EP0391953A1 - Common acute lymphoblastic leukemia antigen - Google Patents

Abstract

The present invention describes complementary DNA encoding all or part of the human acute lymphoblastic leukemia (CALLA) antigen; the encoded polypeptide or a fragment thereof, such as in particular a soluble fragment; methods of making and using the encoded polypeptide; as well as antibodies raised against and capable of binding to CALLA.

Description

COMMON ACUTE LYMPHOBLASTIC LEUKEMIA ANTIGEN

Description

Background of the Invention

The common acute lymphoblastic leukemia antigen (CALLA) has been shown to be a 100KD cell surface glycoprotein. It was initially identified on human lymphoblas tic leukemia cells. Ritz, J. et al ., Nature (London), 283:583-585 (1980). Later work showed that CALLA is also expressed by other lymphoid malignancies, such as lymphoblastic, Burkitts, and nodular poorly differentiated lymphocytic lymphomas, and by early lymphoid progenitors from fetal liver and fetal, pediatric and adult bone marrow and fetal and pediatric thymus. Greaves, M.F. et al., Blood, 61:628-639 (1983); Hokland, P., J. Exp. Med , 157:114-129 (1993); Hokland, P. et al., Blood, 64:662-666 (1984); Hoffman-Fezer, G. et al., Leuk. Res., 6:761-767 (1982) ; Neudorf, JS.M. et al., Leuk. Res., 8:173-179 (1984). Although antibodies against CALLA have been used extensively in both the diagnosis and therapy of these malignancies, the primary structure and function of CALLA are unknown.

The present invention relates to cDNA encoding human CALLA, or a fragment thereof. The cDNA encodes CALLA capable of binding to an anti CALLA antibody (e.g., the monoclonal antibody J5). The primary structure and the function of the encoded human CALLA have been determined. The cDNA has been shown to have near identity with the rat and the rabbit zinc metalloendopeptidase, neutral endopeptidase 24.11 (enkephalinase), and to encode functional neutral endopeptidase activity.

The cDNA or DNA encoding CALLA or a fragment thereof can be inserted into a vector, which can be used to transform cultured cells for the production of CALLA or CALLA fragments. Preferably, CALLA fragments encoded by the DNA are soluble. In order to be soluble, the CALLA fragments of the invention preferably contain none of the hydrophobic transmembrane portion of CALLA or only a portion (generally, six or fewer amino acids) of the transmembrane portion which does not prevent solubilization. CALLA fragments of the present invention need not have perfect homology with the corresponding region of naturally-occurring CALLA, and in some instances (e.g., antibody production), less than perfect or total homology is preferred. In general there will be at least about 50% homology between the amino acid sequence of the naturally-occurring CALLA and the amino acid sequence of CALLA or a CALLA fragment of the present invention.

As a result of the present invention, pure CALLA is now available for use, for example, in the diagnosis and treatment of medical conditions characterized by the presence of cells which express CALLA on their surfaces.

Figure 1 is a graph of the HPLC profile of a CALLA tryptic digest.

Figure 2 shows the restriction maps of CALLA cDNA clones.

Figure 3 is the nucleotide sequence of human

CALLA cDNA and the predicted amino acid sequence of the encoded human CALLA.

Figure 4 is a hydropathicity plot of the translated CALLA cDNA sequence.

Figure 5 is a restriction map of the CALLA cDNA clones used in the pIGTE/N CALLA construct. Figure 6 is a diagramatic representation of the pIGTE/N CALLA construct,

Figure 7 presents a graphic comparison of the relative cell surface CALLA expression on the Nalm-6, J55, A2-3 and A2 - 2 cell lines.

Detailed Description of the Invention

The present invention relates to DNA encoding all or a portion of human common acute lymphoblastic leukemia antigen (CALLA); to the encoded polypeptide, or a fragment thereof; to methods of making and using the encoded polypeptide or a fragment thereof; and to antibodies raised against and capable of binding CALLA. As described briefly in the next sections and in detail in the Examples which follow, cDNA encoding human CALLA which binds to an anti-CALLA antibody has been isolated and sequenced and the encoded polypeptide has been characterized, in terms of its primary structure and its function. The nucleotide sequence of cDNA encoding human CALLA and the predicted amino acid sequence of the encoded CALLA are presented in

Figure 3. As a result of the work described herein, it has been shown that there are striking homologies between the human CALLA cDNA sequence and DNA encoding rat and rabbit neutral endopeptidase 24.11 (enkephalinase) and close identity (i.e., 94%) between the amino acid sequence of human CALLA and the amino acid sequences of rat enkephalinase and rabbit neutral endopeptidase molecules.

Comparison of the amino acid sequence of human CALLA of the present invention with the amino acid sequence of the recently reported human enkephalinase (or neutral endopeptidase) showed that the amino acid sequences of the two polypeptides are virtually identical. (That is, the difference in the two sequences is a single amino acid, which can be the result of sequencing error or genetic polymorphism. Malfroy, B. et al., FEBS Letters, 229: 206-210 (1988)).

Subsequently, as Is also described below and in the Examples, assessment of the human CALLA encoded by the cDNA of the present invention showed that it has neutral endopeptidase activity and that the activity is associated with the cellular membrane fraction and abrogated by a specific neutral endopeptidase inhibitor (i.e., phosphoramidon). This unequivocal identification of CALLA as a functional endop ep t idas e provides insight into its potential role in both normal and malignant lymphoid function and into its possible uses in diagnostic and therap eut ic contexts (e.g., in diagnosis of disease conditions characterized by the presence of cells which express CALLA, in treating or relieving pain, etc.).

Neutral endopeptidase 24.11 enkephalinase is a cell membrane-associated enzyme that cleaves peptide bends on the amino side of hydrophobic amino acids. Hersh, L.B., Mol. Cell. Biochem., 47:35-43 (1982). The enzyme was identified in brain as an enkephalinase because it cleaved the Gly3-Phe4 bond of enkephalins. Malfroy, B. e t al . , Nature, 276:523-526 (1978). The enzyme was subsequently found in many other tissues, including kidney, in which it was present at high levels. Llorens, C. & J.C. Schwartz, Eur. J. Pharmacol, 69:113-116 (1981). In kidney, enkephalinase activity was shown to be identical to that of neutral endopeptidase 24.11, which had been identified several years earlier using the B chain of insulin as substrate. Kerr, M.A. & A.J. Kenny, Biochem. J., 137: 477-488 (1974); Kerr, M.A. & A.J. Kenny, Biochem. J., 137:489-495 (1974). Neutral endopeptidase 24.11 has been shown to react with a variety of physiologically active peptides including chemotactic peptide, substance P and neurotensin, oxytocin, bradykinin, angiotensin I and II and a variety of opioid peptides. Connelly, J.C. et_al., Proc. Natl. Acad. Sci., USA, 82:8737-8741 (1985); Almenoff, J. et al., Biochem. Biophys. Res. Commun., 102:206-214 (1981); Mumford, R.A. et al., Proc. Natl. Acad. Sci., USA ,

78: 6623-6627 (1981); Johnson, A.R., Peptides, 5:789-796 (1984); Gafford, J.R. et al., Biochem, 22:3265-3271 (1983); Hersh, L.B., Neurochem, 43:487-493 (1984). This enzyme has also been shown to hydrolyze the lymphokine, IL-1. Pierart, M.E., J. Immunol., 140:3808-3811 (1988). Neutral endopeptidase 24.11 has been found in numerous tissues other than kidney and brain, including peripheral blood granulocytes, fibroblasts, small intestine and placenta. Connelly, J.C. e t al .,

Proc. Natl. Acad. Sci., USA, 82:8737-8741 (1985); Borkowski, G. et al., Biochem. J., 248:345-350 (1987); Gee, N.S. et al., Biochem, 214:377-386 (1983); Ealfroy, B. et al., FEBS Letters, 229:206-210 (1988). However, lymphoid cells have not previously been shown to possess this enzymatic activity. Bowes, M.A., & A.J. Kenny, Mochem, 236:801-810 (1986). Given that enkephalinase is a zinc binding metalloendopeptidase, it is of interest that the chromosomal location of the CALLA encoding gene is 3q21-27, a region rich in metal binding proteins, including transferrin, lactotransferrin, melanotransferrin, the transferrin receptor and ceruloplasmin. As described in detail in Examples 1-4, human

CALLA protein has been isolated from the Nalm-6 cell line through use of an anti-CALLA monoclonal antibody (J5) and standard techniques. The Nalm-6 cell line was originally derived from a patient with chronic myelogenous leukemia in lymphoid blast crisis and is known to express high levels of surface CALLA. Hurwitz e t al . , Cancer, 23:174 (1970); Goldmacher e t al ., J. Immunol ., 136:320

(1986). Use of the J5 antibody resulted in immunoprecipitation from ¹²⁵I surface-labelled

Nalm-6 cells of a component of 97-100KD in size. Digestion with endoglycos idase F resul ted in a decrease of approximately 10KD in the molecular mass of the immunoprecipitated component. This indi cate s that at least 10% of the molecular mass of CALLA results from N-linked carbohydrates. The isolated protein was subsequently purified to homogeneity, using techniques which are recognized and described in the Examples. NH₂-terminal sequence information was obtained from the intact CALLA protein and from derived tryptic peptides and V8 protease fragments, as shown in Table 1.

