WO2007050057A1

WO2007050057A1 - Identification, expression, functions and uses of mda-bf factors

Info

Publication number: WO2007050057A1
Application number: PCT/US2005/038427
Authority: WO
Inventors: Sue-Hwa Lin; Shi-Ming Tu; Christopher Logothetis; Weiping Luo
Original assignee: University of Texas System; University of Texas at Austin
Current assignee: University of Texas System; University of Texas at Austin
Priority date: 2005-10-24
Filing date: 2005-10-24
Publication date: 2007-05-03
Anticipated expiration: 2008-04-24

Abstract

The disclosure includes paracrine factors such as MDA-BF-1 or MDA-BF-2; an osteoblast stimulating factor having a 45 kDa N-terminal through intron 8 portion of the p180-ErbB3 gene; an antibody raised against purified MDA-BF-1 or a fragment or derivative thereof which binds to MDA-BF-1; and a peptide which binds to MDA-BF-1. The disclosure also includes a method of diagnosing metastatic prostate cancer by testing for the presence of MDA-BF-1. It contains many methods for altering bone growth or other bone-related physiological properties in a subject or treatment of various bone-related diseases in a patient. These methods may include supplying MDA-BF-1 to the patient or removing or interfering with the action of MDA-BF-1.

Description

IDENTIFICATION, EXPRESSION, FUNCTIONS AND USES OF MDA-BF

FACTORS

CROSS REFERENCE TO RELATED APPLICATION

This application claims priority from U.S. Provisional Patent Application Serial No. 60/620,820, filed October 21, 2004, by Sue-Hwa Lin et al . , entitled "Identification, Expression, Functions and Uses of MDA-BF Factors" which is hereby incorporated in its entirety by reference .

GOVERNMENT RIGHTS IN THIS INVENTION

This invention was made in part with U.S. government support from the National Institutes of Health under contract number CA 64856 and CA 86342. The U.S. government has certain rights in this invention.

FIELD OF THE INVENTION

The present invention relates to bone growth factors. It particularly relates to bone growth factors expressed during metastatic cancer, such as osteoblast stimulation factors produced by prostate cancer cells.

BACKGROUND

Prostate cancer is a cancer disease that primarily affects older males. Osseous (bone) pain, neurological complications resulting from spinal cord compression or nerve root impingement, and bone marrow failure due to bone marrow replacement are common manifestations of prostate cancer. Clinically, most androgen-dependent prostate cancer progresses from the site of the prostate gland to the pelvic lymph node, and further to bone. In addition, androgen independent prostate cancer often progresses into severe bone metastases. Thus, bone is the major metastatic site for prostate cancer, and prostate cancer bone metastasis results in many of the cancer- associated symptoms of advanced prostate cancer. There is also a direct correlation between the extent of osseous involvement and patient survival. Patients with advanced prostate cancer and bone metastases have a dismal prognosis; their median survival is less than 6 months. Therefore, bone metastasis is a significant clinical problem and a challenge for prostate cancer treatment.

Currently, there is no highly sensitive or specific test for the diagnosis of bone metastasis in prostate cancer or other cancers . Prostate cancer is invariably fatal once bone metastases occur. In fact, those patients who are more likely to harbor bone metastases are the very patients who have clinically relevant diseases and who succumb to prostate cancer. Generally, prostate cancer patients may benefit from aggressive therapy, such as a radial prostatectomy. After prostatectomy, the median time from the time of prostate serum antigen (PSA) elevation to the development of prostate cancer bone metastases is about 8 years and the median survival period for patients whose bone metastases have become manifested is about 5 years. Hence, there is ample time for the diagnosis of prostate cancer bone metastases and for the use of bone-targeted therapies. Qells in bone, such as osteoblasts and osteoclasts, may be involved in prostate cancer bone metastasis. Prostate cancer bone metastases typically have a blastic phenotype (i.e., excessive bone formation by osteoblasts with increased bone mass at the site of the lesion that can be viewed by a full body bone scan) , rather than a lytic phenotype, distinguishing them from bone lesions in other types of cancers. Interactions between prostate cancer cells and cells of the osteoblast lineage have been shown to contribute to the lethal progression of prostate cancer in the bone. Also, sites of prostate cancer bone metastases often have increased osteoid surfaces, osteoid volumes and mineral apposition rates, indicating increased bone formation. The unique ability of prostate cancer cells to influence osteoblast proliferation suggests that specific biological interactions occur between the prostate cancer cells and the bone environment . Laboratory studies have confirmed that interactions between bone stroma cells and prostate epithelial cells are essential for the development of bone metastases. In particular, paracrine loops involving specific factors (e.g., IGF-I, TGF-β) may promote the proliferation and survival of osteoblasts and prostate cancer cells in the skeleton. Bone-targeted therapies that disrupt such bone- epithelium interactions should improve the clinical outcome of patients by blocking the underlying mechanism of prostate cancer bone metastasis. Understanding the molecular basis for the bone-epithelium interactions will create new opportunities for therapeutic intervention.

Studies on bone metastases have been hampered by the limited availability and small size of patient samples (e.gr., bone biopsy samples) . In addition, many prostate cancer cell lines are not suitable for studying bone metastases since most of the prostate cancer cell lines have undergone changes when adapting cell growth during cell culturing and cannot reproduce bone-epithelial interactions in vitro and may not represent the natural properties of prostate cancer in bone.

On the other hand, bone marrow is an environment which acts like a natural conditioned medium for metastatic prostate cancer cells. In addition, the bone marrow cavity is the site of metastases and thus should have higher concentrations of protein factors involved in bone-epithelium interactions than those located in blood. In fact, protein factors produced in bone marrow as a result of interactions between metastatic prostate cancer cells and bone marrow stromal cells very likely mediate the growth of prostate cancer cells and the abnormal proliferation of osteoblasts in the bone. Identification of these bone metastases factors may provide clues about how prostate cancer progresses in bone.

SUMMARY OF THE INVENTION

Thus, 'there remains a need to identify proteins expressed in metastatic prostate cancer cells and contributing to the clinical behavior and bone tropism of prostate cancer and to further understand the roles of these proteins. In addition, there is a need for providing biological markers that can be used as a basis for diagnosing, prognosticating, and treating bone related diseases, such as bone metastasis in prostate ^■ cancer and other cancers, and osteoporosis.

Aspects of the invention include a paracrine factor such as,_,MDA-BF-I or MDA-BF-2 or an osteoblast stimulating factor, having a 45 kDa N-terminal through intron 8 portion of the pl80--ErbB3 gene. Another embodiment relates to an antibody raised against purified MDA-BF-I or a fragment or derivative thereof which binds to MDA-BF-I.

Another embodiment relates to a peptide which binds to MDA-BF-I. . ^{' ■}' Another aspect of the invention includes a method of diagnosing metastatic prostate cancer disease in a. subject by obtaining a sample from the subject and₍ testing- the sample for the presence of MDA-BF-I protein. Another embodiment relates to a method of stimulating bone growth by introducing MDA-BF-I into a subject .

Still another embodiment includes a method of treating a bone-related disease by stimulating bone growth by introducing MDA-BF-I into a subject.

Another exemplary method relates to a method of treating a disease in a subject by reducing the expression of MDA-BF-I in at least one cell in the subject.

Still another embodiment includes a method of treating a disease in a subject by introducing a modulator for interaction between MDA-BF-I and MDA-BF-I receptor into the subject. Another embodiment of the invention relates to a method of treating a disease in a subject by introducing a peptide corresponding to a sequences from MDA-BF-I into the subject.

Still another aspect includes a method of treating a disease in a subject, by introducing into the subject an

RNA such as dsRNA, siRNA or antisense RNA corresponding to a nucleotide sequence of MDA-BF-I, or any combination thereof .

Another embodiment relates to a method of treating a disease by introducing into a subject an antibody raised against purified MDA-BF-I or a fragment or derivative thereof operable to bind to MDA-BF-I.

Another exemplary methods provides for modulating bone growth in a subject by introducing a modulator for interaction between MDA-BF-I and MDA-BF-I receptor into a subject .

Yet another embodiment includes a method of reducing the expression of MDA-BF-I protein in a cell, comprising introducing into the cell an RNA such as dsRNA, siRNA or antisense RNA corresponding to a nucleotide sequence of MDA-BF-I, or any combination thereof.

Another embodiment includes a siRNA oligonucleotide complementary to a region of an open reading frame of MDA-BF-I DNA.

Finally, another embodiment relates to a method of reducing the production of MDA-BF-I protein in cells, comprising delivering an antisense oligonucleotide complementary to a region of an open reading frame of MDA-BF-I DNA to the cells.

The following abbreviations or designations are commonly used throughout the specification. A short description of these abbreviations is provided below for convenience, but should not be used to limit the scope of the relevant terms. The scope of all terms described below should be consistent with their usage throughout the specification and not merely in this section:

ErbB3 or pl80-ErbB3: a 180 kDa growth factor located in the membrane of some prostate cells. DU145: a brain-derived prostate cancer cell line with osteolytic (bone destroying) features that produces little or no MDA-BF-I.

LNCaP: a lymph-node derived prostate cancer cell line with some osteoblastic (bone growth) features that produces some, but not large amounts of MDA-BF-I.

MDA-BF-I: an osteoblast stimulating (bone growth) factor produced by some metastatic prostate cancer cells that is similar to pl80-ErbB3, but has a molecular weight of only around 45 KDa. MDA PCa 2b: a bone-derived prostate cancer cell line with osteoblastic (bone growth) features that produces a high amount of MDA-BF-I. PC-3 : a bone-derived prostate cancer cell line with osteolytic (bone destroying) features that produces little to no MDA-BF-I.

PMO: primary mouse osteoblast. PSA: prostate-specific antigen. BRIEF DESCRIPTION OF THE DRAWINGS

The present invention may be better understood by reference to the following detailed description, taken in conjunction with the drawings It is to be noted, however, that the appended drawings illustrate only typical embodiments of this invention and are therefore not to be considered limiting of its scope, for the invention may admit to other equally effective embodiments. FIGURE 1 demonstrates SDS-PAGE analysis of total proteins from bone marrow supernatant .

FIGURE 2 is a flow chart of one embodiment of protein purification steps for identification of bone metastasis factors used herein. FIGURE 3 depicts a protein elution profile of WGA- agarose chromatography.

FIGURE 4 demonstrates SDS-PAGE analysis of proteins eluted from WGA-agarose chromatography.

FIGURE 5 demonstrates 2-D gel electrophoresis analysis of proteins eluted from WGA-agarose chromatography using silver stain to detect proteins of interest .

FIGURE 6 depicts a protein elution profile of ConA- agarose chromatography. FIGURE 7 demonstrates SDS-PAGE analysis of proteins eluted from ConA-agarose chromatography.

FIGURE 8 depicts a Western immunoblot of individual patients' bone marrow supernatants probed with polyclonal anti-haptoglobin antibody (DAKO) at 1:2000 dilution and developed by enhanced chemiluminescence .

FIGURE 9 depicts immunohistochemical staining of haptoglobin in metastatic bone lesions of prostate cancer. FIGURE 10 depicts immunohistochemical staining of

MDA-BF-I expression in bone metastasis of prostate cancer probed with an antibody against the extracellular domain of ErbB3.

FIGURE 11 depicts another immunohistochemical staining of MDA-BF-I expression in bone metastasis of prostate cancer probed with an antibody against the extracellular domain of ErbB3. ■ ^■

FIGURE 12 depicts yet another immunohistochemical staining of MDA-BF-I expression in bone metastasis of prostate cancer probed with antibody against the extracellular domain of ErbB3.

FIGURE 13 depicts immunohistochemical staining of MDA-BF-2 expression in bone metastasis of prostate cancer probed with an antibody against the p60 subunit of katanin.

FIGURE 14 compares the mRNA structure and protein structure of MDA-BF-I and pl80-ErbB3.

FIGURE 15 depicts pl80-ErbB3 domain arrangement.

FIGURE 16 depicts the pl80-ErbB3 signal transduction pathway and the proposed MDA-BF-I signal transduction pathway. FIGURE 17 demonstrates a Western blot of MDA-BF-I in enriched bone marrow supernatant samples from individual patients .

FIGURE 18 demonstrates a Western blot comparing the levels of MDA-BF-I in the bone marrow supernatant and the corresponding plasma of patients in the Met group.

FIGURE 19 illustrates antibody recognition sites of four antibodies.

FIGURE 20 demonstrates immunohistochemical staining of MDA-BF-I expression.

FIGURE 2IA demonstrates staining patterns of normal prostate glands for MDA-BF-I expression.

FIGURE 2IB demonstrates staining patterns of normal prostate glands for pl80-ErbB3 expression. FIGURE 21C demonstrates staining patterns of high- grade prostatic intraepithelial neoplasia for MDA-BF-I expression.

FIGURE 21D demonstrates staining, patterns of high- grade prostatic intraepithelial neoplasia for pl80-ErbB3 expression.

FIGURE 21E demonstrates staining patterns of primary prostate cancer for MDA-BF-I expression.

FIGURE 2IF demonstrates staining patterns of primary prostate cancer for pl80-ErbB3 expression. FIGURE 22A demonstrates staining patterns of metastatic prostate cancer in the lymph node for MDA-BF-I expression.

FIGURE 22B demonstrates staining patterns of metastatic prostate cancer in the lymph node for pl80- ErbB3 expression. FIGURE 22C demonstrates staining patterns of metastatic prostate cancer in the lymph node for MDA-BF-I expression.

FIGURE 22D demonstrates staining patterns of metastatic prostate cancer in the lymph node for pl80- ErbB3 expression.

FIGURE 23A demonstrates staining patterns of metastatic prostate cancer in the bone for MDA-BF-I expression. FIGURE 23B demonstrates staining patterns of metastatic prostate cancer in the bone for pl80-ErbB3 expression.

FIGURE 23C demonstrates staining patterns of metastatic prostate cancer in the bone for MDA-BF-I expression.

FIGURE 23D demonstrates staining patterns of metastatic prostate cancer in the bone for pl80-ErbB3 expression.

FIGURE 23E demonstrates staining patterns of metastatic prostate cancer in the bone for MDA-BF-I expression.

FIGURE 23F demonstrates staining patterns of metastatic prostate cancer in the bone for pl80-ErbB3 expression. FIGURE 24 demonstrates a Western blot demonstrating the secretion of MDA-BF-I into conditioned medium of some prostate cancer cell lines.

FIGURE 25 demonstrates a Western blot illustrating the expression and secretion of a recombinant MDA-BF-I protein into conditioned medium of PC-3 cells. FIGURE 26 demonstrates the in vitro effect. of ■ recombinant MDA-BF-I on PC-3 cell growth.

FIGURE 27 demonstrates the in vivo effect of recombinant MDA-BF-I on PC-3 cell growth. FIGURE 28 depicts SDS-PAGE analysis of the purified recombinant MDA-BF-I protein.

FIGURE 29 demonstrates stimulation of cell proliferation of primary mouse osteoblasts by purified MDA-BF-I. F₁IGURE 30 demonstrates that there is no stimulation of ce^Ll proliferation of mouse fibroblasts by purified MDA-BF-I.

FIGURE 31 demonstrates the proposed role of MDA-BF-I in osteoblastic progression of prostate cancer cells in bone.

FIGURE 32 demonstrates the effect of MDA-BF-I on prostate cancer cell-osteoblast interaction in an ijp vivo animal model .

FIGURE 33 demonstrates a histological examination of the effect of MDA-BF-I on prostate cancer cell-osteoblast interaction in vivo.

FIGURE 34 shows the activation of p42/p44 MAPK in primary mouse osteoblasts by MDA-BF-I.

FIGURE 35 shows the activation of AKT , phosphorylation in primary mouse osteoblasts by MDA-BF-I.

FIGURE 36 shows the stimulation of IKB-Q degradation, in primary mouse osteoblasts by MDA-BF-I.

FIGURE 37 demonstrates RT-PCR analysis of the expression of Runx2 in primary mouse osteoblasts. FIGURE 38 depicts the expression of green fluorescence protein by pEGFP when transfected into primary mouse osteoblasts .

FIGURE 39 summarizes the proposed model of activation of signaling pathways by MDA-BF-I.

FIGURE 40 shows that MDA-BF-I does not stimulate tyrosine phosphorylation of ErbB2 in LNCaP cells.

FIGURE 41 shows cross-linking of 1125 -MDA-BF-I to the BF-I receptor by disuccinimidyl suberate. FIGURE 42 summarizes one embodiment of the purification scheme for the BF-I receptor.

FIGURE 43 is a graph showing one embodiment of screening a phage display library with an increase in specificity for binding MDA-BF-I to selected phages after four runs of screening.

FIGURE 44 is a graph demonstrating selective binding of 7 phages screened to MDA-BF-I as compared to binding to BSA using a phage without insert as a control .

FIGURE 45 shows the results of RT-PCR of MDA-BF-I as compared to pl80-ErbB3 and actin in four prostate cancer cell lines.

FIGURE 46 shows the levels of MDA-BF-I expression in individual PC-3 clones transfected with pRSN-FLAG-MDA-BFl . The conditioned media from these clones were examined for the levels of MDA-BF-I by immunoprecipitation with anti- FLAG-agarose and Western blotted with Ab-10, which recognizes the 45 kDa MDA-BF-I protein. Mixed conditioned medium from unselected transfectants (mix) was used as a control. . - FIGURE 47 shows that expression of MDA-BF-I in PCr3 cells and new bone formation. Arrows indicate newly formed woven bone. T indicates tumor cells. Magnification 4Ox.

FIGURE 48 shows ³H thymadine uptake and hence growth of primary mouse osteoblasts (PMOs) in response to MDA- BF-I.

FIGURE 49 shows cell number of PMOs in response to MDA-BF-I.

- ^'FIGURE 50 shows the effects of infection with Ad- MDA-BF-I as compared to Ad-Luc in PMOs. Multiplicity of infection is indicated.

FIGURE 51 shows the effects of MDA-BF-I on PMO differentiation through von Kossa stain of PMOs in differentiation with or without MDA-BF-i at days 0, 4 and 6. FIGURE 52 shows the effects of MDA-BF-I on PMO differentiation by quantitization of von Kossa staining using an NIH image program at 0, 4 and 6 days.

FIGURE 53 shows histological analysis of lesions induced by intra-femoral injection of PC/MDA-BF-1 cells. Bone tumor is indicated by arrow. Magnification 4x.

FIGURE 54 shows histological analysis of lesions induced by PC/MDA-BF-1 cells. T indicates tumors. Magnification 4Ox.

FIGURE 55 shows activated osteoblasts in woven bpne induced by PC/MDA-BF-1. Arrows indicate osteoblasts.

FIGURE 56 shows immunohistochemical staining of MDA- BF-I in tumors produced by PC/neo and PD/MDA-BF-1.

FIGURE 57 shows the locations of three siRNA sequences in MDA-BF-I. FIGURE 58 shows expression of MDA-BF-I and 0180- ErbB3 in MDA PCa 2b cells transfected with siRNA. The conditioned media from these transfectants was immunoprecipitated with Ab-8 and Western blotted with Ab- 10. MDA PCa 2b cells, which constitutively express MDA- BF-I, were used as the control. To examine the levels of pl80-ErbB3, the transfectants were solubilized with Triton X-100. The solubilized cell supernatants were immunoprecipitated with monoclonal antibody 2F12 against the cytoplasmic domain of pl80-ErbB3, and Western blotted with C-17, a polyclonal antibody against pl80-ErbB3.

DETAILED DESCRIPTION

: 'Aspects of the invention provide methods of identifying bone metastasis factors and the expression, functions and uses of these bone metastasis factors .

In one embodiment, bone metastasis factors purified from bone marrow supernatant are provided. In another embodiment, the study of the functions of a bone metastasis factor is provided. Still another embodiment provides detection of a protein having osteoblast stimulating activity in a cancer cell. In yet still ^' another.embodiment, a metastatic factor which exhibits osteoblast stimulating activity is purified. In. yet another embodiment, a protein which is purified from bone marrow supernatants of patients with a metastatic cancer disease and absent in patients with a primary cancer disease is provided. Further, a recombinant bone metastasis factor may be purified from the conditioned media of a prostate cancer cell line. Using a variety of approaches including proteomics, biochemical, immunohistochemical and molecular methods, the invention provides the identification of five novel bone. metastasis factors (MDA-BF-I to 5) from bone marrow supernatant of prostate cancer patients with bone metastasis. Of particular interest is MDA-BF-I, which appears to be a portion of pl80-ErbB3. The invention demonstrates for the first time that MDA-BF-I is a paracrine factor with osteoblast-stimulating properties in vivo. Thus, MDA-BF-I is mechanistically implicated in prostate cancer progression.

Specifically, to understand the mechanisms of prostate cancer bone metastasis, a method of the invention to identify prostate cancer "Metastasis

Proteome" proteins that are involved in bone metastasis of prostate cancer was developed. To ensure that the findings were clinically relevant, disease-relevant samples were used, e.g. , bone marrow superna,tants from prostate cancer patients with bone metastasis and those patient.s without bone metastasis. Using. these methodologies, MDA-BF-I to 5 were identified.

Although bone-epithelium interactions are thought to be involved in the pathogenesis of prostate cancer bone metastases, the factors involved in these interactions were not previously known. Using methods of the present invention to identify bone metastasis-related factors, these factors were purified from patients.' bone marrow samples. Thus, factors identified with this .approach are highly clinically relevant. A protein, purification strategy to enrich low-abundance proteins in bone marrow supernatant was developed. Five differentially expressed proteins, termed MDA-bone metastasis factors MDA-BF-I, MDA-BF-2, MDA-BF-3, MDA-BF-4, and MDA-BF-5, have been identified so far. The first factor identified, MDA-BF-I, was characterized to be a secreted form of a growth factor receptor, ErbB3 , with an apparent molecular weight of 45 kDa. Thus, in one aspect, the invention provides identification and purification of MDA-BF-I, and studies on the expression and function of MDA-BF-I, a soluble form of ErbB3 protein. Unlike ErbB3, which is a 180-kDa transmembrane growth factor receptor, MDA-BF-I has only the first half of the extracellular domain of ErbB3 and has an apparent molecular mass around 45 kDa. The existence of MDA-BF-I challenges the notion that ErbB3 is merely a transmembrane receptor protein and suggests that the ErbB3 family of proteins has other functions and additional roles that soluble isoforms of the ErbB3 family of proteins may play.

Additionally, unlike ErbB3 , MDA-BF-I stimulates osteoblast proliferation in vitro and bone formation in vivo, suggesting that MDA-BF-I is involved in osteoblastic progression of prostate cancer in bone. Thus, MDA-BF-I is the first bone-epithelium interacting protein isolated from human bone marrow.

As described herein, several lines of evidence suggest that MDA-BF-I is a paracrine factor secreted by metastatic prostate cancer cells that mediates osteoblast proliferation in prostate cancer bone metastasis. First, Western blots showed that MDA-BF-I is present only in the bone marrow supernatant of prostate cancer patients with bone metastasis but not those without bone metastasis.

Second, immunohistochemical analysis showed that MDA-BF-I is not expressed in normal prostate epithelial cells and is only produced by the metastatic prostate cancer cells in the bone and lymph node. Third, a bone-derived prostate cancer cell line with osteoblastic features (MDA

PCa 2b) produces a high amount of MDA-BF-I, while a bone- derived prostate cancer cell line with osteolytic features (PC-3) does not. Fourth, recombinant MDA-BF-I induces osteoblast proliferation but not prostate cancer cell proliferation in vitro. Fifth, preliminary studies show that expression of MDA-BF-I in the osteolytic prostate cancer cell line PC-3 generated an osteoblastic response in vivo when these cells are injected into bone. In comparison, MDA-BF-I does not affect PC-3 cell growth in vivo when injected subcutaneously .

These observations demonstrate that MDA-BF-I has osteoblast stimulating activity. Accordingly, it is likely that, unlike ErbB3 which is a receptor anchored to or embedded in cell membranes, MDA-BF₇I₁Is a_. soluble factor secreted by metastatic prostate , caηce.r cel.ls to stimulate osteoblast proliferation and which regulates the interactions between osteoblasts and prostate cancer cells, in bone.

