WO1998054357A1 - SCREENING METHODS TO IDENTIFY COMPOUNDS WHICH EFFECT OSTEOCLAST-SPECIFIC PHOSPHOTYROSYL PHOSPHATASE (PTP-oc) ACTIVITY - Google Patents

Abstract

A screen for osteoclast-specific, anti-bone resorption and/or bone resorption-stimulating pharmaceutical drug therapies to treat patients with bone wasting diseases, including osteopenia and osteoporosis and low bone turnover disease, such as osteopetrosis and aplastic renal bone disease, respectively, is provided. The screening method, as set forth in the figure, targets the signal transduction mechanism of bone resorption and employs a novel osteoclastic phosphotyrosine phosphatase (PTP-oc), which is an important regulator of the signal transduction mechanism of osteoclasts, as a specific target in high intensity screening tests to identify low molecular mass, potent and specific inhibitors and/or stimulators of this osteoclastic enzyme activity.

Description

"SCREENING METHODS TO IDENTIFY COMPOUNDS WHICH EFFECT OSTEOCLAST- SPECIFIC PHOSPHOTYROSYL PHOSPHATASE (PTP-OC) ACTIVITY."

FIELD OF THE INVENTION

This invention is related to the use of a novel, newly discovered, osteoclast-specific phosphotyrosine phosphatase (PTP-oc) as a target in screening tests for the development of new forms of (1) osteoclast-specific anti-resorptive drugs for the use of development of new pharmaceutical therapeutic treatments for bone-wasting diseases (e.g., osteopenia and osteoporosis) and (2) osteoclast-specific resorption- stimulating agents to treat low bone turnover diseases, such as osteopetrosis and aplastic renal bone disease.

BACKGROUND OF THE INVENTION

Osteopenia is a general term used to describe any bone- wasting disease in which the patient has a low bone mass. Osteoporosis, a bone wasting disease characterized by a significant loss of bone mass, is a common bone disorder affecting millions of people throughout the world. The skeleton of osteoporotic patients becomes weakened to an extent that is unable to bear the normal stresses imposed on it. The effects of the disease are generally seen in weight-bearing skeleton, particularly the spine and hips, which can fracture O 98/54357

even in the absence of trauma.

Bone mass in vertebrates is determined by the balance of bone resorption and formation rates. The bone turnover process is initiated by an increase in bone resorption. Under normal situations, an increase in bone resorption is coupled with a subseguent compensatory increase (in equal magnitude) in bone formation rate to ensure that no net bone mass will be lost. A defect in this coupling of bone formation to resorption, i.e., an increase in the bone resorption without the compensatory increase in bone formation, is the primary cause of bone loss in humans. Accordingly, patients with bone-wasting diseases, such as osteopenia and osteoporosis, are characterized by a defect in the bone coupling mechanism in that a loss of bone mass results from increased bone resorption exceeding bone formation. In these patients, the greater the basal bone turnover rate (i.e., bone resorption rate), the greater the bone mass loss would be. Consequently, a viable way to treat patients with bone-wasting diseases (including osteopenia and osteoporosis) is to reduce the basal bone resorption rate with an anti-resorptive drug.

Several anti-resorptive therapies have been developed and three (i.e., estrogen, calcitonin, and alendronate, a third generation of bisphosphonate) have received the FDA approval for the therapeutic use for osteoporosis. Many other anti- resorptive therapies (e.g., Raloxifene, an estrogen agonist in bone, and several fourth generation bisphosphonates) are currently being developed. While most of these therapies are effective in preventing bone loss, these therapies have pitfalls. For example, calcitonin is not orally active, and requires administration through nasal spray or injection. Administration of bisphosphonates, such as alendronate, causes significant gastrointestinal side-effects, and must be administered with water in an empty stomach. Estrogens cause side-effects including breast cancer. These agents also produce substantial side-effects in non-skeletal tissues. Moreover, the molecular mechanism(s) whereby these agents act to inhibit bone resorption is largely unknown. Without the knowledge of their molecular mode(s) of actions, it has been difficult to optimize the therapies and/or to avoid undesirable side-effects. As a result, these problems diminish the therapeutic effectiveness and the tolerance of patients to these therapies. Consequently, there remains a need for an orally active and bone-specific anti-resorptive therapies for bone-wasting diseases.

The approach of the present invention is novel and differs from the currently available anti-resorptive therapies in that the invention targets at the signal transduction mechanism of bone resorption and that it makes use of a recently discovered novel osteoclast-specific enzyme, osteoclastic phosphotyrosine phosphatase (PTP-oc), which is a key regulator of bone resorption, as a screening test for the development of novel, potent, and orally active, anti-resorptive agents. Moreover, because this enzyme activity is specific for osteoclasts, agents specifically affecting this enzyme activity would also be specific for osteoclasts. Therefore, the anti-resorption drugs developed by the approach described in this invention should be specific for osteoclasts, and thus, should have minimal side-effects on other cell types and organ tissues. Accordingly, it would be an ideal therapy for bone-wasting diseases.

Bone resorption is mediated by osteoclasts, which are multinucleated giant cells formed by fusion of mononucleated precursor cells of monocyte-macrophage hematopoietic lineage. The bone resorption process is controlled by two key parameters: 1) the number, and 2) the activity of mature osteoclasts. Osteoclast activity is regulated by a number of osteo-regulatory hormones and effectors. l,25(OH)₂D₃, PTH, and prostaglandin E₂, have been shown to increase both the activity and number of mature osteoclasts; whereas estrogen, calcitonin, and bisphosphonates, reduced both osteoclastic activity and number. In humans, osteoclast activity as well as the number of osteoclasts regulate the bone resorption rate.

With respect to the molecular mechanism of the bone resorption process, recent observations indicated that at least two members of the cytosolic protein tyrosine kinase (PTK) family are involved in normal regulation of osteoclast formation and activity: First, studies with the op/ op variant of murine osteopetrosis, a metabolic bone disease characterized by a defective bone resorption process due to the lack of osteoclasts, have shown that production of macrophage-colony stimulating factor (M-CSF) and activation of its receptor are required for normal osteoclast formation [Yoshida et al . , Nature 345:442-444, 1990]. The M-CSF receptor, which is encoded by the proto-oncogene c-fms, is a PTK. Thus, these findings indicate that the c-fms PTK activity is essential for osteoclast formation. Second, targeted disruption of the c-src proto-oncogene in mice led to the unexpected development of another form of osteopetrosis, in which the bone contains an abundant number of totally inactive osteoclasts [Soriano et al . , Cell 64:693-702, 1991]. The c-src proto-oncogene also encodes a PTK, known as ppβO⁰""⁰. The abrogation of the c-src proto-oncogene expression in osteoclasts resulted in the failure of formation of ruffled borders, which are highly specialized areas in the cytoplasmic membrane of osteoclasts where bone resorption occurs. Consistent with the idea that pp60^c-"^rc PTK is involved in the ruffled border formation, it has been reported that active osteoclasts expressed high level of pp60^c"^βrc activity on the ruffled border membranes. The PTK activity of pp60^c""^rc correlated with the number of osteoclasts which were actively resorbing bones. Inhibitors of c-src PTK activity completely abolished bone resorption in bone slices [Yoneda et al . , Mol. Endocrinol. 7:1313-1318, 1993]. These observations clearly indicate that the PTK activity of pp60^c"^βrc in osteoclasts are absolutely required for the functional activity of mature osteoclasts. Because the osteoclast activity is an important determinant of bone resorption, the PTK activity of pp60^c_Bro in osteoclasts would be important in the regulation of basal bone resorption rate.

The regulation of the PTK activity of pp60^c"^βrc is determined by its tyrosine phosphorylation status. The PTK activity is negatively regulated by phosphorylation of tyr-527 in the carboxy-ter inal tail of the proto-oncogene. In contrast, the phosphorylation of tyr-416 results in the stimulation of its PTK activity. Dephosphorylation of tyr-527 results in a change in conformation, leading to the autophosphorylation of tyr-416 and further stimulation of its PTK activity. Accordingly, the phosphorylation status of tyr- 527 is the primary determinant of the PTK activity of pp60^c"^βrc. Thus, agents that increase the phosphorylation level of tyr-527 but not that of tyr-416 of the pp60^c"^βrc in osteoclasts are expected to inhibit the PTK activity of pp60^c""^rc and bone resorption. Consistent with the speculation that the alteration of the tyrosyl phosphorylation status of pp60^c""^r in osteoclasts is involved in the regulation of bone resorption are the findings that treatment of isolated osteoclasts with calcitonin reduced the PTK activity of pp60^c"^βrc without an effect on the pp60^c"^βr protein level. Conversely, PTH significantly increased both the PTK and protein level of pp60^c" ^βrc in isolated osteoclasts [Yoneda et al . , Mol. Endocrinol. 7:1313-1318, 1993]. We very recently also showed that O 98/54357

bisphosphonates, such as clodronate, and calcitonin each significantly increased the tyrosyl phosphorylation level of pp60^c_βrc; whereas the bone resorption stimulators, PTH, prostaglandin E2, and l,25(OH)₂D₃, each reduced the tyrosyl phosphorylation of pp60^c_βr in isolated osteoclasts. Hence, the tyr-517 phosphorylation level of pp60^c"^βrc in osteoclasts is a key determinant of bone resorption rate in vertebrates.