CALLA cDNAs were isolated from a Nalm-6 cell line λgt10 library, using redundant oligonucleotide probes. Shipp, M.A. e t al . , Proc. Natl. Acad. Sci . , USA, 85:4819-4823 (1988). The CALLA cDNA sequence predicts a 749 amino acid integral membrane protein which has a single 24 amino acid hydrophobic segment which could function as both a transmembrane region and a signal peptide. The extracellular protein segment is made up of the COOH-terminal 700 amino acids, which include six potential N - l inke d glyc o sylat ion sites. The 25 NH₂-terminal amino acids remaining after cleavage of the initiation methionine form the cytoplasmic tail. Comparison of DNA encoding human CALLA and of the amino acid sequence of human CALLA with data in the most recent GenBank release (#56) resulted in the surprising finding that human CALLA has 94% homology at the amino acid level with neutral endopeptidase 24.11 (enkephalinase), a membrane-bound zinc metalloendopeptidase cloned from rat brain and rabbit kidney. Devault, A. et al., EMBO J.,

6:1317-1322 (1987). CALLA- transfected cell lines were subsequently produced and a sensitive enzymatic assay was used, In conjunction with a specific neutral metallopeptidase inhibitor, to demonstrate that CALLA is a functional form of this membrane-bound enzyme.

Expression of Human CALLA in Murine Transfectants To determine whether CALLA derived from a lymphoblastic leukemia cell line has functional neutral endopeptidase 24.11 activity, a construct (pIGTE/N CALLA ) containing the CALLA open reading frame from the leukemic cell line, Nalm-6, under the control of an Immunoglobulin promoter and enhancer was engineered and transfected into the murine myeloma cell line, J558, which lacks cell surface CALLA expression.

Following G418 selection, CALLA+ J558 cells were identified by phenotyping with the J5 anti-CALLA monoclonal antibody, sorted and cloned by limiting dilution. Two J5+ subclones, A2-3 and A2-2, which had high and low levels of CALLA expression, respectively, were chosen for further analysis. Figure 2 represents a comparison of relative CALLA fluorescence of these two CALLA+ s t ab l e transfectants, the parental CALLA multiple myeloma line, J558, and the CALLA+ acute lymphoblastic leukemia line from which the CALLA cDNA was isolated, Nalm-6. As indicated, J558 lacks detectable cell surface CALLA expression (mean channel fluorescence 0). A2-3 and A2-2 express cell surface CALLA. Note that A2 - 3 expresses a substantially higher number of CALLA sites per cell than does A2 - 2 (mean channel fluorescence 116.8 and 7.6, respectively). Comparison of the mean channel fluorescence of A2 - 3 and A2-2 with the mean channel fluorescence of Nalm-6 (172.6), indicates that A2-3 expresses CALLA at a level 67.7% that of Nalm-6, but A2-2 expresses CALLA at only 4.4% the level of Nalm-6.

Analysis of Neutral Endopeptidase Activity

The neutral endopeptidase activity associated with the A2-3, A2-2, J558 and Nalm-6 lines, was assessed by means of two approaches. A sensitive fluorometric assay based upon the cleavage of the substrate glutaryl-Ala-Ala-Phe-4-methoxy-2-naphthylamide was used, as described in Example 7. Table 2 illustrates the results of the enzymatic assay performed using whole cell suspensions of the individual cell populations. Cell suspensions of Nalm-6, A2-3, A2-2 and J558 exhibited neutral endopeptidase activity of 4789, 1906, 446 and 0.2 nanomoles of product per hr per 10⁶ cells, respectively. The value of 0.2 nanomoles per hr per 10⁶ cells is within the experimental error of no activity.

The effect of addition of the specific inhibitor phosphoramidon to the Nalm-6, A2-3 and

A2-2 cell suspensions on neutral endopeptidase activity was also assessed. It resulted in a marke d reduction in activity to 53, 52 and 18 nmols/h/10⁶ cells, respectively, (Table 2). Hudgin, R.L. et al., Life Sci. , 29:2593-2601 (1981). As indicated in Table 3, cell lysates from Nalm-6, A2-3 , A2-2, and J558 contained neutral endopeptidase specific activities of 9.96, 2.35, 1.18 and 0.08 nmols/min/mg protein of neutral endopeptidase activity, respectively. When the assays were performed in the presence of phosphoramidon, the neutral endopeptidase activity associated with the Nalm-6, A2-3 and A2-2 cell lysates was dramatically reduced to 0.20, 0.08 and 0.29nmols/min/mg protein, respectively (Table 3). The observation that the apparent activity in J558 cells is insensitive to phosphoramidon inhibition suggests that this low level of apparent activity is not due to neutral endopeptidase 24.11 and within the experimental error of no activity. Subcellular fractionation of the Nalm-6 and A2-3 cells demonstrated that 100% of the Nalm-6 and more than 95% of the A2-3 neutral endopeptidase activity was associated with the membrane fraction (Table 3). Comparison of the levels of neutral endopeptidase activity (Tables 2 and 3) and levels of CALLA expression of the four cell lines (Figure 5) indicated that there was a correlation between neutral endopeptidase activity and cell surface CALLA expression.

The results of the work described herein make it clear that human CALLA is a functional form of active neutral endopeptidase 24.11 (enkephalinase). The fact that the CALLA protein is found on both early normal lymphoid progenitors and their malignant counterparts and that CALLA cDNA from an acute lymphoblastoid leukemia encodes neutral endopeptidase 24.11 activity indicates that the enzyme functions at a critical stage in lymphoid differentiation. This is of particular interest in light of previous studies demonstrating that the cell surface bound enzyme has the potential to mediate a wide range of biological activities in a variety of tissues. For example, neutral endopeptidase 24.11 (enkephalinase) has been shown to inactivate endogenous opioid pentapeptides on neurons in brain, chemotactic peptide (f-Met-Leu-Phe) on polymorphonuclear granulocytes and a variety of regulatory peptides on the surface of proximal tubule epithelial cells of the kidney. Hersh, L.B., Mol. Cell. Biochem, 47:35-43 (1982); Malfroy, B. et al., Nature, 276:523-526 (1978); Connelly, J.C. et al., Proc. Natl. Acad. Sci., USA , 82:8737-8741 (1985); Gafford, J.T. et al., Biochem, 22:3265-3271 (1983). Amino acid sequences of the enzyme derived from three tissue sources (brain, kidney and placenta) in three species as well as. from a human lymphoblastic leukemia cell line have been shown to be virtually identical. These results imply conservation of critical functional domains for zinc binding, substrate binding and catalysis. Shipp, M.A. et al., Proc. Natl. Acad. Sci., USA, 85: 4819-4823 (1988); Malfroy, B. et_al., Biochem. and Biophys. Res. Comm., 144:59-66 (1987); DevauIt, A. et al., EMBO J, 6:1317-1322 (1987); Malfroy, B. et al., FEBS Letters, 229:206-210 (1988). In light of the structural identity of the enzyme in various tissues, the biologic activity of CALLA/neutral endopeptidase 24.11 is likely to be dictated by the availability of specific substrates in individual organs, rather than by the presence of different functional forms of the enzyme. The substrate for the cell surface CALLA/neutral endopeptidase 24.11 of early lymphoid progenitors is not yet known. However, previous studies of neutral endopeptidase 24.11 showed that optimal substrates for the enzyme are small peptides, rather than large proteins. Consequently, it is likely that CALLA/neutral endopeptidase 24.11 may also react with a small regulatory peptide at the cell surface of lymphoid precursors. Such an action could lead to the inactivation of a physiologically active peptide or convert an inactive form of the peptide to an active one. The CALLA substrate for lymphoid precursors may be a previously defined peptide.