Various molecular constructs, reagents, monoclonal and polyclonal antibodies, microscopy techniques, in vitro assays, RNAs, peptides, in vivo bone growth models, and transgenic animals are also provided to address how MDA-BF-I functions as an osteoblast stimulating factor during prostate cancer progression.

For example^', one embodiment of -the invention ■ ^■ provides a receptor which binds to MDA-BF-I. Another embodiment of the invention provides antibodies which are raised against purified MDA-BF-I, including polyclonal antibodies and monoclonal antibodies, among others. In another embodiment, the invention provides peptides which bind to MDA-BF-I, such as those screened from a phage display library, e.g., ASGADGP (SEQ. ID. NO.1) , FGWPLW (SEQ. ID.NO.2) , GGLALQE (SEQ. ID .NO .3) , LKRGITV , ., (SEQ. ID.NO.4) , FASSFVL (SEQ . ID .NO .5) , TLDFPRR (SEQ. ID.NO.6) , ISFPRRW (SEQ. ID.NO.7) , WAGGRF (SEQ. ID.NO.8) , VAGGSFI (SEQ . ID .NO .9) , QGGVRHH (SEQ. ID.NO.10) , GGVRVLD (SEQ.ID.NO.il), FASRVRS

(SEQ . ID .NO .12 ) , QSRVRVA (SEQ . ID .NO .13 ) , PAGRYTD

(SEQ. ID.NO.14) , GRYTTDR (SEQ . ID .NO .15) , SGYVAKM

(SEQ. ID.NO.16) , and SGYAKVS (SEQ. ID.NO.17) . Another embodiment of the invention provides a transgenic mouse or other animal having MDA-BF-I transgene expressed from various promoters.

The invention also provides an in vivo animal model for developing prostate cancer and treating of the developed prostate cancer by the identified bone metastasis factors, including MDA-BF-I.

In one aspect of the invention, a method of identifying a metastasis factor is provided. The method includes purifying a first group of proteins from a sample of a cancer patient with a metastatic cancer disease and purifying a second group of proteins from another sample of another cancer patient with a primary cancer;,disease. Then, protein profiles of the first and the second group of proteins are compared and the metastasis factor present in the first group of proteins and absent in the second group of proteins is identified.

In another aspect, a method of purifying a bone metastasis factor from a tissue sample is provided. The method includes purifying a first group pf proteins from a bone marrow sample of a cancer patient with a metastatic cancer, disease, purifying a second group of proteins from another bone marrow sample of ■ another , . . cancer patient with a primary cancer disease, and comparing protein profiles of the first and the second group of proteins in order to purify the metastasis factor, present in the first group of proteins and absent in the second group of proteins, from the first group of protein. In a further embodiment, MDA-BF-I protein is detected and/or measured in human samples in order to diagnose or prognosticate the development of human disease, such as prostate cancers and osteoporosis. For example, embodiments of the invention provide diagnostic kits, such as a blood test or a bone marrow test, for diagnosing prostate cancer bone metastases. The diagnostic kit can test a sample from bone marrow, a cancer tissue biopsy, blood, plasma, body fluid, and urine, among others.

For example, aspects of the invention provide a method of diagnosing a metastasis-related cancer, disease or .a, bone-related disease. In another aspect, the invention provides a method of prognosticating a survival rate at a specified period of time of a patient having prostate cancer. Both methods include obtaining a sample from a cancer patient and testing the sample for the presence of MDA-BF-I protein. Testing may include using an antibody raised against MDA-BF-I. In a more specific embodiment, it may include the use of an ELISA.

Yet another aspect of the invention provides a method of diagnosing the degree of progression of a prostate cancer. The method includes obtaining a sample from a prostate cancer patient, testing the sample for the presence of MDA-BF-I protein, and correlating the level of MDA-BF-I in the sample with a predetermined level of MDA-BF-I for each of different progression stages of the prostate cancer. ₍

Diagnostic kits or tests of the invention are preferably sensitive enough (often more sensitive than magnetic resonance imaging (MRI) or bone scans) , to be used very early on during prostate cancer bone metastases, s_'9_: i years before prostate cancer bone metastases become evident on MRI or bone scans. The diagnostic tests may include a method sensitive enough to determine low levels of MDA-BF-I in the sample. For example, using a monoclonal antibody in an ELISA (enzyme-linked immunosorbent assay) detection kit. As a result, levels of MDA-BF-I can be correlated with different stages of prostate cancer as the disease progresses to bone metastasis. Prostate cancer patients may thus benefit from early therapeutic intervention, such as hormonal ablative therapy, which has been shown to prolong survival. In addition, these patients may be tested for the expression of MDA-BF-I in prognosticating the reoccurrence of prostate cancer and/or in any metastatic stage . Further, the invention provides receptors for the identified bone metastasis factors, including MDA-BF-I, and_. the_, specific modulators (including any effectors, inhibitors, stimuli, monoclonal antibodies, polyclonal antibodies, peptides, siRNAs, agonists, and/or antagonists, among others) for the interactions between bone metastasis factors and their cognate receptors. Such specific modulators may be used to improve the efficacy of bone-targeted therapy for treating prostate cancer bone metastases, osteoporosis, and other types of cancers, especially those cancers involving bone metastases.

In another embodiment of the invention, therapies? targeting MDA-BF-I to alter the course of prostate cancer progression are provided, for example, using anti-MDA-BF- 1 monoclonal and polyclonal antibodies raised against the purified MDA-BF-I. As another example, peptides identified from phage display libraries may also be used. These antibodies and peptides are also tested for inhibiting the interaction between MDA-BF-I and bone- related cells. In yet another embodiment, siRNAs that down-regulate MDA-BF-I expression can also be used as for therapeutic intervening of a disease.

The invention further provides method of treating a disease, including introducing an antibody raised against purified MDA-BF-I, such as a monoclonal antibody and/or a polyclonal antibody into a subject.

In addition, the invention provides a method of treating a disease and a method of modulating bone growth, including introducing a modulator for interaction between MDA-BF-I and MDA-BP-I receptor into a subject. The modulator may be a monoclonal antibody, a polyclonal antibody, an agonist, an antagonist, an inhibitor, an inducer, siRNA corresponding to the nucleotides sequences of MDA-BF-I, a peptide corresponding to the amino acid sequences of MDA-BF-I, a peptide corresponding to the amino acid sequences of MDA-BF-I receptor, and combinations thereof.

Further, a method of treating a disease includes introducing a peptide corresponding to the sequences of MDA-BF-I into a subject. For example, the peptide may be ASGADGP (SEQ. ID.NO.1) , FGWPLW (SEQ. ID .NO .2) , GGLALQE (SEQ. ID. NO.3) , LKRGITV (SEQ . ID.NO .4) , FASSFVL (SEQ. ID. NO.5) , TLDFPRR (SEQ. ID.NO.6) , ISFPRRW (SEQ. ID. NO.7) , WAGGRF (SEQ. ID.NO .8) , VAGGSFI (SEQ. ID.NO.9) , QGGVRHH (SEQ. ID .NO .10) , GGVRVLD (SEQ.ID.NO.il), FASRVRS (SEQ . ID.NO.12) , QSRVRVA (SEQ. ID.NO.13) , PAGRYTD (SEQ. ID.NO.14) , GRYTTDR (SEQ. ID.NO.15) , SGYVAKM (SEQ . ID .NO .16) , SGYAKVS (SEQ. ID.NO.17) , and combinations thereof.

Also, a method of reducing the expression of MDA-BF- 1 protein includes introducing an antibody raised against purified MDA-BF-I into a subject. Still further, methods of treating a disease, such as cancer, and reducing expression or production of MDA- BF-I protein may include introducing RNA to inhibit expression of MDA-BF-I into a subject or cell. For 5 example, a dsRNA corresponding to nucleotide sequences of MDA-BF-I may be introduced into the subject or cell. The^' dsRNA sequence may be aacgacgctctgcaggtgctgdTdT (SEQ. ID.NO.18) , aactctcaggcagtgtgtcctdTdT (SEQ. ID.NO.19) , and derivatives and homologs thereof. The invention 10 further provides a siRNA oligonucleotide complementary to a region of an open reading frame of MDA-BF-I DNA. The invention also provided methods of delivering an antisense oligonucleotide complementary to a region of an open reading frame of MDA-BF-I DNA to a subject or cell.

IEi . .. In a .further aspect of the invention, a method of stimulating bone growth including introducing MDA-BF-I into a subject is provided. The invention also provides methods of treating a bone-related disease, such as osteoporosis, bone metastasis, osteolytic disease, and

20 prostate cancer bone metastasis. The methods, include introducing MDA-BF-I into a subject. . . .

In another aspect of the invention,- ^" a method of inducing the expression of MDA-BF-I is provided.- The ^• method' includes introducing an expression construct of 25 MDA-BF-I haying . an expression region encoding MDA-BF-I protein into a subject.

EXAMPLES

The following examples are included to demonstrate 30 specific embodiments of the invention. It should be appreciated by those of skill in the art that the techniques disclosed in the examples that follow represent techniques discovered by the inventors to function well in the practice of the invention. However, those of skill in the art should, in light of the present disclosure, appreciate that many changes can be made in the specific embodiments that are disclosed and still obtain a like or similar result without departing from the spirit and scope of the invention.

I. Identification of MDA-BF-I in the bone marrow supematants of patients with metastatic prostate cancer 1. Collection of Bone Marrow Specimens

Specimen collection was performed according to a laboratory protocol approved by the institutional review board of the University of Texas, M. D. Anderson Cancer Center. Bone marrow collection and processing were performed according to the Standard Operational

Procedures of the Department of Genitourinary Medical Oncology. Bone marrow samples from each patient were processed and stored individually. About 10 ml of marrow specimen was drawn from each patient into a heparinized syringe in the Bone Marrow Clinic. All samples were processed the day they were drawn. Each sample was transferred to a 15-ml Falcon Blue Max tube. The samples were centrifuged at 4°C for 20 min at 2500 rpm. Equal amounts of the bone marrow supernatant and cell pellet each were transferred into three 2 -ml SARSTEDT cryotubes (Sarstedt Inc., Newton, NC) . The tubes were stored in a - 85°C Forma Freezer until the samples were used.

All patients had been diagnosed with adenocarcinoma of tfce prostate and had received hormonal ablation therapy. Osseous metastases were evaluated by bone scintigraphy using Technetium-99m labeled methylene bisphosphonate. Bone marrow aspirates from six patients (median age 60) who had no evidence of bone metastasis, as determined by both bone scintigraphy and histological- examination of bone biopsy samples, were used as the control group (Control samples) . These samples were obtained between 8/25/97 and 10/20/98. Bone marrow aspirates from six patients (median age 68) who had extensive bone metastases were used as the metastatic group (Met samples) . These samples were obtained between 2/6/96 and 12/10/96. The Gleason scores in the initial biopsies of patients in the control group ranged from 7-9 (median 8) and in the Met group ranged from 7-8 (median 7) . At the time that the bone marrow samples were drawn, the patients in control group and Met group were in hormonal therapy for 9-82 months (median 37 months) and 18-74. months (median 50 months) , respectively. The serum prostate-specific antigen (PSA) levels ranged from <0.1 to 21.6 ng/ml (median 1.5 ng/ml; average 6.8 ± 8.4) in the control samples and 47.9 to 1691.2 ng/ml (median 211.6; average 498.6 ± 610.8) _fi,n the Met samples. The serum bone-specific alkaline phosphatase levels ranged from 54 to 123 IU/L (median 76.IU/L; average 80.7 ± 23,2) in the Control samples and.91 to 3534 IU/L (median 160; average, 855.3 + 1334.7) in the Met samples. About 2.5 ml of bone marrow sample were pooled, from each individual for total volumes of 16.5 ml each for the Control and Met samples. An aliquot of 100 μl was saved from each sample and an aliquot of 100 μl pooled sample was saved for later analysis.

2. Purification Strategy

Based on the concentration of several, known serum markers for prostate cancer, e.g., prostate-specific antigen (PSA) and carcinoembroynic antigen (GEA) , it was expected that the bone metastasis related factors were present: inflow concentrations in the bone marrow' samples. Assuming that the concentrations of proteins of interest in the bone marrow are around 1-100 ng/ml, which is the level of PSA in prostate cancer patients' serum samples, it was estimated that at least 10 ml of starting material would be needed to purify sufficient protein for sequence analysis. In addition, enrichment of the low-abundance proteins through protein purification was needed before 2 -D gel electrophoresis could be used as a final protein separation step for sequence determination. Therefore, parallel protein purification was used followed by SDS- PAGE or 2 -D gel electrophoresis of samples to identify proteins differentially expressed in the bone marrow of prostate cancer patients with and without bone metastasis. To minimize the number of fractions to be analyzed, affinity chromatography that generates only two protein fractions, bound and flow-through fractions was initially used.

The major obstacle in purifying low abundance proteins from serum or bone marrow is the presence of large amount of serum albumin, which makes up 90-95% of total, serum proteins. FIGURE 1 demonstrates a SDS-PAGE analysis of total proteins from bone marrow supernatant. Bone marrow supernatant was diluted with distilled water to a _. final protein concentration of 1 mg/ml .and various amounts of -_.bone marrow proteins as indicated were loaded on to SDS-PAGE. Purified bovine, serum albumin (Pierce,) of varying, amounts as indicated was loaded for comparison. The gel was stained with Coomassie blue. Albumin (indicated by arrows) and immunoglobins (indicated by open arrows) are the major components of bone marrow supernatant. As shown in FIGURE 1, serum albumin (marked with black arrow) is the major component of the bone' marrow supernatant . Total protein concentration in the bone marrow sample was measured to be around 80 mg/ml . Thus, about 75 mg of albumin is present per ml' of bone marrow.

To test for the feasibility of removing albumin from the bone marrow supernatant, the serum from a healthy donor was incubated with Cibachrome blue affinity matrix and the protein profile of the flow through fraction was tested by SDS-PAGE. Although Cibachrome blue gel matrix can be used to remove albumin from bone marrow supernatant, it also removes a significant amount of other, proteins as ..a result of nonspecific binding, , (data not shown) . Therefore, to avoid the possibility of losing potentially important low abundance proteins, albumin was not removed, but proteins of interest were selectively separated out .

, Because albumin binds to proteins nonspecifically^ presumably through hydrophobic interactions, in one embodiment, the invention provides the use of detergent during protein purification by column .chromatography to purify proteins of interest from samples having large amount of μnwanted albumin. The detergent present during column chromatography will block hydrophobic interactions between contaminant proteins and proteins of interest.

Suitable detergents include, but are not limited to, various grades of Triton and Tween, such as Triton X-100, Tween 20, Tween 80, among others. The concentration of the detergent may vary. For example, a final detergent concentration of about 20% or lower can be used. However, in some cases, a higher detergent concentration may also be used_....For example, 5% Triton X-100 may be , us.ed to block. the binding of low-abundance proteins to albumin. Addition of Triton X-100 precludes the possibility of removing albumin by Cibachrome gel matrix because the binding of albumin to Cibachrome is sensitive to the presence of detergent. However, chromatography methods that are not affected by the presence of detergent can be used. For this purpose, there are several different affinity columns that may be used in the presence of detergents. The effect of Triton X-100 on the ability of lectin-affinity columns to bind glycoproteins was tested. For example, Triton X-100 was added to serum to final concentrations of 1%, 2%, or 5% and the detergent containing serum was incubated with ConA-Sepharose or WGA-Sepharose .

It is found that both ConA-Sepharose and WGA- Sepharose were able to bind glycoproteins in the presence of a detergent, such as about 5% of Triton X-10. In addition, albumin did not co-purify with the glycoproteins in the presence of 5% Triton X-100. The inclusion of 5% Triton X-100 in the serum sample precludes the use of chromatographies that utilize hydrophobic interactions, e.g., Phenyl-Sepharose and Octyl-Sepharose. Further, a method of purifying proteins of interest from a sample with unwanted albumin is provided herein. Suitable sample includes serum, bone marrow supernatant, blood, urine, body fluid, among others. The method includes providing the sample having unwanted albumin, combining the sample with a detergent into a mixture having a final detergent concentration of about 20 % or lower, and purifying protein of interest free of contaminant albumin from the mixture through column chromatography using a buffer containing the detergent.

A purification scheme used for purifying bone metastasis factors is shown herein in FIGURE 2. Because of the complexity of the bone marrow samples, affinity chromatographies that have high selectivity, and thus bind only small amount of proteins, were used first, followed with less selective affinity matrix or ion- exchanger to sequentially deplete the proteins from the samples. Examples of affinity chromatographies include, but are not limited to, wheat-germ agglutinin (WGA)- agarose affinity chromatography, Con-A-agarose affinity chromatography, protein G-Sepharose affinity chromatography, among others. Finally, the remaining samples, which have reduced complexity, were separated by ion-exchangers using stepwise elution to minimize the number of fractions to be analyzed. Examples of ion- exchange chromatographies include, but are not limited to, POROS-HQ, Mono Q, Mono S, Q-Sepharose, and R-Sepharose, among others.

The protein purification steps illustrated in FIGURE 2 were carried out with the inclusion, of 5% Triton X-IOO to the bone marrow supernatants for the, control and Met samples,, and affinity chromatography was used to isolate specific groups of proteins. This approach is different, from the current proteomics approach in that it avoids interference from albumin and allows us to zero in on the relatively low abundance but biologically important proteins present. Exemplary protein purification techniques are described in Scopes, R. K. (1994) Protein Purification: Principles and Practice (Third Edition) , Springer-Verlag, New York, N. Y.

3.,SPS-PAGE and 2-Dimensional Gel Electrophoresis Analysis . Proteins eluted from the chromatography columns were analyzed by SDS-PAGE using 4-12% gradient NuPage gels (Invitrogen Corp., Carlsbad, CA) after samples were ^■ heated at 100⁰C in the sample buffer containing 1 mM DTT for 5 min. Two-dimensional gel analysis: was performed using immobilized pH gradient- isoelectric focusing-based two-dimensional electrophoresis (Amersham Pharmacia Biotech Inc.) according to the manufacturer's recommendation.

The first dimension of the 2 -D gel electrophoresis separation was carried out using immobilized pH strips (IPG strips with a linear separation range of pH 4-7) from Amersham Pharmacia Biotech Inc. The samples were dialyzed against rehydration buffer (8M urea, 0.1% CHAPS, and 0.0002% bromophenol blue) overnight. CHAPS, ampholytes pH 4-7, and DTT were then added to a final concentration of 2%, 0.5%, and 20 mM, respectively. Isoelectric focusing was performed using the IPGphor system from Amersham Pharmacia Biotech Inc. The second dimension was separated on either 4-12% gradient. NuPage gels (Invitrogen Corp.) or 12.5% Ettan DALTII Gel (Amersham Pharmacia Biotech Inc . ) according to the manufacture ' s recommendation . Proteins were stained by either Coomassie blue

(Sigma, St. Louis, MO) or silver stain. In addition, protein_, pongentrations were determined by protein .assays carried out using the Coomassie blue Plus protein assay reagent kit (Pierce, Rockford, IL) according to the manufacturer's procedures. BSA standards (Pierce) were used for calibration.

4. Protein Identification by LC-MS

After separation of proteins on 4-12% SDS-PAGE, the Coomassie blue- or silver-stained protein bands of interest were excised from the gel, digested with modified trypsin (Roche), and analyzed by 'liquid- chromatography-mass spectrometry (LCQ MS/MS; Hewlett- Packard HPIlOO connected to a Thermo-Finnigan LCQdeca ■ electrospray ionization ion trap mass spectrometer) . The acquired MS data were analyzed by SEQUEST software (Thermo-Finnigan) against the National Center for Biotechnology Information (NCBI) database. 5. Wheat-germ Agglutinin (WGA) -Agarose Affinity Chromatography

As an example, WGA-Agarose affinity chromatography, which binds glycoproteins containing sialic acid and N- acetylglucosamine, was used as the first step in our purification scheme. WGA-Agarose (EY-Laboratories, San Mateo, CA) (2 ml) was transferred to a Poly-Prep Chromatography Column (Bio-Rad Laboratory, Hercules, CA) . The WGA-agarose column' was washed with 40 ml of 50 mM Hepes, pH 7.4. The column was further equilibrated with 50 mM Hepes, pH 7.4 containing 5% Triton X-100. Pooled bone marrow Control samples or Met samples were adjusted to a final concentration of 5% TritonX-100 by the addition of equal volume of 10% Triton X-100. The final volume was 33 ml for each sample. The Control or Met samples were then passed through the WGA-agarose column and the columns were washed with 20 ml of 50 mM Hepes, pH 7.4 containing 0.1% Triton X-100. These columris were eluted with 0.3 M N-acetylglucosamine in water. Fifteen 1-ml fractions were collected. FIGURE 3 is a graph showing the protein elution profile of WGA-agarose chromatography of the Control and Met samples. Peak fractions were pooled together and the concentrations of proteins in the pooled fractions were 75 mg/ml for both the Control and Met samples. From 1,200 mg total proteins, WGA-agarose purification yielded about

0.4-0.5 mg of total proteins and approximately 2400-fold enrichment of specific glycoproteins from the starting material . Both Control and Met samples produced similar . _. . . 31 . elutiori profiles as shown in FIGURE 3 and eluted almost the same amount of proteins .

FIGURE 4 depicts SDS-PAGE analysis of proteins eluted from WGA-agarose chromatography. SDS-PAGE analysis revealed a 16 kDa protein present at a higher concentration in the Control sample than in the Met sample. This protein was subsequently identified by mass spectrometry (MS) to be hemoglobin β-chain. The decrease in the level of hemoglobin in Met sample is most likely due to anemia in the advanced metastatic patients. 2 -D gel electrophoresis analysis was used to further separate the proteins in the WGA fractions.

FIGURE 5 demonstrates 2 -D gel electrophoresis analysis of proteins eluted from WGA-agarose chromatography. When silver staining was used to detect proteins on the 2D gel, several proteins that are differentially expressed in the Control versus Met . ■ samples were identified. Two proteins,, as indicated by the circle,, ■ were cut from the gel for protein identification and were determined by mass spectrometry to be, a, secreted form of ErbB3, designated as MDA-BF-I, and a. katanin-like molecule, designated as MDA-BF-2. The pi for both two proteins is about 5.5 and the molecular weights are about 45 kDa. The invention further provides a, study of the expression and function of MDA-BF-I because MDA-BF-I is a secreted protein that is likely a paracrine . factor mediating the interactions between prostate cancer cells and osteoblasts. Further, MDA-BF-I is an isoform of ErbB3 growth factor receptor; however, as demonstrated, herein, MDA-BF-I is functioned differently from ErbB3. The invention thus provides a study on the role of MDA-BF-I in cancer progression. 6. ConA-Sepharose Affinity Chromatography

ConA-Sepharose (EY-Laboratories) (2ml) was transferred to a Poly-Prep Chromatography Column and washed and equilibrated similar to WGA-Sepharose as described above. The flow through fractions from the WGA- agarose affinity column were passed through the ConA- Sepharose column and the columns were eluted with 0.3 M oc-methylmannoside in water. Fifteen 1-ml fractions were collected. The WGA flow-through fraction was further purified through ConA-agarose affinity chromatography, which binds to high mannose glycoproteins, .

FIGURE 6- depicts a protein elution ,profile of ConA- agarose ^• chromatography. A similar elution profile was obtained from Control and Met samples. About 10 mg of proteins ^! (ConA fraction), which is about 20 times to that from WGA-agarose chromatography, were eluted from ConA- agarose. This observation suggests that high mannose type glycoproteins are present in high abundance in the bone marrow supernatant. As a result, only an approximately 125 fold enrichment of this specific gro.up of . glycoproteins from the starting material was obtained from ConA-affinity chromatography.

