In addition, a smaller increase in activity is expected with either LMP-1 alone or BMP-2 alone. Since LMP-1 increases the pool of R Smads and the Co-Smad, this could result in some increase in activated R-Smad/CoSmad complexes in the nucleus (as suggested by other data herein). However it is expected that the largest effect observed in enhancing BMP-2 efficacy would be due to the effect of 30 LMP-1 to reduce degradation of many important proteins in the BMP-2 signaling cascade. LMP-1 is further expected to enhance the BMP-2 increase in the Dlx5 gene expression as suggested by results disclosed herein. Therefore, if the results are as WO 2006/102417 PCT/US2006/010419 -35 expected, one mechanisms by which LMP-1 increases responsiveness to BMP-2 may likely involve reduced degradation of BMPR1A, R-Smads, and Co-Smad4. Alkaline Phosphatase mRNA levels: Alkaline phosphatase activity is an early marker of the osteoblast phenotype. Human MSCs treated with Ad5/35LMP-1 (0, 1, 5, 10 pfu/cell) 5 with and without BMP-2 (100ng/ml) are harvested for RNA at day 8. Cells are washed with PBS once and cell lysates are prepared by sonication. RNA is isolated and alkaline phosphatase mRNA was quantified by RT-PCR with the alkaline phosphatase specific primers. Alkaline phosphatase activity is measured using p-nitrophenyl phosphate as substrate where nzyme activity is expressed as p-nitrophenol produced (nmoles/ml). Data 10 is normalized to 18S. Osterix Message in Human MSCs: Osterix is a novel zinc finger-containing transcription factor required for osteoblast differentiation and bone formation. Human MSCs are treated with Ad5/35LMP-1 with and without BMP-2 and harvested for RNA at day 8. RNA is isolated and osterix mRNA is quantified by RT-PCR with the osterix primers. 15 Data is normalized to 18S. Bacterial Strains and Cloning of cDNAs in bacterial expression vectors: All cloning methods are performed according to standard protocols. Escherichia coli XL1 blue and BL 21-codon plus (DE3)-RP (Stratagene) hosts are maintained on LB agar plates and grown at 37*C in the presence of ampicillin at 100 mg /L. LMP-1, LMP-1t, LMP-2, LMP-3, Smad1 20 and Smad5 cDNAs were cloned into TAT-HA vector. LMP-1 mutants were generated using the primers in Table 3. Table 3 Primers for LMP-1 mutants LMP primer name Primer Sequence (5'-3') SEQ ID NO: hLMP1Mutant A forward cgcccccgccgcggacgcagcacggtacacctttg 26 primer cac hLMP1Mutant A reverse gtgcaaaggtgtaccgtgctgcgtccgcggcgggg 27 primer gcg hLMPlMutant B forward ggcccggccctttggggcggcagcagcagctgaca 28 primer gcgccccgcaac hLMPlMutant B reverse gttgcggggcgctgtcagctgctgctgccgcccca 29 primer aagggccgggcc 25 Smurf1 cDNA is cloned into pTrcHis vector (Invitrogen). For generation of Smurfi AWW2 mutant, the primers in Table 4 are used. Mutagenesis is performed with Quikchange site-directed mutagenesis kit (Stratagene).