Neutral endopeptidase 24.11 has not previously been Identified on lymphoid cells. It has, however, been detected in nodal tissue. Bowes, M.A. and A.J. Kenny, Biochem. J, 236:801-810 (1986). In porcine lymph nodes, the enzyme is found on a subpopulation of adherent cells with the morphological characteristics of fibroblasts. Bowes, M.A. and A.J. Kenny, Biochem. J, 236:801-810 (1986). These cells are most prevalent in medullary areas and are also found In the center of follicles and encircling them. Of interest, these neutral endopeptidase 24.11+ cells are observed to have clusters of lymphoid cells firmly attached to their cell surface. Bowes, M.A. and A.J. Kenny, Biochem. J,

236:801-810 (1986). In tonsil, spleen, thymus and Peyer's patches, the neutral endopeptidase 24.11+ cells are present in a reticular pattern similar to that seen in lymph nodes, where the enzyme is much more abundant. Bowes, M.A., Immunology, 60:2474253 (1987). Recent studies prompted by the identification of neutral endopeptidase 24.11+ reticular cells in lymphoid tissues indicate that the enzyme inactivates IL-1 in vitro and inhibits thymocyte proliferation in a dose-dependent and sp e c i f ical ly inhibitable fashion. Pierart, M.E. e tal., J. Immunol .. 140:3808-3811 (1987). As noted above, CALLA/neutral endopeptidase

24.11 cleaves chemotactic peptide as a substrate. Connelly, J.C. e t al ., Proc. Natl. Acad. Sci., USA, 82:8737-8741 (1985). The presence of CALLA/neutral endopeptidase 24.11 on the cell surface of mature neutrophils suggests that the enzyme may play an important role in the process of down regulating chemotaxis, perhaps by reducing the local concentration of chemotactic peptides. Chroni c treatment with morphine induces a selective and specific increase in brain enkephalinase activity, likewise indicating that the enzyme may regulate the local c onc ent r a t i o n of opioid neurotransmitters and that the concentration of such neurotransmitters may also affect enzyme levels. Malfroy, B. et al., Nature, 276:523-526 (1978).

Earlier studies indicated that specific antibody treatment of CALLA+ lymphoid cells resulted in rapid cell surface redistribution, internalization and degradation of the CALLA antibody complex. Pesando, J.M. et al . , J. Immunol., 131:2038-2045 (1983), Ritz, J. et al., J. Immunol., 125:1506-1514 (1980). The antibody induced modulation of CALLA was noted to resemble the specific down regulation or loss of cell surface receptors Induced by peptide hormones. Goldstein, J.L., Nature (London), 279 : 679-685 (1979); Catt, K.J. et al., Nature (London), 280:109-116. CALLA encodes functional lympho i d neutral endopeptidase 24.11, and, thus, it is likely that its peptide ligand will also modulate cell surface CALLA. This may result in efficient internalization of ligand. It is not yet known whether such a putative p ep t ide affects migration, growth or other functional aspects of immature normal or malignant B cells. However, the unequivocal identification of CALLA as functional neutral endopeptidase 24.11 (enkephalinase) and the availability of specific endopeptidase inhibitors should make it possible to assess the role of CALLA in lymphoid development. Uses of DNA Encoding Human CALLA, CALLA, Antibodies Reavtive with CALLA

DNA encoding human CALLA, CALLA or fragments thereof and antibodies reactive with CALLA have both diagnostic and therapeutic applications.

For example, the cDNA encoding human CALLA can be used to produce CALLA, by means of techniques known to those skilled in the art (such as those detailed in the Examples). For example, DNA having all or a portion of the nuc l e o t i de sequence of

Figure 3 can be introduced into a suitable host cell (generally as a component of a vector), in which it is subsequently produced and from which it can be isolated. The DNA can be introduced into the cells by known techniques (e.g., co-cultivation, electroporation). The DNA introduced in this manner can be isolated, as described herein, or can be DNA encoding functional CALLA (e.g. DNA, referred to as equivalent DNA, which has a sequence sufficiently similar to that represented in Figure 3 to encode functional CALLA). The term cDNA, as used herein, is intended to include DNA, produced or obtained by any means, which has the sequence represented in Figure 3 or equivalent DNA. One way in which CALLA-encoding DNA can be used to produce all or a fragment of CALLA is as follows:

Human cDNA sequence encoding the CALLA or a desired CALLA fragment (e.g., the sequence enco ding only the portion of CALLA which is exposed beyond the transmembrane region) can be inserted into a suitable expression vector, using conventional techniques. For example, a desired cDNA can be inserted into the expression vector described in Ringold U.S. Patent No. 4,656,134, hereby incorporated by reference. The resulting plasmld can then be used to transform mammalian host cells, and the human CALLA fragments Isolated and purified from those cells and/or their medium according to conventional methods.

In their unglycosylated forms, the polypeptides can be produced in a bacterial host (e.g., E. coli). The cDNA encoding the desired fragment can, for example, be inserted into the expression vector described in DeBoer et: al., Proc. Natl. Acad. Sci.,

USA, 80:21 (1983). The resulting plasmid can then be used to transform E. coli cells, and the CALLA fragment isolated and purified according to conventional methods.

All or a portion of the nucleotide sequence of Figure 3 can also be used as probes to detect, identify and/or quantity DNA encoding CALLA in a sample (e.g., cells) of interest. These probes will generally be labelled (e.g., radioactively, enzymatically, etc.) and can be used in standard techniques.

CALLA polypeptides of the present invention can be used, for example, for therapeutic purposes.

Particularly useful in this context will be soluble CALLA fragments (e.g., all or a portion of the portion of the CALLA region which extends beyond the transmembrane region of a cell). Soluble CALLA fragments will be admixed with a pharmaceutically acceptable carrier substance, e.g., saline, and administered by a medically acceptable administration route (e.g., intravenously, intramuscularly, or orally) to patients suffering from medical conditions characterized by the presence of CALLA-bearing cells which are associated with their disease, in an amount sufficient to inhibit proliferation of those CALLA-bearing cells. The soluble CALLA fragments will function by competitively inhibiting the natural binding events of surface-bound CALLA, which events are apparently necessary for the proliferation and/or mobility of such cells (e.g., acute lymphoblastic leukemic cells). The amount of soluble CALLA fragment administered will be determined on an individual basis, but will generally be from approximately 10 μg/kg bodyweight to approximately 50 μg/kg per day. As a result of the present invention, it is possible to make diagnostically useful antibodies which were previously unavailable. Because the amino acid sequence of CALLA has now been provided, any synthetic peptide corresponding to any given region of the molecule can routinely be made, and that peptide used to raise highly specific antibodies. Two antibodies, each specific to a different region of CALLA, can be used in immunoassays, such as a sandwich assay, which require two different antibodies, which bind to two different sites on the target molecule; such assays are described, for example, in David et al., U.S. Patent No. 4,376,110, the teachings of which are hereby incorporated by reference.

Another important class of antibodies made possible by the invention are antibodies reactive with denatured CALLA (i.e., CALLA which has lost i ts native three-dimensional conformation). Such antibodies can be made by immunizing animals with CALLA fragments or CALLA analogs containing amino acid substitutions which prevent normal folding of the protein. These antibodies will be particularly useful for analysis of pathologic specimens which have been fixed in formalin, a process which causes denaturation of any CALLA present on, for example, cancer cells in a lymph node. Antibodies made to native CALLA, which retains its three-dimensional structure, are unreactive with denatured CALLA, and thus cannot be used to detect CALLA-bearing cells in formalin-fixed specimens. The antibodies reactive with denatured CALLA will be labelled (e.g., with fluorochromes) and can be used in immunofluorescence procedures which are standard in the field of pathology.

The understanding of CALLA provided herein can be used to design a "drug" which can be used as a general analgesic. Enkephalins are endogenous opioid-llke pentapeptides present in areas of the central nervous system associated with perception of pain, as well as with other functions. As explained previously, neutral endopeptidase 24.11

(enkaphalinase) has been shown to inactivate endogenous opioid pentapeptides on neurons in the brain. That is, neutral endopeptidase 24.11 (enkephalinase) is known to cleave the Gly³-Phe³ amide bond of the op io id pentapeptides, enkephalins, both in vitro and in vivo. A parenterally active enkephalinase inhib ito r (acetorphan) has been shown to display analgesic properties in human. Malfroy, B. et al., FEBS Letters, 229:206-210 (1988). The rat and the rabbit neutral endopeptidase 24.11 have been shown to cleave enkephalins. As described herein, it has been shown that CALLA has functional neutral endopeptidase activity, as well as striking similarity at the amino acid level to human enkephalinase and to rat enkephalinase and rabbit neutral endopeptidase molecules. Thus, it should be possible to design a substance capable of interfering with this ability of CALLA/neutral endopeptidase 24.11 to cleave opioid pentapeptides, producing a prolonged and enhanced concentration of endogenous opioids, which, in turn, will produce an analgesic effect.