FIGURE, 7 demonstrates SDS-PAGE analysis of .proteins eluted from ConA-agarose chromatography. Ten microliters of samples from each eluted fraction was loaded in each lane. Lanes C2, C3 , C4, and C5 are fraction number 2-5 from Control samples. Lanes M2, M3 , M4, and M5 are fraction numbers 2-5 from Met samples. The proteins were detected by Coomassie blue staining. Upon SDS-PAGE analysis, high levels of proteins with apparent molecular masses of 40 kDa, 17 kDa, and 13 kDa were detected in Met samples, as indicated by arrows. These three protein bands were cut from the gel for protein identification, and protein sequence analysis indicated that they were haptoglobin. The apparent molecular mass of haptoglobin was shown to be around 40 kDa and it is likely that the 17 kDa and 13 kDa fragments are the proteolytic products of intact haptoglobin.

2 -D gel electrophoresis analysis was used to further resolve the proteins. However, the majority of proteins in ConA fractions appear as a smear around 70 kDa and pi of 4-7 (data not shown) . Such a poor resolution in 2-D gel electrophoresis analysis may be due to glycosylation. Nine protein spots that were found present in Met samples but absent from Control samples were selected for protein sequence analysis. Four of these proteins were found to be haptoglobin, two were αl-antitrypsin, one ^■ was α2- macroglobulin, one protein did not have a match in protein database, and one protein matched with an unknown protein (work in progress) . Western blot analysis of the 2-D gel with anti-haptoglobin antibodies indicated that haptoglobin was present in the Met samples as multiple molecular weight and pi forms (data not shown) .

7. Binding of Immunoglobin (Ig) to Protein G- Sepharose

The flow through fractions from the ConA-Sepharose column were incubated with Protein G-Sepharose (Amersham Pharmacia Biotech Inc., Piscataway, ^' NJ) (2 '^'ml) overnight at 4°C with constant mixing. The mixture was transferred to a Poly-Prep column and the flow-through fraction was collected. The flow through fraction was incubated with another 2 ml of Protein G-Sepharose at room temperature for 6' h. The resulting flow through fraction was collected for separation on a POROS-HQ column as described below. 8. Poros-HQ Chromatography

The protein purification was initially performed in the presence of serum albumin and immunoglobulin, both abundant components of bone marrow. To reduce the complexity of the samples before applying to ion- exchanger chromatography, immunoglobulins were removed from the samples by affinity absorption with Protein G- Sepharose. The flow-through fractions from Protein G- Sepharose were then applied to POROS-HQ and proteins were eluted with several salt concentrations in a stepwise fashion. Ppros 50-HQ (Applied Biosystems Inc. Framingham, MA) (2 ml) in a Poly-Prep Chromatography Column was equilibrated with 20 inM Hepes, pH 7.4. The flow-through fractions from Protein G-Sepharose were applied to the POROS-HQ and bound proteins were eluted in a stepwise manner with 25 ml each of 20 mM Hepes, pH 7.4 buffer containing increasing amounts of sodium chloride (50, 100, 200, 500, 1000, and 2000 mM) . Twenty-five 1-ml fractions were collected for each sodium chloride concentration. Again, detailed protocols of protein purification can be found in Scopes, R. K. (1994) Protein Purification: Principles and Practice (Third Edition) , Springer-Verlag, New York, N. Y.

TABLE 1 demonstrates the amounts of proteins recovered from each protein purification step. Serum albumin was present in the 50 mM, 100 mM, and 200 mM sodium chloride bound fractions from the POROS-HQ column chromatography and was the major component of these fractions (data not shown) . The amount of proteins present in the fractions eluted from 1.0 M and 2.0 M sodium chloride was below the detection limit used in this study. As shown in TABLE 1, the enrichment of proteins by Poros-HQ column is low_/ thus, these fractions were not further analyzed.

TABLE 1: Purification of proteins identified as differentially expressed in bone marrow supernatants of prostate cancer patients by comparing results from prostate cancer samples with or without bone metastasis

Purificati Protein Fold Proteins differentially on Steps (mg) enrichme expressed nt

Total 1200 1

WGA- 2400 Katanin (NP__008975) agarose 0.5 ErbB-3 (P21860)

ConA- 125 Haptoglobin agarose 9.6 (NP 005134.1) αl-antitrypsin (P01009) α2 -macroglobulin (NP 000005.1)

POROS-HQ flow- N. D.* through

50 mM 180 7

100 mM 66 18

200 mM 35 34

500 mM 15 80

*N.D. protein concentration was not determined due to the presence of Triton X-100 in this fraction. 9. Western Blotting

Bone marrow supernatants were diluted 10,000 fold distilled water and ten microliters of each sample was loaded onto SDS-PAGE. Proteins separated by SDS-PAGE or 2 -D gel electrophoresis analyses were transferred onto nitrocellulose (Schleicher and Schuell, Hamburg, Germany) using Tris-glycine transfer buffer containing 20% methanol. Blots were probed with diluted antibodies, such as anti-haptoglobin polyclonal antibody (DAKO,

Carpinteria, CA) at a 1:2000 dilution. Peroxidase-labeled anti-rabbit antibody was used as a secondary antibody and proteins were visualized using ECL Western blotting detection reagent (Amersham Pharmacia Biotech) according to the manufacturer ' s procedures . 10. Immunohistochemical Staining

Tissue sections from formalin-fixed, formic acid decalcified, and paraffin-embedded bone biopsy specimens were dewaxed with xylene and rehydrated in graded alcohol. The sections were then treated with 3% H₂O₂ in methanol at room temperature for 15 minutes, washed with phosphate-buffered saline (PBS) , blocked with normal goat serum at room temperature for 30 minutes, and then incubated at room temperature for 1 hour with an antibody against haptoglobin (DAKO, Carpinteria, CA) , extracellular domain of ErbB3 (Neomarkers, Fremont, CA) , or p60 katanin. The antibody binding was detected by using ABC kit (Vector laboratory. Burlingame, CA) according to the manufacturer's instructions with 3, 3'- diaminobenzidine as the chromogen. The immunostained sections were then counterstained with hematoxylin.

11. Haptoglobin in Individual Bone Marrow Samples

Haptoglobin, one of the proteins identified by the protein purification steps of the invention, showed a significant increase and was present at a relatively high level in the Met sample. It was further examined whether the increase of haptoglobin in the pooled Met sample was due to an incidental event from one of the patients or was generally associated with the Met group. The level of haptoglobin in the bone marrow supernatants of individual patient samples was measured by Western blot.

FIGURE 8 depicts a Western immunoblot of bone marrow supernatants from individual patients probed with polyclonal anti-haptoglobin antibody (DAKO) at 1:2000 dilution and developed by enhanced chemiluminescence . Bone marrow supernatants from the individual patients were diluted 10O₇OOO fold with distilled water and ten microliters of each sample was loaded for Western blot analysis. Cl to C6 and Ml to Mβ of FIGURE 8 represent a total of 12 individual patient samples. They were also pooled, 6 of them each, as Control and Met samples for protein purification. As shown in FIGURE 8, the level of haptoglobin

(approximately 43k) was much higher in the patients of the metastatic group, as 5 out of 6 samples showed a significant increase of haptoglobin levels compared to those of individual control patients. This observation indicates that the increase in haptoglobin detected in the pooled Met samples reflects an increase in multiple patients from the Met group.

12. Haptoglobin in Bone Marrow Biopsies

Haptoglobin is mainly synthesized in the liver. However, its synthesis has also been demonstrated in several other normal or malignant cell types. Immunohistochemical studies were performed to identify haptoglobin expressing cells in a bone sample with metastatic prostate cancer from a patient who underwent laminectomy.

FIGURE 9 demonstrates immunohistochemical staining for haptoglobin in metastatic bone lesion of prostate cancer. The haptoglobin was detected predominantly in the extracellular space and was not detected in prostate cancer cells (arrows) , osteoblasts (open arrowhead) and osteocytes (arrowhead) , and bone matrix (asterisk) . As shown in FIGURE 9, haptoglobin immunoreactivity was seen mainly in the extracellular space. The prostate cancer cells and osteoblasts do not express haptoglobin. This observation suggests that the increased level of haptoglobin is not derived from metastatic prostate cancer or bone cells and is likely secreted from other organs such as the liver.

13. MDA-BF-I in Bone Biopsy of Metastatic Prostate Cancer The MDA-BF-I protein identified from the metastatic bone marrow samples is a short isoform of the full length pl80-ErbB3 and contains only the extracellular domain having an apparent molecular mass of around 45 kDa. Full length ErbB3 was originally identified as a transmembrane growth hormone receptor that shares sequence homology with the EGF receptor family. Alternatively-spliced mRNA transcripts of ErbB3 composed of the extracellular domain have been reported. Thus, the MDA-BF-I protein identified herein is a soluble, secreted isoform of ErbB3 th »at contains only the extracellular domain. Several attempts to detect MDA-BF-I using a monoclonal antibody against the extracellular domain of ErbB3 directly in the unfractionated bone marrow samples from individual patients by Western blot failed (data not shown) . This may be due to the fact that MDA-BF-I, which was purified from WGA-affinity chromatography after about 2,400 fold enrichment, as seen in FIGURE 5, was a low abundance protein in the metastatic bone marrow samples.

To identify the cell types that produce MDA-BF-I in the metastatic bone marrow samples, immunohistochemical studies were performed using an antibody against the extracellular domain of ErbB3 to stain MDA-BF-I. This antibody does not stain ErbB3. It is found that MDA-BF-I was strongly expressed in the tumor cells as well as in the osteoblasts, which were induced as a result of invasion of the bone compartment by prostate cancer cells, as shown in FIGURES 10-12. FIGURES 10-12 demonstrate MDA-BF-I expression in bone metastasis of prostate cancer. Antibody specific, for MDA-BF-I was used to stain the tissue cross-sections of bone biopsy sample. As shown in FIGURE 10, MDA-BF-I was detected in the metastatic prostate cancer cells (arrows) in bone. The pre-existing lamellar bone matrix (asterisk) is negative for MpA-BF-I.

¹ As shown in FIGURE 11, MDA-BF-I staining was also detected in the osteoblasts (arrowhead) and their surrounding new bone matrix (asterisk) . MDA-BF-I was found to be present as a diffuse layer around the activated osteoblasts and their newly formed bone .matrix. In addition, • the adjacent metastatic prostrate cancer cells (arrow) were also positive with.._.MDA-BF-I, staining.

As shown in FIGURE 12, strong MDA-BF-I reactivity was detected in both the osteoblasts (arrowhead) and' their surrounding new bone matrix (asterisk) in a bone trabecula away from tumor deposits in the same bone . specimen.

These observations suggest that soluble MDA-BF-'l is expressed from both the metastatic prostate tumor cells and osteoblasts. In contrast, soluble MDA-BF-I was not expressed in the epithelial cells in 20 normal prostate tissue samples and 20 primary prostate^' tumor samples . ' examined (data not shown) . Soluble MDA-BF-I was also not detected in resting osteoblasts in three npn-metastatic bone tissue samples examined (data not shown) , Further, soluble-. MDA-BF-I was expressed in activated osteoblasts present in non-metastatic bone tissues (data not shown) . 14. MDA-BF-2 in Bone Marrow Biopsy of Metastatic . Prostate Cancer

MDA-BF-2 was a low abundance protein also identified from the WGA-agarose purified fraction. Therefore, the presence of MDA-BF-2 could not be detected directly in the unfractionated total bone marrow samples from individual patients by Western blot. MDA-BF-2 is a katanin-like molecule possibly functioning as a microtubule severing protein. Immunohistochemistry was used to identify the cell types that express MDA-BF-2. An affinity-purified antibody against p60 subunit of katanin was used.

FIGURE 13 depicts MDA-BF-2 expression in bone metastasis of prostate cancer. Antibody against p60 subunit -of katanin detected MDA-BF-2 in the tumor cells (arrows) and osteoblasts (arrowheads) ., The bone matrix (asterisk) is negative.

As shown in FIGURE 13, during bone metastasis, MDA- BF-2 was expressed in the tumor cells and activated ^• osteoblasts, which were localized around the bone' trabeculae adjacent to metastatic carcinoma cell nests. ^■ Osteoblasts that were distant from the tumor cells did not show reactivity with katanin antibody, (data not shown). The presence of MDA-BF-2 in metastatic. bone. marrow supernatant and biopsies suggests that MDA-BF₇? may be .produced locally by these cell types at the site of metastasis.

II. MDA-BF-I is a 45-kDa soluble form of ErbB3 that is one of the bone metastases factors involved in bone- epithelium interactions

Another embodiment of the invention provides MDA-BF- 1, which is a secreted form of ErbB3 having a molecular weight- of about 45 KDa. ErbB3 is a known transmembrane growth factor receptor that is similar in sequence to members of the EGF receptor (EGF-R) family and has an apparent molecular mass of 180 -kDa (full length) . The full length pl80 transmembrane ErbB3 (pl80-ErbB3) has all the structural features of a receptor tyrosine kinase and has a 612 -residue extracellular ligand-binding domain, a 32 -amino acid transmembrane region, and a 677-amino acid C-terminal cytoplasmic domain that is very homologous with those of other members of the receptor tyrosine kinase family.

FIGURE 14 depicts a comparison of the structures of MDA-BF-I and pl80-ErbB3 mRNAs and proteins. Ea.ch box represents one exon. MDA-BF-I mRNA is transcribed from exons 1-8, whereas pl80-ErbB3 mRNA is transcribed from exons 1-28. The ligand-binding domain, the transmembrane domain, the kinase domain, and the C-terminal domain of pl80-ErbB3 are indicated. The ligand-binding domain can_, be divided into four subdomains (I-IV) . Subdomains. II and IV are cysteine-rich regions that are, conserved in EGF-R, and_, subdomains I and III generally define the specificity for ligand binding.

FIGURE 15 depicts the pl80-ErbB3 domain arrangement of the ligand binding domain. X-ray crystallography of the extracellular region of pl80-ErbB3 revealed that domain II interacts with domain IV to form a, closed- looped structure. The ligand must interact, with domains I and III to induce the necessary domain arrangement for signaling. , Comparison of the protein structures of MDA-BF-I and pl80-ErbB3 reveals that MDA-BF-I includes subdomains I (exons 1-4) , subdomains II (exons 5-7) and exon 8 of domain III, having 310 amino acids and extending only to half of the ligand-binding domain of pl80-ΞrbB3. The isolation of MDA-BF-I challenges the notion that ErbB3 protein family is merely a family of transmembrane receptors and suggests that the ErbB3 family of proteins has other roles. The fact that the isolated MDA-BF-I does not have a complete domain III indicates MDA-BF-I may not be able to function as a competitor, inhibitor, or modulator for ErbB3 receptor-Heregulin ligand interaction, suggesting its functions other than a receptor. This is the first time that this protein, MDA-BF-I, has been identified in human tissues. cDNAs encoding a 45-kDa isoform of ErbB3 have been previously described from 3'-RACE and human full length expression cloning using mRNA transcripts, without identification of the protein products or any roles of these mRNA transcripts. The cDNA sequence for the alternatively spliced ErbB3 mRNA transcripts can be found in the GenBank Accession numbers BC002706, U88358, and BT007226.

Previously, elevated levels of various ErbB3 mRNA transcripts were detected in a number of human mammary tumor cell lines without identification of the corresponding protein products. Also, overexpression of ErbB3 is associated with a worse prognosis in, patients with endometrioid carcinoma of the ovary. These observations suggest that increased ErbB3 expression may play a role in human malignancies.

Using primers from exon 7 and exon 9 of pl80-ErbB3 and RNA from MDA PCa 2b cells for RT-PCR analysis, a transcript containing intron 8 (90 bp) able to hydridize to an intron 8 probe was detected. This RT-PCR product was not from genomic DNA because it did not contain intron 7 sequences as indicted by its size and failure to hybridize with intron 7 probes as well as the failure to generate this sequence if reverse transcriptase were not included in the PCR reaction. Thus RT-PCR indicates that MDA-BF-I arises from intron 8 retention. Further, because approximately 40% of bone metastasis samples express only MDA-BF-I, but not pl80-ErbB3, the expression of MDA-BF-I does not appear to be dependent upon plδO- ErbB3 expression, making it unlikely that MDA-BF-I is a proteolytic product of p!80-ErbB3. Finally, MDA-BF-I is likely generated from alternative splicing of the ErbB3 full length transcript because production is not decreased by siRNA directed to intron 8 (see examples below) .

This isoform is likely generated from alternative splicing of the ErbB3 full length transcript because production is not decreased by siRNA directed to intron 8 (see .examples below) . , , .

III. MDA-BF-I is distinctly different from pl80- ErbB3 : Signal transduction of pl80-ErbB3 and the possible function of MDA-BF-I The increased levels of MDA-BF-I in prostate cancer patients with bone metastasis raise the important question of the role of MDA-BF-I in prostate cancer progression. To determine the function of ¹MDA-BF-I, it is useful to understand the signal transduction pathway of pl80-ErbB3. ' ^' ,. ^{■■ ■ • > ■} > ^■ . -

FIGURE 16 depicts pl80-ErbB3 signal transduction pathways. Heregulin (HRG) ligand-induced signaling transduction requires both ErbB3 and ErbB2 ^■. pl80-ErbB3 is unique among members of ErbB receptor family in that its tyrosine kinase activity is low. Although the tyrosine kinase subdomain of pl80-ErbB3 is 60-62% similar to the catalytic domains of the other members of the growth factor receptor subfamily, three critical residues that are conserved in the kinase domain are altered in pl80- ErbB3, resulting in a kinase-inactive ErbB3. Therefore, pl80-ErbB3 alone cannot initiate down-stream signaling. Instead, it forms a heterodimer with ErbB2. Although pl80-ErbB3 has no intrinsic kinase activity and ErbB2 is a ligandless receptor, the ErbB3 ligand heregulin (HRG) - induced heterodimerization of ErbB3 with ErbB2 generates a signaling complex that results in increased PI3 -kinase recruitment and activates a signaling pathway that may lead to neoplastic transformation, as shown in FIGURE 16. Unlike ErbB2, which can spontaneously form active homodimers, ErbB3 does not have the ability to initiate downstream signaling by itself.

MDA-BF-I, on the other hand, may interact with a unique receptor present in osteoblast involved in a signal transduction pathway different from pl80-ErbB3, as shown in FIGURE 16. One possible function of MDA-BF-I is to act as a decoy receptor competing for ligand HRG binding to pl80-ErbB3. However, several lines of evidence suggest that the structure of MDA-BF-I is distinctly different from pl80-ErbB3. First, the conformation of MDA-BF-I is different from pl80-ErbB3. X-ray , crystallography of the extracellular region of pl80-ΞrbB3 revealed that domain II interacts with domain IV to form a closed-looped structure, as shown in FIGURE 15. MDA-BF- 1, which lacks domains III and IV, cannot form such a closed-looped structure that is typical of ErbB family membrane receptors. Second, MDA-BF-I does not bind HRG with high affinity. X-ray crystallography of the extracellular region of pl80-ErbB3 revealed that the HRG ligand must interact with subdomains I and III to induce the necessary domain arrangement for signaling. MDA-BF- 1, which lacks domain III, may not bind HRG with high affinity. Indeed, the affinity of MDA-BF-I for HRG is 68 nM, which is 30-fold lower than the affinity of the extracellular domain of pl80-ErbB3 for HRG (2.3 nM) . Consistent with this observation, it was also reported that recombinant protein made from a previously identified cDNA encoding a secreted form of ErB3 , cannot antagonize HRG-induced growth stimulation (Lee et al . Cancer Research, 61:4407-4473, 2001). These observations suggest that MDA-BF-I is distinctly different from plδO- ErbB3, not only in its structure but also in its functions .

One or more embodiments of the invention thus provide the study of the functions of MDA-BF-I. As described herein, it was found that the levels of MDA-BF- 1 protein are increased in prostate cancer patients with bone metastasis and increased expression of MDA-BF-I correlates with metastatic growth of prostate cancer, suggesting that MDA-BF-I may have a role in prostate cancer progression, especially metastatic growth in bone. Thus, in one embodiment, the invention provides that MDA- BF-I is a paracrine factor that mediates bone-epithelium interactions during prostate cancer progression in bone.

In another embodiment, the invention illustrates that MDA-BF-I plays a role in osteoblast proliferation and differentiation, based on several observations, as described in detail throughout this invention. First, MDA-BF-I was isolated from the bone marrow supernatants of prostate cancer patients with osteoblastic bone metastasis and Western blots showed that MDA-BF-I is only present in prostate cancer patients with bone metastasis. Second, MDA-BF-I is present in the conditioned media of prostate cancer cell lines that produce an osteoblastic phenotype in bone, e.g., MDA PCa 2b cell line, but not in osteolytic cell lines, e.g. , PC-3 and DU145 cell lines. Third, recombinant MDA-BF-I induced osteoblast proliferation in vitro, but no prostate cancer cell proliferation. Fourth, preliminary data showed that expressing MDA-BF-I in PC-3 cells, which normally elicits strong osteolytic response in vivo, generated an osteoblastic response in vivo in an osseous prostate cancer animal model. Together, these observations strongly suggest that MDA-BF-I participates in the osteoblastic metastasis of prostate cancer in bone.

IV. Levels of MDA-BF-I in bone marrow supernatants of individual prostate cancer patients

In order to detect MDA-BF-I in patients and monitor MDA-BF-I expression during prostate cancer progression, the bone marrow supernatants from each patient in the Met and Control groups were screened for MDA-BF-I expression by Western blotting. Several attempts to detect MDA-BF-I in unfractionated bone marrow samples by Western blotting with a monoclonal antibody against the extracellular domain of ErbB3 were not successful (data not shown) .

This is because MDA-BF-I is present at very low concentrations in metastatic bone marrow samples before WGA-affinity chromatography, which produced a 2400-fold enrichment. Therefore, the bone marrow supernatants were incubated with WGA-agarose to enrich for the glycoproteins for analysis.

FIGURE 17 demonstrates Western blots of MDA-BF-I in enriched bone marrow supernatant samples from individual patients after WGA-affinity chromatography. As shown in FIGURE 17, more bone marrow samples from patients in the

Met group exhibited high levels of MDA-BF-I than bone marrow samples from patients in the Control group. The levels, however, may vary among individuals. Using Western blot analysis, twenty three of twenty seven samples (85%) from the Met group showed a significant increase of MDA-BF-I levels, whereas only one of twenty samples (5%) from the Control group had a weak MDA-BF-I level .

Thus, embodiments of the invention demonstrate that prostate cancer patients with bone metastasis have higher levels of MDA-BF-I in their bone marrow than those without clinical evidence of bone metastasis, and increase in the levels of MDA-BF-I in the bone marrow of prostate, cancer patients correlates with the development of bone metastasis. Significantly, from ithe .Western blot analysis,_. MDA-BF-I is the only secreted isofqrm, of ,ErbB3 present in the bone marrow of prostate cancer■ patients with bone metastasis because other isoforms of ErbB3 in either control or Met samples were not detected (data not shown) .

1. Use of MDA-BF-I for Detecting And Monitoring Prostate Cancer Bone Metastasis Embodiments of the invention'also¹ provide a ^•- sensitive method"to determine the levels qf MDA-BF-I in' the bone marrow from prostate cancer patients. Levels of MDA-BF-I in' the bone marrow supernatants are first ^' correlated with different stages of the disease. Then, the levels of MDA-BF-I in prostate cancer patients undergoing treatment for bone metastasis can¹ be measured- by diagnostic kits of the invention. The information can then be analyzed for use of MDA-BF-I as marker in predicting the occurrence of and the treatment effects on prostate cancer bone metastasis.

First, the correlation between the levels of MDA-BF- 1 in the bone marrow supernatants and the extent of bone metastasis in individual patients are correlated. Because prostate cancer progresses within the bone marrow cavity, bone marrow represents the microenvironment in which the bone metastases develop; thus, bone marrow supernatant may be more relevant than blood for studying factors that participate in bone metastasis. However,- blood is easier to obtain and the procedure of blood sample collection is much less invasive than bone marrow aspiration. Thus, an assay, such as an ELISA (enzyme- linked immunosorbent) assay, can allow for quantitative detection of MDA-BF-I in a diagnostic kit for a sample, such as blood and bone marrow supernatant samples from prostate cancer patients in order to correlate the levels of MDA-BF-I with the development of bone metastasis. Other tissue samples such as urine, cancer biopsy, plasma, and body fluid, among others can also be used.