WO 2006/102417 PCT/US2006/010419 -36 Table 4 Primers for Smurf1 mutants Smurf1 primer name Primer Sequence (5'-3') SEQ ID NO: hSMURF1WW2 forward gtgtgaactgtgatgaacttaatcaccagtg 30 primer ccaactc hSMURF1WW2 reverse gagttggcactggtgattaagttcatcacag 31 primer ttcacac Expression and purification of recombinant proteins: Bacterial cultures are grown at 37'C 5 until O.D. 600 reaches 0.8. IPTG is added to 200 RM and the culture grown for another 8 hr. Cells are harvested and pellets are suspended in ice-cold lysis buffer (20 mM phosphate buffer, pH 7.0 containing 50 mM Tris-HCl, pH 7.5 and 0.5 M NaCl). The uniform cell suspension is sonicated (Sonicator, Model W-3 85, Heat systems-Ultrasonics, Inc.) using 4 x 15 sec bursts at minimum power-output settings in ice with a 2 min interval 10 between each burst. The lysate is centrifuged at 10,000 g at 4*C and the supernatant applied to Sephacryl S-100/S-200 columns (HiPrep 16 X 60) using AKTA FPLC system with Unicorn 4.0 software (Amersham Pharmacia Biotech) at a flow rate of Iml/min. Fractions (2-4 ml) are collected immediately after the void volume (35 ml). Aliquots, from each fraction are assayed by slot blotting, SDS-PAGE and western blotting. The fractions, 15 identified by western blots are pooled, dialyzed against 20 mM phosphate buffer, pH 7.5 containing NaCl (50 mM) and imidazole (20 mM) and applied to Ni++ affinity resin (Probond, Invitrogen) previously equilibrated with 4 x 10 ml of buffer. Non-specific proteins are washed off the column with 3 x 10 ml of 20 mM phosphate buffer, pH 6.0 containing NaCI (50 mM) and imidazole (20 mM). Affinity-bound proteins are eluted 20 using 3 x10 ml washes with 20 mM phosphate buffer, pH 4.0 containing NaCl (50 mM). Fractions containing the desired protein are pooled (based on western blot) and then concentrated and de-salted using the centriprep devices (Amicon). Proteins are quantitated using BioRad protein assay reagent. The yield of recombinant protein is routinely 0.5 to 1 mg of pure protein from every 2-liter culture. 25 Biotinylation of protein ligands: Purified protein ligands are prepared at 10 mg/ml in 50 mM sodium borate buffer, pH 8.5; 0.5 M NaCl. Various amounts of sulfo-NHS-biotin (100 mM stock in DMSO) are mixed with protein ligand to achieve a molar ratio of Sulfo NHS-biotin/protein ligand of 10.0 in a 100 pl reaction volume. After 2 hr on ice with occasional shaking the reaction is terminated with the addition of lysine to a final 30 concentration of 20 mM. The unreacted free biotin is removed by gel filtration and the WO 2006/102417 PCT/US2006/010419 -37 concentrated labeled ligand stored at -200*C until use. Labeled or unlabeled LMP-1, Smad1, Smurfl and Smad5 are prepared by using TnT coupled in vitro transcription/translation system (Promega). Slot-blot assay: 20 ul of purified Smurfl (50 pig/ml) is blotted onto nitrocellulose in slot 5 blot wells and the wells are blocked with 0.5% Tween 20 in TBST for 30 min. Biotinylated ligand (LMP-1, LMP-2, LMP-3, Smad1 or Smad5) is mixed with varying concentrations of competing proteins or peptides and incubated in slot blot wells with Smurfl for 90 min. Wells are washed and the blots were blocked with TBST containing 0.5% Tween 20. The blots are then incubated with HRP-labeled avidin for 1 hr. After 10 washes the blots are incubated with ECL substrate solution and the membranes are exposed to X-ray film for signal detection. Preparation of peptides having a protein transduction domain (PTD): Peptides are synthesized with a protein transduction domain (PTD) at the c-terminal end, (rrqrrtsldmkr, herein incorporated as SEQ ID NO: 32) according to Mi Z, et. al.;Mol. Other. 2000;2:339 15 47. Osteogenic differentiation of hMSCs: hMSCs at passage 4 are seeded at 3 x 104 cells/well in a 6-well plate. The next day, the cells are infected with Ad35LMP-1 (1-10 pfu/cell) and incubated with and without BMP-2 (100ng/ml). The medium is replaced every 3-4 days and deposition of mineral observed after 2 weeks. To assess mineralization, cultures are 20 washed with PBS and fixed in a solution of ice-cold 70% ethanol for 2-3 hours. Cultures are rinsed with water and stained for 10 minutes with 1ml of 40 mM Alizarin red (pH 4.1). Cultures are rinsed 2-3 times with PBS to reduce non-specific staining, air dried and photographed. 25 Ectopic Bone Formation Experiments Ninety 4-6 week old male athymic rats (mu/iu, Harlan; housed in sterile cages (2 rats/cage) and observed daily) are used to test LMP-1-mimetic compounds for their ability to enhance responsiveness of MSCs treated with sub-optimal doses of BMP-2 (2.5 ug/ml) to induce bone formation in the rat model of ectopic bone formation. Athymic rats are 30 used as they have no immune response to implanted foreign materials. After approval by the IACUC, 4-6 week old rats are anesthetized using 1.5% Isoflurane and the chest scrubbed with Chlorohexiderm spray. This method is consistent with the recommendations WO 2006/102417 PCT/US2006/010419 -38 of the Panel on Euthansia of the AVMA is selected because it is a rapid and painless method of euthanasia for rats. Chest implants: The surgical area is draped using sterile drapes with a hole cut to allow access to the surgical area. Surgery is performed on a draped circulating water heating 5 pad. A 1 cm skin incision is made in 4 locations on the chest, the skin separated from the muscle by blunt dissection and the discs loaded with BMP-2 + cells positioned in separate pockets. Incisions are closed using resorbable sutures. Rats typcially survive 4 weeks after which they are euthanized using canister CO 2 consistent with the recommendation of the AVMA. 10 Controls: Because of the narrow efficacious dose range for LMP-1 it is necessary to test multiple doses of each compound and to include positive control (i.e.TAT-LMP-1) and negative (i.e. cells alone or cells + BMP-2 only) controls. The positive control chosen should not induce bone formation when given at the chosen dose without BMP-2. Each dose or control is tested on multiple sites and there are multiple sites/rat. If compounds 15 are postive, more testing can be performed. Anesthesia: Anesthesia is adminsitered using 1.5% isoflurane prior to the surgical procedure. Prior to making incisions Bupivicaine 0.1-0.3 ml is given by subcutaneous or intramuscular injection around the surgical site. Buprenorphine (Buprenex) (0.05mg/100g) is injected subcutaneously immediately post-operatively and every 8 hrs 20 for 3 days post-op to relieve pain. Ectopic experiments of Smurf1 binding peptides that induce bone formation in vivo With the main hypothesis that interaction of LMP-1 with Smurfl results in increased BMP-2 signaling activity and bone formation, it is expected that peptides that 25 bind Smurf1 and activate the BMP/Smad signaling pathway in vitro can also induce bone in vivo in the rat model of ectopic bone formation. Once peptides that induce bone formation have been identified, these peptides are tested in combination with low doses of BMP-2 to determine whether there are synergistic effects (as seen in vitro) that might lower the required dose of either agent. To evaluate potential synergy, identified herein is 30 a dose of BMP-2 (2.5 ug) that induces bone formation in only 50% of the implants and a lower dose (1 ug) that consistently fails to induce bone formation in the rat ectopic model. These doses are known as the "suboptimal doses." WO 2006/102417 PCT/US2006/010419 -39 Experimental Design: In these studies, multiple doses of each of several peptides are studied. The positive control are MSCs treated with Ad5F35-LMP-1 (5 pfu/cell). MSCs (1 -2M) are mixed with appropriate doses of peptides in a 100 uL total volume and placed on a collagen disc. The disc is implanted subcutaneously on the chest of athymic rats and 5 explanted after 4 weeks. Bone formation is evaluated by palpation, x-ray and semi quantitative scoring of non-decalcified histologic sections (Edwards JT, Diegmann MH, and Scarborough NL. Clin.Orthop. 1998;219-28). It has been previously found that 1-2 million cells transduced with TAT-LMP can induce bone formation in this model, although not consistently. Hence, these studies will attempt to achieve more consistent in 10 vivo bone formation results which will be required for clinical translation. In addition the ability of the successful peptides to enhance the ability of suboptimal doses of BMP-2 to induce ectopic bone formation in this model may also be investigated. Expected Results: It is expected that the peptides that are able to bind Smurfl and enhance BMP-2 signaling will also induce bone in implants containing transduced MSCs 15 in the rat model of ectopic bone formation. Further it is predicted that MSCs treated with lower doses of the same peptides will improve the bone induction by suboptimal doses of BMP-2. Both findings are interpreted as extremely promising strategies for inducing bone in a clinical setting and these strategies would be moved to the more challenging rabbit and non-human primate models. 20 Abbreviations: MSCs, mesenchymal stem cells; hMSC, human mesenchymal stem cells; P-Smad, phosphorylated Smad; R-Smad, receptor Smad: I-Smad, inhibitor Smad; PTD, protein transduction domain; siRNA, small interfering RNA; LMP, LIM mineralization protein; Smurf, Smad Ubiquitin Regulatory Factor; BMP-2, bone morphogenic protein-2; rhBMP-2, recombinant human bone morphogenic protein-2; Jab1, Jun Activation Domain 25 Binding Protein; pfu, plaque forming units; MOI, multiplicity of infection. As used herein, the when referring to treatment of cells, tissues or animals, the terms "BMP-2" and "rhBMP-2" are synonymous. The present invention may be more fully understood by reference to the following non-limiting examples. 30 WO 2006/102417 PCT/US2006/010419 -40 EXAMPLES Example 1: LMP- 1 dramatically and synergistically increases the responsiveness of mesenchymal stem cells (MSCs) to BMP-2. It is known that LMP- 1 induces bone but produces very small amounts of BMPs. 5 In an effort to elucidate the mechanism behind this phenomena, the hypothesis that LMP-1 increases the responsiveness of cells allowing them to respond to lower levels of BMPs with respect to osteoblastice differentiation, was tested. MSC cultures transfected with the chimeric Ad5F35 vector overexpressiong LMP-1 were treated with either LMP-1 (0, 1, 5, 10 pfu/cell) or BMP-2 (100 ng/mL) alone or in 10 combination. Neither BMP-2 nor LMP-1 alone induce any bone nodule mineralization in human MSCs on day 21, but treating MSCs overexpressing LMP-1 (5-10 pfu/cell) with rhBMP-2 (100 ng/mL) induced dramatic bone nodule mineralization as shown by alizarin red staining. Thus, unexpectedly, concurrent exposure to LMP-1 enabled an ineffective dose of BMP-2 to facilitate bone formation suggesting that LMP-1 increases the BMP 15 signaling pathway activity/sensitivity. The chimeric Ad5F35 vector, in which the Ad5 fiber protein is replaced with the Ad35 fiber protein, was used to reduce susceptibility to neutralizing antibodies also tranduces human MSCs more effectively than Ad5 vectors. (Yotnda P, Onishi H, Heslop HE et al. Gene Ther. 2001;8:930-7; Gugala Z, Olmsted Davis EA, Gannon FH, Lindsey RW, and Davis AR.. Gene Ther. 2003;10:1289-96). 20 LMP-1 was also shown to increase the responsiveness of human MSCs to BMP-2 as evidenced by alkaline phosphatase mRNA levels and an increase in enzyme activity. Example 2: LMP Variant (LMP-2) does not induce nodule formation in rat calvarial osteoblast cultures. Secondary rat calvarial osteoblasts do not spontaneously differentiate without 25 exposure to a stimulus such as glucocorticoid (GC). Cells were transfected with plasmids containing three LMP (hLMP-1; hLMP-2 and hLMP-3) variants and assessed for multilayer mineralized nodule formation 14 days after treatment. Control cells received no treatment and all data were normalized to the control group. Overexpression of hLMP 1 resulted in nodule formation (225 nodules) comparable to that seen with GC (275 30 nodules). hLMP-3, which is a truncated version of LMP-1 also induced nodule formation (290 nodules). However, hLMP-2 which lacks a 45aa region failed to induce nodule WO 2006/102417 PCT/US2006/010419 -41 formation (20 nodules), suggesting that the missing 45aa are required for LMP's osteoinductive properties. Example 3: Detection of Smad1 and phosphorylated Smad1 (P-Smadl) Measuring activation of the BMP signaling pathway requires the ability to measure 5 Smad1 and Smad5. SDS-PAGE separated cytoplasmic and nuclear protein blots were probed with Smad-1 specific antibody. A comparison of cytoplasmic and nuclear protein extracts from untreated pleuripotent cells demonstrated that most of the Smad1 (54 kDa) was detected in the cytoplasmic fraction. For detection of P-Smadl (phosphorylated Smad1) in the nucleus, SDS-PAGE 10 blots of nuclear proteins were prepared from MSCs overexpressing LMP-1 at 4 and 8 hours following infection with Ad5F35-LMP-1. The blots were probed with primary antibody specific to P-Smadl. The binding of primary antibody was detected using the HRP-labeled second antibody after signal development by enhanced chemi-luminescence (ECL). A single band at the apparent size of 54 kDa showed an increase of 15 phosphorylated Smad1 in the nuclear fraction as early as 4 hours in cells over-expressing LMP-1 compared to control MSCs not over-expressing LMP-1. Human MSCs treated with BMP-2 also have increased levels of cytoplasmic phosphorylated Smad1 in the presence of overexpressed LMP-1. For detection of P-Smadl (phosphorylated Smad1), in the cytoplasm, SDS-PAGE 20 blots of cytoplasmic proteins were prepared from MSCs overexpressing LMP-1 at 4 and 8 hours following infection with LMP- 1 delivered as Ad5F35-LMP-1. Cells were treated with rhBMP-2 at 1 OOng/mL or 200 ng/mL alone or in combination with LMP-1 (5 pfu/cell). rhBMP-2 treatment at either 1 OOng/mL or 200 ng/mL resulted in a small increase in cytoplasmic phosphorylated Smad1 (64 kDa band). Addition of rhBMP-2 (100 25 ng/mL) + LMP- 1 (5 pfu/cell) however, resulted in a significant increase (over 10 fold increase) in the amount of phosphorylated Smad1 protein which was not seen with LMP-1 at 5pfu/cell alone. These data support the hypothesis that LMP-1 blocks the Smurf1 mediated degradation of unphosphorylated Smad1 resulting in a larger pool of Smad1 and therefore a greater amount of phosphorylated Smad1 is produced for a given amount of 30 BMP-2.