Based on the present invention, it is also possible to design an inhibitor of CALLA which will inhibit/reduce leukemic growth or behavior of cells. For example, CALLA exhibits enzymatic function which might influence (enhance) leukemic growth or behavior. As a result of the present invention, an inhibitor of CALLA, which binds to it and reduces or eliminates its enzymatic activity, can be designed and used to inhibit such growth. An alternative approach is based on the knowledge presented herein and the fact that there appears to be a natural ligand of CALLA which is involved in leukemic cell growth; this putative natural ligand might bind to CALLA and, after cleavage, become active. Thus, it is possible to design an "artificial ligand" which inhibits binding of the natural CALLA ligand (e.g., by binding to CALLA and, thus, preventing CALLA natural ligand binding or by binding to the natural ligand, and, again, preventing CALLA natural ligand binding). The present invention will now be illustrated by the following examples, which are not intended to be limiting in any manner.

EXAMPLE 1 Isolation and Characterization of CALLA protein Isolation

To Isolate CALLA protein, the Nalm-6 line was utilized as a cellular source in conjunction with the anti-CALLA monoclonal antibody J5. Ritz et al., Nature 258:454 (1975). The Nalm-6 cell line was originally derived from a patient with chronic myelogenous leukemia in lymphoid blast crisis and is known to express high levels of surface CALLA. Hurwitz et al., Cancer, 23:174 (1979); Goldmacher et al., J. Immunol ., 136:320 (1986). J5 antibody was found to immunoprecipitate a structure of 97-100KD from ¹²⁵I surface labelled cells. Following digestion with endoglycosidase F, the molecular mass of CALLA decreases by approximately 10KD, indicating that at least 10% of the molecular mass of CALLA results from N-linked carbohydrates Two dimensional electrophoresis demonstrates that the CALLA protein migrates as a single polypeptide exhibiting limited microheterogeneity. Newman et al., J. Immunol .,

126:2024 (1981). Purification

Nalm-6 cells (3 × 10⁷) were radioiodinated using lactoperoxidase and lysed in RIPA buffer containing 1% Triton X-100, 0.15 M NaCl, ImM PMSF,

80 mM iodoacetamide, 0.02 μg/ml tryp s in inhibitor and 0.5 μg/ml each of chymostatin,, leupeptin, pepstatin, and antipain. Fabbi et al , Nature,

312:269 (1984). Ultracentrifuged lysates were precleared sequentially with preimmune rabbit Ig and mouse monoclonal antibody anti-β2 microglobulin coupled to protein A Sepharose, and CALLA was immunoprecipitated with the mouse monoclonal antibody J5 coupled to protein A Sepharose. The anti-CALLA beads were washed with: 1) TBS/1% DOC; 2) TBS/1% DOC/0.05% SDS; and 3) TBS/1% NP40. Bound

CALLA was eluted with SDS sample buffer containing 5% 2 -mercap toethanol. Endo-F glycosidase treatment was carried out, and aliquots of samples were analyzed with and without enzyme treatment on 10% SDS-PAGE. Luescher and Bron, J. Immunol ., 134:1084 (1985).

For preparative isolation of the CALLA protein, 10^{1 1} Nalm-6 cells were lysed for 1 hour at 4°C in

400 ml of lysis buffer as above. The crude lysate was centrifuged at 3000 × g for 20 minutes. The supernatant was made 0.5% in solium deoxycholate and ultracentrifuged for 60 minutes at 150,000 × g.

Iodinated Nalm-6 cells (3 × 10⁷) were treated with

0.5 ml of lysis buffer and added to the large scale preparation. The combined lysates were applied at

32 ml/hr to a 10 ml "pre-clear" immunoabsorbent column containing irrelevant mouse monoclonal antibodies anti-T3 (8C8), anti-Ti₃ (9H5), and anti-β2 microglobulin coupled to protein A-Sepharose CL-4B beads at 5 mg/ml, followed by a 5 ml specific antibody column containing anti-CALLA antibody (J5) coupled to protein A-Sepharose beads at 5 mg/ml. The anti-J5 column was washed sequentially with 50 ml aliquots of 10 mM Tris, pH 8/0.15 M NaCl with 1) 0.05% SDS/0.5% TX-100/1% DOC; 2) 0.5% TX-100, 1% DOC; and 3) 0.5%-Tx-100. Bound material was eluted with 0.1 M glycine, pH 3.0/0.5% Triton X-100 and collected in 1 ml fractions in tubes containing 60 μl.of 1 M Tris, pH 8. Subsequently, 100 μl 10% SDS was added to each tube at 22°C. Fractions containing radioactivity were pooled, made 10% (vol/vol) in glycerol and 2% in NaDodSO₄, heated at 60°C for 20 minutes, and electrophoresed in a 10% preparative (3 mm thick) polyacrylamide gel. A 0.5 cm strip of the gel was dried and autoradiographed, and the rest was stained with

Coomassie blue. The stained band containing CALLA was localized by comparison with the migration of surface labelled CALLA as shown by the autoradiographed strip. The 100KD CALLA band was electroeluted in 50 mM ammonium bicarbonate containing 0.1% SDS, and an aliquot was quantltated and analyzed for purity by NaDodSO₄/PAGE, followed by silver staining. Analysis indicated that the CALLA protein had been purified to homogeneity. EXAMPLE 2 Amino Acid Sequence Analysis of CALLA

Fifty picomoles of purified CALLA were subjected to amino terminal sequencing, and only a 5 picomole signal (XXSESQ) was obtained, suggesting that CALLA was in large part N-terminally blocked to Edman degradation. For this reason, CALLA was digested with trypsin and with V8 protease. Four hundred picomoles of electroeluted CALLA in 50 mM ammonium bicarbonate/0.1% SDS was mixed with V8 protease (Boehringer Mannheim) to give a protein/enzyme ratio of 5:1. After incubation at 37°C for. 1 hour and 22°C for 16 hours, the sample was made 0.1% in trifluoroacetic acid (TFA) and applied to a reverse phase C18 HPLC column to separate fragments as described below. An aliquot of each V8 fragment was analyzed by NaDodSo₄/PAGE followed by silver staining.

One nanomole of electroeluted CALLA was made 0.1M Tris-HCL pH 8/20 mM dithiothreitol/2% SDS and adjusted 60 minutes later to 50 mM in iodacetic acid. The reduced s - carb oxymethylated preparation was then precipitated by the addition of 9 volumes of ethanol at -20°C for 16 hours, dissolved in 0.1M Tris-HCL pH8/2mM CaCl₂, mixed with TPCK trypsin (Cooper) to give a protein/enzyme ratio of 50:1, and incubated for 16 hours at 37°C. Following the addition of TFA to 0.1% and microfuging, the supernatant was subjected to reverse phase HPLC (Vydak, TP54, 4.6 mm X 25 cm, 5 μ C18) at 1 ml/min eluting in 0.1% TFA, using a HP1090 chromatograph equipped with a diode array detector (Hewlett-Packard). A 0-50% acetonitrile gradient over 50 minutes was established and elution fractions of 1 ml were collected. Selected fractions were purified further by HPLC using an acetonitrile gradient of 25-45% over 15 minutes with a flow rate of 1 ml/min and collection of 0.5 ml fractions. CALLA protein, V8 fragments, and tryptic peptides were then analyzed for N-terminal sequence on a gas phase protein sequenator (Applied Biosystems, model 470A) equipped with an in-line 120A PTH analyzer using program 03PTH.

Three pure V8 fragments were identified and sequenced, giving the residues indicated in Table 1.