2. Development of ELISA assay for MDA-BF-I .

A series of monoclonal antibodies were -generated' using purified MDA-BF-I expressed in Sf9 ' cells. Various methods known in the art can be used to generate polyclonal and monoclonal antibodies for the' identified bone factors described herein. Eight monoclonal antibodies that showed positive reaction in ELISA assay • are being characterized for their ability to capture MDA- BF-I' in a sandwich ELISA assay. This ELISA system can^' be used to detect the levels of MDA-BF-I in samples, such as the bone marrow supernatants and blood from prostate cancer patients, among others. The volume and dilutions of the bone marrow supernatant or blood samples required for the ELISA analysis are determined by titration. To test for potential plasma interference in the ELISA assay, plasma or bone marrow from normal patient samples are also used for plotting a standard curve and the level of MDA-BF-I from patient sample is compared^" with the standard. ■

Embodiments of the invention provide that a method of correlating increases in the levels of MDA-BF-I in bone marrow with the development and extent of bone metastasis in prostate cancer patients. Although PSA has been widely used to monitor prostate cancer progression, the levels of PSA do not correlate well with the occurrence and extent of bone metastasis. Thus, MDA-BF-I can be further developed as a valuable marker for monitoring prostate cancer bone metastasis. A .sensitive ELISA assays or other assays such as EIA, etc., can be developed using MDA-BF-I specific monoclonal antibodies of the invention. In case the ELISA sensitivity is not sufficient or there is interference from serum proteins, embodiments of the invention also provide the generation of additional polyclonal antibodies against the purified MDA-BF-I. Polyclonal antibodies against the purified MDA- BF-I are able to recognize multiple epitopes on the molecule and thus have higher avidity.

2.1,- Patient samples , , _,-■■, ■ ; .

¹ _.'Both blood and bone marrow samples from patients with^' prostate cancer are tested for the level of MDA-BF-I₁ performed according to a protocol approved by the Institutional Review Board of M. D. Anderson Cancer

Center. The samples are drawn from University of Texas M. D. Anderson Cancer Center (UTMDACC) Genitourinary Serum Bank, which has a detailed registry Of the- bone marrow ^■ and blood samples collected. For example, the clinical information -and laboratory results of the 42 patients are as follows . Non-Met group (20 patients) : No evidence of bone metastasis. Median age was 62 years. All patients were still alive after a median follow-up of 64 months. Their median serum PSA and alkaline phosphatase levels (at the time of the bone marrow biopsy and aspiration) were 5.6 (range, 0.2 - 13.1 ng/ml) and 73 (range, 45 - 124 IU/L) , respectively. Four patients had their bone marrow procured subsequent to a local irradiative therapy or radical ^"prostatectomy. The rest of the patients had their prostate intact at the time of the bone marrow aspiration. Met group (22 patients) ; with evidence of bone metastasis. Median age was 70 years. Median overall survival was 7 months. The median serum PSA and alkaline phosphatase levels were 158 (range, 0.8 - 10,120 ng/ml) and 565 (range, 103 - 3534 IU/L; normal, 38 - 126 IU/L) , respectively. All patients showed positive bone scan. Seventeen patients had evidence of infiltration of the bone marrow by tumor cells in their bone biopsies and 4 patients had a superscan on their bone scan. Both bone marrow and plasma samples were obtained from 11 of the Non-Met group and 12 of the Met-group patients. These samples will be used to compare the level of MDA-BF-I in their bone marrow versus plasma.

_,2-2 Sample size and statistical analysis

Statistical analysis is used to'compare the levels of MDA-BF-I in the bone marrow and plasma between the 20 Non-Met and 22 Met samples. Using the data from these 42 samples; this experiment will then be confirmed- -with an appropriate¹ 'number of samples from each group- to detect a clinically meaningful difference in the levels of MDA-BF- 1 between the groups with 80% power and a significance level of 5%. The expression levels of MDA-BF-I are" compared between the two groups using appropriate parametric (i.e., two-sample t-test), or, non-parametric methods (i.e., chi-square, Fisher-exact, or Wilcox on rank-sum tests) . 2.3 Selection of Monoclonal Antibodies

As described earlier, the invention provides several selected monoclonal antibodies that react with purified MDA-BF-I on ELISA. Embodiments of the invention also provide characterization of these monoclonal antibodies for their ability to (1) Inhibit the binding of ¹²⁵I-MDA- BF-I to osteoblasts/ and (2) inhibit MDA-BF-I induced osteoblast proliferation.

Purified MDA-BF-I can be radio-labeled with ¹²⁵I and used as the binding ligand. First, recombinant MDA-BF-I

(about 5 μg) in PBS is labeled with Na¹²⁵I (0.5 mCi) using iodogen-coated tube. After about 2 minutes at 23°C, the mixture is separated on a Sephadex G-50 column. The specific activity of the radiolabeled MIDA-BF-I (¹²⁵I-MDA- BF-I)₁ is then measured by a gamma counter and the- labeled MDA-BF-I can be used for receptor binding assays. Primary mouse osteoblasts plated in 6-well plates -,are assayed at confluence.. The cells are incubated for 2 hours at 4°C with 1 x 10⁵ cpm ¹²⁵I-MDA-BF-I and increasing concentrations of monoclonal antibodies. Nonspecific binding is determined by using monoclonal antibodies with matching subtype. The cells are then washed three times with ice-cold binding buffer and lysed in 0.5 ml of 0.1 N NaOH and 0.1% SDS for 30 min and radioactivity is determined using a gamma-counter.

Inhibition of MDA-BF-I-induced osteoblast proliferation by these monoclonal antibodies can also be tested._, Primary mouse osteoblasts (PMO). are cultured in Dulbecco's Modified Eagle's (DME) medium . supplemented with Q₁.5% of fetal bovine serum (FBS) and incubated with MDA-BF-I (50 ng/ml) and increasing concentrations of monoclonal antibodies for 40 hours. Cell proliferation can be assessed by a cell proliferation assay, for example by^' measuring [³H] -thymadine incorporation, among others. Monoclonal antibodies with matching subtype are used as controls .

V. MDA-BF-I levels in the plasma of prostate cancer patients with bone metastasis

FIGURE 18 compares the levels of MDA-BF-I in the bone marrow supernatant and the corresponding plasma of patients in the Met group. Unlike the bone marrow supernatants (shown as BM in FIGURE 18) , the plasma samples (shown as plasma in FIGURE 18) had very little or no MDA-BF-I.¹ -This result suggests that MDA-BF-I is ' produced locally by cells in the bone marrow. > ;

VI. MDA-BF-I expression in prostate cancer progression Because MDA-BF-I was detected in bone marrow supernatants, the invention also provides a method of determining which cells in the bone marrow produce MDA- BF-I.- In addition, the invention also provides a method of monitoring and diagnosing the expression of MDA-BF-I at various' stages of prostate cancer progression. First, immunohistochemical staining of prostate^" tissues frpm various stages of prostate cancer progression was performed_....-Various prostate tissues can be used, such as tissues from a primary site, lymph node, and bone metastasis, among others. Although other soluble forms of ErbB3 may be present in the tissues, Western blot analysis of bone marrow samples showed that MDA-BF-I (45 kDa) is the major soluble isoform present in the bone marrow. Thus, the soluble form of ErbB3 detected by immunohistochemistry is MDA-BF-I.

1. Human Prostate Tissue Samples Formalin-fixed paraffin embedded tissue samples were selected from the Tissue Bank of The University of Texas M. D. Anderson Cancer Center. These samples were selected to represent a spectrum of localized and metastatic prostate cancer, including radical prostatectomy specimens (n=20) , lymph nodes with prostate cancer metastases (n=19) , and bone with prostate cancer metastases (n=26) . The bone samples were derived from laminectomy and resection of the femur and humerus. These bone specimens were decalcified in formic acid according to the standard protocol established in the Department of Pathology at M. D. Anderson.

TABLE 2 is a table showing the clinical and pathological characteristics of some of the prostate samples tested. Of the 20 prostatectomy samples/ 10 were from patients with pathologic stage B, five from stage C, and five from stage D.

TABLE 2 : Clinical and Pathological Characteristics of Some Prostate Samples Tested

2. Antibodies Used in the Study to Determine Antibody Specificity

Rabbit polyclonal antibody C-17 and mouse monoclonal antibody RTJ.2, both against the ErbB3 cytoplasmic domain sequence, were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA) and used to detect the expression of the pl80-ErbB3. Because RTJ.2 gave less background staining than did C-17, the staining patterns of RTJ.2 are reported.

Rabbit polyclonal antibody Ab-10, against amino acids 71-86 in the ligand-binding domain of human ErbB3, and mouse monoclonal antibody Ab-8, against the extracellular domain of human ErbB3, were purchased from NeoMarkers, Inc. (Fremont, CA) and used in the immunohistochemistry study for the detection of MDA-BF-I. Consecutive slides from the same specimen were used for ease of comparison. Similar staining patterns were observed with Ab-10 and Ab-8 in prostate samples from radial prostatectomies and laminectomies. Because AbIO gave a slightly more intense staining than did Ab8, the staining patterns of AbIO are reported.

FIGURE 19 illustrates antibody recognition sites of these four antibodies. Antibodies C-17 and RTJ.2 only detect pl80-ErbB3. Because MDA-BF-I is a fragment of pl80-ErbB3, immunoreactivity observed only with AbIO or Ab8, but not with RTJ.2 or C-17, was interpreted as specific for MDA-BF-I. When the staining was positive with both AbIO and RTJ.2, both MDA-BF-I and pl80-ErbB3 were considered expressed.

3. Immunohistochemistry of MDA-BF-I expression

Four-micrometer-thick sections were dewaxed with xylene and rehydrated in graded alcohol. Sections were then treated with 3% H₂O₂ in methanol for 15 minutes at room temperature. Slides were washed with phosphate- buffered saline, blocked with normal goat_..serum at room temperature for 30 minutes, and then incubated at room temperature ' for 1 hour with the primary antibody against the extracellular or the cytoplasmic domain of ErbB3. To test for -the specificity of AbIO staining, in some cases, AbIO (1 μg/ml) was incubated with histidine-tagged purified MDA-BF-I (5 μg/ml) at 4°C overnight before applying to the sample. The antibody binding was detected by using an ABC kit (Vector Laboratories, Burlingame, CA) with 3, 3' -diaminobenzidine tetrachloride as the chromogen according to the manufacturer's instructions. The immunostained sections were then counterstained with hematoxylin, dehydrated, and mounted. Sections that were immunostained in a similar fashion but without the primary antibody were used as negative controls. The immunpstaining was considered positive when more than 50% of the tumor cells were immunoreactive . Slides were read independently by two pathologists, and the evaluation was concordant in 90% of the readings. In the cases where there were differences in the readings, a consensus was reached after a concurrent review by the two pathologists . FIGURES 20A-20B demonstrate immunohistochemistry of MDA-BF-I expression. As shown in FIGURE 2OA, staining of a prostate tissue by AbIO showed that the stroma and smooth muscle in the blood vessel walls were positive. In addition, as shown in FIGURE 2OB, the immunoreactivity could be blocked by recombinant MDA-BF-I, suggesting that AbIO immunoreactivity is specific to MDA-BF-I.

Although MDA-BF-I is a truncated form of pl80-ErbB3, it was recently found out that several antibodies against the extracellular domains of p!80-ErbB3, e.g., Ab-10, Ab- 8, and several monoclonal antibodies that wer,e generated against purified MDA-BF-I, recognize only MDA-BF-I but not p!80-ErbB3 in both Western blot and, immunohistochemistry analysis. As was discussed above, X- ray crystallography analysis indicates that the structure of MDA-BF-I is distinctly different from tha,t of pi8,0- ErbB3. Thus, it is thought that some of the antibodies against MDA-BF-I may bind to conformation epitopes specific to MDA-BF-I and thus can only recognize MDA-BF- 1.

4. MDA-BF-I Expression in Normal Prostate Cells

Embodiments of the invention provide a method to test and compare the expression of MDA-BF-I and full- length ErbB3 in various prostate cancer related tissues, including samples from normal prostatic glands, high- grade prostatic intraepithelial neoplasia (PIN) , primary prostate cancer, ^• lymph node metastasis, and bone metastasis. Other tissues can also be tested. The results are summarized in TABLE 3 and the immunohistochemistry results are shown in FIGURES 19 and 20.

TABLE 3: Results of comparing expression of MDA-BF-2 and full-length pl80-ErbB3 in various prostate cancer-related tissues. Histology MDA-BP-I pl80-ErbB3 MDA-BF-I positive positive positive ' only

Normal 0/20 (0%) 0/20 (0%) 0 prostatic glands

PIN HG^a 0/10 (0%) 0/10 (0%) 0

PCa (primary) 0/20 (0%) 0/20 (0%) 0

Lymph node 18/19 (95%) 12/19 (63%) 6/19 (32%) metastasis

Bone 23/26 (88%) 12/26 (46%) 11/26 (42%) metastasis

FIGURES 2IA-2IF demonstrate staining patterns of normal prostate glands (FIGURES 21A-21B) , high-grade prostatic intraepithelial neoplasia (Figures^' 21C-21D) , and primary prostate cancer (FIGURES 21E-21F) for MDA-BF- 1 and pl80~ErbB3. FIGURES 21A, 21C, and 21E indicate staining patterns for MDA-BF-I expression and FIGURES 2IB, 21D, and 21F indicate staining patterns for pl80-ErbB3 expression. As shown in FIGURE 2IA and 24B, normal prostate glandular epithelia (arrow) and basal cells (arrowhead) did not exhibit- staining for MDA-BF-I (FIGURE 21A) or pl80-ErbB3. (FIGURE 21B). These results., suggest that neither ; pi80, -ErbB3 nor MDA-BF_rl is expressed in normal prostate, epithelia. ,, ■ , . ,, • . ,, . ,

^■ In' contrast, prostatic stroma (asterisk) showed intense, diffuse, cytoplasmic staining for MDA-BF-I expression (FIGURE 21A) . In addition, no staining of epithelial cells, basal cells, or stroma for pl80-ErbB3 was observed (FIGURE 21B). Similarly, .smooth, muscle bundles in the vessels also showed intense staining for MDA-BF-I, but not pl80-ErbB3 (data not ₍ shown) . These results suggest that MDA-BF-I is produced in the_. stroma and smooth .muscle cells of normal prostate, .but not in the glandular epithelia or basal cells of normal prostate, and pl80-ErbB3 is not expressed in any of the normal prostate cells.

4. MDA-BF-I Expression in Primary Prostate Cancer

As shown in FIGURES 21C and 2ID, prostatic glands with high-grade prostatic intraepithelial neoplasia (PIN) did not exhibit staining for MDA-BF-I or pl80-ErbB3. FIGURES 2IE and 2IF also showed that all primary prostate tumors were stained negative, regardless of the Gleason scores or stages . In contrast, the stroma surrounding the high-grade prostatic intraepithelial neoplasia and primary prostate cancer were positive for MDA-BF-I expression, but negative for pl80-ErbB3 expression (FIGURES' 2lC-2JF) . These results suggest that MDA-BF-I is expressed in the stroma of prostate tumors. ,

5. MDA-BF-I Expression in Lymph Node Metastasis of Prostate Cancer.

As shown in TABLE 2 and FIGURE 22, the majority (95%) of the metastatic prostate cancer cells in lymph node stained positive for MDA-BF-I expression while a subset (63%) also stained positive for pl80-ErbB3 expression. Metastatic prostate cancer cells in eighteen of the nineteen lymph node samples examined were stained positive for MDA-BF-I expression. Overall, metastatic prostate cancer cells in twelve of the nineteen lymph node samples stained positive with pl80-ErbB3, while seven of the nineteen samples were negative for pl80- ErbB3 staining. By subtraction, there were six specimens that stained positive for MDA-BF-I only (TABLE 3) . FIGURES 22A-22D demonstrate staining patterns of metastatic lymph node prostate cancer for MDA-BF-I and p!80-ErbB3 expression. In FIGURES 22A and 22C, metastatic cancer cells (arrows) in lymph nodes are positive for MDA-BF-I expression whereas the adjacent lymphoid tissue (L) is negative. In addition, the smooth muscle fibers in the blood vessel walls of the lymph nodes also stained positive for MDA-BF-I expression (data not shown) , similar to the observations in normal prostate tissue.

FIGURE 22B and 22D demonstrate staining patterns for pl80-ErbB3 expression in metastatic lymph nodes cancer cells, where some cells are stained positive (FIGURE 22B) and others are stained negative (FIGURE 22D) for pl80-ErbB3 expression. The adjacent lymphoid tissue is negative for pl80-ErbB3 expression. Lymphoid tissues next to the metastasized prostate cells did not stain for either MDA-BF-I or pl80-ErbB3 expression (FIGURES 22A and 22B) .

These observations suggest that both MDA-BF-I and pl80-ErbB3 are up-regulated in prostate cancer cells that have metastasized to lymph nodes. Further, of the eighteen lymph node metastasis specimens that stained positive for MDA-BF-I expression, ten were from patients who had previously received hormone therapy (TABLE 2) . Among the twelve lymph node metastasis specimens that stained positive for pl80-ErbB3 expression, eight were from patients who had previously received hormone therapy. Thus, the increased expression of MDA-BF-I and pl80-ErbB3 in lymph node metastasis does not seem to correlate with hormone treatment.

6. MDA-BF-I Expression in Bone Metastasis

FIGURE 23 demonstrates staining' patterns of bone metastatic prostate cancer cells for MDA-BF-I and pl80- ErbB3 expression. FIGURES 23A, 23C, and 23E are _'the results for MDA-BF-I staining. Metastatic cancer cells (arrows) in desmoplastic stroma in the bone marrow are positive for MDA-BF-I expression. Stroma is focally and weakly positive. Activated osteoblasts (arrowheads) and the adjacent matrix (asterisks) in the bone marrow are also positive for MDA-BF-I expression. FIGURES 23B, 23D, and 23F (are the results for plδO- ErbB3 staining. Bone marrow metastatic cancer cells (arrows) and activated osteoblasts (arrowheads) in some cases are negative (FIGURE 23B) , weakly positive (FIGURE 23D) , or strongly positive (FIGURE 23F) for pl80-ErbB3 expression. The adjacent matrix (open circle) shows no staining for pl80-ErbB3 expression.

As shown in FIGURES 23A-23F, metastatic prostate¹ cancer cells in twenty three of twenty six bone specimens examined ^■ were stained positive for MDA-BF-I expression. In contrast, staining of the prostate cancer cells for pl80-ErbB3 expression in bone metastasis was somewhat variable: only twelve of twenty-six cases were stained positive for pl80-ErbB3 expression. , ,

Twenty-five of twenty-six bone metastasis specimens were from prostate cancer patients who had previously received androgen ablation therapy. In the only bone specimen from a patient who did not undergo hormone therapy .when the bone specimen was obtained, immunostaining for MDA-BF-I was positive in both tumor cells and osteoblasts and this specimen did not stain for pl80-ErbB3 (data not shown) . Thus, it appears that the up-regulation of MDA-BF-I or pl80-ErbB3 in bone metastasis does not correlate with hormone treatment.

In the bone compartment, the resting osteoblasts, which are bone-rimming cells with flattened morphology and intermittent distribution around the bone, did not exhibit staining for either MDA-BF-I or pl80-ErbB3 expression. ,In contrast, activated osteoblasts, which are bone-rimming cells with polygonal or cuboidal .shapes that line the bone in a continuous fashion and are found along the edge of the newly formed trabecular bone, showed positive staining for MDA-BF-I expression (FIGURES 23A, 23C, and 23E) . The results suggest that MDA-BF-I is up- regulated in the activated osteoblasts. These activated osteoblasts also showed varying intensities, from negative to strong (FIGURES 23B, 23D, and 23F) , of pl80- ErbB3 'staining. The newly formed bone matrix also showed positive staining for MDA-BF-I expression (FIGURES 23C and 23E) , but not for pl80-ErbB3 expression (FIGURES 23D and 23F) . These observations suggest that MDA-BF-I is secreted by activated osteoblasts.

7. MDA-BF-I Expression in Metastatic Sites Other, Than Bone , , ;., . > ._.. _.

Although^' prostate cancer mainly metastasizes to 'bone (about 80%), metastases to other organs, '^■ e.g.,^' liver, lung, and brain, are also seen in prostate cancer_' , patients. Method of the invention also includes detection of MDA-BF-I expression in a disease, such as a cancer disease or a metastatic disease in many other organ sites. As an example, detection of MDA-BF-I is also tested in prostate cancer that was metastasized to lung (three specimens), liver, adrenal gland, brai.n , (two, specimens each) , and skin, adipose, and chest wall (one specimen each) . It was found that MDA-BF-I was not expressed in prostate cancer cells in these organs (data not showμ) . Thus, expression of MDA-BF-I is specific to prostate cancer in bone and lymph nodes. 8., MDA-BF-I Expression in Activated Osteoblasts of

Non-cancer Related Bone Fractures

Embodiments of the invention also examine whether MDA-BF-I is expressed in activated osteoblasts that are unrelated to prostate cancer bone metastasis. Bone specimens obtained from patients with non-cancer related bone fractures were stained for the expression of MDA-BF- 1. It was found that activated osteoblasts in these specimens stained positive for MDA-BF-I. Thus, expression of MDA-BF-I in activated osteoblasts is not specific to prostate cancer, suggesting that MDA-BF-I is a bone- specific protein.

9. Summary of MDA-BF-I Expression in Prostate Cancer MDA-BF-I was detected in metastatic prostate cancer cells which are progressed to lymph node and bone, and in activated osteoblasts of the bone marrow. In> addition, ^■ using immunohistochemical staining, MDA-BF-I was not expressed in epithelia of normal prostate and localized prostate cancer. TABLE 3 summarizes MDA-BF-I staining in these samples. These observations suggest that MDA-BF-I may have a role in osteoblastic progression of prostate cancer in bone.

VII. Secretion of MDA-BF-I protein _' into the conditioned media (CM) of prostate cancer cell lines correlates witht the osteoblast-stimulating activity of these cell, lines _, , , ,_ι i

¹ ' ""Embodiments of the invention also examine the ^■ expression of MDA-BF-I protein in several established prostate cancer cell lines. LNCaP, PC-3 and DU145 are known established human prostate cancer cell lines isolated from lymph node, bone, and brain metastasis, respectively. MDA PCa 2b is a new prostate cancer cell line established by Dr. Nora Navone from,, a prostate cancer patient with bone metastasis. MDA PCa 2b cells, like LNCaP cells, secrete prostate specific antigen (PSA) , which confirms their prostate origin. TABLE 4 summarizes the osteoblastic or osteolytic properties of the various prostate cancer cell lines and demonstrates the results of in vivo animal models of osseous prostate cancer. It is known that intra-bone injection of MDA PCa 2b and LNCaP cells produces moderate and weak osteoblastic responses, respectively, in addition to osteolytic lesion. On the other hand, intra- bone injection of PC-3 and DU145 cells produces only osteolytic reactions.

TABLE 4: Comparison of Phenotypes of different prostate cancer cell lines when injected into cells

Also, conditioned media (CM) were prepared from these human prostate cancer cell lines and used for immunoprecipitation (IP) /Western analysis. Immunoprecipitations (IP) with Ab-8 antibody and Western blotting with Ab-10 antibody were used to determine whether MDA-BF-I protein is expressed in various prostate cancer cell lines. One million cells were grown in 5 ml of serum-free medium for 2 days.

FIGURE 24 shows the detection of MDA-BF-I secreted into the conditioned media of each prostate cancer cell line. Cell lines that have osteoblastic phenotypes, e.g., MDA PCa 2b and LNCaP, express and secrete MDA-BF-I protein into the medium. For example, MDA PCa 2b expresses the most MDA-BF-I and LNCaP express MDA-BF-I but to a lesser level .