The pp60^c"^βr is a ubiquitous protein and plays important roles in cellular functions and activities in most, if not all, cell types in vertebrates. Any effectors of pp60^c"^βrc would alter the enzyme activity not only in osteoclasts but also in other cell types. Accordingly, inasmuch as there is evidence that the pp60^c-"^rc PTK is a key player in the signal transduction mechanism of the osteoclastic activity, the pp60^c"^βrc PTK would probably not be an appropriate screening target for specific anti-resorptive drugs. However, phosphorylation of tyr-527 is phosphorylated by Csk (c-src kinase) PTK; whereas tyr-416 phosphorylation is mediated primarily by autophosphorylation. The enzymes responsible for catalyzing the removal of the phosphate moiety from the tyrosyl residues of pp60^c""^rc are the phosphotyrosine phosphatases (PTPs). Hence, the cellular steady state tyrosine phosphorylation level (or PTK activity) of pp60^{c" βrc} is controlled by the balance of the activity of Csk and PTPs. Accordingly, increases in steady state phosphorylation level of tyr-527 of pp60^"βrc in osteoclasts can be achieved by increasing the Csk activity, or by inhibiting one or more PTP activities that are responsible for dephosphorylating tyr-527, or both. On the other hand, it should be noted that, as for _pp60^c-Brc, the Csk PTK is a ubiquitous protein. Thus, Csk would probably also not be a suitable screening target for specific anti-resorptive agents.

There are at least two distinct PTP supergene families: one family consists of cytosolic enzymes with a single catalytic domain; and the other family are transmembrane "receptor-like" PTPs, which structurally are composed of extracellular, transmembrane and cytoplasmic domains. The cytoplasmic portion of all of the known transmembrane PTPs contains two tandem catalytic domains (i.e., PDI and PDII), with the exception of HPTPβ and DPTPIOD, both of which contain only a single catalytic domain. Recent studies suggest that the membrane-proximal catalytic domain (PDI) is catalytically active, whereas the membrane-distal catalytic domain (PDII) by itself has no measurable enzymatic activity but may have regulatory functions. While the cytoplasmic core phosphatase domains of PTPs are highly conserved, the extracellular domains of the receptor-like PTPs are unrelated to each other. We have recently discovered a novel PTP (PTP-oc) in osteoclasts that regulates the phosphorylation level and PTK activity of pp60^c" ^erc. The PTP-oc belongs to the transmembrane PTPs. However, PTP-oc is structurally unique in that, unlike the known transmembrane "receptor-like" PTPs, it has a very short extracellular portion, lacks a signal peptide proximal to the N-terminus, and contains only a single ' PTP catalytic domain' . The PTP-oc expresses primarily in osteoclasts, but can also be found in spleen cells, the brain, and the kidney. We have evidence that PTP-oc plays a key role in regulating the tyrosyl phosphorylation levels (and thus the activity) of several key signal transduction proteins, including pp60^o"^βrc, in osteoclasts. In this regard, the enzyme is found to be associated as tight complexes with several signaling transduction proteins, e.g., pp60^c_βrc, mitogen-activated protein kinases (MAPK, also known as Erk) , GTPase-activating proteins (GAPs), and docking proteins, Grb2 and she. PTP-oc dephosphorylated pp60^c-"^rc in vitro, a finding that is consistent with the interpretation that PTP-oc is a key regulator of the phosphorylation level, and also perhaps the activity, of pp60^c_ "^rc in osteoclasts. Consequently, it follows that specific inhibitors of PTP-oc will inhibit osteoclastic activity and as such will be specific inhibitors of bone resorption. Conversely, specific activators of PTP-oc will be specific activators of bone resorption. Accordingly, it is an object of the present invention to the use PTP-oc as a screening target to a) identify specific low molecular weight inhibitors of PTP- oc for the development of osteoclast-specific anti-resorptive agents to be used as pharmacologic therapies for treatments of bone wasting diseases with high bone turnover; and b) identify specific activators of PTP-oc for the development of agents to stimulate bone resorption to treat low bone turnover diseases such as osteopetrosis and aplastic renal bone disease.

SUMMARY OF THE INVENTION

The present invention provides a screening test that is useful for identification of osteoclast-specific anti-bone resorption and/or bone resorption-stimulating pharmaceutical therapies. This screening test is novel in that the screening procedure is based on the signal transduction mechanism that regulates osteoclastic activity. The screening test employs the recently discovered novel osteoclastic PTP (PTP-oc), which is a key regulator of the signal transduction pathway(s) of the osteoclastic activity, as the screening target for osteoclast- specific, anti-resorptive or resorption-stimulating agents. The present invention involves the use of recombinant PTP-oc in high intensity screening tests to search for specific and potent inhibitors and/or activators of this enzyme activity. The identified specific and potent PTP-oc inhibitors and/or activators will then be subsequently evaluated for the efficacy as an inhibitor of bone resorption in organ cultures, animal studies and finally in human clinical trials. This approach should allow for the development of 1 ) orally active and osteoclast-specific anti-resorptive pharmaceutical compound(s) to be used in the treatment of bone-wasting diseases with high bone turnover, such as osteopenia, osteoporosis, Paget' s disease, hyperparathyroidism, osteolytic metastasis, etc; or 2) orally active and osteoclast-specific bone resorption stimulating compound(s) to treat low bone turnover diseases, such as osteopetrosis and aplastic renal bone disease.

BRIEF DESCRIPTION OF DRAWINGS

Fig. 1 is a schematic protocol of a screening assay for PTP-oc inhibitors and activators.

Fig. 2 is a flow diagram of the protocol used to prepare unique, specific cDNA probes for screening for PTP-oc in a rabbit osteoclast cDNA library using the reverse transcriptase polymerase chain reaction approach.

Fig. 3 lists the nucleotide sequence of the unique, osteoclast-specific cDNA probe that was used to molecularly clone PTP-oc cDNA from rabbit osteoclast cDNA library.

Fig. 4 is a flow diagram of the approach for the screening of PTP-oc in a rabbit osteoclast cDNA library.

Fig. 5 is a flow diagram of the sequencing protocol for rabbit PTP-oc cDNA.

Figs. 6a and 6b list the nucleotide sequence and predicted amino acid sequence of the rabbit PTP-oc.

Fig. 7 is a flow diagram of the procedure for producing glutathione S-transferase (GST) -cPTP-oc fusion protein.

Fig. 8 shows the SDS-polyacrylamide gel electrophoresis analysis of recombinant GST-cPTP-oc fusion protein. In this figure, the purified recombinant proteins were electrophoresed by means of 10% SDS-polyacrylamide gel electrophoresis and stained with Coomassie blue dye. Molecular size markers are indicated in kilodalton (kD) . Lanes 1 and 4 show 10 μl of the bacterial lysate before IPTG induction; lanes 2 and 5 show 10 μl of bacterial lysate after the IPTG induction; and lanes 3 and 6 re[resemt 5 μl of the recombinant glutathione S- transerase (GST) and 5 μl of the recombinant GST-CPTP-oc protein, respectively, after purification by glutathione- agarose chromatography.

Fig. 9 shows the expression of PTP-oc mRNA in rabbit bone- derived cells. In this study, 2-20 μg of total RNA from rabbit osteoclasts, osteoblasts, and bone-marrow stromal cells were subjected to electrophoresis in a 1.1% formaldehyde-agarose gel and transferred to a MagnaGraph filter. The RNA blot was hybridized with a [³²P] labelled 5 'PTP-oc cDNA probe (top panel). The same blot was then stripped and rehydridized with GAPDH cDNA (middle panel). The ethidium bromide-stained 18 S 28 S ribosomal RNAs are included to indicate RNA loading (bottom panel) .

Fig. 10 illustrates the tissue-specificity of the expression of PTP-oc mRNA in various rabbit tissues. In this study, 20 μg of total RNA each from rabbit hemopoietic cells, spleen, thymus, heart, brain, liver, lung, kidney, muscle, or stomach were subjected to northern blotting analysis. The blot was hybridized with the 5 '-PTP-oc cDNA probe (i.e., northern blotting analysis) (top panel), then with the GAPDH cDNA probe (middle panel). The ethidium bromide-stained 18 S and 28 S rRNA are also included (bottom panel). Fig. 11 shows that the expression of PTP-oc gene in osteoclasts is regulated by osteo-regulatory effectors in vitro . In this study, isolated rabbit osteoclasts were treated with vehicle control (C) , 10 nM l,25(OH)₂D₃ (1,25D), 10 nM parathyroid hormone (PTH), 1 μM prostaglandin E2 (PGE2), 30 nM calciton (CT) , and 1 μM clodronate (BP), respectively, for 24 hours. Cell extracts (50 μg protein each) of each treatment group were prepared and western immunoblot analysis ( in duplicates) were performed with the anti-PTP-oc antibody. Molecular mass standards (Mr Std) are indicated by the arrows on the left side, and the PTP-oc protein band is shown by the right arrow.