WO 2006/102417 PCT/US2006/010419 -42 Example 4: Investigation of increased Smad1 induced by BMP and LMP-1. To determine whether the increased level of Smad1 protein might be due to increased expression of Smadi, MSCs were treated with rhBMP-2 (100 ng/mL) or Ad5F35-LMP-1 (5 pfu/cell). Control cells were untreated. After 4 or 8 hrs total RNA was 5 harvested and Smadi mRNA was measured by real time RT-PCR. LMP-1 and BMP-2 each increased Smad1 mRNA by 4 hours, 5 fold and 6.2 fold, respectively. At 8 hours Smad1 mRNA was increased by 4.5 fold and 2 fold for LMP-1 and BMP-2 respectively. The data suggest that LMP-1 causes increased nuclear levels of P-Smadl via increased expression of Smad1 mRNA, which results in increased Smad1 protein that can be 10 phosphorylated by the BMP receptor kinase. Given the short time frame of P-Smadl accumulation in the nucleus shown above however, this mechanism is not the likely explanation for thesshort term (4 hr) increase. It is more likely a result of decreased degradation of P-Smads that is responsible for the short term regulatory mechanism. 15 Example 5: LMP-1 and BMP-2 synergistically increase expression of Smad-regulated genes. In order to measure immediate downstream markers of BMP signaling, mRNA levels of Dlx5, a gene known to be induced by the BMP-Smads was measured. MSCs were untreated or treated with BMP-2 (100 ng/mL), Ad5F35-LMP-1 (5 pfu/cell), or both. 20 After 24 hr total RNA was harvested and Dlx5 mRNA levels measured by real time RT PCR. Data are expressed as fold change in Dlx5 mRNA. Untreated cells exhibited 1 fold increase. Cells treated with BMP-2 demonstrated 6 fold increase, while cells treated with LMP-1 demonstrated 2 fold increase. A synergistic effect was seen on treatment with both LMP-1 (at 5 pfu/cell) and BMP-2 (at 100 ng/mL), showing 25 fold increase. 25 While protein levels were not measured in this study, methods are well known and include Western analysis and ELISA. Example 6: Measurement of other Smad-regulated targets Measurement of the mRNA and protein levels of Smad6, and BMP-2 can also be performed. For these studies, human MSCs are plated at 3 x 10 4 cells/well in 6-well 30 plates, grown overnight, and treated with Ad5F35-LMP-1 (0, 5, 10 pfu/cell), BMP-2 (100 ng/ml) or both agents in combination or in series. The control treatment would be the vector, Ad5F35-GFP. After 1, 2, 4, 8, 12, 18, 24, 36, 48, and 72 hrs, cells are harvested WO 2006/102417 PCT/US2006/010419 -43 for analysis of total RNA or cytoplasmic and nuclear protein. After reverse transcription, real-time PCR is performed using primers specific for human BMP-2 and Smad6. Similarly, secreted, cytoplasmic and nuclear proteins can be analyzed by ELISA using commercially available antibodies to the proteins of interest, with alkaline phosphatase 5 conjugated secondary antibodies. Specificity of the antibody can be determined by Western blot to assure that there is only one immunoreactive species. Intracellular BMP-2 is expected to show several immunoreactive bands corresponding to various cleavage steps during processing of pro-BMP-2. The processed BMP-2 protein will also likely be found largely in the medium, as it is rapidly secreted. 10 Measurement of BMP-2 transcripts can be used for determining the relative effect of LMP-1 and BMP-2 on BMP-2 gene expression, as the majority of the protein measured in the medium from BMP-2 treated cells may be exogenously added as a treatment. Example 7: LMP-1 is associated with an 85 kDa nuclear protein. Recombinant LMP-1 was labeled with SBED-biotin transfer reagent and incubated 15 as bait with nuclear proteins. Biotin transfer to target proteins was accomplished by photo activation and decoupling was performed by reduction of bound protein-partners. Enrichment of biotinylated proteins was performed using neutravidin-beads. Biotinylated proteins were separated by SDS-PAGE and detected on Western blots using HRP-labeled neutravidin and ECL. The corresponding coomassie stained bands were excised for 20 tryptic digestion and MALDI-TOF analysis. LMP-1 was seen to associate with three protein bands; one 85 kDa band and two smaller bands. The smaller size bands were sequenced and represent cytoskeletal proteins likely involved with cellular localization of full length LMP-1 and not its osteoinductive properties. Based on partial sequence analysis, an 85 kDa band was investigated further (see below). 25 Example 8: Peptide mass profile analysis of the 85 kDa protein binding to LMP-1. The SDS-PAGE separated LMP-1 binding protein bands were in-gel digested by 5% (w/w) trypsin. Molecular mass of tryptic-peptides was obtained by MALDI-TOF and analyzed by PepIdent. The resulting 85 kDa candidate protein matches from the data base were analyzed. Smurfl and its splice-isoform variant showed the best ranking with peptide 30 mass profile. Although the tryptic digestion was done on doublet protein bands, the unique identity was still able to be determined due to identical peptide profiles of both isoforms of WO 2006/102417 PCT/US2006/010419 -44 Smurfi in tryptic digestion. By comparison the peptide profile obtained did not match the peptide mass profile of Smurf2. Example 9: The 85 kDa protein reacts with Smurfl/2 antibody. Identity of the gel purified LMP-1 binding protein was verified by performing a 5 Western blot using antibody that binds both Smurf1 and Smurf2. The doublet containing two protein bands, one running at 86 kDa and the other running at 84 kDa, were both immuno-reactive and probably represent two known splice variants of Smurf1 of similar molecular size rather than Smurfi and Smurf2. Based on molecular size, immuno reactivity with Smurfl/2 antibody in Western blots, and tryptic-peptide profiles, the 10 putative LMP-binding protein was identified as Smurf1. Example 10: Immunoprecipitation endogenous Smurfl and LMP-1 in MSCs. Human MSCs were plated at 3 x 10 4 cells/well in 6-well plates, grown overnight, and treated with Ad5F35-LMP-1 (0, 5, 10 pfu/cell), or Ad5F35-GFP. After 2, 4, 8, 12, 24, and 48 hrs, cells were harvested and nuclear proteins mixed with LMP-1 or Smurfl 15 antibody and subjected to immunoprecipitation. Eluted proteins were separated by SDS PAGE and Western analysis performed using LMP-1 or Smurf1 antibodies. Nuclear protein extracts of untreated MSCs were incubated with LMP-1 antibody and immunoprecipitated using protein-A beads. The immunoprecipitated proteins were concentrated and analysed in Western blots with LMP-1 and Smurfl/2 antibody, 20 separately. Both endogenous LMP-1 and Smurf1 were present in the complex immunoprecipitated with the LMP-1 antibody from untreated MSCs. This observation was confirmed by detection of LMP-1 when