In contrast, also shown in FIGURE 24, cell lines that have osteolytic phenotypes, e.g., PC-3 and DU145, express and secrete very low levels of MDA-BF-I under these experimental conditions. On the other hand, all four prostate cancer cell lines tested express similar amounts of pl80-ErbB3 (data not shown) . Thus, the levels of MDA-BF-I protein in the conditioned media correlate with the osteoblast-stimulating activity of these cell lines.

Because MDA-BF-I mRNA is generated from intron retention in the ErbB3 gene, PCR primers containing MDA- BF-1-specific sequences were used to detect MDA-BF-I mRNA. The forward primer sequence was GGATCCGGACTTGGC

TGGGCTCCCTTCACC (SEQ . ID .NO .20 ) ; the reverse primer sequence was GCGGCCGCTTAATGATGATGATGATGATGATQ CCCACCTTGGGACATAGTCCCCC (SEQ. ID.NO.21) . The, reverse primer included. a Notl site (underlined) and a sequence encoding s,even histidine residues (italics) for subcloning of MDA-BF-I into a baculoviral expression system for purification from Sf9 cells. The expected RT- PCR product is 1036 bp and corresponds to full length MDA-BF-I. MDA-BF-I transcripts were detected in MDA PCa 2b and LNCaP cells, which are derived from PCa patients with bone an lymph node metastasis, respectively, as shown in FIGURE 45. This expression correlates with the Western blots described above.

TABLE.7. provides a list of various cells lines or xβnqgrafts derived from prostate cancer metastases or primary prostate cancer specimens. These osteoblastic and osteolytic features of these cells lines, where known, are indicated. These cell lines may all be useful in further exploration of the effects of MDA-BF-I, or other MDA-BF factors and prostate cancer factors affecting bone growth.

TABLE 7: Cell lines or xenografts derived from prostate cancer metastases or primary prostate cancer specimens

VIII. Recombinant MDA-BF-I protein expression and recombinant MDA-BF-I protein do not affect the growth of prostate cancer cells in vitro and in vivo Because osteolytic PC-3 cells express very low levels of MDA-BF-I, expression constructs can be used to generate recombinant MDA-BF-I protein in these cells. However, other cell lines can also be used. Host cells transformed with nucleic acid sequences encoding the bone metastasis factors of the invention may be cultured under conditions suitable for expression and recovery of the recombinant bone metastasis factors from cell cultures. The recombinant the bone metastasis factp,rs_;of the invention produced may be secreted or contained intracellularly. They may be expressed as soluble compounds or agents, or as insoluble aggregates or inclusion bodies. For example, expression vectors containing polynucleotides that encode the bone metastasis factors of the invention may be designed to contain signal sequences which help to direct secretion of the bone metastasis factors through a prokaryotic or, preferably, eukaryotic cell membrane and into extracellular environments or culture media. As another example, a host cell line may be chosen for its ability to modulate the expression of the inserted sequences or to process the expressed proteins or peptides in the desired fashion. Such modifications of the polypeptide include, but are not limited to, acetylation, carboxylation, glycosylation, phosphorylation, lipidation, and acylation. Post-translational processing which cleaves a "prepro" form of the protein may also be used to facilitate correct insertion, folding and/or function. Different host cells such as CHO, HeLa, MDCK, HEK293, and WI38, which have specific cellular machinery and characteristic mechanisms for such post-translational activities, may be chosen to ensure the correct modification and processing of the foreign protein. An expression vector may contain necessary elements for transcription and/or translation of the inserted coding sequences. Expression vectors and systems known in the art may be employed for producing full length or only portions of the polypeptides of the bone metastasis factors of the invention. For long-term, high-yield production of recombinant proteins, stable expression of the DNA construct of the bone metastasis factors is preferred. Recombinant constructions known in the art may 61 ^{■ •} , be used to join all or portions of the nucleotide coding sequences for the bone metastasis factors. The; polypeptide domains can be used to facilitate the purification of the bone metastasis factors of the invention. Such purification-facilitating domains include, but are not limited to, metal chelating peptides such as histidine-tryptophan modules that allow purification on immobilized metals, protein A domains that allow purification on immobilized immunoglobulin, and the domain utilized in the FLAGS extension/affinity purification system (Immunex Corp., Seattle, Wash.). In addition, it may be useful to include cleavable linker sequences between the coding sequences and the purification facilitating domains, such as those specific for Factor XA or enterokinase (Invitrogen, San Diego, Calif.) to facilitate purification and separate the purification-facilitating domains after purification.

As an example, an expression plasmj.4_/ ,pRSl>T-MDA-BF-l containing a G418 selection marker, i,s> construςted and used to transfect , PC-3 cells. pRSN contains,^ RSV promoter for gene expression and is used herein for making stably transfected prostate cancer cell lines. MDA-BF-I is FLAG-tagged to allow monitoring the level of MDA-BF-I , expression from the expression construct. G418- resistant MDA-BF-I transfectants (PC/MDA-BF-1) were generated and individual clones of G418-resistant MDA-BF- 1 transfectant are being selected. PC-3 cells , transfected with pRSN vector (PC/neo) was used as a control. Figures 25-27 demonstrates the effect of MDA-BF-I on cell growth and proliferation after transfecting the expression plasmid into prostate cancer cell lines having osteolytic phenotypes and expressing low level of MDA-BF-I, e.g., PC-3. Methods well known to those skilled in the art may be used to construct cloning vectors containing appropriate transcriptional and translational control elements and DNA sequences . Exemplary techniques are described in Sambrook, J. et al. (1989) Molecular

Cloning, A Laboratory Manual, Cold Spring Harbor Press, Plainview, N.Y., Ausubel, F. M. et al . (1989) Current Protocols in Molecular Biology, John Wiley & Sons, New York, N. Y., and Green, E. et al . (1997) Genome Analysis, A Laboratory Manual, Cold Spring Harbor Press, Plainview, N. Y.

FIGURE 25 demonstrates the detection of, secreted MDA-BF-I protein. in the conditioned medium of the, transfected PC/MDA-BF-1 cells using Western blot analysis, as compared to negative control of PC-3 cells transfected with vector (PC/neo) . Conditioned media from PC/Neo and PC/MDA-BF-1 cells were immunoprecipitated with anti-FLAG- agarose and Western blotted with Ab-10.

As shown in FIGURE 25, recombinant MDA-BF-I is secreted into the conditioned medium of PC/MDA-BF-1 transfectants . The results confirm the expression and secretion of a recombinant MDA-BF-I in osteolytic prostate cancer,,cell lines, such as PC-3,, which generally does not secrete MDA-BF-I protein into medium under the experimental conditions described herein (FIGURE 24) .

The- levels of MDA-BF-I expression in stable PC/MDA- BF-1 transfectants are further screened in order to demonstrate any dose-dependent stimulation of osteoblastic response by MDA-BF-I. At least three types of independent PC-3 clones secreting the highest (PC/MDA-

BF-1 (Hi)), medium (PC/MDA-BF-1 (Med) ), and lowest (PC/MDA- BF-1 (Lo) ) levels of MDA-BF-I can be used for further study. PC-3 cells transfected with pRSN vector alone (PC/neo) is used as a negative control.

FIGURE 26 demonstrated the measurement of the effect of MDA-BF-I on PC-3 cell growth/proliferation' in vitro. ^• Cells (2x 10⁵) were plated onto 6-well plates and the cell numbers were counted every 24 hrs . As shown in FIGURE 26, there was no significant difference in the growth rate between the PC/neo and PC/MDA-BF-1 cells.

FIGURE 27 demonstrates the result of an in vivo animal model used to test the effect of MDA-BF-I. For example, to test the effect of MDA-BF-I on PC-3 cell growth in vivo, PC/neo and PC/MDA-BF-1 were injected subcutaneously into the flanks of nu/nu,nude mice at,,l x 10⁶ cells/site. _.Twelve sites were injected ,,for each cell line and tumors were measured weekly after, injection and, monitored over four weeks. Average tumor sizes (in mm³) ± S. E. from each cell line are shown. As shown in FIGURE 27, expression of MDA-BF-I does not have significant effect on the growth of an osteolytic prostate cancer cell line, PC-3, in vivo in a subcutaneous tumor model.

Further, additional experiments have shown that different levels of MDA-BF-I expression in- PC-3 cells have different effects. Additional PC-3 cells were, . . transfected with pRSN-FLAG-MDA-BF-I . _.Several, independent PC-3 clones were, then generated by G418,_.selection,. The, . conditioned media from these clones. were examined for levels of MDA-BF-I by immunoprecipitation with .anti -FLAG- agarose. and .Western blotting with the Ab-10 antibody- Conditioned, media from unselected transfectants was mixed and used, a control. As show in FIGURE 46, different transfected PC-3 clones expressed various levels of MDA- BF-I as compared to unselected transfectants. Clone #16 expressed high levels. Clone #18 expressed intermediate levels. Clones #1, #10 and #10 expressed only low levels of MDA-BF-I. Clones #16, #18 and #19 were selected for further study.

Clone #16 was injected into mouse femurs to determine the effects of MDA-BF-I on osteoblast-prostate cancer cell interactions in vivo. Mice were killed after 4 weeks. Histological analysis of the lesion shows that there was significant new bone formation along the cortical bone in the area with tumor growths. In FIGURE 47, a fixed, decalcified, sectioned femur from a mouse that received clone #16 injections was stained with hematoxylin and eosin. Arrows indicated newly formed woven bone. T indicated tumor cells.

IX. MDA-BF-I plays a role in osteoblast proliferation, and differentiation

1. Preparation of Primary Mouse Osteoblasts and Measurement of Osteoblast-stimulating Activity

The invention provides a method of measuring osteoblast stimulating activity through a cell proliferation assay using various osteoblasts available, for example, primary mouse osteoblasts (PMO) prepared from newborn mice calvaria. Suitable cell proliferation assay includes a radio-labeled thymadine incorporation assay, counting of cell numbers during cell growth (cell count versus time), a 3- (4, 5-dimethylthiazol-2-yl) -2, 5- diphenyltetrazolium bromide (MTT) assay, and a BrdU incorporation assay, among others. In addition, the ability of the bone metastasis factors of the invention, e.g., MDA-BF-I, to stimulate cell proliferation should not be limited to osteoblast and can also be tested in other types of cells. Primary mouse osteoblasts have been shown to respond to bone-derived human prostate cancer cell line, MDA PCa 2b, in a co-culture system by stimulating proliferation and differentiation. The ability of ,MDA-BF^l to. stimulate osteoblast proliferation was determined by measuring their ability to stimulate DNA synthesis in primary osteoblasts.

To obtain primary osteoblasts, calvaria were removed from 2 -day-old to 5-day-old mice and digested three times with collagens P (0.1 mg/ml) at 37°C. To enrich for osteoblasts, the supernatants from the first and second digestions were discarded. The third sμpernatant , containing released osteoblasts, was collected, for further study. The cells were cultured and used at passages two or, three. To measure osteoblast-stimulating activity, osteoblasts were plated in 12 -well plates at 5000 ,gells/well. The cells were allowed to attach overnight and purified recombinant MDA-BF-I protein was added to the cells and incubated for about 40 hour. [³H]- thymadine (1 μCi/well, 25 Ci/mmol) was added to each well and incubated at 37°C for about 4 hour. The cells were washed with 5% trichloroacetic acid and solubilized with 0.25 N NaOH. The radioactivity in triplicate samples was measured by scintillation counting. 2. Effect of, MDA-BF-I on the Proliferation of Mouse Primary Osteoblast Cells

To prepare MDA-BF-I for functional studies, another expression construct for recombinant MDA-BF-I protein was prepared. A histidine-tagged MDA-BF-I recombinant protein was expressed in Sf9 insect cells using a baculoviral expression system. Seven histidines were added to the C- terminus of MDA-BF-I to facilitate protein purification-. The forward and reverse primer sequences- were ■ . : GGATCCGGACTTGGCTGGGCTCCCTTCACC (SEQ . ID.NO .20) and GCGGCCGCTTAATgArGArGATgArGArGArGCCCACCTTTGGGACATAGTCCCCC (SEQ. ID.NO.21) , respectively. The reverse primer includes a Not I site (underlined) and a sequence encoding seven histidine residues (italic) for purification of recombinant MDA-BF-I expressed in Sf9 cells. The RT-PCR product was 1036 bp, the size of the complete MDA-BF-I. Recombinant MDA-BF-I protein was purified from cell pellets by nickel affinity chromatography. FIGURE 28 demonstrates SDS-PAGE analysis of the pμrified_. recombinant MDA-BF-I protein, showed as a single band.with an apparent molecular weight of about 45 kDa. This purified recombinant protein was used to. determine the effects of MDA-BF-I on osteoblast proliferation. To test the effect of serum on primary mouse osteoblast (PMO) proliferation, primary mouse osteoblasts were cultured in Dulbecco's Modified Eagle's (DME) medium supplemented ^'with various concentrations (0:5%-10%) of fetal bovine serum (FBS) and incubated, with and without MDA-BF-I ,(50 ng/ml) . Cell proliferation was assessed by measuring . [³H] -thymadine incorporation. It was found /that MDA-BF-I increased [³H] -thymadine incorporation with all the serum concentrations tested. However, low serum concentrations gave lower basal ³H-thymadine incorporation. Thus, about 0.5% of serum, e.g., FBS, is used herein, however, other concentrations and other serum may also be used.

FIGURES 29. and 48-50 show that purified recombinant MDA-BF-I protein stimulates proliferation, of primary mouse, osteoblasts. Higher levels of ³H-thymadine incorporation and thus higher levels of proliferation are observed with increased concentrations of recombinant . _: MDA-BF-I from about 5 ng/ml to 100 ng/ml. However, as shown in FIGURE 30, purified recombinant MDA-BF-I protein can not stimulate mouse fibroblast proliferation, as demonstrated in NIH3T3 cell line. Both cell types responded to growth stimulation by PDGF, which was used as a positive control. Since MDA-BF-I stimulates osteoblast proliferation but not fibroblast proliferation, the results indicate that MDA-BF-I is an osteoblast- stimulating factor.

FIGURE 31 demonstrates the proposed role of MDA-BF-I in osteoblastic progression of prostate cancer cells in bone. Since MDA-BF-I is identified herein from _ι bone marrow supernatant of prostate cancer , patients, thus,, soluble MDA-BF-I protein is likely initially _; secreted from bone .metastatic prostate cancer cells to affe.ct,. ,, ■ osteoblast_, proliferation, resulting in increased, bone mass at the site of the lesion.

This likely mode of action is based on several observations. First, MDA-BF-I was isolated from the bone marrow supernatants of prostate cancer patients _, with osteoblastic bone metastasis and Western blot showed that MDA-BF-I is only present in prostate cancer patients with bone metastasis., - . .,

Second, immunohistochemistry studies showed that¹ MDA-BF-I,is ^' expressed and secreted by 'activated' osteoblasts^ in- bone metastasis. ' ^■ •- ' ' ^•

Third, recombinant MDA-BF-I is present in the conditioned media of prostate cancer cell lines that produce osteoblastic phenotype in bone, e.g. , MDA PCa 2b, but not in osteolytic cell lines, e.g., PC-3 and DU145 cells. _, . ^{■• ■ ■} . ^{•• '}

Fourth, recombinant MDA-BF-I does not affect prostate cancer cell growth in vitro and in vivo, suggesting that prostate cancer cells may not be the cell type it interacts with.

Fifth, recombinant MDA-BF-I induced osteoblast proliferation in vitro but not prostate cancer cell proliferation.

Sixth, expressing recombinant MDA-BF-I in PC-3 cells, which normally elicits a strong osteolytic response, . in vivo, generated an osteoblastic response in vivo in an osseous prostate cancer animal model . Together, these observations strongly suggest that MDA-BF-I participates in the osteoblastic metastasis of prostate cancer in bone. Thus, the invention examines the function of MDA-BF-I during osteoblast proliferation and differentiation. In one embodiment, increases in the expression of MDA-BF-I in PC-3 cells resulting in higher osteoblastic activity of PC-3 cells ,in bone are tested in vivo in an animal model . In another embodiment, gene silencing technique to knock down the expression of MDA- BF-I in MDA-BF-I expression cell lines, e.g., _; MDA PCa 2b, is performed to demonstrate any decreases in the osteoblast-stimulating activity of MDA-BF-I in bone.

3. Effect of MDA-BF-I on the Differentiation of Mouse Primary Osteoblast Cells

To test whether MDA-BF-I affects osteoblast ^; differentiation,¹ PMOs were incubated with or without '50 ng/ml MDA-BF-I in MEM plus 2% FBS for 4 days to allow cells- to reach confluence. The medium was¹ 'then changed to differentiation medium containing 100' μg/ml ascorbic acid, 5 mD -β-glycerol phosphate, and 2% FBS. The differentiation also contained or lacked 50 ng/ml of MDA- BF-I.^' The medium was changed every 2 days. Fixation and staining with von Kossa stain showed that MDA-BF-I increased the mineralization of PMOs, as shown in FIGURES 51 and 52.

X. Elucidate the effects of MDA-BF-I on osteoblast- prostate cancer cell interactions in vivo in an osseous prostate cancer animal model.

1. Expression of MDA-BF-I in PC-3 Cells Stimulate New Bone Formation

Embodiments of the invention determine the effect of expressing MDA-BF-I in PC-3 cells on bone-prostate cancer interactions in vivo by injecting transfected PC/MDA-BF-1 cells into the mouse femur. PC-3 cells were transfected with pRSN (PC/neo) or pRSN-MDA-BF-1 (PC/MDA-BF-1) and selected with G418. Male severe combined immunodeficiency- disease (SCID) mice from Harlan (Indianapolis, IN) were anesthetized. Cells (10⁵) in 3 μl of medium were injected into the right femur of each SCID mouse . The same volume of medium was injected into the left femur of each mouse as a control . Radiographs of the bones were made biweekly. FIGURE 32 showed X-ray images of _r femurs injected with control PC/neo or PC/MDA-BF-1 cells six weeks after injection. Left; femurs were injected with medium, and used as controls. At 6-weeks, severe osteolytic effects (upper arrow) were seen in femurs injected with control PC/neo cells, which were manifested as radiolucent areas in the X-ray of bone. At 6 weeks post-injection, severe bone lysis was observed in all five mice injected with PC/neo cells.

In contrast, de novo bone formation, which was shown as radio-opaque areas throughout the femur, was observed in PC/MDA-BF-1 injected bone. In FIGURE 32,. increased bone density (lower arrow) is seen in femurs injected with PC/MDA-BF-1 cells. Of the five mice injected with PC/MDA-BF-1, significant increase in bone mass was observed in the femurs of three mice, suggesting that expression of MDA-BF-I in PC-3 cells stimulates new bone formation.

FIGURE 33 showed the result of histological examination of lesions induced by intra-femoral injection of PC/MDA-BF-1 cells into SCID mice. There is an increase in the number of bone trabeculae around the tumor (T, tumor cells) , a phenotype similar to those observed in prostate cancer patients with bone metastasis. These results indicate that MDA-BF-I is able to stimulate osteoblast proliferation or function. Similar effects were not seen using the control PC/neo cells , Histological analysis showed that the PC/neo control induced a severe osteolytic response in the bone with increased numbers of osteoclasts (data not shown) . In contrast. PC/MDA-BF-1 cells did not induce osteolysis and produced tumors surrounded by woven bone as shown in FIGURES 53 and 54. The phenotype is similar to that observed in prostate cancer patients with bone metastasis. Moreover, activated osteoblasts were seen in the woven bone area as seen in FIGURE 55. Immunohistochemical analysis of the MDA-BF-I tumor lesions shown in FIGURE 56 confirmed the expression of MDA-BF-I in tumors produced by intrabone injection of PC/MDA-BF-1 but not PC/neo control cells.

The in vivo new bone growth as demonstrated herein provides evidence that MDA-BF-I may be used to treat various bone-related diseases, such as bone metastasis in various cancers and osteoporosis, among others. MDA-BF-I can be used especially to those diseases with osteolytic lesions and/or bone loss in order to stimulate osteoblast proliferation and new bone growth.

2. Effect of MDA-BF-I on Osteoblast/Prostate Cancer Cell Interactions in vivo Animal models are critical for understanding the interactions between prostate cancer cells and the bone microenvironment in vivo. The invention provides the study on osteoblastic progression of prostate cancer cells in bone. Intra-tibia or intra-femur injection of human prostate cancer cells and the SCID-hu model of metastasis are in' vivo mouse models that have been ^l developed for studying prostate cancer progression in bone. For example, the results using the intra-bone injection method are shown in FIGURE 32, thus demonstrating that this method produces reliable results.

Cell lines as listed in TABLE 5 are designed ,to, be injected into mouse femurs to determine the effe,ct of MDA-BF-I on osteoblast/prostate cancer . cel,l , interactions in vivo., Radiographs of the injected bones , are. taken biweekly and,- immediately before the mice are._, sacrificed. Blood from the mice is obtained from a small incision in the main tail vein and collected biweekly. Urine samples are collected immediately before the mice are sacrificed. Mice are killed at various time points and their femurs (both the injected bone and its non-injected counterpart) are separated from the bodies and dissected from surrounding tissue.

TABLE 5: Co-culture and intra-bone studies

Phenotypic characterization of mice includes (1) radiographic analysis of radiodensity of long bones; (2) histological examination of the injected limbs for tumor growth and new, bone formation; (3) histomorphometric analysis, which will be described below, for evidence of increased bone mass and osteoblasts; (4) measurement of serum markers for osteoblasts and tumor cells _v aV Evaluation of tumor area After radiological (Faxitron) analysis of the skeletons, bones are dissected out and standard histological analyses of the injected and control femurs are performed. Briefly, the dissected bones are fixed in 4% formaldehyde for 6 hour at 4°C. Undecaϊcified bones are embedded in methylmethacrylate, and sections with about 5 μm longitudinal thickness (including^" at least the distal epiphysis, ήaetaphysis, and diaphysis) are' prepared' with a rotation -microtome (Jung, Heidelberg, Germany). Sections are stained with hematoxylin/eosin and evaluated under a Zeiss microscope. The area of each injected femur is measured by digitized image analysis with a microcomputer. The tumor area is outlined manually and measured with the image program. Four histological sequential but not consecutive sections are analyzed per femur, and the mean tumor area is calculated. Histology analyses of the samples are performed by an experienced pathologist . b. Histomorphometrical analysis

Histomorphometrical analyses are also performed on the injected and control limbs (four histological sequential but not consecutive sections are analyzed per femur) . To perform histomorphometrical- analysis, the bones are dissected out and fixed and sectioned as described. For quantification of osteoblast numbers, some of the sections are stained with 1% toluidine blue according to standard protocol . Both osteoblasts number and osteoblast surface (Ob. S/BS, %) are measured.

For quantification of osteoclast numbers, some sections are subjected to an enzymatic assay for ,, detection of tartrate resistant alkaline, phosphatase staining (TRAP) , an enzyme specifically expressed by osteoclasts in the bone marrow. Both ,OStZe₁QcIa₁St; number and osteoclast surface are measured._. For histomorphometrical analysis, sections, are treated by the von Kossa reagent that stains mineralized bone matrix. Three ..parameters are measured using the semi-automatic mode present in the Osteomeasure system: bone volume

(calculated as the total bone marrow volume occupied by trabecular bone (BV/TV, %) , trabecular number (evaluated as the number of trabeculae present in the field analyzed for the bone volume) , and trabecular thickness (calculated as the averaged thickness of the trabequlae present in the field analyzed for the bone volume) . All bone-specific parameters are measured and expressed in units following the guidelines established by the ASBMR histomorphometry nomenclature committee. c. Measurement of serum markers

For MDA PCa 2b cells, the levels of prostate serum antigen (PSA) in blood are measured to monitor the growth of human MDA PCa 2b in mouse bone. Serum is separated from the blood and PSA is measured by using a microparticle enzyme immunoassay (Imx PSA assay, Abbott Laboratories) . Serum osteocalcin concentration and alkaline phosphatase activity, which reflect hone formation at the systemic level, are measured with standard radioimmunoassays (Sigma) and colorimetric assays (Sigma) , respectively. Mouse serum tartrate- resistant acid phosphatase activity is measured using mouse TRAP assay (SBA Sciences, Turku, Finland) . d. Sample size and statistical analysis

The incidence of tumor growth in bone was reported to be 100% and 80% for PC-3 and MDA PCa 2b, respectively. The results indicated that about 60% of mice injected with PC/MDA-BF-1 cells showed new bone growth, as represented by the increase in bone mass under radiographs. It is calculated that about eight mice are required to detect a significant difference between the test and control groups using a two sample t-test with 80% power at a two-sided significance level. Thus, to obtain at least eight animals for analysis for each time point, ten mice need to be sacrificed at each time point.