Fig. 12 is a flow diagram of the procedure used to support the conclusion that PTP-oc affects the c-Src-MAPK signal transduction pathway in osteoclasts.

Fig. 13 shows the western immunoblot of cellular phosphotyrosine proteins in rabbit osteoclasts after treatment with various osteo-regulatory effects. The treatment and dosage of the osteo-regulatory effects were identical to those shown in Fig. 11. Cell lysates of isolated osteoclasts (500 μg ) after treatment with corresponding effectors for 24 hours were immunoprecipitated with the anti-PTP-oc antibody. The immunoprecipitated proteins were subjected to western immunoblot analysis using a polyclonal anti-phosphotyrosine antibody [purchased from Promega (Madison, Wisconsin)]. The molecular size marker standards (Mr Std) are shown by the right arrows .

Fig. 14 shows that c-Src, Erk-1 (also known as p44 MAPK) , and Grb-2 were co-immunoprecipitated with PTP-oc. Cell extracts (200 μg each) of isolated rabbit osteoclasts were immunoprecipitated (IP) with the anti-PTP-oc antibody. The immunoprecipitates were then resolved on 10% SDS-polyacrylamide gel electrophoresis and transblotted onto an immobilon membrane. The membrane was then immunoblotted (IB) with the mixture of specific antibodies against c_Src, Erk-1, and Grb-2. The locations of the c-Src, Erk-1, and Grb-2 are indicated by the right arrows.

Fig. 15 illustrates that PTP-oc was co-immunoprecipitated with phosphotyrosine proteins, c-Src, and Erkl . Osteoclast cell extracts (200 μg each) were first immunoprecipitated with anti-phosphotyrosine antibody (left panel), anti-Erkl antibody (middle panel), and anti-c-Src antibody (right panel), respectively. The immunoprecipitated proteins from each treatment were immunoblotted (IB) with the anti-PTP-oc antibody.

DESCRIPTION OF SPECIFIC EMBODIMENTS

The present invention provides a unique screening approach to identify specific and potent low molecular mass agents, that can be used to develop orally active, potent, and osteoclast- specific anti-resorptive therapies for bone-wasting diseases, such as but not limited to osteopenia, osteoporosis, Paget' s diseases, hyperparathyroidism, osteolytic metastasis; and/or orally active, potent, and osteoclast-specific resorption- stimulating therapies for low bone turnover diseases such as osteopetrosis and aplastic renal bone disease. Specifically, the invented method targets at the molecular mechanism of bone resorption and involves the use of a novel osteoclast-specific PTP (PTP-oc), which is a key regulator of the signal transduction pathway that control osteoclastic activity, as a screening target to identify specific low molecular mass (i.e., orally active) inhibitors (for anti-resorptive therapies) and activators (for resorption-stimulating therapies) of PTP-oc.

The invention is based on two important principles. First, PTP-oc is a novel, osteoclast-specific enzyme. Second, this osteoclastic enzyme (PTP-oc) is an important regulator of the signal transduction mechanism of osteoclastic activity; and therefore, is an important regulator of the bone resorption process in vivo.

The principle that the PTP-oc is a novel osteoclastic enzyme is described in the June 1996 article entitled "Molecular cloning and expression of a unique rabbit osteoclastic phosphotyrosyl phosphatase" , Biochem, J. (1996), 515-523 (Printed in Great Britain) and is based on the following new discoveries: 1. The molecular cloning and nucleotide sequence determination of the PTP-oc revealed that PTP-oc differs from the known PTPs in nucleotide and predicted amino acid sequences and in its molecular structure. 2. PTP-oc has a unique molecular structure. PTP-oc is a transmembrane PTP. However, it differs from the known transmembrane PTPs in that 1) it has a very short extracellular domain, 2) it lacks a signal peptide proximal to the N-terminus, 3) the intracellular portion of the enzyme contains only a single ' PTP catalytic domain' , and 4) it is the smallest transmembrane PTPs (i.e., approximately 50,000 daltons) known to-date. The concept that the PTP-oc is specific for osteoclasts is based on the observations that the PTP-oc mRNA (3.8 kb) is expressed primarily in osteoclasts and in the spleen, which contains osteoclast precursors. Although PTP-oc mRNA is also expressed in the kidney and the brain, these two tissues are also found to express a much larger (6.5 kb) slicing variant mRNA species, which is absent in osteoclasts and the spleen. Other studies [Thomas et al . , J. Biol. Chem. 269:19953-19962, 1994] indicate that the 6.5 kb (which is referred to as GLEPP1) is the predominant species in the kidney. Thus, it is envisioned that the kidney and the brain produce the larger protein species (approximately 80,000 daltons) corresponding to the 6.5 kb, whereas the osteoclasts and precursors synthesized the smaller (50,000 daltons) mature PTP-oc protein. The concept that PTP- oc is a regulator of the signal transduction mechanism that regulates osteoclastic activity is based on the recent in vitro discoveries which indicate that 1) the expression of PTP-oc mRNA and mature protein in isolated osteoclasts is regulated by the osteo-regulatory hormones and effectors; 2) the PTP-oc in osteoclasts forms complexes with the tyrosyl phosphorylated forms of several key signal transduction proteins, such as pp60^c-"^rc, MAPKs, GAPs, and Grb2 and She; 3) the amounts of tyrosyl phosphorylated species of these signaling proteins that are associated with the PTP-oc are regulated by the same osteo- regulatory hormones and effectors which are known regulators of bone resorption; and 4) the tyrosyl phosphorylated pp60^c"^βrc peptide is a substrate for the recombinant PTP-oc in vitro .

The screening protocol for inhibitors and activators of PTP-oc is schematically shown in Fig. 1. To obtain continuous supplies of a large quantity of mature and functionally active PTP-oc protein for such high intensity screening tests, the recombinant PTP-oc protein can be prepared by a suitable (known or newly developed) recombinant DNA protocol using the in-frame PTP-oc cDNA as an insert or probe DNA and an appropriate (known or newly discovered) expression vector in a suitable host (bacteria or otherwise) system. An example of the procedure for preparing the recombinant DNA is described in Example 3 below while the procedure for obtaining full length PTP-oc cDNA is described in specific Example 1 below. To facilitate the purification of the recombinant PTP-oc protein, the recombinant enzyme can be synthesized as a fusion protein (for example but not limited to, glutathione-S-transferase fusion proteins, PTP- oc-His_x fusion proteins) using standard procedures for enzyme isolation and purification. For example recombinant PTP-oc fusion proteins can easily be isolated with excellent yields with a single chromatographic purification step using standard techniques (e.g., (a) glutathione-sepharose column). The recombinant, functionally active, PTP-oc is then used in a high intensity screening test to identify specific inhibitors using an appropriate in vitro enzymatic assay. In that regard, there are many ways to assess PTP-oc enzymatic activity; and several examples are given in detail in Example 6 below. Specific inhibitors of PTP-oc are defined as compounds that specifically inhibit the enzymatic activity of PTP-oc (i.e., its ability to hydrolyze the phosphotyrosine moiety) with a sufficiently high affinity and specificity; and specific activators are those compounds that stimulate the PTP activity of PTP-oc with high affinity and specificity.

Following identification of specific inhibitors (or activators) by the high intensity screening test (e.g. Example 6 ) , such inhibitor and activators may be tested for their ability to inhibit (or stimulate) bone resorption in osteoclast cultures and bone organ cultures, in animal (e.g., rodents, dogs, etc.) studies, and eventually in clinical trials for humans .

The invention now being generally described, the same will be better understood by reference to the following specific examples which are provided for purposes of illustration only and are not to be considered limiting of the invention unless specified. EXAMPLE 1 MOLECULAR CLONING OF PTP-OC A study to molecularly clone a unique osteoclastic PTP was performed with a λZAPII rabbit osteoclast cDNA library. Rabbit osteoclast cDNA library was constructed with poly-A containing RNAs isolated from purified rabbit osteoclasts using standard procedures. Our rabbit osteoclasts were prepared with a procedure (procedure I) that was modified from the one developed by Tezuka et al . [Biochem. Biophys. Res. Commun. 186:911-917, 1992]. Briefly, long bones and calvaria were dissected from 100-150 g baby rabbits in a sterile hood. The isolated bones were then minced into small pieces with a pair of sterile scissors in a DMEM culture medium. The resulting bone chips were gently vortexed for one minute and allowed to sediment for four minutes. The procedure was repeated three times. This procedure releases osteoclasts from bone chips into the culture medium. After the final sedimentation, the cell in the supernatant which contained osteoclasts and stromal cells was plated in 10 ml each of DMEM supplemented with 10% fetal bovine serum in a 90-mm culture dish. Cells were then allowed to attach to the culture dish for three hours. The culture medium was then changed to fresh DMEM supplemented with 10% fetal bovine serum; and the cultures were allowed to grow overnight at 37°C in an atmosphere containing 5% C0₂. After 24 hours, the non-adherent, non-osteoclasts were removed by changing the culture medium with fresh DMEM containing 10% fetal bovine serum. Three hours later, stromal cells were removed by a treatment with 0.001% pronase E and 0.02% EDTA for 10 minutes. The remaining cells (i.e., the osteoclasts) were maintained in DMEM supplemented with 10% fetal bovine serum. Rabbit osteoclasts were identified by their multinucleated feature and by positive staining for tartrate-resistant acid phosphatase activity. The resulting purified rabbit osteoclasts are of greater than 95% purity.