Smurf1 antibody was used for the immunoprecipitation. Other studies demonstrated interaction of these proteins in the nuclear fraction. It is 25 unknown whether the Smurfl/LMP-1 interaction will also be observed in the cytoplasm but these studies can be performed according to the methods described herein. Example 11: LMP-1 interacts with Smurfi via the Smurf1 WW domain. Blots of SDS-PAGE resolved nuclear proteins from MSCs showed a predominant band at 85 kDa when probed with Smurfl WW-domain antibody in Western blots. Pre 30 incubation of LMP-1 (10 uM) with these blots inhibited Smurf1 antibody binding. Binding competition between LMP-1 protein and WW-antibody towards the same target sequence suggests that the interaction with LMP-1 occurs at the WW-domain(s) of Smurf1. The WO 2006/102417 PCT/US2006/010419 -45 WW-domains are the region that enables Smurfl to bind to critical BMP signaling Smad1 and Smad5 and target them for degradation. These data support the hypothesis that LMP 1 blocks Smurf1 from interacting with and targeting Smad1 and Smad5. Not only did LMP-1 inhibit binding, truncated LMP variants also inhibit Smurf 5 WW-domain antibody binding to Smurf1. Blots of SDS-PAGE resolved nuclear proteins from untreated MSCs were incubated with the antibody specific for the Smurfl WW domain. Both full length recombinant LMP-1 and recombinant LMP-lt (a C-terminal truncated version of LMP-1), when preincubated on the blots, were able to prevent the WW-domain antibody from binding to Smurf1. These data suggest that the LIM domains 10 (located at the C-terminus of LMP-1) are not needed for the direct interaction with Smurfl. Example 12: Characterization of the binding properties of Smurfl, LMP-1, Smadl, and Smad5 using purified proteins to confirm competitive binding of LMP-1. Recombinant Smad1 and Smad5 are prepared by bacterial expression and 15 purification. Smads and LMP-1 are then fluorescently labeled. The Enzyme Linked Sorbent Assay (ELSA) described herein is performed in microtiter plates. Binding curves for evaluating the competitive binding to Smurfl of fluorescently labeled LMP-1 with unlabeled Smadl or Smad5 are then obtained. Conversely, the competitive binding of fluorescently labeled Smads with unlabeled LMP-1 may also be measured. Fluorescence is 20 monitored using a Biolumin 960 combined fluorescence and absorbance microplate reader. Scatchard analyses, well known in the art, can serve to assess the binding affinity, dissociation constants, number of binding sites, and stoichiometry for each protein involved in the interaction. The results can then predict competition between LMP-1 and Smads for binding Smurfl at 25 one or two sites. Competitive binding at two sites might occur, as Smurfl contains two WW domains and it is uncertain whether all or some of the proteins bind one or both sites. If Smad5 does not accumulate in the nucleus it might not be expected to compete with LMP-1 for Smurfl binding because it might be assumed that LMP-1 binds Smurfl at a different WW domain than Smad5. 30 Example 13: Identification of two regions within LMP-1 with high affinity for Smurf1 WW-domains WO 2006/102417 PCT/US2006/010419 -46 On finding that full length LMP-1 binds Smurf1 WW domains, comparative sequence analysis of three LMP- 1 variants was performed and identified a 45 amino acid "osteoinductive region" that is present in the two LMP-1 isoforms that induce bone formation, but is not present in a third isoform that fails to induce bone. These LMP 5 variants (LMP-1, LMPlt and LMP-3) were created which contain a unique peptide sequence, (AADPPRYTFAPSVSLNKTARPFGAPPPADSAPQQNG; SEQ ID NO: 33) within a larger osteoinductive region which includes two Smurfl WW-domain interaction sites (WW interacting site A comprising the sequence ADPPRYTFAP; herein referred to as SEQ ID NO: 34; and WW interacting site B comprising the sequence GAPPPADSAP; 10 herein referred to as SEQ ID NO: 35). The WW domain interacting sites are absent in a non-osteogenic LMP variant (LMP-2). The 45 amino acid osteoinductive region appears to be critical for the bone-forming activing of LMP. Two LMP variants (LMP lt and LMP-3) are truncated at the carboxy-terminus, and do not contain the LIM domains, but do induce bone formation. Thus, it is the 45 amino acid osteoinductive region and not the 15 LIM domains that are required for bone formation (Liu, Y, et. al.; J. Bone Min. Res., 2002:17:406-414). Slot blots prepared with recombinant Smurf1 and hybridized with biotin-labeled LMP variants demonstrated that only the LMP variants containing the WW-domain interaction sites (LMP-1, LMP-lt and LMP-3) were able to bind to Smurf1. 20 To determine which of the two WW-domain interaction sites were required for the binding of LMP with Smurf1, two mutant LMP-1 proteins that are mutated in either WW interaction site A (Prolines at positions 100 and 101 of the 457 amino acid parent protein being converted to Alanine, herein incorporated as SEQ ID NO: 36; LMP-1AWWA) or site B (Prolines at positions 122, 123 and 124 of the 457 amino acid parent protein being 25 converted to Alanine, herein incorporated as SEQ ID NO: 37 ; LMP-1AWWB) were prepared. The mutations remove proline residues that are required for interaction with the Smurfl WW domain and disrupt the PY motif in each of the two sites. LMP-1AWWA-SEQ ID NO: 36; <200> SEQUENCE CHARACTERISTICS: 30 <210> SEQ ID NO 36 <211> LENGTH: 457 <212> TYPE: PRT <213> ORGANISM: Homo sapiens <400> SEQUENCE: 36 35 Met Asp Ser Phe Lys Val Val Leu Glu Gly Pro Ala Pro Trp Gly Phe 1 5 10 15 WO 2006/102417 PCT/US2006/010419 -47 Arg Leu Gln Gly Gly Lys Asp Phe Asn Val Pro Leu Ser Ile Ser Arg 20 25 30 Leu Thr Pro Gly Gly Lys Ala Ala Gln Ala Gly Val Ala Val Gly Asp 35 40 45 5 Trp Val Leu Ser Ile Asp Gly Glu Asn Ala Gly Ser Leu Thr His Ile 50 55 60 Glu Ala Gln Asn Lys Ile Arg Ala Cys Gly Glu Arg Leu Ser Leu Gly 65 70 75 80 Leu Ser Arg Ala Gln Pro Val Gln Ser Lys Pro Gln Lys Ala Ser Ala 10 85 90 95 Pro Ala Ala Asp Ala Ala Arg Tyr Thr Phe Ala Pro Ser Val Ser Leu 100 105 110 Asn Lys Thr Ala Arg Pro Phe Gly Ala Pro Pro Pro Ala Asp Ser Ala 115 120 125 15 Pro Gln Gln Asn Gly Gln Pro Leu Arg Pro Leu Val Pro Asp Ala Ser 130 135 140 Lys Gln Arg Leu Met Glu Asn Thr Glu Asp Trp Arg Pro Arg Pro Gly 145 150 155 160 Thr Gly Gln Ser Arg Ser Phe Arg Ile Leu Ala His Leu Thr Gly Thr 20 165 170 175 Glu Phe Met Gln Asp Pro Asp Glu Glu His Leu Lys Lys Ser Ser Gln 180 185 190 Val Pro Arg Thr Glu Ala Pro Ala Pro Ala Ser Ser Thr Pro Gln Glu 195 200 205 25 Pro Trp Pro Gly Pro Thr Ala Pro Ser Pro Thr Ser Arg Pro Pro Trp 210 215 220 Ala Val Asp Pro Ala Phe Ala Glu Arg Tyr Ala Pro Asp Lys Thr Ser 225 230 235 240 Thr Val Leu Thr Arg His Ser Gln Pro Ala Thr Pro