TABLE 1

CALLA AMINO ACID SEQUENCE FROM INTACT PROTEIN, V8 FRAGMENTS AND TRYPTIC PEPTIDES

. . . . 5 . . . . 10 . . . . 15 . . . . 20

TP l. A L Y G T T S E T A T (W) (K)

II. 1. Y A C G G W L K

2. L Y V E A A F A G E S (K)

3. L L P G L D L N H K

III. L I Q N M D A T T E P C T D F F K

IV. 1. D L Q N L H S W

2. H V V E D L I A Q I

V. D G D L V D W W T Q Q S A S (N)

VI. L L P D I Y G (W) P V A T E N (W) E Q (K)

VII. E V F I Q T L D D L T W M D A E T

VIII. L N N E Y L E L N Y K K D E Y F E N l l Q N (L)

V8P

I . L N Y K E D E Y F E N I I Q N L

I I . I A F A T A K P E D - N D P I L L - N

I I I . - - S E S Q M D I T D I N T P (S)

NT - - S E S Q

V8 peptide III contains a sequence Identical to that obtained from the intact CALLA protein, indicating that V8 peptide III is derived from the partially blocked NH₂-terminus. Reverse phase HPLC separation of the product of the tryptic digest yielded at least 80 different peaks (Figure 1); 8 of these were selected for further HPLC purification with an alternative gradient, yielding eleven tryptic peptides with the sequences shown in Table 1. Tryptic peptide VIII and V8 peptide I contain an overlapping amino acid sequence LNYKEDEYFENIIQN . (The lysine residue in TPVIII was likely not cleaved by enzymes because tryptic digestion of lysine residues followed by C-terminal acidic amino acids is ineffective. Cunningham et al., Biochem., 12:4811 (1973)).

EXAMPLE 3 DNA Library Construction and Screening Nalm-6 RNA was polyA selected and utilized to construct a λgt10 cDNA library according to standard protocols. Gubler, V. and Hoffman, B.S., Gene,

25:263 (1983); Myers et al., PNAS, 77:1316 (1984); Huynh e t al . , In:DNA Cloning Techniques, Oxford Press (1984). 10 μg of polyA+ RNA primed with oligo(dT) was copied with reverse transcriptase; RNAse H and DNA polymerase I were utilized to synthesize the second strands. After methylation of endogenous EcoRI sites, the cDNA was blunt-end ligated to EcoRI linkers and digested with EcoRI. cDNA was size fractionated on a BioGel A50 column to remove excess linkers and cDNAs less than 700 BP in size, and subsequently ligated into the EcoRI site of λgt10 and packaged in vitro.

EXAMPLE 4 Isolation and Characterization of cDNA

Clo ne s E ncodi ng CALL A In order to isolate cDNA clones encoding CALLA, redundant oligonucleo tide probes corresponding to amino acid sequences from the tryptic and V8 peptides were used to screen the Nalm-6 λgt10 cDNA library. Redundant oligonucleotide probes were synthesized on an Applied Biosystems model 381A DNA synthesizer and labelled at their 5" ends. Filters were hybridized with the appropriate oligonucleotide probe for 16 hours in 6X SSC-/15X Denhardts/10 μg/ml. SSDNA/0.1% SDA/0.05% NaPiPO₄ and washed in 6XSSC/0.1% SDS. Hybridization and final wash temperatures were 44ºC and 48°C for oligonucleotide #1, 26°C and 30°C for oligonucleotide #2 and 38°C and 42°C for oligonucleotide #3.

Oligonucleotide #1 (3' TTY-CTY-CTR-CTY-ATR-AARCTY-TT 5', Y = T or C, R = A or G), which corresponds to eight amino acids in both TP VIII and VPI (Table 1), was used in the initial screening. Positive clones were re-screened with a second oligonucleotide (3' TTR-TTR-CTY-ATR-RAN-CT 5', N = A, T, C or G) which corresponds to six amino acid residues NH2-terminal to those represented in o l igonuc leo t ide #1 ( Tab le 1 ) . cDNA ins erts from clones containing sequences complementary to both probes were further analyzed. Each double positive clone contained an identical EcoRI fragment approximately 1.6Kb in size and certain clones contained additional EcoRI fragments of smaller size. The clone containing the longest cDNA insert (approximately 3.5Kb) was subcloned into the m13 vector mp18 and is referred to as clone 1.2 (12.1). An overlapping clone, 1.1 (5'4-1), which contains the complete NH₂-terminus was identified by rescreening the Nalm-6 cDNA library with oligonucleotide #3 (3' GTY-TAC - CTR-TAD -TGN-CTR-TA 5' D = A, G or T) corresponding to s even amino ac ids from V8 pep tide III (Tab le 1 , Figure 3). In order to confirm the CALLA cDNA sequence obtained from clones 1.1 and 1.2, a second set of overlapp ing CALLA cDNA clones ( 2.1, 2.2 and 2.3, Figure 2) was characterized. Clones 1.1 and 2.1 begin 100BP 5' to the initiation methionine ATG . In Figure 2, the open reading frame from clones 1.1, 1.2 and 2.1, 2.2 is indicated. The 3' untranslated sequence from clones 1.2, 2.2, and 2.3 is 1472 BP in length, ending in a polyA sequence. Clone 3.1 (1-8) is identical to the other CALLA cDNAs at its 5' end as shown, but contains an additional 1775 bp of 3' untranslated sequence ending in a polyA tail.

Three phages, containing, respectively, cDNA segments 1.1, 1.2, and 3.1, described above, were deposited, under the terms of the Budapest Treaty, on December 2, 1987 In the American Type Culture Collection, Rockville, Maryland, and given, respectively, ATCC accession Nos. 40395, 40396, and 40397. Applicants assignee, Dana-Farber Cancer Institute, hereby acknowledges its responsibility to replace these cultures should they die before the end of the term of a patent issued hereon, five years after the last request for a culture, or thirty years, whichever is the longer, and its responsibility to notify the depository of the issuance of such patent, at which time the deposits will be made irrevocably available to the public.

Until that time, the deposits will be made available to the Commissioner of Patents under the terms of 37 C.F.R. §§ 1.14 and 35 U.S.C. §1.12.

The nucleotide sequence shown is a composite of clones 1.1, 1.2, and 3.1 (Figure 3). Position 1 (11bp 5' to the ATG) represents the position at which clones 1.1 and 2.1 become identical. The translated CALLA cDNA sequence predicts the indicated 750 amino acid protein with six potential N-linked glycosylation sites (Asn-Xaa- Ser/Thr*).

The eleven CALLA tryptic peptides and three CALLA V8 fragments identified by protein micro - sequencing are underlined in the translated CALLA cDNA sequence. The single 24 amino acid hydrophobic segment (amino acids 27-50) is underlined.

The individual EcoRI inserts from clones 1.1 and 1.2 were sequenced by the Sanger dideoxy chain termination method ; the universal M13 primer and sequence specific oligonucleotide primers were utilized. Sanger et al . , Proc. Natl. Acad, Sci. ,

USA., 74:5463 (1977). Bglll/Hinglll fragments from clone 1.2 were also sequenced In order to confirm the orientation and positions of individual EcoRI fragments. Clones 2.1, 2.2, 2.3 and 3.1 were also sequenced. The sequence was assembled with the PC Gene database program and analyzed utilizing the National Institutes of Health 1987 Genetic Sequence Databank (GenBank, Research Systems Div., Bolt, Branch and Newman, Cambridge, MA) release 56.0, and the Protein Identification Resources Databank.

Clone 1.2 (Figure 2, BP 199 to 3734, Figure 3) has an open reading frame and a translated amino acid sequence which contains all eleven of the tryptic peptides but only two of the three V8 peptides determined by microsequencing of the CALLA protein (Table 1, Figures 2 and 3). The translated sequence of clone 1.2 lacks the NH₂-terminal residues identified from both the intact CALLA protein and V8 peptide III (Table 1, Figure 3). The independently derived CALLA cDNA sequences from 1.1/1.2 and 2.1/2.2/2.3 are identical from BP 1 to 3734, Including 11 nucleotides of 5' untranslated sequence. Each clone contains an additional unique approximately 100 bp of 5' sequence, possible representing alternative 5' splicing. An additional overlapping clone (3.1) was identified which corresponds to the previously characterized clones from its 5' end (bp 2321) through bp 3733; thereafter, clone 3.1 contains an additional 1775 bp of sequence which ends in a polyA tail (Figures 2 and 3). The complete nucleotide sequence derived from CALLA clones 1.1, 1.2 and 3.1 contains an open reading frame of 2250 bases (positions 12-2261) beginning with an ATG methionlne codon which is preceded by a TAG termination codon at bp 6-8 in-frame. After 2250 bp of nucleotides encoding an open reading frame, there is a TGA termination codon at positions 2262-2264 and 3244 bp of 3' untranslated sequence ending in a polyA tail at bp 5508. A canonical polyadenylation signal (AATAAA) is found 25 bp upstream of the polyadenylation site in clone 3.1. Additional canonical polyadenylation signals are found in the 3' untranslated region at bp 3088-3093, 3371-3376, 3792-3797, and 4406-4411. The translated CALLA cDNA sequence includes 170 of the 182 amino acid residues identified by microsequencing the intact CALLA protein, the derived tryptic peptides, and the V8 protease fragments (Table 1, Figure 3), providing conclusive evidence that the cDNA represents authentic CALLA.