A reverse-transcriptase polymerase chain reaction (RT-PCR) was performed to prepare a unique and specific cDNA probe (Procedure II) to be used to screen for PTP-oc in a cDNA library. The protocol is shown in Fig. 2. Two degenerate oligonucleotide primers (the sense oligo: 5' - AARUGYSMNCARUAYUGGCC and the anti-sense oligo: 5' - CCNAYRCCBGCRCTRCAGT) which corresponded to two conserved motifs [i.e., KC(A/D/H)QYWP and ( I/V) HCSAG(V/A)G, respectively] within the ' PTP catalytic domain' of mammalian PTPs were synthesized by the Center of Molecular Biology and Gene Therapy at Loma Linda University (Loma Linda, California), and were used in RT- PCR with total rabbit osteoclast RNA as the template. Total RNA was isolated from isolated rabbit osteoclasts with a single-step guanidinium thiocyanate extraction procedure described by Chomczynski and Sacchi, [Anal. Biochem. 162:156- 159, 1987]. Reverse transcription reaction (procedure Ila) was initiated by adding 200-500 units of Superscript II reverse transcriptase (or other reverse transcriptases ) to a 20 μl reaction mixture containing 50 mM Tris/HCl (pH 8.3-8.8), 75 mM KC1, 3 mM MgCl₂, 0.5 μg to 2 μg of total rabbit osteoclast RNA, 4-10 μM of the antisense oligo primer, 20-50 units of RNAsin, 10 mM dithiothreitol, and 0.5-2 mM each of the four dNTP. The reaction was carried out at a temperature no less than 36°C and no higher than 46°C for 10 to 30 minutes. The resulting cDNA was purified by standard phenol-chloroform extraction, followed by standard ethanol-ammonium acetate precipitation. The purified cDNA was used in the PCR.

The PCR (procedure lib) was carried out in a 50 μl reaction mixture containing 10 mM KC1, 20 mM Tris/HCl (pH 8.3- 8.8), 10 mM (NH₄)₂S0₄, 2 mM MgS0₄, 0.1% Triton X-100, 1-5 units of Vent (exo^") DNA polymerase [or other thermally stable DNA polymerase], 0.8 mM-1 mM each of dNTP, 0.1 mg/ml bovine serum albumin, 4 μM each of the sense and the antisense oligos, and the purified cDNA product generated from the reverse transcriptase reaction. Before PCR, the reaction was subjected to a hot start for 5 minutes at 100°C, followed by 2 minutes at 49°C, before the addition of Vent (exo^") polymerase. The polymerase chain reaction was performed for 25-35 cycles of extension for 1 minute at 74°C, denaturing for 1 minute at 94°C, and annealing for 1 minute at 49°C using standard procedures. The resulting PCR products were purified by phenol-chloroform extraction, blunt-ended by Klenow and subcloned into the Smal- digested pUC119 cloning vector. The amplified plasmids containing the PCR products were isolated by either the O 98/54357

modified alkaline lysis method [Hackett et al.. An introduction to recombinant DNA techniques: basic experiments in gene manipulation, 2nd ed, 1988] or with a QAIGEN (Chatsworth, California) plasmid isolation kit . The complete nucleotide sequence of the insert of the isolated plasmids were determined by the standard dideoxynucleotide chain-termination method described by Sanger et al . , [Proc. Natl. Acad. Sci. U.S.A. 74:5463-5468, 1977]. The complete nucleotide sequence of the insert (i.e., the PCR product) is shown in Fig. 3.

The insert nucleotide sequence information indicates that this PCR product encoded a novel member of the PTP family. The insert of the subclone of this PCR product was used as the specific probe to screen the λZAPII rabbit osteoclast cDNA library (procedure III). A schematic approach is shown in Fig. 4. The PCR probe was labelled with [α-³²P]CTP by the random prime labeling approach often used in laboratory procedures described by Feinberg and Volgelstein, [Anal. Biochem. 132:6- 13, 1983]. Twenty-five cDNA library phage plates (in agar) , each with 20,000 plaques/100 mm², were prepared. "Nitrocellulose filter lifts" of each plate were prepared by gently pressing a nitrocellulose filter paper onto the surface of each cDNA library phage agar plate, which allowed transferring some of the cDNA phage from each bacterial clone onto the nitrocellulose filter. Each "filter lift" was washed with 5 X SSC, 0.5% SDS and 1 mM EDTA (pH 8) at a temperature between 36°C to 46°C for two hours to lyse the bacteria before hybridization. O 98/54357

[SSC: 0.15 M NaCl, 0.015 M sodium citrate]. The washed nitrocellulose filters were hybridized overnight by heating them at a temperature between 36°C to 46°C in a hybridization buffer consisting of 5X SSPE, 5 X Denhardt' s solution, 0.1% SDS, 100 μg/ml salmon sperm DNA and 10⁶ cpm/ml of the [³²P]labelled PCR cDNA probe. [SSPE consisting of 0.15 M NaCl/10 mM sodium phosphate (pH 7.4 )/l mM EDTA] . The filters were washed at least four times with 2 X SSC/0.1% SDS at temperature between 20°C and 37°C, each for 10-20 minutes, followed by washing once in 1 X SSC/0.1% SDS at 60-70°C for 1-2 hours. The final two washes were with 0.2 X SSC/0.1% SDS at 60-70°C each for 1-2 hours. The positive clones containing the [³²P] were then isolated from the corresponding clones in the original cDNA library plate and the cDNA inserts were recovered from λZAPII phage vector by in vivo excision according to the instruction of the manufacturer Strategene (La Jolla, California) . The cDNA inserts were purified and subcloned into pBluescript SKII(-) plasmid (purchased from Strategene) and inserted into a competent Escherichia coli strain using standard procedures. The pBluescript-derived plasmids containing osteoclastic cDNA inserts were amplified, and the insert was isolated for cDNA sequencing determination. The complete sequence of both strands of the cDNA insert in the plasmid were sequenced by the didoxy chain-termination method using the T3 and T7 universal primers as the initial sequencing primers. The sequencing process was repeated with additional oligonucleotide primers; the sequences of which were clerived from the determined sequence in previous cDNA sequencing cycles. This process was continued until the complete nucleotide sequence of the cDNA of this PTP, referred to as PTP-oc, was determined and confirmed by sequencing both DNA strands. The strategy for the PTP-oc cDNA sequencing protocol is shown schematically in Fig. 5. The complete nucleotide sequence of the cDNA of this PTP and the corresponding predicted amino acid sequence are shown in Fig. 6a and 6b.

EXAMPLE 2 DESCRIPTION OF STRUCTURAL CHARACTERISTICS OF PTP-oc Following standard laboratory techniques, a single open reading frame (ORF) of 1215 bp was identified in the above described cDNA insert, 3623 bp in size. The PTP-oc cDNA sequence had two in-frame methionine codons (ATG) , separated by nine nucleotides at the beginning of the ORF. The nucleotide sequence around the first ATG conforms to the canonical translation initiation signal, (A/G)CCATGG. The putative initiation methionine codon is preceded by several in-frame termination codons. The ORF was preceded by a 5' -untranslated region (UTR) of 391 bp in length and followed by a relatively long (1917 bp) 3' -UTR with stop codons in all three ORFs and a putative polyadenylation signal (AATAAA) . Subsequent analysis indicated that it is a full-length cDNA clone. Southern analysis of rabbit genomic DNA with specific restriction enzymes indicated that the PTP-oc gene exists as a single copy gene .