For each cell line listed in FIGURE 33, the numbers of mice needed and the schedule for monitoring bone growth,_. are calculated. For example, about thirty animals for each cell line are needed for ten mice to be sacrificed at four, six, and eight-weeks for those prostate cancer cell lines with osteolytic phenotypes, such as PC-3 cells. Also for osteoblastic prostate cancer cells, such as MDA PCa 2b cells, which have a slower growth rate in bone as compared to osteolytic cell lines (e.g., PC-3), about thirty animals for each cell line are needed for ten mice to be sacrificed at eight, twelve, and sixteen-weeks. Our statistical analyses focus on developing models to accurately quantitate the association of prostate cancer progression and bone response with factors/biological markers, such as PSA/ tumor size, bone volume, trabecular number, trabecular thickness, osteoblast and osteoclast number. A statistical model is selected for use in analyzing the results. To evaluate the biological significance of these markers, based on changes in their presence or absence of quantitative levels, statistical regression modelings are needed to formulate and test hypotheses. For binary (osteoblastic or osteolytic, control or treatment, etc.) and ordinal outcomes (three or more ordered values) , nonparametric methods such as recursive partitioning trees or multivariate adaptive regression splines are used as exploratory tools. Logistic regression is also considered, with appropriate goodness-of-fit tests. For continuous outcomes (tumor size, bone parameters) , analysis of variance models and other non-parametric analogues are applied first. In addition, the general linear model and, when appropriate, nonlinear regression models are then applied. For continuous outcomes measured repeatedly (time course experiments) , all data, including covariate effects, are fitted using mixed models. For most of these models, diagnostics based on the examination of residuals are available. If appropriate parametric models cannot be constructed or non-parametric techniques ascertained for an analysis, bootstrap and permutation techniques are used for estimation testing. XI. Regulation of MDA-BF-I expression in bone- derived prostate cancer cell lines

1. siRNAs ^•• • To further confirm the role of MDA-BF-I in the osteoblastic response, down-regulation of MDA-BF-I expression can be performed in MDA-BF-I expressing cells, using small interference RNAs (siRNAs) analyses, for example, in prostate cancer cell lines with osteoblastic phenotypes, such as MDA PCa 2b, among others. siRNAs, either chemically synthesized or expressed intracellularly via a polymerase III -based transcription system, are used to target genes in cell cultures. Selection of specific sequences to be targeted for down- regulation is critical for using the siRNA approach.

TABLE 5 demonstrates various expression constructs containing coding sequences or siRNA for MDA-BF-I that can be prepared in MDA PCa 2b cells for in vivo animal studies to measure bone growth of each construct. Several chemically synthesized nucleotide siRNA oligonucleotides are used and demonstrate transient downregulation of MDA- BF-I expression in MDA PCa 2b cells. Among them two sequences are identified to have significant effects and are suitable targets, i.e., siRNAl : aacgacgctctgcaggtgctgdTdT (SEQ. ID. NO.18) and SiRNA2 : aactctcaggcagtgtgtcctdTdT (SEQ . ID .NO .19) .

Both pSilencer 2.1-MDA-BF-l siRNA and pSilencer-2.1- U6neo plasmid (as a negative control) are used to transfect MDA-BF-I expressing prostate cancer cells, e.g., MDA PCa 2b cells to suppress/regulate MDA-BF-I expression. Suitable sequences of siRNA include, but are not limited_. to, siRNA 1 and siRNA2, as ϋs,ted above.

For long term (months) silencing of the MDA-BF-I gene, additional sequences which generate MDA-BF-I- specific hairpin¹ siRNA oligonucleotides are designed, synthesized, and cloned into the vector pSilencer 2.1- U6neo according to the manufacturer's protocol (Ambion, Inc.) to generate pSilencer 2.1-MDA-BF-l siRNA.

For selecting stable transfectants, cells are selected with G418, and resistant clones are cultured (MDA PCa2b/siRNA-BF-l) . Down-regulation of MDA-BF-I ' expression in each G418 resistant clone is tested by- Western blot analysis of the conditioned media from each clone. Clones that show significant decreases in MDA-BF-I levels are selected for further studies. It has been shown that this strategy suppresses long term target protein expression, such as for more than 2, months_,.

Because MDA-BF-I siRNA may also down÷,regulate pl80- ErbB3 pr other _;.genes, the specificity, of inhibition by siRNA is also tested. As shown in TABLE 5, the invention also provides siRNA-treated cells reconstituted with a siRNA-resistant MDA-BF-I constructs (rMDA-BF-1) . This kind of construct introduces silent mutations in the DNA sequences of MDA-BF-I by changing 4-5 nucleotides and ,. generate expressing plasmids, rMDA-BF-1, that are resistant to siRNA inhibition, no longer a. target of MDA- BF-I siRNA. , ..., _|(| ,, .,< _.

To facilitate the transfection of rMDA-BF.-l plasmid constrμct, .a lentiviral vector (Invitrpgen) inserted, with rMDA-BF-1 is constructed and used to infect, MDA-, PCa2b/siRNA-BF-l cells to generate MDA PCa2b/siRNA/r-MDA- BF-I cells, as shown in TABLE 5. Then, MDA PCa 2b, MDA PCa2b/siRNA control, MDA PCa2b/siRNA-BF-l, and MDA PCa2b/siRNA-BF-l/rMDA-BF-l are injected into mouse femurs to be tested in vivo. Stable transfectants expressing MDA-BF-I siRNA by transfection of pSilencer 2.1-MDA-BF-l siRNA are generated..Alternatively, to increase the efficiency ,in the delivery of siRNA into MDA PCa2b cells, retroviral vectors containing MDA-BF-1-specific siRNA can also be generated. The use of retroviral vectors for efficient transfection and long-term expression of siRNA was recently reported and these vectors are commercially available (IMGENEX Corp, San Diego) .

The expected growth responses for these in vivo studies are summarized in TABLE 5. Similar to the results of PC-3 cells in FIGURE 33, increased in number and thickness of trabeculae bone in femurs injected with PC/MDA-BF-1 cells are compared with controls. The results may suggest that downregulation of MDA-BF-I expression in MDA PCa 2b cells may decrease their ability to cause osteoblastic growth in bone.

The over-expression of MDA-BF-I in PC-3 cells may have several possible explanations. First, only the osteoblast function is affected by increased expression of MDA-BF-I- This would suggest that MDA-BF-I affects only osteoblasts. Second, both osteoblastic and osteoclastic activities are affected by over-expression of MDA-BF-I. This would suggest that MDA-BF-I also affects directly or indirectly the osteoclast function. In the latter case, MDA-BF-I may induce osteoblasts to secrete other factors and these factors in turn influence the osteoclasts. For example, both PDGF and TGF-β have been shown to have an effect on osteoclast differentiation through activation of IL-6 production in osteoblasts. Third, only osteoclastic activity is affected by MDA-BF-I. However, this possibility is unlikely because MDA-BF-I has an effect on osteoblast proliferation in vitro (FIGURE 29) and in vivo (FIGURE 32) . Fourth, there is no significant difference between MDA-BF-I-modified and control cells. This may occur in mice injected with PC/MDA-BF-I (Lo) clone, as the effect of MDA-BF-I on osteoblasts may be dose-dependent..

Long-term (several months) silencing of the MDA-BF-I gene has been achieved by cloning MDA-BF-I-specific hairpin siRNA oligonucleotides into the vector, pSuppressor Retro, according to the manufacturer's protocol (Imgenex) , to generate pSuppressorRetro-MDA-BF-.l shRNA. MDA PCa 2b cells were stably transfected with either pSuppressorRetro-MDA-BF-1 shRNA or pSuppressorRetro as a control. Cells were selected with G418 and_. resistant clones, PCa2/shRNA-BF-l were cultured. Three sequences . in MDA-BF-I were targeted for siRNA. mediated gene silencing. These regions, shown in FIGURE 57, include two sequences from the N-terminus, which are common between MDA-BF-I and pl80-ErbB3, and one sequence from the C-terminus, which is the intron 8 sequence unique to MDA-BF-I.

Downregulation of MDA-BF-I in the . G4.18-resistant clones was assessed by immunoprecipitatipn followed with Western blotting of the conditioned media from these, clones,. As show in FIGURE 58, all three shRNAs inhibited MDA-BF-I expression, with the third shRNA directed, to intron 8 exhibiting the most significant inhibition. Consistent with intron 8 being unique to MDA-BF-I, expression of pl80-ErbB3 was not affected by the. third shRNA. However, the second shRNA decreased pl80-ErbB3 levels significantly.

The animal models provided herein may mimic the. events in human prostate cancer since. they allow analyzing the ability of MDA-BF-I to modulate the osteoblastic response of prostate cancer metastasis , in vivo over_%a longer period of time as compared to . convention methods. The results described herein strongly indicate that MDA-BF-I is a novel osteoblastic factor. The fact that soluble MDA-BF-I protein is purified from disease relevant clinical samples and in vivo animal model results exhibit phenotypes remarkably similar to those observed in patient specimen provides novel insights into the mechanism of prostate cancer bone metastasis. Thus, MDA-BF-I can be used as a biomarker for various pathological relevant events, e.g., osteoblast proliferation, bone metastasis, and prostate cancer progression, among others.

2. peptides from phage display library

2.1 The Use of MDA-BF-I as a Therapeutic Target of Prostate Cancer Bone Metastasis ^• ■ ■ ^•

Since the expression of MDA-BF-I in prostate cancer cells stimulates new bone formation, it is proposed that blocking the interactions between MDA-BF-I and its receptor on osteoblasts inhibits its function on osteoblasts. Toward this goal, several approaches are used to develop reagents for blocking the in vivo functions of MDA-BF-I. For example, anti-MDA-BF-1 monoclonal antibodies are generated to .inhibit the interactions between MDA-BF-I and osteoblasts. As another example, a method of screening phage display libraries is used to identify peptides that inhibit the interactions between MDA-BF-I and osteoblasts. These results can be used to design small molecule inhibitors to inhibit the effect of MDA-BF-I on osteoblasts.

2.2 Identification of MDA-BF-I Binding Peptides by Phage Display Inhibition of the interaction of MDA-BF-I with osteoblasts is one of the strategies to ^ι treat' the^* increased bone mass observed in prostate cancer patients with bone metastasis. To select for peptides that can inhibit the interactions between MDA-BF-I and its corresponding receptor on the osteoblasts, a phage display peptide library was screened using recombinant full-length MDA-BF-I. Methods known in..the art can be used to generate phage display library and screen fro peptides that interact with the identified bone factors described herein.

A phage display random peptide library displaying the insert CX₇C (where X is any amino acid and C is a^' cysteine residue) was used. Purified recombinant MDA-BF-I protein at about 1 μg stored in about 50 μl phosphate buffered saline were immobilized on microtiter wells overnight at 4°C. Phage binding assays on purified protein were carried out using methods known in the art.

At least four rounds of screening were performed and selected phages ^'were sequenced. The enrichment of phage on MDA-BF--1 ■ was monitored by counting the number of transduping .units recovered from the MDA-BF-1-coated wells versus the number recovered from wells coated with BSA as control protein.

FIGURE 43 demonstrates an observed pronounced ^: enrichment for phage binding to MDA-BF-I. The DNA inserts of 265 randomly chosen phage clones recovered from the¹ 4th round were sequenced. Seventeen phages- recovered with high frequency during the screening were,obtained and the peptide sequences of the displayed epitopes are listed in TABLE 6.. These 17 phages were further characterized for their specificity. All of the 17 phages bind , to MDA-BF-I specifically. Among the 17 phages, the, binding specificity of 7 phages (phage #l-#7) to MDA-BF-I is shown in FIGURE 44. TABLE 6: Peptide identified from screening a phage display library

ASGADGP (SEQ. ID.NO.1)

FGWPLW (SEQ. ID.NO.2)

GGLALQE (SEQ. ID.NO.3)

LKRGITV (SEQ. ID.NO.4)

FASSFVL (SEQ. ID.NO.5)

TLDFPRR (SEQ. ID.NO.6)

ISFPRRW (SEQ. ID.NO.7)

WAGGRF (SEQ. ID.NO.8)

VAGGSFI (SEQ. ID.NO.9)

QGGVRHH (SEQ. ID.NO.10)

GGVRVLD (SEQ.ID.NO.il)

FASRVRS (SEQ. ID.NO.12)

QSRVRVA (SEQ. ID.NO.13)

PAGRYTD (SEQ. ID.NO.14)

GRYTTDR (SEQ. ID.NO.15)

SGYVAKM (SEQ. ID.NO.16)

SGYAKVS (SEQ. ID.NO.17)

Epitope-mimic peptides identified from phage displayed peptide libraries can then be synthesized. These peptides can be tested for their ability to inhibit the binding of ¹²⁵I-MDA-BF-I to osteoblasts and MDA-BF-I induced osteoblast proliferation as described above.

Thus, embodiments of the invention provides monoclonal antibodies and peptide sequences derived from phage display panning to inhibit the binding of MDA-BF-I to its receptor and block the biological activity of MDA- BF-I. XII. Investigate the mechanism of MDA-BF-I-mediated osteoblast proliferation

1. p42/44 MAPK in Osteoblasts Is Activated by MDA- BF-I Because MDA-BF-I stimulates primary mouse osteoblasts (PMO) proliferation in vitro (FIGURE 29) and induces an increase in bone mass in mouse in vivo (FIGURE 32) , primary mouse osteoblasts are used to further study signal transduction pathways mediated by MDA-BF-I. Other osteoblasts or other cell types can also be used. The effects of MDA-BF-I on MAPK phosphorylation are tested ^■ first. Primary mouse osteoblasts in serum-free α-MEM were treated with or without 50 ng/ml MDA-BF-I₁ for various durations. Cell extracts (about 50 μg/lane) were analyzed by Western blot analysis with antibodies against phosphorylated p42/44 MAPK and total MAPK.

FIGURE 34 demonstrates treatment of primary, mouse, osteoblasts with MDA-BF-I stimulates p42/44MAPK, phosphorylation. Maximum stimulation was detected after ,5 minutes of treatment, and the response returned to a basal level at 20 min. The total p42/44MAPK protein, levels did not change significantly. . _., ■ . ■ .

2. MDA-BF-I Activates Akt Kinase^: dji' Osteoblasts

FIGURE 35 demonstrates that the treatment of primary mouse osteoblasts with MDA-BF-I also stimulates the phosphorylation of Akt kinase. An increase in phosphorylated Akt kinase was detected after ¹IO min of MDA-BF-I treatment and was sustained at least for 4 hour.

3. MDA-BF-I Promotes IκBα Degradation in PMO

NF-KB regulates the expression of genes involved in apoptosis. The activity of NF-KB is tightly controlled by inhibitory IKB proteins that bind to NF-κB complexes and sequester NF-κB in the cytoplasm. Cytokines or stress factors promote the serine phosphorylation of IκBα and facilitate its polyubiquitination and proteosome-mediated degradation. Degradation of IκBα liberates bound NF-KB subunits and allows NF-κB translocation to the nucleus, where it activates a series of genes. To determine whether MDA-BF-I activates NF-KB, IκBα levels were measured at various times after MDA-BF-I addition by immunoblotting with anti-IκBα antibody.

FIGURE 36 demonstrates that MDA-BF-I transiently decreased cellular levels of IκBα, indicating that MDA- BF-I may stimulate NF-KB activation by promoting IκBα degradation. The kinetics of induction of IκBα degradation are consistent with the regulation of IκBα transcription by activated NF-KB. Because the NF-KB pathway protects cells against proapoptotic agents and also promotes cell growth, this result suggests that MDA- BF-I stimulates cell growth/survival through activation of NF-KB signal transduction pathway. Thus, MDA-BF-I can be used in a method to stimulate cell growth/survival by providing a subject with a therapeutic amount of MDA-BF-I or derivatives or homologs thereof.

4. MDA-BF-I activates the osteoblast-specific transcription factor Runx2

Runx2 is an osteoblast-specific transcription factor and a regulator of osteoblast differentiation. Runx2 is essential for bone formation, because homozygous Runx2 - /- mice show a complete lack of functional osteoblasts. Previous studies had showed that conditioned medium from MDA PCa 2b cells induced osteoblast differentiation through a Runx2 -dependent pathway. Because MDA PCa 2b cells express MDA-BF-I, MDA-BF-I was tested to see whether it stimulates Runx2 expression. Primary mouse osteoblasts were treated with MDA-BF-I for various duration of time, RNA was prepared, and the expression of Runx2 was detected by RT-PCR.

FIGURE 37 demonstrates that treatment of primary mouse osteoblasts with MDA-BF-I increases the expression of Runx2 in a time-dependent manner. This result suggests that MDA-BF-I induces osteoblast differentiation through Runx2. Thus, MDA-BF-I can be used in a method to induce osteoblast differentiation by providing a subject with a therapeutic amount of MDA-BF-I or derivatives or homologs thereof . 5 ■ Transfection efficiency of primary mouse osteoblasts

The invention also provides the studies of the effect of MDA-BF-I on the activities of several promoter activities by transfecting various promoter constructs into primary mouse osteoblasts. First, the transfection efficiency of primary mouse osteoblasts' was tested. Primary mouse osteoblasts were transfected with a reporter plasmid containing CMV driven green fluorescence protein (pEGFP, Clontech, CA) . Four commercially available transfection reagents, i.e. lipofectamine, calciμm phosphate, transfectam, and lipofectamine 2000, were tested.

FIGURE 38 demonstrates that primary mouse osteoblasts can be transfected with plasmid having green fluorescence protein (GFP) as indicator for expression level and, among the transfection reagents tested, lipofectamine 2000 results in the best transfection efficiency with about 30% of primary mouse osteoblasts positively expressing with GFP.

6. Effects of MDA-BF-I on. primary mouse osteoblast proliferation Our results showed that MDA-BF-I stimulated [³IJ] - thymadine incorporation in primary mouse osteoblasts. Other approaches are also used to confirm this observation,. Primary mouse osteoblasts are cultured in DME medium supplemented with 0.5% fetal bovine serum. Cells are treated with or without MDA-BP-I and their proliferation are assessed by (1) counting the cells; (2) using 3- (4, 5-dimethylthiazol-2-yl) -2, 5- diphenyltetrazolium bromide (MTT) dye to measure metabolic rates; and (3) assessing BrdU incorporation to measure DNA replication rates. The osteoblasts are expected to grow significantly faster in the presence of recombinant MDA-BF-I.

7. Signal transduction pathway (s) that mediates MDA- BF₁-I' s effect on osteoblast proliferation The invention illustrates that MDA-BF-I promotes osteoblastic responses in bone. In addition, understanding the signal transduction pathways mediated by MDA-BF-I helps to delineate the mechanism by which MDA-BF-I mediates osteoblast proliferation/differentiation and how MDA-BF-I acts for developing therapeutic strategies targeting MDA-BF-I.

Inhibition of specific signaling pathways mediated by growth factors is an effective therapeutic approach for cancer treatment. As an example, STI571, an inhibitor which specifically inhibits the tyrosine kinase activities of bcr-abl, PDGF receptor, and c-kit, has been shown to be effective in the treatment of cancers involving these kinases. In addition, elucidation of the signaling pathways helps to identify structure and function of the receptor for MDA-BF-I mediated responses.

FIGURE 39 demonstrates the proposed activation of various signal transduction pathways by MDA-BF-I. As described above, MDA-BF-I stimulates [³H] -thymadine incorporation into osteoblasts, and activates the MAPK, Akt, Ikb/NF-kB, and Runx2. Thus, it is hypothesized that MDA-BF-I promotes a sustained osteoblast response during bone metastasis, by activating the MAPK, NF-kB, or Akt pathways to. modulate osteoblast . .. . proliferation/differentiation. The invention .further provides methods jto study (1) which signal, transduction pathway (s) mediates MDA-BF-I' s effect on osteoblast proliferation and (2) which pathway (s) mediates MDA-BF- l's effect on Runx2, an osteoblast specific transcription factor. Primary mouse osteoblasts are used as an example to study MDA-BF-I mediated signal transduction pathways because (1) MDA-BF-I stimulates primary mo.use osteoblast proliferation and Runx2 activation and (2) primary mouse osteoblasts are not immortalized and thus may accurately reflect the in vivo responses of osteoblasts.

₍ The MAPK pathway is an important pathway in. ,the , growth and transformation of many cell types. Most. mitogenic signaling pathways that are stimulated by receptor tyrosine kinase activation, such as the activation of the transmembrane proteins ErbB2/ErbB3, converge on the MAPK cascade. An important osteoblast- stimulating factor, endothelin-1, which stimulates , • osteoblast activity by binding to, cell sμrface_^G protein- coupled endothelin receptors (ET_A) / alsp. transduces signals through a pathway that converges on the MAPK cascade. Importantly, activation of MAPK has been shown to activate osteoblast-specific gene Runx2. Our observation that MDA-BF-I stimulates Runx2 expression prompted us to hypothesize that MDA-BF-I may activate the MAPK pathway to affect osteoblast growth and differentiation.

The invention also analyzes the effect of MDA-BF-I on the Akt and NF-KB pathways. The serine-threonine kinase Akt is a downstream target of PI-3K. The PI-3K/Akt pathway produces cell-survival signals in response to several growth factors. Akt have been shown to promote cell survival by phosphorylating multiple targets, including the Bcl-2 family member BAD, i the apoptosis- inducing enzyme caspase-9, and the Forkhead transcription factor FKHRLl, which regulates Fas ligand gene expression. Recently, several studies showed that Akt can also activate NF-KB to promote cell survival and interferon α/β promotes cell survival by activating NF-KB through PI3-kinase and Akt. In addition, it was demonstrated that HER-2/neu blocks tumor necrosis factor- induced apoptosis via the Akt/NF-κB pathway. Thus, activation of Akt/NF-KB pathway is another major signaling pathway that affects cell growth and survival.

7.1 To determine whether MAPK activation is necessary for MDA-BF-1-mediated proliferation Inhibitors of MAPK signaling are used to show inhibition of MDA-BF-I mediated primary mouse osteoblast proliferation. For example, PD98059 and UO126 are such inhibitors that can be used to block MDA-BF-I' s effect on primary mouse osteoblast proliferation. PD98059 is a broad range inhibitor of MEK kinases

(i.e., it inhibits MEKs 1-7) and inhibits phosphorylation of MEK kinases by upstream kinases, e.g., Raf kinases (Raf-1, Raf-A, Raf-B) . U0126 is an inhibitor that binds MEKl and MEK2 to noncompetitively inhibit the phosphorylation of p42/44MAPK and has little effect on other kinases, such as MEKK and Raf. U0126's inactive analogue U0124 is used as a negative control for U0126.

Primary mouse osteoblasts are treated with (1) medium only; (2) MDA-BF-I (50 ng/ml) ; (3) MDA-BF-I plus 10 μM of U0126; (4) MDA-BF-I plus 10 μM of UQ124; (5) U0126 only; and (6) U0124 only, for about 24 hour in 0.5 % FBS. Cell proliferation is measured as described above. Similar experiments are performed with 50. μM of , PD98059.