Thr Pro Leu Gln 30 245 250 255 Ser Arg Thr Ser Ile Val Gln Ala Ala Ala Gly Gly Val Pro Gly Gly 260 265 270 Gly Ser Asn Asn Gly Lys Thr Pro Val Cys His Gln Cys His Lys Val 275 280 285 35 Ile Arg Gly Arg Tyr Leu Val Ala Leu Gly His Ala Tyr His Pro Glu 290 295 300 Glu Phe Val Cys Ser Gln Cys Gly Lys Val Leu Glu Glu Gly Gly Phe 305 310 315 320 Phe Glu Glu Lys Gly Ala Ile Phe Cys Pro Pro Cys Tyr Asp Val Arg 40 325 330 335 Tyr Ala Pro Ser Cys Ala Lys Cys Lys Lys Lys Ile Thr Gly Glu Ile 340 345 350 Met His Ala Leu Lys Met Thr Trp His Val His Cys Phe Thr Cys Ala 355 360 365 45 Ala Cys Lys Thr Pro Ile Arg Asn Arg Ala Phe Tyr Met Glu Glu Gly 370 375 380 Val Pro Tyr Cys Glu Arg Asp Tyr Glu Lys Met Phe Gly Thr Lys Cys 385 390 395 400 His Gly Cys Asp Phe Lys Ile Asp Ala Gly Asp Arg Phe Leu Glu Ala 50 405 410 415 Leu Gly Phe Ser Trp His Asp Thr Cys Phe Val Cys Ala Ile Cys Gln 420 425 430 Ile Asn Leu Glu Gly Lys Thr Phe Tyr Ser Lys Lys Asp Arg Pro Leu 435 440 445 55 Cys Lys Ser His Ala Phe Ser His Val 450 455 LMP-1AWWB- SEQ ID NO: 37 WO 2006/102417 PCT/US2006/010419 -48 <200> SEQUENCE CHARACTERISTICS: <210> SEQ ID NO 37 <211> LENGTH: 457 <212> TYPE: PRT 5 <213> ORGANISM: Homo sapiens <400> SEQUENCE: 37 Met Asp Ser Phe Lys Val Val Leu Glu Gly Pro Ala Pro Trp Gly Phe 1 5 10 15 Arg Leu Gln Gly Gly Lys Asp Phe Asn Val Pro Leu Ser Ile Ser Arg 10 20 25 30 Leu Thr Pro Gly Gly Lys Ala Ala Gln Ala Gly Val Ala Val Gly Asp 35 40 45 Trp Val Leu Ser Ile Asp Gly Glu Asn Ala Gly Ser Leu Thr His Ile 50 55 60 15 Glu Ala Gln Asn Lys Ile Arg Ala Cys Gly Glu Arg Leu Ser Leu Gly 65 70 75 80 Leu Ser Arg Ala Gln Pro Val Gln Ser Lys Pro Gln Lys Ala Ser Ala 85 90 95 Pro Ala Ala Asp Pro Pro Arg Tyr Thr Phe Ala Pro Ser Val Ser Leu 20 100 105 110 Asn Lys Thr Ala Arg Pro Phe Gly Ala Ala Ala Ala Ala Asp Ser Ala 115 120 125 Pro Gln Gln Asn Gly Gln Pro Leu Arg Pro Leu Val Pro Asp Ala Ser 25 130 135 140 Lys Gln Arg Leu Met Glu Asn Thr Glu Asp Trp Arg Pro Arg Pro Gly 145 150 155 160 Thr Gly Gln Ser Arg Ser Phe Arg Ile Leu Ala His Leu Thr Gly Thr 165 170 175 30 Glu Phe Met Gln Asp Pro Asp Glu Glu His Leu Lys Lys Ser Ser Gln 180 185 190 Val Pro Arg Thr Glu Ala Pro Ala Pro Ala Ser Ser Thr Pro Gln Glu 195 200 205 35 Pro Trp Pro Gly Pro Thr Ala Pro Ser Pro Thr Ser Arg Pro Pro Trp 210 215 220 Ala Val Asp Pro Ala Phe Ala Glu Arg Tyr Ala Pro Asp Lys Thr Ser 225 230 235 240 Thr Val Leu Thr Arg His Ser Gln Pro Ala Thr Pro Thr Pro Leu Gln 40 245 250 255 Ser Arg Thr Ser Ile Val Gln Ala Ala Ala Gly Gly Val Pro Gly Gly 260 265 270 Gly Ser Asn Asn Gly Lys Thr Pro Val Cys His Gln Cys His Lys Val 275 280 285 45 Ile Arg Gly Arg Tyr Leu Val Ala Leu Gly His Ala Tyr His Pro Glu 290 295 300 Glu Phe Val Cys Ser Gln Cys Gly Lys Val Leu Glu Glu Gly Gly Phe 305 310 315 320 Phe Glu Glu Lys Gly Ala Ile Phe Cys Pro Pro Cys Tyr Asp Val Arg 50 325 330 335 Tyr Ala Pro Ser Cys Ala Lys Cys Lys Lys Lys Ile Thr Gly Glu Ile 340 345 350 Met His Ala Leu Lys Met Thr Trp His Val His Cys Phe Thr Cys Ala 355 360 365 55 Ala Cys Lys Thr Pro Ile Arg Asn Arg Ala Phe Tyr Met Glu Glu Gly 370 375 380 Val Pro Tyr Cys Glu Arg Asp Tyr Glu Lys Met Phe Gly Thr Lys Cys 385 390 395 400 WO 2006/102417 PCT/US2006/010419 -49 His Gly Cys Asp Phe Lys Ile Asp Ala Gly Asp Arg Phe Leu Glu Ala 405 410 415 Leu Gly Phe Ser Trp His Asp Thr Cys Phe Val Cys Ala Ile Cys Gln 420 425 430 5 Ile Asn Leu Glu Gly Lys Thr Phe Tyr Ser Lys Lys Asp Arg Pro Leu 435 440 445 Cys Lys Ser His Ala Phe Ser His Val 450 455 10 Recombinant Smurfi was run on SDS-PAGE, transferred to membrane, and incubated with buffer, LMP-1 protein, LMP-2 protein (negative control, no osteoinductive region), LMP-1AWWA or LMP-1AWWB. After washing, Smurfl WW domain antibody was applied and then detected with HRP-labeled secondary antibody. Slot blots of competitive binding assays were analyzed as before and it was found that 15 only WW-domain interaction site B was required for the interaction with Smurf1. The specificity of these interaction sites was confirmed by demonstrating that LMP-1 was no longer able to bind to a mutant form of Smurf1 which had its WW2 domain deleted (SmurflAWW2). Example 14: Peptide mimics of LMP-1 20 To design and test molecules that mimic the LMP- 1 interaction with Smurfl, known crystallography data and homology modeling of related homologues similar to the WW-2 domain in Smurf1 and the two potential interacting motifs in LMP-1 (WW-A and WW-B) as defined by iSPOT were used (Brannetti, B. and Helmer-Citterich, M., Nucleic Acids Res. 2003; 31:3709-3711). The MODELLER program was used to assign structure 25 to the two interacting elements and to model templates. The designed Smurfl WW-2 template showed a close (up to 70%) match to the coordinates of crystallographic data available for homologous WW domains from the protein data bank (PDB) (Fiser, A. and Sali, A., Methods Enzymol. 2003; 374:461-491). DOCKING and SSA (Surface Solvant Accessability) programs were then used to define the key residues in each binding partner 30 (Morris, G. M., Goodsell, D. S., Huey, R., and Olson, A. J., J. Comput. Aided Mol.Des., 1996, 10:293-304). A directory of commercially available low molecular weight and cell penetrable chemicals were screened using the LUDI program with both complimentary screening (WW-2) and analogue screening (WW-A or WW-B) to identify candidate compounds (Honma, T., Med. Res. Rev. 2003, 23:606-632).We used computational 35 mutagenesis to eliminate non-specific compounds and cross-matched the complimentary and analogue compound lists to arrive at 75 candidate compounds for each WW WO 2006/102417 PCT/US2006/010419 -50 interacting domain of LMP-1. The specificity of the LMP motif that interacts with Smurf1 was confirmed and it was determined that the activity of the full length LMP protein could be replicated by a small peptide. We synthesized small peptides comprising various portions of the osteoinductive region, 5 (KPQKASAPAADPPRYTFAPSVSLNKTARPFGAPPPADSAPQQNGQPLRPLVPDA SKQRLM; herein referred to as SEQ ID NO: 38, BOLDED fragments represent the WW domain interacting sites, with ADPPRYTFAP representing Site A, SEQ ID NO: 34 and GAPPPADSAP representing Site B, SEQ ID NO: 35) containing one, two, or none of the putative LMP WW-domain interacting sites. The Table below lists the sequences of 10 Peptide 1, 3, 5, and 7 that were designed to the osteogenic region of LMP-1. These peptides were chemically synthesized, HPLC purified for in vivo bone formation studies. Only peptide 7 is acetylated at N-terminus, and was therefore not expected to alter the peptide function. Table 5 15 Peptide mimics Peptide SEQ ID Amino Acid Sequence # NO: 1 39 H 2 N-APSVSLNKTARPFGAPPPADSAGGRRQRRTSKLMKR-CONH, 3 40 H 2 N-KPQKASAPAADPPRYTFAPSVSGGRRQRRTSKLM l-CONH 2 5 41 H 2