The ORF predicted a protein of 405 amino acid residues with a calculated molecular mass of 47,276 daltons. The hydropathy index computation analysis identified a single stretch of strongly hydrophobic amino acid residues (indicated in Figure 6 by underlining) , a characteristic of a transmembrane domain. These 25 hydrophobic residues are followed by several basic residues (i.e., KKK) , that are consistent with the stop-transfer signal associated with the membrane-spanning domains. These characteristics are compatible with the conclusion that PTP-oc is a member of the transmembrane PTP family. However, unlike the transmembrane " receptor-like" PTPs , the PTP-oc lacks at the N-terminus a signal peptide that targets the protein to the secretory pathway. In addition, the extracellular domain is very short (8 amino acid residues) and the cytoplasmic region is comprised of only 372 residues. Several consensus motifs, including the two regions that were used to design degenerate oligonucleotides for RT-PCR, were identified in the cytoplasmic region. The protein has no SH2 domain. The ' signature sequence' of the active site (i.e., HCSAGVGRTG) , which is highly conserved among PTPs in vertebrates was found within the putative 'catalytic domain' of PTP-oc. Thus, a cDNA of a unique PTP (i.e., PTP-oc) has been cloned and completely sequenced from an osteoclast cDNA library. Although the cDNA described herein above was originated from the rabbit, homologues of PTP-oc cDNA from other species, including humans, can easily be isolated and used for the production of the recombinant PTP-oc (Example 3). Human osteoclasts can be prepared according to Fujikawa et al . [Endocrinology 137:4058-1060, 1996]; and rodent osteoclasts may be isolated according to Amano et al . [Calcif. Tissue Int. 57:367-370, 1995]. Osteoclast cDNA libraries of other species can thus easily be prepared. Accordingly, PTP-oc homologs from other species, including humans, can easily be cloned.

EXAMPLE 3 PRODUCTION OF PTP-oc FUSION PROTEIN A bacterial fusion protein (referred to as GST-cPTP-oc) that contained a putative catalytic domain of the PTP-oc linked to the glutathione-S-transferase (GST) protein was produced. The procedure to produce cPTP-oc fusion protein is shown schematically in Fig. 7. Briefly, the PTP-oc 0.9 kb cDNA fragment encoding the PTP catalytic domain (nucleotides 665- 1538) was prepared using standard laboratory techniques. The fragment was cloned in-frame to the gene coding for GST in the pGEX-5X-3 vector using standard techniques (procedure IV) . In this case, because EcoRI site was missing, EcoRl linkers were added at the 5' end of the catalytic domain in the PTP-oc cDNA clone to introduce a novel .EcoRI site. The 0.9 kb DNA fragment containing the putative catalytic domain was released from the O 98/54357

J?coRI-mutated PTP-oc cDNA clone by digestion with restriction enzymes, EcoRl and PvuII according to standard procedures. The pGEX-cPTP-oc construct was generated by ligation of the £.co.RI/SΛaI-linearized pGEX-5X-3 vector to the 0.9 kb cPTP-oc (PTP-oc catalytic domain) cDNA fragment. The reading frame of the pGEX-cPTP-oc construct was confirmed by DNA sequencing using standard procedures. The vector for full-length PTP-oc- pGEX construct can be prepared with similar approach using different restriction enzymes. Vectors for PTP-oc and other fusion proteins (e.g., PTP-oc-His.. fusion protein) can also be produced with a similar approach.

To produce fusion proteins (procedure V) , the pGEX-5X-3 (producing GST protein as a negative control) and the pGEX- cPTP-oc plasmids were each transfected into HB101 Escherichia coli cells. Other competent host cells can also be used. Tenfold diluted overnight cultures of transformed HB101 Escherichia coli cells were grown to mid-logarithmic phase at 30°C or 37°C, and expression of fusion proteins was induced with 0.1 mM isopropyl β-o-thiogalactopyranoside (ITGP) for 5-6 hours. After induction, the bacterial cultures were harvested by centrifugation at 3000 x g for 10 minutes. Cell pellets were resuspended in a 1/50 culture volume of the lysis buffer consisting of 20 mM Hepes (pH 7.6), 100 mM KC1, 0.2 mM EDTA, 20% (v/v) glycerol, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 10 μg/ml leupeptin, 10 μg/ l pepstatin, and 10 μg/ml antipain. The GST and GST-PTP-oc fusion proteins were batch-purified using a glutathione- sepharose 4B column. The purified proteins could be stored either at 4°C or -20°C until assayed for enzymatic activity.

SDS/PAGE analysis of the fusion proteins containing PTP-oc revealed that a protein band corresponding to the predicted size of the GST-cPTP-oc fusion protein was seen in the lysate of IPTG-induced pGEX-cPTP vector-containing transformants (lane 5 and 6 in Fig. 8). The GST-cPTP-oc fusion protein was purified with glutathione-sepharose affinity chromatography (lane 6 in Fig. 8). GST protein was produced and purified as a control vector following induction of pGEX-5X-3 vector- containing bacteria by IPTG (lanes 1-3 in Fig. 8).

Although this specific example describes the production of cPTP-oc fusion proteins, the same approaches have been used successfully to produce full-length GST-PTP-oc fusion proteins. However, because the highly hydrophobic nature of the full- length GST-PTP-oc (i.e., PTP-oc is a transmembrane protein), the resulting fusion protein was shown to be insoluble, such that non-ionic detergents, such as Triton X-100, and phospholipids are required to assist the solubilization of the full-length PTP-oc protein products.

EXAMPLE 4 EVIDENCE THAT PTP-oc IS A PHOSPHOTYROSINE PHOSPHATASE To confirm that PTP-oc indeed encodes a PTP, the PTP-oc fusion proteins were tested for phosphotyrosine phosphatase _«,_{« «}„ O 98/54357

activities. Because previous studies showed that PTPs have significant activities on para-nitrophenyl phosphate (pNPP), tyrosyl-phosphorylated peptides, and phosphotyrosyl proteins in vitro, the phosphatase activity of the fusion protein was measured at neutral pH using three artificial substrates: (a) pNPP, (b) tyrosyl-phosphorylated Raytide, and (c) tyrosyl- phosphorylated histone. The enzyme assays used in this study are described in detail in Example 6 below and in Wu et al . [Biochem. J., 316:515-523, 1996]. The GST-cPTP-oc fusion protein was highly efficient in dephosphorylating pNPP, and that the dephosphorylation was proportional to the amount of the fusion protein and to the reaction time. GST alone had no activity on pNPP or any of the phosphotyroslne containing substrates. Similarly, the GST-cPTP-oc, but not the GST, was effective at dephosphorylating the [³²P]tyrosyl-phosphorylated Raytide in a time- and dose-dependent manner. While the PTP fusion protein displayed strong phosphatase activity toward [³²P]tyrosyl-phosphorylated histone, it was totally ineffective in dephosphorylating [³²P]seryl-phosphorylated casein under the same conditions. These findings confirm that the PTP-oc gene indeed encodes a PTP activity.

EXAMPLE 5 PTP-OC IS A UNIQUE, OSTEOCLAST-SPECIFIC PTP Comparison and alignment of the predicted amino acid sequence of PTP-oc with the existing protein sequences of PTP- oc with the existing protein sequences in the SWISS-PROT data bank using the FASTA program indicated that the PTP-oc is a novel protein. In addition, unlike most known mammalian transmembrane PTPs (with the exception of HPTPβ), which have two tandem catalytic domains in their cytoplasmic portions, the PTP-oc contains only a single ' catalytic domain' . Because the catalytic domains of cytosolic and transmembrane PTPs are highly conserved, the core phosphatase domain of this PTP-oc was compared with those of the known human receptor-like and cytosolic PTPs. The predicted catalytic domain of PTP-oc also shares a very low degree of (less than 50%) sequence identity with that of the known human PTPs. This low level of identity is unlikely to be attributed to species variance, since species differences usually account for less than 10% sequence variance among PTP homologues. For example, human RTPTγ and its murine homologues are 95% identical in amino acids with 90% identity in nucleotide sequence. Thus, the sequence comparison indicates that PTP-oc is a novel PTP.

The molecular structure of PTP-oc is unique and differs from other known mammalian PTPs in four structural properties: 1) PTP-oc is one of the smallest 'transmembrane' PTPs, since it is much smaller (approximately 50,000 daltons) than most, if not all, transmembrane PTPs, whose molecular size is in the range of 80,000 to 250,000 daltons. 2) Unlike other transmembrane PTPs, it lacks a signal peptide proximal to the N-terminus, which targets the protein to secretory vesicle. 3) Unlike the known transmembrane PTPs (except HPTPβ), the intracellular portion of the PTP-oc contains only a single catalytic domain rather than two tandem repeats. 4) In contrast to the transmembrane PTPs whose extracellular domain is relatively large (several hundred residues in length) and contains multiple glycosylation sites, immunoglobulin-like and/or fibronectin-III domains, or carbonic anhydrase-like domains, the extracellular domain of PTP-oc is very short (8 residues in length) and lacks these domains. Taken together, these distinct molecular features indicate that PTP-oc is a novel transmembrane PTP.