7.2 To determine whether PI-3K/Akt activation- is necessary for MDA-BF-I mediated proliferation Inhibitors of PI-3K/Akt signaling are used to show inhibition of MDA-BF-I mediated primary• mouse osteoblast proliferation.^' For example, LY294002 and Wortmannin' are used to block MDA-BF-I' s effect on primary mouse osteoblast proliferation. ^■ - ,^{, ■ ■, • •} Primary mouse osteoblasts are treated with (1) medium oniy; (2) MDA-BF-I (50 ng/ml) ; (3) MDA-BF-I plus 10 μM of LY294002; (4) LY294002 only; (5) MDA-BF-I plus 10 μM of wortmannin; and (6) wortmannin only, for about 24 hour in the presence of 0.5% FBS and cell' proliferation is measured. Both inhibitors are expected to suppress the effect of MDA-BF-I on primary mouse osteoblast proliferation. _:

In addition, ^• dominant-negative' protein' kinases, e.g., Akt (dnAkt) or PTEN (a tumor suppressor that blocks the PI3K-Akt_, pathway) , are used to confirm whether, PI3-K/Akt pathway is required for MDA-BF-I' s effect on primary _; mouse osteoblast proliferation. Specifically, primary, SG mouse osteoblasts are transfected with a dnAkt expression vector construct or a PTEN expression vector construct by the liposome-mediated gene transfer technique. The transfectants with various expression constructs are treated with or without MDA-BF-I in culture, and their proliferation rates are measured. The results are also compared to a control vector construct .

7.3 To determine whether NF-kB activation is essential for MDA-BF-I-mediated proliferation

Inhibitors of NF-KB signaling are used to show inhibition of MDA-BF-I mediated primary mouse osteoblast proliferation. For example, PS-341 and curcumin are used to block MDA-BF-I 's effect on primary mouse osteoblast. PS-341 is a synthetic IKK/NF-KB inhibitor, a potent proteosome inhibitor that blocks degradation of IKB.

Curcumin is a natural NF-KB inhibitor, a dietary agent that suppresses activation of NF-KB.

Primary mouse osteoblasts are treated with or without PS-341 (1 μM) or curcumin (50 μM) for about 24 hours in 0.5% FBS. Cell proliferation is measured by [³H]- thymadine incorporation, cell number count, and/or MTT assays as described above.

In addition, "super-repressors" of NF-KB in which the phosphorylation sites have been removed to prevent degradation are used to determine whether the effect of MDA-BF-I on cell proliferation can be blocked. One example of the NF-κB super-repressor is a N-terminal deletion mutant of IKB-α designated IκB-α ΔN. Overexpression of IKB-α ΔN has been shown έo block NF~κB activity in several cell types. Primary mouse osteoblasts are infected with adenovirus expressing IκB-α ΔN or a control virus before treatment with MDA-BF-I and cell proliferation is measured. Alternatively, primary mouse osteoblasts are transfected with IκB-α ΔN expression vector or control vector by the liposome-mediated gene transfer technique. ^{' '■}

8. Signal transduction pathway that mediates MDA-BF- 1 ' s effect on Runx2

8.1; To determine whether MAPK activation is necessary for MDA-BF-I mediated up-regulation of Runx2 To determine whether MAPK activation is necessary for MDA-BF-I -mediated up-regulation of Runx2, inhibitors for MAPK signaling, e.g., PD98059 and U0126, are used to block MDA-BF-I' s effect on Runx2 expression. Also, U0126's inactive analogue U0124 is used as a negative control for U0126.

Primary mouse osteoblasts are treated with (1) medium only;' (2) MDA-BF-I (50 ng/ml) ; (3) MDA-BF-I plus 10 μM Of U0126; (4) MDA-BF-I plus 10 μM of U0124; (5) U0126 only; and (6) U0124 only, for about !&^■ 'hour in 0.5 % FBS. The RNA and protein levels of Runx2 in the primary mouse osteoblasts are determined by Northern and Western blot, respectively. The phosphorylation . status of Runx2 is then measured.

8.2 To determine whether PI-3K/Akt activation is necessary for MDA-BF-I mediated upregulation of Runx2

Inhibitors for PI-3K/Akt signaling, e.g. , LY294002 and wόrtmarinin, are used to block MDA-BF-I' s effect on Runx2 expression. ^• Primary mouse osteoblasts' are treated with^" (1> medium only; (2) MDA-BF-I (50 ng/ml) ; (3) MDA- BF-I plus 10 μM of LY294002; (4) LY294002 only; (5) MDA- BF-I plus 10 μM of wortmannin; and (6) wortmannin only, for about 16 hour in 0.5% FBS. The RNA and protein levels of Runx2 are determined by Northern and Western blot, respectively. The phosphorylation status of Runx2 is also measured. ^'

In addition, dominant-negative protein kinases, e.g., Akt (dnAkt) or PTEN, are used to further confirm whether PI3-K/Akt pathway is required for MDA-BF-I' s effect on Runx2 expression.

8.3 To determine whether NF-kB activation is essential for MDA-BF-I-mediated up-regulation of Runx2

Inhibitors of NF-KB signaling are used to show inhibition of MDA-BF-I mediated up-regulation of Runx2. For example, PS-341 and curcumin are μsed to block MDA- BF-I' s effect on Runx2 expression using primary mouse osteoblasts. Primary mouse osteoblasts are' treated with ' or without PS-341 (1 μM) or curcumin (50 μM) ^' for about 24 hour in 0.5% FBS. The RNA and protein levels of Runx2 are determined by Northern and Western blot, respectively. The phosphorylation status of Runx2 is also measured.

In addition, "super-repressors" of NF-KB, such as IKB-OC ΔN, are' used to determine whether"the effect of MDA-BF-I on Runx2 -expression can be blocked. Primary mouse osteoblasts are infected with adenovirus expressing IκB-α^'ΔN or control virus before treatment with MDA-BF-I and RNA and protein levels of Runx2 are determined by Northern and Western blot, respectively. The ' ^{' : •} phosphorylation status of Runx2 is then measured. Alternatively, primary mouse osteoblasts are^' transfected with ΪKB-OC ΔN expression vector or control vector by the liposome-mediated gene transfer technique. , 9. Effects of MDA-BF-I on Runx2 activation Our preliminary results showed that MDA-BF-I stimulated the transcription of Runx2 in primary mouse osteoblasts (FIGURE 37) . It was reported that MAPK pathways activate and phosphorylate the osteoblast- specific transcription factor Runx2. Phosphorylation of Runx2 by MAPK is required for FGF2 induction of osteocalcin (OCN) gene expression in a mouse osteoblast cell line. Thus, it is likely that MDA-BF-I stimulates the increase in the expression of the Runx2 as well as the phosphorylation of Runx2 in primary mouse osteoblast.

9.1 To determine whether MDA-BF-I stimulates Runx2 phosphorylation

Primary mouse osteoblasts are labeled with [³²P] - orthophosphate in phosphate- free medium with or without increasing concentrations of MDA-BF-I (10, 25, 50, and 100 ng/ml) ,with or without U0126 (10 um) for 6 hour. Whole-cell extracts will then be used for immunoprecipitation with antibody against mouse Runx2. MDA-BF-I is expected to stimulate Runx2 phosphorylation and that the stimulation is expected be inhibited by U0126.

§.2 To determine whether MDA-BF-I activates the transcriptional activity of Runx2

A well -described Runx2 target gene is^' osteocalcin (OCN) . Runx2 binds to osteoblast-specific element 2 (OSE2>, a cis-acting sequence in the promoter of the murine 0CN2 gene that is required for OCN expression in osteoblastic cells. Treatment of primary mouse osteoblast with MDA-BF-I is provided to determine whether (1) endogenous OCN expression; (2) OSE2-dependent OCN promoter activity, and (3) in vitro binding of Runx2 to the OSE2 sequence are increased. 9.3 To test the effect of MPA-BF-I on endogenous OCN expression

Primary mouse osteoblasts are seeded at a density of 50,000 cells/cm² in 100-mτn dishes and cultured in 10% FBS medium for about 24 hour. The cells (~4 x 10⁶) are then switched to 0.5% FBS with or without MDA-BF-I (50 ng/ml) for additional 12 hour. RNA is isolated with TriZol reagent according to the manufacturer's protocol (Invitrogen). PoIy(A)+ RNA is purified, and Northern blotting will be performed with mouse OCN cDNA.

Alternatively, quantitative RT-PCR using OCN specific primers is performed.

9.4 To study the effect of MDA-BF-I on OCN2 promoter activity , , .: , , , , . . Primary mouse osteoblast are transfected. with a^1' • plasmid containing a 1.3-kb OCN2 promoter- driving a ^• luciferase reporter gene by using Lipofectamine 2000 (Invitrogen) , which provides high transfeσtion efficiency (about ^■ 30%)^: in primary mouse osteoblast (FIGURE ' 3'8). Primary 'mouse osteoblasts transfected with the 1.3-kb

0CN2 reporter construct are treated with and without MDA- BF-I (50 ng/ml) in 0.5% FBS for about 16 hour. Cells are then harvested and assayed for luciferase activity,

9.5 To determine whether the MDA-BF-I acts on OCN2 promoter through Runx2 ;•■,. ,-.,•, ., ..■ ;

The effect 'of MDA-BF-I is determined on the -mutant OCN2 promoter 6OSE2mut/34-luc, which contains' a 2-bp mutation. in the^' Runx2 binding site. Primary mouse¹ osteoblasts are transfected with the plasmid 6OSE2/34- lue, which is a reporter plasmid containing six copies of OSE2 upstream of a minimal 34-bp OCN2 promoter, or with 6OSE2mut/34-luc. After transfection, cells are treated." with 50 ng/ml MDA-BP-I for about 16 hour in 0.5% F§S, harvested, and assayed for luciferase activity.

9.6, To determine whether MDA-BF-I increases in 'vitro binding of Runx2 to OSE2 ' Nuclear extracts are prepared from primary mouse osteoblasts treated with and without MDA-BF-I. An electrophoretic mobility shift assay (EMSA) is performed by incubating nuclear proteins with a ³²P-labeled oligonucleotide containing OSE2 sequence. Competition experiments with non-labeled oligonucleotide and supershift experiments with antibodies against Runx2 are performed to confirm the binding specificity. MDA-BF-I is thought to activate OCN expression through _t increased, phosphorylation of Runx2 and binding of Runx2 to the OSE2 in the OCN promoter.

10. Sample size and statistical analysis

Primary mouse osteoblasts are isolated from the ^•••' • calvaria of 2 to 5-day-old pups and put into culture. Approximately, about one litter of ten to twelve ^■mouse pups produces 8-9xlO^s cells, forty to sixty^' pups^1' generally yield -3-5'- xlO⁶ primary mouse osteoblasts, and the' cell number generally doubles after four days¹ in cell culture. With "careful attention to animal husbandry, ^' about three new litters are harvested every week to have sufficient • primary mouse osteoblasts for the studies of MDA-BF-I at the cellular and molecular levels. All experiments are repeated a minimum of three times, and the results recorded qualitatively and quantitatively. Summary statistics including mean, standard deviation, median, and ranges are computed for cell proliferation and expression Runx2 through MAPK, Akt and IκB/NF-kB pathways. Furthermore, proliferation and Runx2 expression are compared between groups with various concentrations of antagonists (six replicates each) using the Student t- test, Wilcox on rank-sum test, Kruskal-Wallis test or analysis of variance (ANOVA) in the case of continuous data or by Pearson's chi-square statistic or Fisher's exact test for categorical data.

It is possible that MAPK pathway regulates Runx2 activity, while Akt or IkB/NF-kB pathway regulates cell proliferation/survival. It is also possible that the NF- kB pathway may directly affect Runx2 gene expression. It is anticipated that multiple pathways can be activated by MDA-BF-I to coordinate gene expression programs necessary for osteoblast proliferation and differentiation in vivo. After elucidation of the key signaling pathways, ■ future¹ studies can be used to investigate MDA-BF-I targeted genes in primary osteoblasts, e.g. , standard gene/protein expression profiling strategies, such as DNA microarrays or protein-arrays. The studies help to elucidate how MDA-BF-I functions as a novel bone- epithelium interacting factor and provide insights into how MDA-BF-I functions as a novel and potent osteoblast. , stimulating factor in promoting prostate canςer bone, metastasis. , ^■ . . ; , XIII. Purify/ identify, and clone the BFlrreceptor

¹1. The-BFl-Receptor is Not ErbB2- ^• ' •

That MDA-BF-I activates p42/44MAPK and^'Akt, and stimulates IKB-α degradation in primary mouse osteoblasts suggests that MDA-BF-I functions through a receptor. One of the candidate receptors for MDA-BF-I is ErbB2. The

ErbB3 ligand heregulin (HRG) induces heterodimerization of ErbB3 with ErbB2 , which generates a signaling complex that increases PI3-kinase recruitment. Although MDA-BF-I is not an effective competitor for binding of HRG to pl80-ErbB3, the possibility that MDA-BF-I interacts with ErbB2 directly and activates a signaling pathway through ErbB2 may not be ruled out completely. To test this possibility, the invention also evaluates the effect of MDA-BF-I on ErbB2 activation.

To test whether MDA-BF-I can activate ErbB2 in primary mouse osteoblasts, primary mouse osteoblasts with 50 ng/ml MDA-BF-I were treated for various durations and phosphorylations of ErbB2 and MAPK were measured with an anti-phosphotyrosine antibody (mAb 4G10, UBI) and an anti-phospho-MAPK antibody, respectively. It was found that treatment of PMO with MDA-BF-I does not induce tyrosine phosphorylation of ErbB2 (data not shown) , suggesting that MDA-BF-I does not interact with ErbB2. However, MDA-BF-I was able to induce phosphorylation of p42/44 MAPK in primary mouse osteoblasts, as shown in FIGURE 34. Thus, MDA-BF-I activates MAPK without inducing ErbB2 tyrosine phosphorylation in primary mouse osteoblasts.

Previous studies have shown that stimulation of LNCaP with HRG-β induces cross-talk between ErbB3 and ErbB2 that leads to tyrosine phosphorylation of ErbB2 and activation of MAPK. To test whether MDA-BF-I can activate ErbB2 in LNCaP cells, LNCaP cells were treated with about 50 ng/ml of MDA-BF-I for various amount of time and measured phosphorylation of ErbB2 and MAPK. Treatment of LNCaP' with HRG-β was used as a positive control.

FIGURE 40 shows that HRG-β induced ErbB2 tyrosine phosphorylation and MAPK activation in LNCaP cells. However, treatment of LNCaP cells with MDA-BF-I did not have an effect on ErbB2 phosphorylation or MAPK activation. Similar results were observed with PC-3 and MDA PCa 2b cells (data not shown) . The results from Figures 34 and 40 suggest that MDA-BF-I does not interact with the ErbB2 receptor. In addition, it is possible that the BF-I-receptor is present in primary mouse osteoblasts but not in LNCaP, PC-3, and MDA Pea 2b cells.

Further, it was also found that primary mouse osteoblasts do not express pl80-ErbB3, as demonstrated by immunoprecipitation followed by immunoblotting with anti- ErbB3 antibodies (data not shown) . Thus, the BFl-receptor is neither ErbB2 nor ErbB3. It is concluded tha^ MDA-BF- I₁ mediates osteoblast proliferation through a novel membrane receptor designated as the BF-I receptor. 2. Characterization of the BFl-receptor

2.1 Radiolabeling of ligands

A receptor binding assay is used to detect the BFl- receptor. Purified MDA-BF-I was radiolabeled with ¹²⁵I and used as the binding ligand. Recombinant MDA-BF-I (5 μg) in PBS was labeled with Na¹²⁵I (0.5 mCi) using iodogen- coated tube. After 2 min at 23°C, the mixture was separated on a Sephadex G-50 column. The specific activity of the radiolabeled MDA-BF-I (¹²⁵I-MDA-BF-I) was measured to be 1 x 10^s cpm/μg and the labeled MDA-BF-I was used for receptor binding and cross-linking assays.

2.2 Binding of MDA-BF-I to primary mouse osteoblasts

Primary mouse osteoblasts were plated in 6-well plates and assayed at confluence. The cells were incubated for 2 h at 4°C with 1 x 10⁵ cpm ¹²⁵I-MDA-BF-I and increasing concentrations of unlabeled MDA-BF-I.

Nonspecific binding was determined in the presence of 100-fold excess of unlabeled ligand. The cells were washed three times with ice-cold binding buffer and lysed in 0.5 ml of 0.1 N NaOH and 0.1% SDS for 30 min and radioactivity was determined using a gamma-counter. Our preliminary binding studies suggest that there is about 0.8 pmole BFl-receptors in 10⁶ PMO and the affinity of MDA-BF-I to BFl-receptor is 3 nM. Such a receptor abundance and ligand affinity is sufficient for the isolation of the receptor by biochemical approaches as described below. 2.3 Cross-linking of MDA-BF-I with its receptor

Cross-linking of MDA-BF-I to its receptor will allow us to determine the size of the receptor before it is purified. This information is very helpful during the purification of the receptor to screen for' enrichment of a protein of a^' specific size. Monolayers' of ^■ PMO or LNCaP cells '^'in PBS were incubated on ice for 2 h^' with ¹²⁵I-MDA- BF-1 -with or without excess cold MDA-BF-I. The chemical cross-linking reagent disuccinimidyl suberate (DSS) was then added to a final concentration of 1 mM and the mixture incubated at room temperature for 5 min. The cross-linking reaction was stopped by the addition of IM NH₄Cl, and the supernatant containing the free ligand was removed. The cells were then lysed in SDS-sample buffer and analyzed by SDS-PAGE. The receptor cross-linked to ¹²⁵I-MDA-BF-I was detected by autoradiography.

. FIGURE _:41 showed that a radiolabeled- protein with a molecular weight around 300-kDa was detected. Addition of excess amount of cold MDA-BF-I protein blocks the binding of ¹²⁵I-MDA-BF-I to this protein, suggesting that the labeling is specific. In addition, the 300-kDa radiolabeled-protein was not detected in LNCaP cells. This observation is consistent with the absen.ce of MDA- BF-I signal transduction in LNCaP cells (FIGURE 40) and the observation that LNCaP cells do not express BFl- receptor. By subtracting the mass of MDA-BF-I from the total molecular mass, the apparent molecular mass of BFl- receptor is around 250-kDa. It is concluded that MDA-BF-I has osteoblast stimulating activity both in vitro (FIGURE 29) and in vivo (FIGURE 32) and MDA-BF-I mediates its osteoblast stimulating activity through a receptor or receptors (BF- 1 receptors) . Our studies have ruled out the possibility that MDA-BF-I receptor is ErbB2. Thus, the BFl receptor is thought to be a novel receptor protein that has not been identified. One other possibility is that it is a known protein whose function in osteoblasts has not been previously appreciated. Purification and identification of BFl receptor provide a direct answer to its identity.

2.4 Protein purification and identification strategy

To purify the BFl receptor, in one embodiment, a method of biochemically fractionating membrane extracts and assaying the activity of BF-I receptor is provided. Enrichment of BFl receptor through this purification process increases the probability of obtaining purified receptor by MDA-BF-I affinity chromatography. Thus, biochemical fractionation is used to enrich for the BFl receptor. Fractions enriched for MDA-BF-I binding activity can be detected using receptor binding assays to quantify their activity.

When the activity is enriched about 50 to 100 fold, the enriched fractions are bound to affinity matrix containing immobilized MDA-BF-I (MDA-BF-I ligand-

Sepharose) . Affinity matrix without immobilized MDA-BF-I is used as a control and performed in parallel (control- Sepharose) . The proteins eluted from the ligand-Sepharose and control -Sepharose were then analyzed on 2-D gels that separate proteins according to their isoelectric points and sizes.

Proteins present in samples from affinity-Sepharose but not in control-Sepharose are candidate BF-I receptors and are excised from the gel. The excised proteins are digested in gel with trypsin and directly sequenced by using tandem mass spectrometry (MS/MS) . Comparing the protein contents of purified control-Sepharose and ligand-Sepharose fractions by 2 -D gel electrophoresis permits identification of proteins that bind specifically to MDA-BF-I ligand-Sepharose.

2.5 Estimation of abundance of BFl-receptor

Using receptor binding assays, it is suggested that there is about 0.8 pmol BFl-receptor in one million primary mouse osteoblasts. The abundance of BFl-receptor in primary mouse osteoblasts is similar to that of the insulin receptor and C-CAMl cell adhesion molecule, which was isolated and cloned previously. Purifying these membrane proteins, about 500 to 5000-fold purification is required. Because only picomoles to femtomoles of protein are needed for microsequencing, for a protein of 250 kDa, about 250 ng or less is sufficient for sequence analysis.

2.6 Determination of receptor affinity: Binding of MDA-BF-I to primary mouse osteoblasts are measured first. For measuring the binding of MDA-BF-I to the osteoblast membranes, primary mouse osteoblasts are homogenized with a Dounce homogenizer. The cell lysate is centrifuged at 8,000xg to remove the unbroken cells and mitochondria. The supernatant is centrifuged at 100,000xg for about 1 hour to obtain the membrane fraction. The membranes are incubated for about 2 hour at 4°C with ¹²⁵I- MDA-BF-I, and the receptor affinity is measured. ^' Scatchard analysis will be performed by using the computer program LIGAND.

2.7 Receptor purification 2.7.1 Receptor binding assay for solubilized receptor

¹²⁵I-MDA-BF-I binding to detergent-solubilized receptor is also measured as the assay for the measurement for binding of MDA-BF-I to osteoblast membrane, except that the receptor-ligand complex is precipitated with polyethylene glycol .

2.7.2 Detergents for receptor solubilization

Several detergents are tested for their ability to sσlubilize the BFl-receptor from osteoblast membranes while preserving its ability to bind to MDA-BF-I. For example, for membrane protein solubilization and reconstitution, CHAPS, octylglucoside, Ci₂E_s, Triton X- 100, and NP-40 can be used to solubilize the BFl- receptor. In addition, digitonin, BIG-CHAPS, ' and- other detergents, e.g. cholate, deoxycholate, Tween.20,

Zwittergens are commercially available.

2.7.3 Protein fractionation ^{' ' ■} -

Because most membrane proteins are ^■ glycosylated, ^' lectin-affinity matrix is used to enrich for BFl- receptor. For example, the lectin-affinity matrices commonly used to purify membrane receptors are WGA- Agarose (which binds to glycoproteins containing N- acetylglucosamine and neuraminic acid) and ConA-Sepharose (which binds glycoproteins containing maήnose and glucose) . Other column chromatographies can also be used, such as those that separate proteins by charge or size. FIGURE 42 depicts the purification scheme for purifying BFl-receptor . First, the detergent-soluble fraction is bound to WGA-agarose affinity column and eluted with 0.25 M N-acetylglucosamine. Roughly, only 2- 5% of the proteins can bind to the WGA-agarose column.

Both the flow-through and bound fractions are assayed for the presence of BFl-receptor by the receptor binding assay. The bound fractions of the WGA-agarose column are applied to MDA-BF-I ligand-Sepharose, prepared by coupling MDA-BF-I to CNBr-activated Sepharose .

On the other hand, the flow-through fractions of the WGA-agarose affinity column are applied to a ConA- Sepharose affinity column and eluted with 0.25 M cornethylmannoside. Then, the bound fractions of the ConA- Sepharose column are applied to gel filtration chromatography or MDA-BF-I ligand-Sepharose column, or both, sequentially. , ,

Alternatively, the flow-through fraction from the ConA-Sepharose column is loaded onto a Mono-Q anion exchange column in 50 mM Tries and eluted sequentially with 0.1 M, 0.2 M, 0.3 M, 0.5 M, and 1 M of NaCl. Fractions from each step are assayed for BFl-receptor activity, and the receptor-containing fractions are further purified by gel filtration chromatography before it is applied to the MDA-BF-I ligand-Sepharose affinity column .

After enrichment by lectin-affinity, ion-exchanger, or gel filtration chromatography, half of the partially purified receptor fraction will be incubated with ligand- Sepharose. The other half will be incubated with control-Sepharose, which is CNBr-Sepharose coupled with BSA. The proteins bound to the ligand- and control- Sepharose will be eluted from the affinity matrix by using Gentle-elute (Pierce) . The samples will be dialyzed to remove the salt present in the Gentle-elute.