N

ASAPAADPPRYTFAPSVSLNKTARPFGAPPPADSAPQQNGGGRRQ

RRTSKLMKR-CONH

2 7 42 Acetyl GAPPPADSAPQQNGQPLRPLVPDASKQRLMGGRRQRRgKLMKR

NH

2 Competition studies (across the range of 0.00luM to 1.0uM, and including 0.001, 0.01, 0.1 and luM) of the peptides revealed that the two peptides that contained intact 30 WW-domain interacting Site B were able to compete with full length LMP-1 in a dose dependent manner as well as with Smad1 and Smad5 for binding with Smurf1 at concentrations of 1.0uM and 1OuM.

WO 2006/102417 PCT/US2006/010419 -51 This result is consistent with the mutational analysis above establishing Site B as the more critical site in LMP. To determine if the ability of the small LMP-derived peptides to competitively bind Smurf1 has physiologic significance, the peptides were linked to a protein transduction domain (PTD) that facilitates entry into cells (Sequence: 5 RRQRRTSKLMKR, herein incorporated as SEQ ID NO: 32). We found that only peptide 7 had the ability to induce bone formation in a rat ectopic bone formation assay. Example 15: Osteoinductive region fusion peptides To verify that the 36 amino acid sequence (AADPPRYTFAPSVSLNKTARPFGAPPPADSAPQQNG; SEQ ID NO: 33) is 10 responsible for the LMP- 1 interaction with Smurf1, an "osteoinductive region-fusion peptide" (PTD-LMP-1/OR) containing a PTD that readily enters cells (RRQRRTSKLMKR, herein incorporated as SEQ ID NO: 32) was designed. Initial studies were performed to determine a dose of PTD-LMP-1/OR that enhanced the effect of 100 ng/ml BMP-2 to mineralize hMSCs. Once two effective doses, 15nM and 17.5nM, 15 were established, hMSCs were plated at 3 x 104 cells/well in 6-well plates, grown overnight, and treated with PTD-LMP-1/OR, BMP-1 (100 ng/ml) or both agents. The control was PTD-PGal. After 1, 2, 4, 8, 12, 24, 36, 48, and 72 hrs, cells were harvested and total RNA plus cytoplasmic and nuclear protein fractions prepared. The nuclear protein fraction from the untreated sample was analyzed for the ability of purified PTD 20 LMP-1/OR to compete with Smurf1 antibody binding to Smurf1 on a Western blot. All protein samples were analyzed for the presence of P-Smadl and P-Smad5 by Western blot using appropriate antibody. All RNA samples were analyzed by real-time RT-PCR for mRNA levels of Dlx5, Smad6, and BMP-2. Similarly, protein fractions were analyzed by ELISA using commercially available antibodies to Dlx5, Smad6, and BMP-2. 25 The results illustrate that the 36aa peptide has all the ability of full length LMP-1 to compete with WW domain antibody binding, to induce increased nuclear levels of P Smad1 and P-Smad5, and to increase expression of BMP/Smad regulated genes. It is noted that PTD-fusion peptide derivatives of one of the osteoinductive LMP isoforms (LMP-3) have been shown by others to induce bone formation (Pola E, Gao W, Zhou Y et 30 al. Gene Ther. 2004;11:683-93).

WO 2006/102417 PCT/US2006/010419 -52 Example 16: Full length and truncated LMP-1 variants form bone in ectopic and orthotopic models in vivo. Bone marrow cells and leukocytes were transfected with the cDNA of full length or truncated LMP-1 lacking LIM domains. Collagen discs containing the engineered cells 5 were implanted subcutaneously into athymic rats. Histology of the implants showed ectopic and orthotopic bone formation. New bone trabeculae lined with osteoblasts were seen in the implants containing cells overexpressing truncated LMP-1 (LMP-1t) and the absence of new bone in implants containing control cells. This demonstrates that genetically engineered cells can form bone in vivo and confirms that the LIM domains of 10 LMP-1 are not needed for bone formation. Example 17: Bone marrow cells transfected with LMP-1 induce bone in the rat model of spine fusion Spine fusion is a more challenging bone formation model than an ectopic site. Rats were implanted with carrier soaked with bone marrow cells which had been 15 transfected with the cDNA of LMP- 1 in an active or inactive form. Lateral radiogroaphy of the animals at 4 weeks indicated solid spine fusion when the marrow cells expressed active LMP-1 protein but no evidence of bone formation was seen when the marrow cells had been transfected with an inactive form of LMP-1. These data indicate that genetically engineered bone marrow cells can induce rat spine fusion. 20 Example 18: A TAT-LMP-1 fusion protein enters the nucleus of cells and forms bone in vivo Fusion proteins that contain an 11 aa sequence, found in the HIV protein Tat, are known to readily enter cells within minutes (Nagahara H, Vocero-Akbani AM, Snyder EL et al. Nat.Med.1998; 4:1449-52). This method of protein delivery is superior to adenoviral 25 gene transfer because it avoids immune response issues. To confirm that a LMP fusion protein would enter cells, a FITC labeled LMP-1 Tat fusion protein (TAT-LMP-1) was synthesized and fluroescence monitored for localization in the MSCs when treated at doses of 10, 25 and lOOnM. The fusion protein readily entered nucleated blood cells in a dose dependent manner with 2, 25 and 90% of the cells positive for the label, respectively. 30 Western blots from another experiment confirmed the ability of TAT-LMP-1 to enter the nucleus. X-rays showed good bone formed on collagen discs in rats implanted with cells treated for 30 min with TAT-LMP-1. The bone induction with TAT-LMP-1 was not WO 2006/102417 PCT/US2006/010419 -53 always consistent between experiments and thus this was not an optimal "stand alone" strategy for bone formation. A better understanding of the mechanism of action of LMP- 1 and a designer small molecule may allow it to perform more consistently as a "stand alone" initiator of bone formation or as a means of increasing the responsiveness to BMPs. 5 Example 19: LMP-l siRNA In an effort to selectively inhibit expression of specific genes in order to elucidate aspects of the proposed LMP-1 mechanism of action, siRNA were designed to target LMP-1. Use of small inhibitory RNAs (siRNAs) to destroy specific mRNAs has become the method of choice for specifically silencing expression of genes (Martinez J, 10 Patkaniowska A, Urlaub H, Luhrmann R, and Tuschl T. Cell 2002;110:563-74; Maeda S, Hayashi M, Komiya S, Imamura T, and Miyazono K.. EMBO J. 2004; 23:1-12). Conditions for siRNA treatment were first optimized and MC3T3-El cells were transfected with siRNA (1 Opmol) to LMP- 1 disclosed herein and according the methods disclosed herein. LMP- 1 mRNA levels were reduced by 90% 48 hrs after siRNA treatment 15 as compared to no treatment. Reduced bone nodule mineralization was also observed in addition to the decrease in LMP-1 mRNA levels. Example 20: Smurfl siRNA In an effort to demonstrate the effects of reducing the level of functional Smurfl to support the hypothesis that LMP-1 acts by decreasing the amount of Smurf1 available to 20 bind to Smadl/5, siRNA to Smurf were designed and tested for target reduction in TE85 human osteosarcoma cells. Cells were treated according the methods taught herein and total RNA was harvested and Smurfl mRNA levels measured by RT-PCR using primers specific for Smurfl. Smurf2 mRNA levels were also measured using specific primers and these levels did not change. There was a single product of each primer set which was 25 sequenced to confirm its identity. The results indicate that Smurf1 siRNA produced a dose-dependent decrease in Smurfl mRNA levels. Example 21: Osterix Message in Human MSCs Osterix is a novel zinc finger-containing transcription factor required for osteoblast 30 differentiation and bone formation. Human MSCs were treated with Ad5/35LMP-1 (0, 1, 5, 10 pfu/cell) with and without BMP-2 (1 Ong/mL) and harvested for RNA at day 8. RNA was isolated and osterix mRNA quantified by RT-PCR with the osterix primers.