To examine whether expression of PTP-oc mRNA is specific to osteoclasts but not other bone-derived cells, a 1.6 kb 5' cDNA probe was used as a probe in the Northern blot analysis of total RNA isolated from osteoclasts, osteoblasts, and bone- marrow stomal cells of 10-day-old rabbits. Figure 8 shows that a major transcript of approximately 3.8 kb was found in osteoclasts, but not in osteoblasts or stromal cells, indicating that the expression of the PTP-oc mRNA is specific for osteoclasts and not for other bone-derived cells. It should be pointed out that much less osteoclast RNA (lane 1 in the bottom panel of Fig. 9) than RNAs from osteoblasts and stromal cells (lanes 2-5 in the bottom panel of Fig. 9) has been loaded on to this gel. Accordingly, the lack of the 3.8 kb transcript in osteoblasts and bone-marrow stromal cells was not due to insufficient RNA loading. These findings also indicated that this PTP-oc mRNA was expressed in high abundance in osteoclasts. Considering the size of the poly(A) tail in mouse mammalian mRNA (which is approximately 200 bp in length) , the mRNA size of the PTP-oc cDNA clone (3.8 kb) corresponded well with the determined length of the PTP-oc cDNA clone (3623 bp) . Thus, the isolated PTP-oc cDNA is a full-length cDNA.

To investigate the tissue-specific expression pattern of the PTP-oc mRNA, the same 5' cDNA probe was used as a probe in multi-tissue Northern blot analysis. Among the tissues tested so far, a transcript of 3.8 kb was found only in the brain, kidney and spleen, in addition to osteoclasts (Fig. 10). However, the 3.8 kb transcript in the spleen was less abundant than that in the brain and kidney. The brain and kidney also expressed, in addition to this 3.8 kb transcript, a related, but larger (6.5 kb) transcript. The spleen, like osteoclasts, did not express this 6.5 kb transcript (Fig. 10). Neither the 3.8 kb nor the 6.5 kb transcript was found in the other test tissues, i.e., hematopoietic blood cells, thymus, heart, lung, liver, stomach, and muscle (Fig. 10). To confirm the findings of the tissue-specific expression of PTP-oc, Northern blot analysis using the 3' probe, which corresponded to the 3' end of the PTP-oc cDNA was obtained; and identical results to those obtained with the 5' probe were obtained. Consequently, these findings indicate that the expression of the PTP-oc mRNA is restricted to the osteoclasts, the spleen, the kidney, and the brain. However, it should be noted that the kidney and brain (but not spleen and osteoclasts) expressed, in addition to the 3.8 kb PTP-oc mRNA transcript, also a larger 6.5 kb related mRNA transcript. Osteoclast progenitor cells are derived from the splenic hematopoietic stem cells. Accordingly, that the PTP-oc mRNA is present in the spleen and osteoclasts suggest that the expression of the PTP-oc (and not the 6.5 kb mRNA transcript) is unique to osteoclasts and precursor cells.

A 6.5 kb transcript, referred to as GLEPPl, has been cloned by Thomas et al . [J. Biol. Chem. 269:19953-19962, 1994] from a rabbit kidney cDNA library. Comparison of the nucleotide and predicted amino acid sequences of the kidney GLEPPl with those of the PTP-oc revealed that the intracellular domains of these two isoenzymes shared approximately 90% sequence identity, indicating that these two proteins are related. The kidney enzyme differs from the osteoclastic PTP- oc in that the kidney enzyme contains a large extracellular domain, comprising eight repeats of an FN-III-like motif and 15 putative N-glycosylation sites, whereas PTP-oc contains only a very short (8 residues) extracellular domain. Most importantly, PTP-oc contains a unique region of 28 amino acids (residues 66-93 of Fig. 6a and 6b, shown as bold and italic letters) inserted in the intracellular domain, which is absent in the kidney isoenzyme. Thus, PTP-oc may be a truncated version of the kidney GLEPPl, presumably resulting from an alternative splicing at the 5' end of the gene. The presence of a splicing variant in osteoclasts and precursor cells (spleen cells) could play a tissue-specific regulatory role. To support this conclusion, Thomas et al . have reported that the 6.5 kb mRNA and the corresponding protein product are the predominant species in the kidney. Accordingly, it is likely that the larger GLEPPl is synthesized by the kidney and brain whereas the synthesis of the smaller PTP-oc protein is unique to osteoclasts and its precursors.

EXAMPLE 6 ENZYMATIC ASSAYS FOR PTP-oc ACTIVITY In order to identify inhibitors and/or activators of PTP- oc, an enzymatic assay for PTP-oc is needed for a screening test. In a typical enzymatic screening assay, a phosphotyroslne containing protein, peptide, or compound is used as the substrate, and the rate of hydrolysis of the phosphotyrosine residue, which may be determined by measuring the release of inorganic phosphate or the amount of dephosphorylated tyrosine residue, is followed. Inhibitors of PTP-oc are those compounds which inhibit the ability of the recombinant PTP-oc to hydrolyze the phosphotyrosine containing substrate; whereas activators are those compounds which stimulate the recombinant PTP-oc mediated hydrolysis of the phosphotyrosine-containing substrate. Because the phosphorylated tyr-527 of pp60^c"^β,:c appears to be a physiological substrate for PTP-oc, a suitable substrate would be a synthetic peptide corresponding to the tyr-527 site of pp60^c-"^rc, in which the tyr-527 residue is tyrosyl phosphorylated; this substrate is referred to as c-src peptide in this application. Other phosphotyrosine containing proteins or peptides may also be used. In addition, most mammalian PTPs, including PTP-oc, are able to hydrolyze low molecular weight compounds , such as p-nitrophenyl phosphate (PNPP), phosphotyrosine, etc., which resemble the phosphotyrosine moiety, at both neutral pH and an acidic pH; these low molecular weight phosphotyrosine containing compounds can also be used as the substrate at both neutral and acidic pHs. The activity of PTP-oc is stimulated by reducing agents, such as β-mercaptoethanol or dithiothreitol. Thus, the assay buffer may contain a reducing agent. The ionic strength of the buffer must be sufficiently high to maintain a neutral pH. The buffer can be Hepes, Tris, or any buffers whose pKa is at around neutral pH. For an acidic pH, sodium citrate or sodium acetate, or other buffers with an acidic pKa, may be used. Because of the hydrophobic nature of the enzyme (i.e., it is a transmembrane protein), salts, such as NaCl, and non-ionic detergents, such as Triton X-100, and/or phospholipids may be needed to increase the solubility of the recombinant enzyme. The enzyme activity is monitored by the hydrolysis of phosphotyrosine residue. Thus, the enzyme activity may be monitored by following the rate of release of inorganic phosphate, or the production of dephosphorylated products. Convenient radioactive assays can be radioactive (e.g., with [³²P] labelled phosphotyrosyl protein or peptide as the substrate), fluorometric (using a fluorescent phosphotyrosyl containing substrate), spectrophometric (using a chromophore- containing substrate, such as PNPP or phosphotyrosine) , or immunological (e.g., using specific antibody against phosphotyrosine) . The enzymatic assay may be automated to be carried out by robotic equipment.

A typical screening assay for inhibitors or activators of PTP-oc is as follows: an aliquot of the test compound is added to an assay mixture containing 100 mM Hepes, pH 6.5-7.5, 100 mM NaCl, 10 mM dithiothreitol, 5 mM EDTA, 0.1% Triton X-100, 1 mg/ml bovine serum albumin (to increase stability of highly purified fusion proteins), and [³²P] labelled tyrosyl phosphorylated c-src peptide (20,000 cpm per assay). The tyrosyl phosphorylated c-src peptide is prepared by phosphorylating the synthetic c-src peptide with Csk tyrosyl kinase as previously described [Chan et al . , J. Biol. Chem. 261:9890-9895, 1986]. The reaction is initiated with the addition of a fixed amount of the recombinant PTP-oc, and the dephosphorylation reaction is carried out at 37°C for 10-20 min. The reaction is terminated by addition of a fixed volume of ice-cold 10% trichloroacetic acid, and [³²P]inorganic phosphate is either extracted with ammonium molybdate or with acidified charcoal [0.9^{^} M HC1/90 mM sodium pyrophosphate/2 mM NaH₂P0₄, and 1.4% (w/v) Norit A], and counted in a liquid scintillation counter. A control in which a solvent instead of the compound to be screened is added is included for comparison. An effective inhibitor would be the compound which significantly inhibits the release of [³²P] inorganic phosphate from the [³²P]tyrosyl phosphorylated c-src peptide; whereas an activator would significantly stimulate the release of [³²P] inorganic phosphates compared to the vehicle control. A dose-response curve for each compound will be obtained to determine relative affinity. To determine specificity, the ability of the test compound to inhibit or activate the PTP activity of the recombinant PTP-IB (a non-specific PTP) is included for comparison.

Many other assays have also been used successfully to assess the enzymatic activity of mammalian PTPs, including PTP- oc. The following specific examples are provided for purposes of illustration and not be considered limitations of the invention.