The CALLA cDNA predicts a 750 amino acid protein with a polypeptide core MW of 85.5KD and six potential N-linked glycosylation sites

(Asn-Xaa- Ser/Thr) located at amino acid positions 144, 284, 310, 324, 627. Although the number of potential N-linked glycosylation sites that are actually utilized is unknown, the MW of CALLA following removal of N-linked sugars is in good agreement with that predicted from the CALLA cDNA. The translated CALLA cDNA sequence contains 12 cysteine residues with spacing that is not reminiscent of immunoglobulin domains or other known structures.

The translated CALLA cDNA sequence has a single hydrophobic 24 amino acid segment at positions 27-50 with the characteristics of a transmembrane region. This hydrophobic segment is preceded by five basic residues, suggesting that CALLA is oriented such that the NH₂ terminus constitutes the cytoplasmic tail. The translated CALLA sequence does not contain an initial hydrophobic NH₂ terminal segment that could function as a signal peptide. In fact, comparison of the translated cDNA sequence with that derived from the Intact protein and N-terminal V8 fragment (V8III) reveals that the only NH₂- terminal proteolytic processing which occurs during CALLA synthesis results In removal of the initiation methionine (Table 1, Figure 3). Hydrophobicity values were calculated by the PRSTRC algorithm assigned over windows of five amino acids (Ralph eta l . , CABIOS 3:211 (1987)). In Figure 4, positions above the baseline are hydrophobic. Amino acid positions are indicated below the plot and the transmembrane region is shaded. In contrast to previous studies suggesting that

CALLA is a non-integral membrane protein, these results indicate that CALLA is a type II integral membrane protein with a short (25 amino acid) NH₂ terminal cytoplasmic tail, a single 24 amino acid hydrophobic region that functions as both an uncleaved internal signal sequence and a transmembrane segment, and a large (700 amino acids) carboxy terminal extracellular domain. Wickner, W.T. and Lodish, K.F., Science, 230:400 (1985). The fact that all six putative N-linked glycosylation sites are located in the carboxy terminal segment is consistent with this interpretation, as are the ¹²⁵I surface labelling studies, given that there are no tyrosine residues in the NH₂-terminal segment. Searches of the GenBank data base (release

52.0) and data base from Protein Identification

Resource (release 12.0) reveal that CALLA Is a novel protein with no significant internal duplications or previously characterized active sites or consensus binding site sequences. The region immediately preceding and including the transmembrane segment of CALLA has partial identity (14 amino acids out of 31 amino acids with no gaps) with that of another type II membrane protein, pro-sucrase isomaltase, raising the possibility that the dual function transmembrane segments of certain type II proteins may have common features. Hunziker et al., Cell, 46:227 (1986).

Similar post translational cleavage of NH₂-terminal methionine residues has been shown for other class II membrane proteins.. Nunziker e t al . , Cell , 46:227 (1986); Holland e t al . , Proc. Natl. Acad. Sci., USA, 81:7338 (1984).

RNA samples were prepared and analyzed by

Northern blotting as previously described. Shipp,

M.A. and Reinherz, E.L. J. Immunol., 139:2143 (1987). Filters were hybridized with a ³²P labelled

1.6Kb EcoRI CALLA cDNA fragment which includes bp

541-2227, an uninterrupted open reading frame

(Figure 3). Filters were washed in 2 × SSC/0.1%

NaDodSO₄ at 25°C for 30 minutes and 65°C for 30 minutes, and 0.2 × SSC/0.1% NaDodSO₄ at 65°C for 60 minutes.

RNAs from a panel of CALLA+ and CALLA- lymphoid cell lines and primary tumors were probed in

Northern analysis with an approximately 1.6Kb EcoRI fragment from the 1.2 CALLA cDNA clone. Twenty micrograms of total RNA from the following cell types were analyzed: Nalm-6; Raji, a CALLA + Burkitts lymphoma cell line; Molt -4, a CALLA+ T cell acute lymphoblas tic leukemia (ALL) cell line; HSB, Jurkat and J77, CALLA- T cell leukemia cell lines; unstimulated peripheral blood mononuclear cells (PBMC); mitogen triggered PBMC ; LAZ 221, a CALLA+ ALL cell line, and LAZ 388, a CALLA- EBV transformed lymphoblas toid line from the same donor as LAZ 221. A 10 μg sample of poly A selected Nalm-6 RNA was also analyzed. Cell lines defined as CALLA positive (by immunofluorescence using the J5 antibody) Including Nalm-6, the ALL cell lines Laz 221 and Molt-4, and the Burkitts lymphoma cell line Raji contain two major CALLA mRNAs: a 3.7Kb and a 5.7Kb CALLA transcript. In contrast, CALLA negative sources, including three T cell tumor lines, Jurkat J77. (a second Jurkat clone), and HSB, and EBV transformed lymphoblas toid line, Laz 388, and resting and mitogen-stimulsted peripheral blood lymphocytes, lack these transcripts. In addition to the 3.7 and 5.7 Kb mRNAs, there are other low abundance CALLA transcripts in J5 positive cells. For example, in polyA+ RNA from Nalm-6, additional less abundant CALLA-related mRNAs of 4.5, 3.1, and 2.7Kb are also detected by Northern blotting. It Is likely that the 5.7Kb transcript results from utilization of the downstream canonical polyadenylation site located in clone 3.1. An anti-sense oligonucleotide derived from the 3' untranslated region unique to clone 3.1 (bp 4865-4903) identifies only the larger 5.7 Kb transcript, consistent with this interpretation. The 3.7 Kb transcript probably results from utilization of the polyadenylation signal at bp 3792, and the minor mRNAs of 4.5, 3.1 and 2.7 Kb from utilization of polyadenylation signals located at 4406-4411, 3371-3376, and 3088-1093, respectively.

Anti-CALLA monoclonal antibody reactivity has also been observed on normal granulocytes, a subpopulation of bone marrow stromal cells, certain elements in kidney, fetal intestine, and breast, and certain non-lymphoid tumor cells. Braun e t al . , Blood, 61:718 (1983); Cossman et al., J. Exp. Med., 157:1064 (1983); Keating et a l . , Bri t. J. Haematol,,

55:623 (1983); Pesando et al., J. Immunol ., 131.:2038 (1983); Metzgan et al, J. Exp. Med, 154:1249

(1981); Carrel et al., J. Immaunol., 130:2465 (1983). With the exception of granulocytes where peptide mapping has shown, relatedness of the granulocyte protein to the lymphoid CALLA protein the nature of antibody reactivity with these cellular sources has been ill-defined. McCormack, R.T., J. Immunol ., 137:1075 (1986). In order to determine whether CALLA transcripts from non-lymphoid sources are similar to those in lymphoid cells, total RNA was isolated from granulocytes, a CALLA+ fibroblast line (HB-100), and two additional CALLA+ tumor cell lines, a melanoma line (G361) and a colon carcinoma line (CLL-227) and probed by northern blot. Pesando et al., J. Immunol ., 131:2038 (1983). Twenty micrograms of each cell type was analyzed: HB-100, a CALLA+ fibroblast strain; G361, a CALLA-melanoma line; CLL-227, a CALLA+ colon carcinoma line; Nalm-6 (for comparison); and granulocyte. As is the case in Nalm-6 cells, the 3.7Kb transcript is the major CALLA mRNA in HB-100 and CCL-227 cells. In contrast, the 3.7 and 5.7Kb CALLA transcripts are both abundant in G-361 cells whereas the 5.7 Kb m RNA is the major CALLA transcript in granulocytes. The basis for this tissue specific differential expression of the 3.7 and 5.7Kb CALLA transcripts in different CALLA+ cell types is not presently known. Nevertheless, the fact that the same major CALLA messages are expressed in these non-lymphoid sources indicates that J5 recognizes authentic CALLA on these cells and not another protein with a common J5 epitope. The small differences in size of lymphoid and granulocyte CALLA proteins (95-100 versus

95-110KD) presumably relate to different patterns of

N-lined glycan addition. Cossman e t al., J. Ex p .

Med., 157:1064 (1983); McCormack, et al., Immunol ., 137:1075 (1986); Pesando et al., Blood 67:588 (1986). That six potential N-linked glycosylation sites are identified in the predicted 85.KD core polypeptide, but not all are likely to be utilized in the Nalm-6 CALLA protein, is consistent with this interpretation.