WO 2006/102417 PCT/US2006/010419 -54 Data was normalized to 18S. These data illustrate that LMP-1 increased the responsiveness of human MSCs to BMP-2 as evidenced by increase in osterix message. The data are summarized in Table 6. Table 6 5 Increase in Osterix mRNA Treatment No BMP- LMP-1 LMP-1 LMP-1 LMP-1 treatment 2 (5pfu/cell) (10 (5pfu/cell) (10 pfu/cell) +BMP-2 pfu/cell) +BMP-2 Fold 1.0 88.3 0.4 0.4 178.6 424.6 Increase in Osterix mRNA Example 22: Overexpression of LMP-1 increases BMPRIA (ALK3) levels in human MSCs. Human MSCs were treated for 8 hours with rhBMP-2 (100 ng/mL) alone or with 10 Ad5F35-LMP-1 (5 pfu/cell) and cytoplasmic proteins were enriched for the plasma membrane fraction and resolved by SDS-PAGE separation. The blot was then probed with antibody specific for the Type IA BMP receptor (BMPR1A/ALK3) and a predominant band at the expected size for BMPR1A (55 kDa) was observed. rhBMP-2 treatment (100ng/mL) resulted in an expected increase in BMPR1A over untreated control 15 cells. However, when rhBMP-2 was given in the presence of LMP-1 a significantly greater increase in the BMPR1A levels was observed. These data support the hypothesis that LMP-1 interrupts the Smurfl/Smad6 mediated degradation of BMP receptors. Example 23: LMP- 1 Interacts with Jun Activation Domain Binding Protein (Jab 1). 20 The yeast-two-hybrid (Y2H) system (Clontech) was used to identify other proteins which could interact with LMP-1. Positive clones were selected based on Y2H screening of a bone marrow library. The sequencing and database matching of 10 putative positive clones identified Jab 1 as a likely candidate binding partner for LMP- 1. It was then determined if this association occurred in cells. Immunoprecipitation of cytoplasmic 25 proteins using LMP-1 antibody beads demonstrated that Jab1 was found in complexes WO 2006/102417 PCT/US2006/010419 -55 with LMP-1. Although these data demonstrate that an association between Jab 1 and LMP-1 occurs in cells, they do not establish direct binding of the two proteins. Example 24: LMP-1 binds to Jab1 To demonstrate that LMP-1 binds Jab1 directly, cytoplasmic proteins from human 5 osteoblastic TE-85 cells were separated by SDS-PAGE and blots were probed with Biotin LMP-1. The bound biotin-LMP-1 was detected using neutravadin-HRP. Two bands were present on the blot and demonstrated that LMP-l is capable of binding directly to both Smurf1 (85 kDa) and Jab1 (38 kDa). The identity of these two bands were confirmed by staining with antibody specific to Smurf1 and Jab . These blots provide evidence that 10 LMP-1 interacts directly with Jab l supporting the hypothesis that LMP-1 may interrupt the binding of Jab 1 to one of its targets (Smad4). Example 25: Identification of binding interactions with Jab 1. While not identified herein, the sites or protein domains necessary for interaction between LMP- 1 and Jab 1 can be determined in the same manner as those for the 15 interaction between LMP-1 and Smurf1 (see Examples 13 and 14). Once identified, peptides containing those sites or domains may be designed modulate the interaction of the proteins. Example 26: LMP-1 overexpression increases nuclear Smad4 levels in human MSCs. To demonstreate the effects of LMP-1 overexpression and rhBMP-2 treatment on 20 Smad4, human MSC's were treated with rhBMP-2 (100 ng/mL), or LMP-1 (Ad5F35, MOI=5) for 8 hours. SDS-PAGE separated nuclear protein blots were probed with Smad4 specific antibody. A 66 kDa band representing nuclear Smad4 was seen to increase 8 hours after LMP-l treatment. A nonspecific band (running above the 66kDa band) was also seen. These data support the hypothesis that LMP-1, presumably via its interaction 25 with Jab 1, decreases the targeting of Smad4 for proteasomal degradation thereby resulting in increased Smad4 levels. As was expected, rhBMP-2 alone did not affect the nuclear Smad4 levels as it does not interact with Jabl. Collectively these data support the direct interaction of LMP-1 and Jab 1 in cells and the fact that LMP-1 can increase the levels of Smad4 in cells. This may represent a third regulatory point for LMP-1 to modulate 30 cellular responsiveness to BMPs since Smad4 is required for nuclear transport of activated R-Smads. Example 27: Screening of LMP-1 mimics WO 2006/102417 PCT/US2006/010419 -56 To identify small molecules that mimic the effect of LMP-1 on induction of bone formation, and have properties that make it more clinically convenient, multiple compounds may be screened which contain or mimic the more important LMP-1 interacting domain (B). These compounds are tested for their ability to compete with 5 LMP-1 for Smurf1 binding. In these studies, recombinant Smurfl is applied to wells of a 96 well plate. After removing excess Smurfl, test compounds (in excess) are pre-incubated with Smurfl followed by incubation of the Smurf1 complexes with recombinant biotin-LMP-1. Biotin LMP-1/Smurfl is detected using Strepavidin-alkaline phosphatase. Absorbance is 10 determined at 405 nm using the SpectraMax M2 microtiter plate reader. Compounds that block Biotin-LMP-1 binding may then be re-screened using appropriate lower amounts of the compound to determine the IC 50 (dose that prevents 50% of the maximum LMP-1 binding). Those compounds with the lowest IC 50 are considered the most efficacious and can then be screened for cellular effects. The IC 50 is used to determine the dose of each 15 compound that we will screen. Example 28: Screening compounds in vitro Candidate compounds that most efficiently inhibit binding of LMP-1 to Smurf1 in binding competition assays are further evaluated for their ability to mimic LMP-1 in cells. One appropriate cell line are mesenchymal stem cells (MSCs) and appropriate endpoints 20 include an increase BMPR1A levels and an increase in luciferase production from a Smad activated reporter construct. In these studies, human MSCs are plated at 3 x 104 cells/well in 6-well plates, grown overnight, and treated with candidate compounds at the IC 50 dose and two doses above and below that dose. Ad5F35LMP-1 and Ad5F35GFP are applied as positive and 25 negative controls. After 1, 2, 4, 8, 12, 24 and 48 hrs, cells are harvested and plasma membrane enriched fractions prepared. Fractions are subjected to SDS-PAGE and Western analysis using BMPR1A specific antibody. In the second set of experiments MSCs plated as above and transfected with a 9xGCCG/Smad-activated luciferase reporter construct are incubated with all compounds (at the successful dose) that increased 30 BMPR1A in the first experiment. These studies are performed in the presence or absence of 100 ng/mL BMP-2.