2.8 2 -D gel analysis and protein sequencing of the identified receptor The proteins from the ligand-Sepharose purification are analyzed by 2 -D gel electrophoresis. Samples from the control-Sepharose are also analyzed in parallel. The separated proteins are visualized with Coomassie Blue staining (Gelcode, Pierce Chemical Company) or silver staining. The protein spots present in the ligand- Sepharose but not in control-Sepharose are cut from the gel and digested directly in gel with trypsin. The tryptic fragments are then sequenced by MS/MS. With MS/MS, protein sequence information can be obtained from picomole to femtomole amounts of protein. The first-stage MS separates peptides according to their mass-to-charge ratios. Those separated peptides are then subjected to ^■ second-stage MS, in which they are bombarded with high energy and degraded (cleavage at peptide bonds) to give a distinctive MS spectrum. The MS spectrum is then analyzed to give peptide sequence information. The peptide sequences thus obtained are compared with those in the protein data bank for possible matches.

XIV. Generation of Transgenic Animal Expressing MDA- BF-I ■ , _, . ;

Genetically-engineered host cells^' can be further ^■ used to produce- non-human transgenic animals¹. A ' transgenic^" animal ' is preferably a mammal', for- example a rodent, such as- a rat or mouse, in which one or more of the cells of the animal include a transgene. A transgene usually contains exogenous DNA integrated into the genome of a cell from which a transgenic animal develops and which remains in the genome of the mature animal in one or more cell types or tissues of the transgenic an,imal . These- animals are useful for studying the function of the identified bone factor and identifying and evaluating modulators of kinase protein activity. Other examples of transgenic animals include non-human primates, sheep, dogs, cows, goats, chickens, and amphibians.

A transgenic animal can be produced by introducing nucleic acid into the male pronuclei of a fertilized oocyte, e.g., by microinjection, retroviral infection, and allowing the oocyte to develop in a pseudopregnant female foster animal . Any of the nucleotide sequences of the identified bone factors can be introduced as a, transgene into .the genome of a non-human animal,, such as a mouse. Any of the regulatory or other sequences useful in expression vectors can form part of the transgenic sequence. This includes intronic sequences and polyadenylation signals, if not already included, A ■ tissue-specific regulatory sequence (s) ,, such as bone- specific regulatory sequence (s) can . be ..operably linked to the transgene to. direct expression of the transcjene. to particular . cells . . , ,

^• Methods¹ for generating transgenic animals via 'embryo manipulation and microinjection, particularly animals such as mice, have become conventional in the art _. and are described, for example, in U.S. Pat. Nos . 4,736,866 and 4,870,009, both by Leder et al., U.S. Pat. No. 4,873,191 by Wagner et al. and in Hogan, B., Manipulating the Mouse Embryo, (Cold Spring Harbor Laboratory Press,/ Cold Spring Harbor, N. Y. , 1986). Similar methods are used , for . production of other transgenic animals. A transgenic founder. animal can be identified based upon the presence of the transgene in its genome and/or expression of transgenic mRNA in tissues or cells of the animals. A transgenic founder animal can then be used to breed additional animals carrying the transgene. Moreover, transgenic animals carrying a transgene can further be bred to other transgenic animals carrying other transgenes. A transgenic animal also includes animals in which the entire animal or tissues in the animal have been produced using the homologously recombinant host cells described herein. In another embodiment, transgenic non-human animals can be produced which contain selected systems that allow for regulated expression of the transgene. One example of such a system is the cre/loxP recombinase system of bacteriophage Pl. For a description of the cre/loxP recombinase system, see, e.g., Lakso et al. PNAS 89:6232- 6236 (1992) . Another example of a recombinase system is the FLP recombinase system of S. cerevisiae (O 'Gorman et al. Science 251:1351-1355 (1991). If a cre/loxP recombinase system is used to regulate expression Transgenic animals that express MDA-BF-I in specific types of cells, e.g., osteoblasts, are generated- These transgenic animals allow the examination of the effect of MDA-BF-I on bone development in vivo. Because metastatic prostate cancer cells may secrete multiple factors to modulate osteoblast growth and differentiation in a paracrine fashion, an autocrine mechanism in this transgenic model can test the role of MDA-BF-I without the involvement of other factors from metastatic prostate cancer cells. Establishing animal models with MDA-BF-I is provided to verify the role of MDA-BF-I in vivo and also allows therapy strategies that target MDA-BF-I to be tested in vivo.

1. Osteoblast-specific promoters As an example, transgenic mice that express MDA-BF-I specifically in osteoblasts were generated. Expression of targeted protein in osteoblasts requires an osteoblast- specific promoter. Only few genes are known to be osteoblast-specific . Osteocalcin is the most osteoblast- specific gene known and its expression is not only cell- specific but also time-specific and stage-specific. Indeed, its expression is virtually absent before birth, restricted to differentiated osteoblasts which are able to produce a bone extracellular matrix, and is absent in osteoblast progenitors.

A 1.3-kb fragment of the mouse osteocalcin [0Gr2].. promoter contains all the regulatory elements necessary to confer differentiated osteoblast-specifid and post- natal-specific expression to a reporter gene . in vivo _;has been reported by Frendo et al.. This osteocalcin promoter fragment constitutes a unique resource to address osteoblast fμnction. This promoter has been sμccessfully used in expressing a fragment of Cbfal that , conferred , dominant, negative Cbfal function in vivo and, in expressing soluble colony-stimulating factor-1 to rescue the osteopetrotic defect in op/op mice.

Another example of osteoblast-specific promoter is a type I collagen promoter. Type I collagen is the .most abundant protein of the bone extracellular matrix, accounting for 90% of the matrix protein qontent. The type I collagen^ genes are expressed in osteoblastic cells at all stages during development and throughout life._. It has been showed that Cbfal, a key regulator of , osteoblast-specific gene expression, is one of the factors controlling osteoblast-specific expression of type I₁ collagen genes. Dacquin et al. has tested several promoter fragments for their ability to drive efficient protein expression, specifically Cre recombinase , in osteoblasts. Three osteoblast_.-specific promoters were tested. The first promoter is the 2.3-kb proximal fragment of the αl(I)- collagen promoter, which is expressed at high levels in osteoblasts throughout their differentiation. The second promoter is the 1.3-kb fragment of 0G2 promoter, which is active in differentiated osteoblasts. The third promoter is an artificial promoter derived from the OG2 promoter that showed higher activity than natural 0G2 promoter. They found that the only promoter able to drive Cre recombinase expression at a level sufficient to induce recombination in. osteoblasts is the αl (I) -collagen promoter. Thus, αl (I) -collagen promoter is used herein for in vivo osteoblast-specific over-expression of the transgene of the invention, such as the osteoblast- stimulating factor, MDA-BF-I.

2. Expression Constructs of MDA-BF-I with ■ ^'■ ■ osteoblast-specific promoters ^■"■ ^: ' • ^■• •

As an example, an expression construct containing the 2.3-kb αl (I) -collagen promoter to drive MDA-BF-I expression was generated, αl (I) -col-MDA-BF-1-IRES-LacZ plasmid was constructed by inserting the 2.3-kb αl(I)- collagen promoter fragment and the cloned MDA-BF-I cDNA into a plasmid pIRES-LacZ which contains IRES-LacZ sequences. This αl (I) -col-MDA-BF-1 plasmid was used to generate transgenic mice.

First of all, the activity and cell-specificity of αl (I) -col-MDA-BF-1 was tested before injecting the plasmid into mouse embryo. The αl (I) -col-MDA-BF-1 plasmid was transfected into ROS17/2.8 osteoblasts and F9 carcinoma cells. The F9 cells have no feature of, . osteoblasts and is used as a control . The expression of MDA-BF-I was measured by Western blot. A CMV promoter- driven MDA-BF-I plasmid, which is not cell-type specific, was also transfected into these two cell lines, ROS17/2.8 osteoblasts and F9 carcinoma cells, as positive control for trarisfection.^'

3. Generation and analysis of transgenic mice

The plasmid αl (I) -col-MDA-BF-I was used to generate transgenic mouse lines. The construct αl (I) -col-MDA-BF-I was linearized, purified by two rounds of agarose gel electrophoresis and injected into the pronuclei of fertilized B6D2 mouse oocytes, which were then implanted in the oviducts of pseudopregnant CDl foster mothers for development to term. The generation of transgenic mice was performed by Institutional Transgenic mouse core facility in M. D. Anderson Cancer Center. Transgenic mice were genotyped by PCR on tail genomic DNA' using MDA- BF-1-specific primers. Seven founder lines were generated.

4. Tissue distribution of the transgene

The tissue distribution of the transgene is examined, for example, by RT-PCR. Total RNA from different tissues of transgenic mice was extracted with RNAeasy reagent ^■ (Qiagen) . cDNA was synthesized from about ϊ μg^;of ^:

DNasel-treated total RNA with random primers by using the Superscript II RNaseH reverse transcriptase kit ■ (Invitrogen, Inc.) . PCR was performed using MDA-BF-I- specific primers. Real-time PCR analysis can be used to quantify the levels of transgene expression among different transgenic mouse lines. Western blots and immunostaining are then used to confirm that the protein is expressed from transgene. For western blot analysis, total cellular protein from tissues is prepared by sonication in RIPA buffer - supplemented with protease and phosphatase inhibitors. For immunohistochemistry, tissues from the transgenic mouse are collected, frozen in OCT medium or fixed in paraformaldehyde and embedded in paraffin. Animal specimens are sectioned at 4-5 μm and immunostained with antibody. Specimens from bone are fixed in formalin, decalcified with 10% formic acid, and embedded in paraffin. All the sections from paraffin-embedded tissues are dewaxed with xylene and rehydrated in graded alcohol before proceeding with immunostaining procedures. The sections are then treated with 3% H₂θ2 in methanol at room temperature for 15 minutes, washed with phosphate- buffered saline (PBS) , blocked with normal goat serum at room temperature for 30 minutes, and then incubated at room temperature for 1 hour with antibody. The antibody binding will be detected by using ABC kit (Vector laboratory, Burlingame, CA) according to the manufacturer's instructions with 3, 3' -diaminobenzidine as the chromogen. The immunostained sections are also counterstained with hematoxylin. 5. Phenotypic characterization of transgenic mice

Transgenic mice are analyzed for the presence of increased number of osteoblasts or bone mass . Phenotypic characterization of transgenic mice includes measuring (1) the growth- rate as compared to their -wild-type littermates; (2) gross change in body shape, e.g. , hunched back, etc. , due to the abnormality' in bone development; (3) radiography analysis of radio-density of long bones, vertebral bodies, ribs, and the skull; (4) histological analysis for evidence of increased bone mass, osteoblasts, and osteoid; (5) histochemical staining for tartrate-resistant acid phosphatase (TRAP) for the possible involvement of osteoclasts. To study the cellular mechanisms, histpmorphometric analysis is performed. The bone specimens are decalcified in 10% sodium EDTA in 0.1 M phosphate buffer (pH 7.0) at 4°C for 4 days. The samples are then dehydrated through standard graded alcohol solutions and embedded in methylmethacrylate resin. Tissues are sectioned longitudinally using microtome and the sections are also stained for tartrate-resistant acid phosphatase activity followed by thionin green counte,rstaining. Bone and osteoclast surfaces can be traced. The .trabecular bone volume, amount of newly formed bone matrix, such as osteoid volume and osteoid-covered surfaces, osteoclast numbers can be calculated using Osteomeasure software (Osteometries, Atlanta, GA) . Markers of known osteoblast differentiation and osteoclast activity are measured to corroborate with the histomorphometric analysis. Serum osteocalcin concentration and alkaline phosphatase activity, which reflects bone formation at the syst.emic level, are measured. In addition, expression of bone matrix proteins, reflecting the osteoblastic activity, is ^' also measured. For osteoclast activity, urinary deoxypyrodinoline crosslinks can be monitored.

6. Phenotypic characterization of transgenic mice in the absence of androgen

Because bone metastasis of prostate cancer frequently occurs after hormone therapy, it is possible that androgen has a role in the induction- of^'osteoblastic response. It has been hypothesized that -increased osteoclastic bone resorption due . to androgen deprivation causes a more fertile environment for bone metastasis, which results in the enhanced growth of tumor cells in the bone compartment. Therefore, to investigate the role of androgen deprivation in the development of osteoblastic response, transgenic mice undergoing orchiectomy or sham surgery can be prepared. Radiographs, bone mineral density and bone histomorphometry are used to characterize the phenotypic changes of these transgenic mice . Several possible phenotypes may arise from the over- expression of the osteoblast-stimulating factor. First, only ,the osteoblast function is affected by the transgene. Second, both osteoblastic and osteoclastic activities are affected by over-expression of the transgene. Third, only osteoclastic activity is affected by the transgene. Fourth, there is no detectable phenotype .

Overexpression of MDA-BF-I in vivo in osteoblasts may lead to increased proliferation of osteoblasts resulting in osteosclerotic phenotype. However, because MDA-BF-I may have an effect on osteoclast differentiation, it is equally likely that both osteoblast and osteoclast activities may be affected. It is also likely that the osteoclastic effect of MDA-BF-I is the dominant effect in vivo as to have an overall osteolytic phenotype. ,

In the case that there is no osteoblastic or osteoclastic phenotype in the transgenic mice, the effect of androgen deprivation on the bone remolding of the transgenic mice can be evaluated. If there were still no discernable difference between the transgenic and wild- type mice, it may be concluded that these bone metastasis factors alone do not have significant influence in bone development and they may corroborate with other factors for the osteoblastic response observed.

Because αl (I) -collagen is expressed in osteoblastic cells at all stages during development and throughout life, the MDA-BF-I transgene may be expressed. early in development and the transgenic mice may develop phenotype early. If the phenotype of transgenic mice occurs too early as to have an effect on mouse development, overexpressing MDA-BF-I using other promoter, such as the 0G2 promoter, can also be tested. OG2 promoter is derived from osteocalcin that is expressed postnatally in differentiated osteoblast and as such 0G2 prompter can provide stage-specific and cell-specific expression of candidate gene in osteoblast as a possible alternative promoter to use.

Although only exemplary embodiments of the invention are specifically described above, it will be appreciated that modifications and variations of these examples are possible without departing from the spir,i_.t „ and intended scope of the invention.

Claims

CLAIMS :

1. A paracrine factor comprising MDA-BF-I or MDA-

BF-2.

2. An osteoblast stimulating factor comprising a 45 kDa N-terminal through intron 8 portion of the pl80- ErbB3 gene .

3. An antibody comprising an antibody raised against purified MDA-BF-I or a fragment or derivative thereof which binds to MDA-BF-I.

4. An antibody according to Claim 3, wherein the antibody is a polyclonal antibody.

5. An antibody according to Claim 3, wherein the antibody is a monoclonal antibody.

6. A peptide comprising a peptide which binds to MDA-BF-I.

7. A peptide according to Claim 6, further comprising the peptide screened from a phage display library by binding to MDA-BF-I.

8. A peptide according to Claim 6, further comprising a peptide having a sequence selected from the group consisting of ASGADGP (SEQ. ID.NO.1) , FGWPLW

(SEQ. ID.NO.2) , GGLALQE (SEQ . ID.NO.3 ), LKRGITV (SEQ. ID.NO.4) , FASSFVL (SEQ. ID.NO.5) , TLDFPRR (SEQ. ID.NO.6) , ISFPRRW (SEQ. ID.NO.7) , WAGGRF (SEQ. ID.NO.8) , VAGGSFI (SEQ . ID.NO .9) , QGGVRHH (SEQ. ID.NO.10) , GGVRVLD (SEQ.ID.NO.il), FASRVRS (SEQ. ID.NO.12) , QSRVRVA (SEQ . ID .NO .13) , PAGRYTD (SEQ. ID.NO.14) , GRYTTDR (SEQ . ID.NO .15) , SGYVAKM (SEQ. ID.NO.16) , SGYAKVS (SEQ . ID .NO .17) , and any combinations thereof.

9. A method of diagnosing metastatic prostate cancer disease in a subject comprising: obtaining a sample from the subject; and testing the sample for the presence of MDA-BF-I protein.

10. A method according to Claim 9, wherein the sample is selected from the group consisting of a cancer tissue biopsy, a prostate cancer biopsy, bone marrow, blood, plasma, urine, and any combinations thereof.

11. A method according to Claim 9, further comprising measuring an amount of MDA-BP-I protein in the sample .

12. A method according to Claim 11, further comprising comparing the amount of MDA-BF-I protein in the sample to at least one amount of MDA-BF-I protein in a standard or set of standards .

13. A method according to Claim 12, wherein the set of standards further comprises standards corresponding to non-cancerous tissue, non-metastatic cancerous tissue, and metastatic cancerous tissue.

14. A method according to Claim 11, further comprising determining the correlation between the amount of MDA-BF-I in the sample and the extent of metastasis of the prostate cancer disease in the subject.

15. A method according to Claim 11, further comprising correlating the amount of MDA-BF-I in the sample with a predetermined amount of MDA-BF-I for each of different stages of prostate cancer disease.

16. A method according to Claim 9, wherein testing the sample further comprises using an antibody raised against MDA-BF-I or a fragment or derivative thereof operable to bind to MDA-BF-I.

17. A method of stimulating bone growth, comprising introducing MDA-BF-I into a subject.

18. A method according to Claim 17, wherein MDA-BF- 1 is introduced through a way selected from the group consisting of oral delivery, subcutaneous injection, intra-bone injection, and any combinations thereof.

19. A method according to Claim 17, wherein MDA-BF- 1 is introduced from a cell with an expression construct having an expression region encoding MDA-BF-I protein.

20. A method of treating a bone-related disease, comprising stimulating bone growth by introducing MDA-BF- 1 into a subject.

21. A method according to Claim 20, wherein the bone-related disease is osteoporosis.

22. A method according to Claim 20, wherein the bone-related disease is selected from the group consisting of: bone metastasis, osteolytic disease, prostate cancer bone metastasis, and any combination thereof .

23. A method according to Claim 20, wherein MDA-BF- 1 is introduced through a process selected from the group consisting of oral delivery, subcutaneous injection, intra-bone injection, and any combination thereof.

^{■' ■ ■ ■} '24. A method according to Claim 20, wherein MDA-BF- 1 is introduced from a cell with an expression construct having an expression region encoding MDA-BF-I protein.

25. A method of treating a disease in a subject, comprising reducing the expression of MDA-BF-I in at least one cell in the subject.

26. A method according to Claim 25, further comprising introducing into the subject an antibody raised against purified MDA-BF-I or a fragment or derivative thereof operative to bind to MDA-BF-I.

27. A method according to Claim 26, wherein the antibody is selected from the group consisting of: monoclonal antibodies, polyclonal antibodies, and any combinations thereof.

28. A method according to Claim 25, wherein the disease is selected from the group consisting of: bone metastasis, osteoblastic disease, prostate cancer bone metastasis, prostate cancer lymph node metastasis, and any combination thereof.

29. A method according to Claim 25, wherein reducing expression of MDA-BF-I comprises introducing a double stranded RMA, siRNA, or antisense RNA into the cell.

30. A method of treating a disease in a subject comprising introducing a modulator for interaction between MDA-BF-I and MDA-BF-I receptor into the subject.

31. A method according to Claim 3Q wherein the modulator is selected from the group consisting of monoclonal antibodies, polyclonal antibodies, antibody derivatives or fragments operable to bind MDA-BF-I, agonists, antagonists, inhibitors, inducers, siRNA, dsRMA or antisense RNA corresponding to the nucleotides sequences of MDA-BF-I, peptides corresponding to the amino acid sequences of MDA-BF-I, peptides corresponding to the amino acid sequences of MDA-BF-I receptor, and combinations thereof.

32. A method according to Claim 32, wherein the disease is selected from the group consisting of: bone metastasis, osteoblastic disease, prostate cancer bone metastasis, prostate cancer lymph node metastasis, and any combination thereof.

33. A method of treating a disease in a subject, comprising introducing a peptide corresponding to a sequences from MDA-BF-I into the subject.

34. A method according to Claim 33, wherein the peptide is selected from the group consisting of ASGADGP

(SEQ. ID.NO.1) , FGWPLW (SEQ. ID.NO .2) , GGLALQE

(SEQ. ID.NO.3) , LKRGITV (SEQ. ID .NO .4) , FASSFVL

(SEQ. ID.NO.5) , TLDFPRR (SEQ. ID .NO .6) , ISFPRRW

(SEQ. ID.NO.7) , WAGGRF (SEQ . ID.NO .8) , VAGGSFI (SEQ. ID.NO.9) , QGGVRHH (SEQ. ID.NO.10) , GGVRVLD (SEQ.ID.NO.il), FASRVRS (SEQ. ID.NO.12) , QSRVRVA (SEQ . ID .NO .13) , PAGRYTD (SEQ . ID .NO .14) , GRYTTDR (SEQ. ID. NO.15) , SGYVAKM (SEQ. ID.NO.16) , SGYAKVS (SEQ. ID.NO.17) , and any combinations thereof.

35. A method according to Claim 34, wherein the disease is selected from the group consisting of bone metastasis, osteoblastic disease, prostate cancer bone metastasis, prostate cancer lymph node metastasis, and any combinations thereof.

36. A method of treating a disease in a subject, comprising introducing into the subject an RNA selected from the group consisting of: dsRNA, siRNA or antisense RNA corresponding to a nucleotide sequence of MDA-BF-I, or any combination thereof.

37. A method according to Claim 36, wherein the RNA comprises a sequence selected from the group consisting of: aacgacgctctgcaggtgctgdTdT (SEQ. ID.NO.18) , aactctcaggcagtgtgtcctdTdT (SEQ. ID.NO.19) , derivatives and homologs thereof, and any combination thereof.

38. A method according to Claim 36, wherein the disease is a disease selected from the group consisting of bone metastasis, osteoblastic disease, prostate cancer bone metastasis, prostate cancer lymph node metastasis, and any combination thereof.

39. A method of treating a disease, comprising introducing into a subject an antibody raised against purified MDA-BF-I or a fragment or derivative thereof operable to bind to MDA-BF-I.

40. A method according to Claim 39, wherein the antibody is selected from the group consisting of monoclonal antibodies, polyclonal antibodies, and any combinations thereof.

41. A method of modulating bone growth in a subject, comprising introducing a modulator for interaction between MDA-BP-I and MDA-BF-I receptor into a subject.

42. A method according to Claim 41, wherein the modulator is selected from the group consisting of monoclonal antibodies, polyclonal antibodies, fragments and derivates thereof operable to bind to MDA-BF-I, agonists, antagonists, inhibitors, inducers, siRNA corresponding to the nucleotides sequences of MDA-BF-I, peptides corresponding to the amino acid sequences of MDA-BF-I, peptides corresponding to the amino acid sequences of MDA-BF-I receptor, and any combinations thereof .

43. A method of reducing the expression of MDA-BF-I protein in a cell, comprising introducing into the cell an RNA selected from the group consisting of dsRNA, siRNA or antisense RNA corresponding to a nucleotide sequence of MDA-BF-I, or any combination thereof.

44. A method according to Claim 43, wherein the RNA comprises a sequence selected from the group consisting of aacgacgctctgcaggtgctgdTdT (SEQ. ID.NO.18) , aactctcaggcagtgtgtcctdTdT (SEQ. ID.NO.19) , and derivatives and homologs thereof.

45. A method according to Claim 44, wherein RNA is introduced from an expression construct.

46. A siRNA oligonucleotide complementary to a region of an open reading frame of MDA-BF-I DNA.

47. An siRNA oligonucleotide according to Claim 46, wherein the oligonucleotide has a length of less than 200 nucleotides .

48. An oligonucleotide according to Claim 46, wherein the oligonucleotide is selected from the group consisting of aacgacgctctgcaggtgctgdTdT (SEQ. ID.NO.18) and aactctcaggcagtgtgtcctdTdT (SEQ. ID.NO.19) .

49. A method of reducing the production of MDA-BF-I protein in cells, comprising delivering an antisense oligonucleotide complementary to a region of an open reading frame of MDA-BF-I DNA to the cells.