DNA Blot Hybridization Analysis

DNA was prepared and analyzed by Southern blotting according to standard protocols. Maniatis e t al ., Molecular Cloning, Cold Spring Harbor, NY,

(1982). BamHI and Hindlll digested genomic DNAs from a panel of CALLA+ and CALLA- sources (Nalm-6, Laz 221, Laz 388, and PBL) were Southern blotted and hybridized with the approximately 1.6Kb EcoRI CALLA cDNA fragment. Ten micrograms of high molecular weight DNA from PBL; Laz 388; Laz 221; and Nalm-6 was digested with Hindlll or BamHI, and subjected to Southern blot analysis using the 1.6Kb CALLA cDNA probe . There are thre e BamHI genomic fragments (sizes: two bands greater than 35Kb and one 62.5Kb) and eight Hindlll genomic fragments (sizes: 35,26, 17, 8.4, 4.8, 3.7, 2.9, and 2.5Kb) that hybridize with the approximately 1.6Kb CALLA cDNA probe. There appears to be no rearrangement of the gene in CALLA+ malignancies detectable with these restriction enzymes. Furthermore, the restriction pattern suggests that the CALLA gene is most likely present as a single copy gene in the human haploid genome.

The molecular cloning of CALLA and its identification as a type II transmembrane glycoprotein does not make it possible to infer, in a direct way, its role in lymphoid function. Proteins in this class have diverse functions ranging from receptors to membrane bound enzymes and include the transferrin receptor, the asialoglycoprotein receptor, influenza viral neuraminidase, gamma glutamyl transpeptidase, prosucrase-isomaltase complex and the invariant chain of HLA .

The fact that CALLA is an integral membrane protein is consistent with previous studies demonstrating rapid cell surface redistribution, internalization, and degradation of the CALLA- antibody complex following specific antibody treatment of CALLA+ cells at 37°C. Antibody induced modulation of CALLA was noted to be similar to that seen with cell surface receptors such as surface Ig and T3-Ti; it also resembled the specific down regulation or loss of cell surface receptors induced by many peptide hormones. The fact that CALLA has a relatively short cytoplasmic tail makes it unlikely to serve a direct signal transduction function. CALLA expression appears on uncommitted TdT+ lymphoid progenitors and generally declines as the cells display evidence of B cell or T cell commitment, suggesting that the antigen may function in the earliest stages of lymphoid differentiation. The fact that CALLA is also expressed by a subpopulation of bone marrow stromal cells raises the possibility that CALLA may participate in the microenvironment necessary for early lymphoid maturation.

EXAMPLE 5 CALLA Open Reading Frame Construct

In order to generate a CALLA cDNA containing the entire open reading frame (bp 12-2261), two recombinant λgt10 clones containing CALLA cDNA fragments bp -132 to +583 (clone 1.1) and bp +127 to +3723 (clone 2) were utilized (Figure 1). The numbering system for the two λgt10 clones is derived from Shipp, e t al ., in which bp 1 is located 11 bp 5' to the initiation methionine. Shipp, M.A. et al., Proc. Natl. Acad. Sci. USA, 85:4819-4823 (1988). DNA from clone 1.1 was digested with EcoRI and Aval, yielding a 0.435Kb fragment (Figure 5). Aliquots of DNA from clone 2 were digested with Aval and Clal, yielding a 0.499Kb fragment, or with Clal and ApaI, yielding a 1.546Kb fragment, respectively (Figure 5). The plasmid vector pBluescript SK(+) (Stratagene) was digested with EcoRI and Apal. The 0.435Kb EcoRI-Aval, 0.499Kb Ava-Clal, 1.546Kb Clal-Apal CALLA cDNA fragments and the EcoRI-Apal cut plasmid vector were gel-purified, l i gate d and used to transform DH5α+ Eschericia coli (Bethesda Research Labs).

Recombinants were identified on the basis of blue/white color selection on LB plates containing 100 μg/ml ampicillin, 80 μg/ml X-gal and 20 mM lPTG. Following a large scale plasmid preparation, the reconstructed CALLA cDNA containing the intact open reading frame was excised from the plasmid vec to r us in»g Drl - Apal and b lunted on its 3 ' end wi th T4 DNA polymerase. In order to generate a 5' Sail site, the resulting 5' and 3' blunt-ended CALLA cDNA fragment was ligated into EcoR5 digested pBluescript SK(+) which contains a SalI site in the polylinker.

Following transformation, recombinants were identified and analyzed for orientation using a panel of diagnostic restriction endonucleases. An appropriate clone was further analyzed by sequencing the 5' and 3' ends and the EcoRI-Aval, Aval-Clal, Clal-Apal junctions of the CALLA insert using the dideoxy chain termination method. Sanger, F. e t al . , Proc. Natl. Acad. Sci. USA, 74:5463-5467 (1977).

In order to obtain the CALLA open reading frame with the SalI site from the polylinker, the fragment was excised with Smal and Apal which cleaved at polylinker sites 3' of the CALLA fragment arid 5' of the CALLA fragment and SalI site, respectively. The Apal end was blunted with T4 DNA polymerase and the modified CALLA open reading frame insert was ligated into EcoR5 cut pBluescript SK(+). Following transformation, recombinants containing a resulting 3' SalI end from the polylinker were identified with diagnostic SalI digestions. Following a large scale plasmid preparation of a representative clone, the CALLA open reading frame was excised with SalI, and ligated into the Xhol site of pIGTE/N (Figure 6). The pIGTE/N plasmid (Figure 6) contains the human immunoglobulin promoter, both human and murine immunoglobulin enhancers, an SV40 intron and polyadenylation signal, and the gene for

CMV-neomycIn resistance. Following transformation, recombinants were analyzed for orientation using a panel of diagnostic restriction endonucleases. A plasmid, designated pIGTE/N CALLA_s, which contains the CALLA open reading frame in the sense orientation was obtained.

EXAMPLE 6 Generation of CALLA+ Cell Lines

The pIGTE/N CALLA construct was transfected into the CALLA- murine myeloma cell line, J558, by electroporation as previously described. Patten, H. et al., Proc. Natl. Acad. Sci. USA 81:7161-7165 (1984). Briefly, 2 × 10⁷ cells were washed once in ice cold phosphate buffered saline (7.7 mM K₂HPO₄, 145 mM NaCl, 2.3 mM KH₂PO₄) and resuspended in 500 μl of ice cold phosphate buffered saline containing 50 μg of pIGTE/N CALLA_s The cell/DNA solution was transferred to an electroporation cuvette and, after 5 minutes on ice, electroporated with 2000 V. Following a 10 minutes room temperature incubation, cells were diluted in 10 ml of RPMI supplemented with 10% fetal calf serum/2% glutamine/100 U/ml penicillin and 100 μg/ml streptomycin and cultured for 48 hours at 37°C in 5% CO₂. Thereafter, cells were maintained in RPMI supplemented with 10% fetal calf serum/2% glutamine/100 U/ml penicillin and 100 μg/ml streptomycin and 900 μg/ml of G418 for 14 day s and phenotyped for CALLA expression using the anti-CALLA monoclonal antibody, J5. Ritz, J. et al.. Nature, 283:583-585 (1980). Subsequently, J5+ cells were selected by fluo r e s c enc e activated cell sorting (Epics V), using standard techniques, and cloned by limiting dilution at 0.5 cells/well in G418-containing media. Ritz, J. et al., Nature (London), 283:583-585 (1980).

EXAMPLE 7 Expression of Human CALLA in Mur ine Transfectants

Following G418 selection, as described in Example 6, CALLA+J558 cells were identified by phenotyping with the J5 anti-CALLA monoclonal antibody, sorted and cloned by limiting dilution. Two J5+ subclones, As-3 and A2-2, which had high and low levels of CALLA expression, respectively, were chosen for further analysis. Figure 2 presents a comparison of relative CALLA fluorescence of these two CALLA+ stable transfectants, the parental CALLA multiple myeloma line, J558, and the CALLA+ acute lymphoblastic leukemia line from which the CALLA cDNA was isolated, Nalm-6. As indicated, J558 lacks detectable cell surface CALLA expres s i on (mean channel Fluorescence 0). The two J5+ sub clone s (A2-3 and A2-2) express cell surface CALLA. Note that A2-3 expresses a substantially higher number of CALLA sites per cell than does A2-2 (mean channel fluorescence 116.8 and 7.6, respectively). Comparison of the mean channel fluorescence of A2-3 and A2-2 with the mean channel fluorescence of Nalm-6 (172.6), indicates that A2-3 expresses CALLA at a level 67.7% that of Nalm-6 and A2-2 expresses CALLA at only 4.4% the level of Nalm-6.