WO 2006/102417 PCT/US2006/010419 -57 The results will allow for identification of compounds and doses of compounds that will mimic the effect of LMP-1 on the particular endpoint. It is also expected that doses of candidate compounds that successfully increase BMPRIA will increase luciferase activity somewhat when applied alone, but will greatly enhance the effect of BMP-2 on 5 the luciferase activity. Example 29: Screening compounds in vivo Compounds found to activate the BMP/Smad signaling pathway in vitro can then be screened in vivo for effects on induction of bone in the rat model of ectopic bone formation described herein. 10 Example 30: Synergy screening Once compounds that induce bone formation have been identified, those compounds are then tested in combination with low doses of BMP-2 to determine whether there are synergistic effects (as demonstrated in other examples herein) that might lower the required dose of either agent. 15 To evaluate potential synergy, we have identified herein a dose of BMP-2 (2.5 ug) that induces bone formation in only 50% of the implants and a lower dose (1 ug) that fails to induce bone formation in the rat model. In the synergy studies, multiple doses of each of several candidate compounds are tested in cell culture studies of enhancement of BMP-2 signaling in MSCs. These studies 20 test the ability of each compound to enhance the efficacy of a suboptimal dose of BMP-2 (2.5 ug) to make bone in more than 50% of the implants. The positive control is MSCs treated with TAT-LMP-1 (0.625 nM); negative controls include MSCs alone and BMP-2 (2.5 ug) alone. MSCs (1-2M) are mixed with appropriate doses of compounds in a 100 uL total volume and placed on a collagen disc. The disc is implanted subcutaneously on the 25 chest of athymic rats and explanted after 4 weeks. Bone formation is evaluated by palpation, x-ray and semi-quantitative scoring of non-decalcified histologic sections (Edwards, J. T., Diegmann, M. H., and Scarborough, N. L., Clin. Orthop., 1998, 219-228). The most promising compounds are tested twice more to determine the best one or two compounds to be used in future experiments in higher animals. 30 If the LMP- 1 enhancement of BMP-2 efficacy to induce bone cannot be emulated by the chemical compounds that are screened, the need to also mimic the LMP-1/Jabl interaction can be examined as the control of Smurf4 levels could be a rate-limiting step.

Claims (37)

1. A method for inducing deposition and maturation of bone in a subject having compromised bone conditions comprising a co-therapeutic treatment regimen wherein at least one LMP agent and a subtherapeutically effective dose of at least one BMP agent are administered to a subject in need of such treatment.

2. The method of Claim 1, wherein said at least one LMP agent is administered before said at least one BMP agent; or said at least one LMP agent is administered after said at least one BMP agent; or said at least one LMP agent is administered simultaneously with said at least one BMP agent.

3. The method of Claim 2, wherein when administered simultaneously said at least one LMP agent and said at least one BMP agent is administered as an osteogenic composition.

4. The method of any one of the preceding claims, wherein said administration is to a targeted site.

5. The method of Claim 4, wherein the targeted site is the intervertebral space, a facet joint, site of a bone fracture, bones of the mouth, chin and jaw, and an implant site.

6. The method of any one of the preceding claims, wherein said administration is selected from the group consisting of intrathecal injection, subcutaneous, intravenous, intraperitoneal, intramuscular injection, in an implant or combinations of the above.

7. The method of any one of the preceding claims, wherein said therapeutically effective dose of BMP agent is less than 1.5mg/mL; or less than 0.15 mg/mL.

8. The method of any one of the preceding claims, wherein said at least one LMP agent is selected from the group consisting of a functional fragment of an LMP protein, a functional fragment of an LMP protein with a PTD attached, an LMP protein with a PTD attached, an LMP protein without a PTD attached, a functional fragment of an LMP protein variant, a functional fragment of an LMP protein variant with a PTD attached, an LMP protein variant 59 with a PTD attached, an LMP protein variant without a PTD attached , an oligonucleotide sequence encoding any of the above, and an LMP gene.

9. The method of Claim 8, wherein LMP is LMP1 and biologically active fragments thereof; or LMP1t and biologically active fragments thereof; or LMP-3 and biologically active fragments thereof.

10. The method of Claim 8, where the PTD is the HIV-TAT protein.

11. The method of any one of the preceding claims, wherein said at least one BMP agent is selected from the group consisting of a functional fragment of a BMP protein, a functional fragment of a BMP protein with a PTD attached, a BMP protein, a functional fragment of an BMP protein variant, a functional fragment of a BMP protein variant with a PTD attached, a BMP protein variant, an oligonucleotide sequence encoding any of the above, and a BMP gene.

12. The method of Claim 8, wherein the LMP protein fragment is peptide 7 comprising the following sequence: Acetyl-GAPPPADSAPQQNGQPLRPLVPDASKQRLMGGRRQRRTSKLMKR-NH2.

13. The method of Claim 12, wherein BMP is BMP-2; or rhBMP-2.

14. The method of any one of the preceding claims, wherein the conditions are selected from the group consisting of broken bones, bone defects, bone transplant, bone grafts, bone cancer, joint replacements, joint repair, fusion, facet repair, bone degeneration, dental implants and repair, bone marrow deficits and other conditions associated with bone and bony tissue.

15. The method of Claim 14, wherein the bone defect is a gap, deformation or a non-union fracture in a bone; or due to dwarfism

16. The method of Claim 14, wherein the bone degeneration is due to osteopenia or osteoporosis. 60

17. The method of Claim 14, wherein the joint replacement is selected from the group consisting of vertebral, knee, hip, tarsal, phalangeal, elbow, ankle or other articulating joint; or wherein the joint repair is selected from the group consisting of vertebral, knee, hip, tarsal, phalangeal, elbow, ankle, sacroiliac joint.

18. A method for decreasing the time required to form new bone in the presence of a BMP agent comprising co-administering at least one LMP agent.

19. The method of Claim 18, wherein a BMP agent is selected from the group consisting of endogenous BMP protein, exogenous BMP protein, exogenous BMP protein variants.

20. An osteogenic composition comprising at least one LMP agent; at least one BMP agent; and optionally, at least one pharmaceutically acceptable carrier.

21. An osteogenic composition comprising at least one LMP agent comprising at least one Smurf binding agent; and at least one BMP agent; and optionally, at least one pharmaceutically acceptable carrier.

22. The osteogenic composition of Claim 21, wherein the at least one pharmaceutically acceptable carrier is selected from the group consisting of cells, vectors, gels, microspheres, macromolecules, biocompatible foams, biocompatible matrices, and implants.

23. An osteogenic composition comprising at least one LMP agent comprising at least one phosphorylated Smad 1 competitive binding agent and at least one BMP agent.

24. An osteogenic composition comprising at least one LMP agent comprising at least one phosphorylated Smad 5 competitive binding agent and at least one BMP agent. 61

25. The osteogenic composition of Claim 21, wherein said at least one Smurf binding agent is selected from the group consisting of LMP-1 protein, and LMP-1 protein fragment, siRNA specific for Smurf mRNA.

26. The composition of Claim 21, wherein Smurf is selected from the group consisting of Smad Ubiquitin Regulatory Factor-1 (Smurf1) and Smurf 2.

27. The osteogenic composition of Claim 23, wherein said at least one phosphorylated Smad 1 competitive binding agent is selected from the group consisting of an LMP-1 protein, and LMP-1 protein fragment, siRNA specific for phosphorylated Smad 1.

28. The osteogenic composition of Claim 24 wherein said at least one phosphorylated Smad 5 competitive binding agent is selected from the group consisting of an LMP-1 protein, and LMP-1 protein fragment, siRNA specific for phosphorylated Smad 5.

29. A process for engineering bone tissue comprising combining at least one LMP agent and at least one BMP agent with a cell selected from the group consisting of osteogenic cells, pluripotent stem cells, mesenchymal cells, and embryonic stem cells.

30. The engineered bone tissue produced by the process of Claim 28.

31. A method for inducing bone formation in a subject comprising administering the engineered bone tissue of Claim 38.

32. An osteogenic composition comprising at least one LMP agent comprising at least one phosphorylated Smad 4 competitive binding agent and at least one BMP agent.

33. The osteogenic composition of Claim 32, wherein said at least one phosphorylated Smad 4 competitive binding agent is selected from the group consisting of an LMP-1 protein, and LMP-1 protein fragment, siRNA specific for phosphorylated Smad 4. 62

34. A process for engineering bone tissue comprising combining at least one LMP agent comprising at least one phosphorylated Smad 4 competitive binding agent and at least one BMP agent with a cell selected from the group consisting of osteogenic cells, pluripotent stem cells, mesenchymal cells, and embryonic stem cells.

35. The engineered bone tissue produced by the process of Claim 34.

36. A method for inducing bone formation in a subject comprising administering the engineered bone tissue of Claim 35.

37. A method for inducing deposition and maturation of bone in a subject having compromised bone conditions comprising administering to the subject at least one LMP agent comprising at least one JAB1 inhibiting agent and a therapeutically effective dose of at least one BMP agent.