1) Determination of hydrolysis of tyrosyl phosphorylated peptide with continuous spectrophometric and/or fluorometric assays. The chemical synthesis of tyrosyl phosphorylated peptides has been described by Zhang et al . [Anal. Biochem. 211:7-15, 1993]. The assay conditions are identical to those described in above. An appropriate amount of tyrosyl phosphorylated peptide is preincubated at the assay temperature for at least 15 minutes, and the enzymatic reaction is initiated with addition of an aliquot of recombinant PTP-oc. Because the dephosphorylated peptides have a much greater extinction coefficient at 282 nm than the tyrosyl phosphorylated peptides, the PTP-oc-catalyzed hydrolysis of tyrosyl phosphorylated peptide is determined by continuously measuring the increase in absorption at 282 nm in a spectrophotometer until completion. The slope of the progression curve is used to calculate the enzymatic activity. Because the fluorescence property of the tyrosine moiety of the dephosphorylate peptides, identical assay conditions can also be followed fluorometrically by continuously monitoring the increase in tyrosine fluorescence at 305 nm with excitation at 280 nm.

2) PNPP as the substrate: The reaction mixture consists of 20 mM Tris-HCl (pH 7.0-7.5), 0.1% β-mercaptoethanol (or 5 mM dithiothreitol), 0.05%-0.2% Triton X-100, 0.1 mg/ml bovine serum albumin, an aliquot of recombinant PTP-oc, and 5-20 mM PNPP. NaCl (100 mM) may be added. The reaction (i.e., hydrolysis) is carried out at a temperature from 25°C to 37°C and for 10 to 60 minutes. The reaction is terminated by addition of NaOH, and the rate of hydrolysis is determined by measuring the absorbance at 410 nM, and a molar extinction coefficient of 1.78 x 10⁴ M^.cm"¹ is used to calculate the amount of p-nitrophenolate ion produced. Because most PTPs have activity towards PNPP at an acidic pH, the reaction can also be performed in a buffer with a pH of 5 to 6.

3) Phosphoprotein as the substrate: Tyrosyl phosphorylated proteins, such as, histone, casein, IgG, pp60^c"^erc, etc. can be used. The protein substrate is phosphorylated with [³²P]ATP by a PTK, such as pp60^c"8rc. An example for the preparation of tyrosyl phosphorylated proteins is shown by Chan et al . [J. Biol. Chem. 261:9890-9895, 1986]. [³²P]Labelled tyrosyl phosphorylated proteins are then separated from [³²P]ATP by a number of approaches, such as precipitation with trichloroacetic acid followed by dialysis, gel filtration, ion- exchange chromatographies, etc . Dephosphorylation of tyrosyl phosphorylated proteins is performed under the same conditions as those described for [³²P]c-src peptide in above. To increase the sensitivity of the assay, a greater amount of radioactivity (e.g., up to 50,000 cpm) may be needed. The reaction is terminated by addition of trichloroacetic acid and bovine serum albumin as carrier proteins. Following centrifugation, the release of [³²P] inorganic phosphate is determined by spotting an aliquot of the supernatant onto a Whatman (or related) filter paper, and the radioactivity is measured by liquid scintillation counting or Cerenkov counting. The amount of hydrolysis is then calculated based on the specific radioactivity of the [³²P]ATP used to prepare the tyrosyl phosphorylated proteins .

4) Phosphotyrosine as the substrate: the hydrolysis of phosphate residue from the phosphotyrosine can be assessed by measuring the formation of tyrosine residue. In this assay, the reaction mixture contains 5 to 20 mM phosphotyrosine and other aforementioned cofactors. The reaction can be carried out in a buffer with a pH between 5 and 7.5. The reaction can be carried out at a temperature between room temperature and 37°C. The release of tyrosine is determined according to the method of Lowry et al . [J. Biol. Chem. 193:265-275, 1951]. However, the rate of hydrolysis can also be measured by determining the release of inorganic phosphate, which is assayed with a colorimetric assay [e.g., Fiske, C.H., and Subbarow, Y. J. Biol. Chem. 66:375-400, 1925].

EXAMPLE 7 REGULATION OF PTP-OC EXPRESSION BY OSTEO-REGULATORY AGENTS Two studies were performed to evaluate the concept that the expression of PTP-oc in isolated osteoclasts is regulated by the osteo-regulatory hormones. In the first study, the effects of the 24-hour treatment with an effective dose of prostaglandin E2 (3 nM) , a bone resorption activator, and calcitonin (10 nM) , a potent bone resorption inhibitor, on the expression of PTP-oc mRNA in isolated rabbit osteoclasts were measured by quantitative RT-PCR, using a sense PTP-oc 19- oligomer and an antisense PTP-oc 20-oligomer, respectively, as the PCR primers. The PCR reaction was carried out for 26 cycles using the Vent (Exo^") polymerase: denaturing for 45 seconds at 94°C, annealing for 45 seconds at 60°C, and extension for 2 minutes at 74°C. The resulting PCR products on a Northern blot were probed with a [³²P] labelled PTP-oc antisense 23-oligomer. To assess the integrity and the loading of RNA sample in the reverse transcriptase reaction, each of the reverse transcribed samples was also PCR amplified for 24 cycles for the glyceraldehyde-3-phosphate dehydrogenase transcript (GAPDH, a house keeping gene) using a commercial amplifier kit. The relative level of PTP-oc PCR transcript was normalized against that of the GAPDH PCR transcript. The quantitative RT-PCR analysis in this study revealed that the prostaglandin E2 treatment increased PTP-oc mRNA expression in isolated osteoclasts by 2- to 3-fold; whereas calcitonin showed no significant effects. Thus, it is concluded that the PTP-oc gene expression in osteoclasts is regulatory in that bone resorption activators (e.g., prostaglandin E2 ) , but not the inhibitors (e.g., calcitonin), stimulate the PTP-oc mRNA expression in isolated osteoclasts.

The second study was to confirm that the osteo-regulatory effectors exerted corresponding changes in cellular PTP-oc protein level in osteoclasts. To accomplish this goal, a specific polyclonal antibody against the PTP-oc fusion protein was generated on contract by Cocalico Biologicals, Inc. (Rearristown, Pennsylvania) . The resulting anti-PTP-oc antibody was shown to be specific for PTP-oc protein, and was used in an immunoprecipitation and immunoblotting experiment to identify and quantify the relative level of the PTP-oc protein after the 24-hour treatment with or without bone resorption activators [i.e., 1 μM prostaglandin E2, 10 nM parathyroid hormone, and 10 nM l,25(OH)₂D₃] or inhibitors [i.e., 30 nM calcitonin and 1 μM clodronate]. Fig. 11 shows that the three bone resorption activators each significantly increasecl the amount of cellular PTP-oc by 2- to 10-fold; whereas neither resorption inhibitors significantly altered the cellular PTP-oc protein level. Accordingly, the results of these two studies together indicate that the expression of PTP-oc gene in osteoclasts is increased by bone resorption stimulators. Thus, these findings suggest that the PTP-oc may play a role in the regulation of osteoclast activity and bone resorption.

EXAMPLE 8 ASSOCIATION BETWEEN PTP-OC AND SIGNAL TRANSDUCTION PROTEINS It is reasoned that the physiological substrate (s) of PTP- oc and/or those phosphotyrosyl proteins whose activity is regulated by PTP-oc are likely to form complexes with PTP-oc. Accordingly, a co-immunoprecipitation study was conducted to determine whether PTP-oc might form complexes with cellular phosphotyrosyl proteins in rabbit osteoclasts (Fig. 12). Briefly, cellular proteins of the isolated rabbit osteoclasts were immunoprecipitated with the PTP-oc antibody. The phosphotyrosyl proteins that were co-immunoprecipitated with the PTP-oc were identified by western immunoblot analysis using a commercial monoclonal anti-phosphotyrosine antibody. Fig. 13 shows that at least nine major phosphotyrosine proteins, with apparent molecular size ranging from 20 to over 200 kilodaltons, were co-immunoprecipitated with PTP-oc by the anti-PTP-oc antibody, indicating that the PTP-oc might form complexes with these phosphotyrosyl proteins. Because PTP-oc does not contain SH2 domain [SH2 domain has a high binding affinity for the phosphotyrosine moiety], the association between PTP-oc and the phosphotyrosine proteins are not due to interactions between the phosphotyrosine residues and SH2 domains, but rather may be related to the specific binding (e.g., substrate binding) of these phosphotyrosyl proteins to the PTP-oc. It should also be noted that the determined molecular size of some of these proteins was very similar to that of the known signal transduction proteins (e.g., pp60^c"^βrc, ppl20^raβGAP; pp44^MΛPK; pp42^MAPK; She; and Grb-2). Thus, these findings are consistent with the notion that some of these key signal transduction proteins may be physiological substrates of PTP-oc, and/or that PTP-oc interacts with the key signal transduction proteins to affect osteoclast activity.

To determine the identity of the phosphotyrosyl proteins that are the key signal transduction proteins, rabbit osteoclastic cellular proteins were prepared with standard laboratory procedures. Also, using standard procedures, these proteins were immunoprecipitated with PTP-oc by the anti-PTP-oc antibody prepared with standard procedure, followed by standard western immunoblot analysis techniques with commercial antibodies against each specific signal transduction protein. Some of these phosphotyrosyl proteins that were co- immunoprecipitated with PTP-oc were indeed pp60^c"^βrc, ppl20^rasCAP; pp₄₄ ^MAPκ. _pp42^MPK; she; and Grb-2. This conclusion resulted from the finding that each respective specific antibody reacted strongly with a protein band with an apparent molecular weight corresponding to each respective signal transduction protein. As an example. Fig. 14 shows the immunoblot of the anti-PTP-oc immunoprecipitate with a mixture of anti-Src, anti-Erkl, and anti-Grb2 antibodies. These are significant findings since these phosphotyrosyl proteins are key components of the MAP kinase signal transduction pathway, which is the major signal transduction pathway that control cell activities as well as proliferation and differentiation in mammalian cells.

To definitively confirm the notion that these key signal transduction proteins were each associated with the PTP-oc in osteoclasts, the immunoprecipitation-immunoblotting study was performed in the reverse order. That is, the rabbit osteoclast cellular proteins prepared with standard procedures were first immunoprecipitated with each specific antibody against respective signal transduction protein, and the presence of PTP-oc in the immunoprecipitate was verified with standard western immunoblots using the anti-PTP-oc antibody. Fig. 15 demonstrates that in each case, there was a single, strongly reactive, protein band with a molecular size of approximately 50,000 daltons, which corresponded to that of PTP-oc. Therefore, these discoveries confirm the conclusion that PTP-oc forms complexes with the signal transduction proteins, such as p_P60^c-^βrc, ppl20^raβGAP; pp44^MΛPK; pp42^MAPK; She; and Grb-2 and that PTP-oc is involved in the regulation of the MAP kinase signal transduction pathway in osteoclasts.

EXAMPLE 9 REGULATION OF TYROSYL PHOSPHORYLATION OF KEY SIGNAL PROTEINS If these signal transduction proteins were involved in the regulation of bone resorption process, it would be anticipated that agents that alter bone resorption rate, in addition to affecting the expression of PTP-oc, could also have an impact on the tyrosyl phosphorylation level and/or the relative amount of each of these signal transduction proteins that are co- immunoprecipitated with the PTP-oc. Accordingly, a study was performed to determine whether bone resorption activators [i.e., parathyroid hormone, prostaglandin E2, and l,25(OH)₂D₃] or bone resorption inhibitors [i.e., calcitonin and clodronate] could have an effect on the tyrosyl phosphorylation level and/or the relative amount of these signal transduction proteins that are associated with the PTP-oc. In this regard, the relative amount of each signal transduction protein that were co-immunoprecipitated with the PTP-oc was determined by the immunoprecipitation of osteoclast extracts with the anti- PTP-oc antibody followed by western immunoblotting using the corresponding specific antibody against each signal transduction protein after a 24-hour treatment with each effectors. However, because these osteo-regulatory effectors had significant effects on the expression of PTP-oc level in osteoclasts (see Example 7), the relative level of each signal transduction protein was normalized against the PTP-oc protein level. Of the observed changes, the effects of these agents on the pp60^c"βrc and pp62 GAP-associated proteins, two important regulators of signal transduction mechanisms in mammalian cells, are of the most interest. Accordingly, the relative level of each of these proteins (per PTP-oc level) that were co-immunoprecipitated were significantly reduced (to as low as 20% of the vehicle-treated control) by the 24-hour treatment with 1 μM prostaglandin E2, 10 nM parathyroid hormone, or 10 nM l,25(OH)₂D₃; whereas the level of these two proteins (adjusted for PTP-oc level) was increased (3- to 5-fold) by the 24-hour treatment with 30 nM calcitonin or 1 μM clodronate.

To analyze the effects of these agents on tyrosyl phosphorylation level of the signal transduction proteins that were co-immunoprecipitated with PTP-oc, rabbit osteoclast extracts, after treatment with each effector for 24 hours, were immunoprecipitated with PTP-oc. The level of the tyrosyl phosphorylated species of each signal transduction protein was detected by western immunoblot analysis using a monoclonal anti-phosphotyrosine antibody. The relative level of each signal transduction protein was quantified by laser densitometry and normalized against the amount of each respective immunoprecipitated signal transduction protein. It was shown that 1 μM prostaglandin E2, 10 nM parathyroid hormone, and 10 nM l,25(OH)₂D₃, each markedly reduced its tyrosyl phosphorylation level of both pp60"^βrc and pρ62 GAP- associated protein. In contrast, 30 nM calcitonin and 1 μM clodronate each significantly increased the tyrosyl phosphorylation level of both pp60^c"^βrc and pp62 GAP-assόciated protein by as much as 4-fold of the vehicle-treated control. Therefore, these studies indicate that in general, bone resorption stimulators reduce and bone resorption inhibitors increase, both the "recruitment" and the tyrosyl phosphorylation level of pp60^c_βrc and pp62 GAP-associated protein in osteoclasts. Thus, these studies demonstrate that PTP-oc might have a regulatory role in the regulation of these two signal transduction proteins in osteoclasts.

It is important to note that pp60^c"^βrc and pp62 GAP- associated protein differ from most other signal transduction proteins of the MAP kinase signaling pathway in that tyrosyl phosphorylation of these two proteins leads to an inhibition of the MAP kinase pathway. Accordingly, reduction of tyrosyl phosphorylation level of these two proteins would result in the activation of the signal transduction pathway. Consequently, the findings that bone resorption inhibitors increase, while bone resorption activators reduce, the tyrosyl phosphorylation level of these proteins are consistent with the notion that the activation of pp60^c_Brc (and/or the pp62 GAP-associated protein) and the MAPK signal transduction in osteoclasts may be in part responsible for the stimulation of bone resorption. The findings that bone resorption activators reduces, and the activators stimulates, the "recruitment" of these two signal transduction proteins to PTP-oc and their tyrosyl phosphorylation level are intriguing. Because PTP-oc contains no SH2 domain, the "recruitment" of phosphotyrosyl proteins to PTP-oc may be related to targeting the substrate to the active site for hydrolysis. It is postulated that the pp60^c"^βrc and pp62 GAP-associated protein are the physiological substrates for PTP-oc. Consistent with this speculation, the PTP-oc fusion protein was capable of dephosphorylating [³²P] labelled c- src peptide corresponding to the tyr-527 in vitro . The bone resorption stimulators, such as prostaglandin E2, increases both the expression of PTP-oc and its enzymatic activity. Accordingly, treatment with bone resorption activators would lead to an increase in the dephosphorylation of the pp60^c"^Brc and pp62 GAP-associated protein. Upon dephosphorylation, these two proteins would lose their affinity for binding to PTP-oc. Therefore, the stimulation of bone resorption would result in reduction in both " recruitment" to PTP-oc and the tyrosyl phosphorylation level of these two signal transduction proteins, which are consistent with the findings of the experiments shown in this example. However, it should be noted that pp62 GAP-associated protein has been suggested to be physiological substrate of pp60^c"^βrc. Thus, it is possible that the pp60^c_βrc is the physiological substrate of PTP-oc. Consistent with this speculation, tyrosyl phosphorylated c-src peptide can be effectively dephosphorylated by the cPTP-oc recombinant protein.

All publications mentioned in this specification are indicative of the level of skill of those skilled in the art to which this invention pertains and are herein incorporated by reference to the same extent as if each individual publication or patent application was specifically and individually indicated to be incorporated by reference.

The invention now being fully described, it will be apparent to one of ordinary skill in the art that many changes and modifications can be made thereto without departing from the spirit or scope of the following claims.

Claims

1. A screening assay for identifying substances that stimulate or inhibit bone resorption, comprising the steps of: obtaining a test substance; contacting the test substance with an osteoclast specific phosphotyrosyl phosphatase (PTP-oc) that effects bone resorption by effecting signal transduction proteins involved in signal transduction pathways in osteoclasts such that phosphates are released or dephosphorylated compounds are produced; and monitoring the phosphate release or the dephosphorylated compound production.

2. The assay of claim 1 wherein: the PTP-oc is added to a solution containing the test substance and a radioactively tagged phosphate of a compound which is released upon contact of the test substance with the PTP-oc; and monitoring the level of radioactivity of the released phosphate.

3. The assay of claim 2 wherein: the radioactively tagged compound is a tyrosyl phosphorylated protein.

4. The assay of claim 1 wherein: the PTP-oc is added to a solution containing the test substance and a chromophore phosphotyrosyl containing compound; and the monitoring monitors the production of dephosphorylated products spectrophotometrically.

5. The assay of claim 1 wherein: the PTP-oc is added to a solution containing the test substance and a fluorescent phosphotyrosyl containing compound; and the monitoring monitors the production of dephosphorylated products fluorometrically.

6. The assay of claim 1 wherein: the PTP-oc is added to a solution containing the best substance and PNPP; and the monitoring monitors the amount of p-nitrophenolate ion produced.

7. The assay of claim 1 wherein: the PTP-oc is added to a solution containing the test substance and phosphotyrosine; and the monitoring monitors the release of inorganic phosphate with a colormetric assay.