EXAMPLE 8 Enzymatic Assay for Neutral Endopeptidase 24.11 Activity As described previously, search of the CALLA cDNA sequence against GenBank (release 56, 6/88) revealed striking homologies with the rat and rabbit neutral endopeptidase 24.11, commonly referred to as "enkephalinase". The DASHER program (D.V. Faulkner, Molecular

Biology Computer Resource Center, Dana Farber Cancer Institute) was used to compare CALLA cDNA segments of 600 bp each, with a 100 bp overlaps, to 600 bp segments with 100 bp overlaps of each sequence in the GenBank data base. Relates segments of the CALLA cDNA, rat enkephalinase and rabbit neutral endopeptidase 24.11 (bp 1-600, 501-1100, 1001-1600, 1501-2100 and 2001-2600) had homology scores of 121 through 236. In the DASHER program, homology scores greater than 12 are thought to be of potential significance. Thus, the high scores noted herein are indicative of near identity. Subsequent comparison of amino acid sequences of human CALLA and rat enkephalinase and rabbit neutral endopeptidase molecules showed 94% ident i ty in each case. Shipp, M.A. e t al ., Proc. Natl. Acad. Sci.,

USA , 8 5 : 4819 - 48 23 ( 19 8 8 ) ; Mal f roy , B . e t al . , Biochem. and Biophys. Res. Comm. , 144:59-66 (1987);

Devault, A. et al., EMBO J., 6:1317-1322 (1987).

Analysis of the recently reported human homologue shows virtual identity with CALLA; the latter two sequences differ by one amino acid, which is either the consequence of a sequencing error or genetic polymorphism. Malfroy, B. et al., FEBS Letters,

229:206-210 (1988). Neutral endopeptidase 24 . 11 activity was meas ure d f luorome tr ical ly in a coupled assay using glutaryl-Ala-Ala-Phe-4-methoxy-2-naphthylamide (Enzyme Systems Products) as substrate. Orlowski, M. & S. Wilk, Biochem., 20:4942-4950 (1981). Cleavage of this substrate by neutral endopeptidase 24.11 yields Phe-4-methoxy-2-naphthylamide which, in the presence of aminopeptidase activity, is converted to the fluorescent product 4-methoxy-2-naphthylamine. Reaction mixtures contained 0.1 mM substrate, 100mM MES

2[(N-M-orpholino)ethanesulfonic acid], pH 6.5, 0.3 M NaCl, 0.5 milliunits of purified rat brain aminopeptidase and enzyme in a final volume of 100 μl. McLellan, S. et al., J. Neurol, (1988). Reactions were initiated with enzyme and followed at 30°C at an excitation wavelength of 340 nm and an emission wavelength of 425 nm using an Aminco-Bowman spectrofluorometer equipped with a thermostated cell holder and strip chart recorder. Inhibition by phosphoramidon (2 μM) was used to confirm the specificity of the assay. Hudgin, R.L. et a l . , Life Sci., 29:2593-2601 (1981). Protein was determined by the BCA method. Smith, P.K. et al., Anal. Biochem., 150:76-85 (1985).

Cell extracts were prepared by resuspending washed cells in 200-400 μl of 20 mM MES, pH 6.5, containing 1% octyl glycoside

(1-0-n-octyl-β-D-glycopyranoside, Sigma). After a 1-hour room temperature incubation, the samples were centrifuged for 30 min in an Eppendorf centrifuge and the supernatant taken for enzyme assays. Whole cell suspensions were washed once in RPMI and then utilized at a concentration of 10³-10⁴ cells per 100 μl reaction mixture. For the preparation of the membrane fraction, washed cells were homogenized in

1 ml of Tris buffered saline using a teflon glass homogenizer and the membrane fraction isolated by centrifugation at 40,000 x g for 30 minutes. This fraction was resuspended in the MES/octylglycos ide buffer and treated as described above.

Table 2 illustrates the results of the enzymatic assay performed using whole cell suspensions of the individual cell populations. Cell suspensions of Nalm-6, A2-3, A2-2 and J558 exhibited neutral endopeptidase activity of 4789, 1906, 446 and 0.2 nanomoles of product per hr per 10⁶ cells, respectively. TABLE 2. DETERMINATION OF NEUTRAL ENDOPEPTIDASE ACTIVITY IN WHOLE CELL SUSPENSIONS

Specific activity (nmols/hr/10⁶ cells)

Cell suspension phosphoramidon phosphoramidon

Nalm-6 4,789

A2-3 1,906

A2-2 446 18

J558 0.2 0.2

The value of 0.2 nanomoles per hr per 10 cells is within the experimental error of no activity.

Addition of the specific inhibitor phosphoramidon to the Nalm-6, A2-3 and A2-2 cell suspensions markedly reduced neutral endopeptidase activity to 53, 52 and 18 nmols/h/10⁶ cells (Table 2). Hudgin, R.L. et al. , Life Sci. , 29:2593-2601 (1981)

TABLE 3 DETERMINATION OF NEUTRAL ENDOPEPTIDASE ACTIVITY IN TOTAL CELL LYSATES

Cell source Specific activity (nmols/min/mg protein) total cell lysate total cell lysate membrane fraction

+ phosphoramidon

Nalm-6 9.96 0.02 9.96 A2-3 2.35 0.08 2.12 A2-2 1.18 0.29 ND J558 0.08 0.05 ND

ND = not determined As indicated in Table 3, cell lysates from Nalm-6, A2-3, A2-2, and J558 contained neutral endopeptidase specific activities of 9.96, 2.35, 1.18 and 0.08 nmols/min/mg protein of neutral endopeptidase activity, respectively. When the assays were performed in the presence of phosphoramidon, the neutral endopeptidase activity associated with the Nalm-6, A2-3 and A2-2 cell lysates was dramatically reduced to 0.20, 0.08 and 0.29 nmols/min/mg protein, respectively (Table 3). The observation that the apparent activity In J558 cells is insensitive to phosphoramidon inhibition suggests that this low level of apparent activity is not due to neutral endopeptidase 24.11 and is within the experimental error of no activity. Subcellular fractionation of the Nalm-6 and A2-3 cells demonstrated that 100% of the Nalm-6 and more than 90% of the A2-3 neutral endopeptidase activity was associated with the membrane fraction (Table 3). Comparison of the levels of neutral endopeptidase activity (Tables 2 and 3) and levels of CALLA expression of the four cell lines (Figure 7) indicated that there was a correlation between neutral endopeptidase activity and cell surface CALLA expression.

Equivalents

Those skilled in the art will recognize, or be able to ascertain, using no more than routine experimentation, many equivalents to the specific embodiments of the invention described specifically herein. Such equivalents are intended to be encompassed in the scope of the following claims

Claims

CLA IMS

1. Isolated DNA encoding human CALLA capable of binding to an anti-CALLA antibody or a fragment of said DNA.

2. Isolated DNA of Claim 1 encoding a soluble human CALLA fragment.

3. Isolated DNA of Claim 2 encoding a soluble human CALLA fragment having functional endopeptidase activity.

4. Isolated DNA of Claim 3 having all or a portion of the nucleotide sequence of Figure 3.

5. Human CALLA cDNA contained in a phage selected from the group consisting of: the phage deposited at the ATCC under accession number 40395; the phage deposited at the ATCC under accession number 40396; and the phage deposited at the ATCC under accession number 40397.

6. A soluble human CALLA fragment capable of binding to an anti-CALLA antibody.

7. A soluble CALLA fragment of Claim 6, said fragment including none or only a portion of the transmembrane portion of CALLA.

8. A soluble human CALLA fragment of Claim 6 which is a neutral endopeptidase.

9. A protein or a polypeptide having the amino acid sequence of Figure 3 or of a portion thereof.

10. Essentially pure neutral endopeptidase expressed by human lymphoblastic leukemia cells.

11. A vector comprising a DNA sequence encoding the soluble CALLA fragment of Claim 6.

12. A cell containing the vector of Claim 11.

13. A phage containing human CALLA cDNA, selected from the group consisting of: the phage deposited at the ATCC under accession number 40395; the phage deposited at the ATCC under accession number 40396; and the phage deposited at the ATCC under accession number 40397.

14. A cell containing a phage selected from the group consisting of: the phage deposited at the ATCC under accession number 40395; the phage deposited at the ATCC under accession number 40396; and the phage deposited at the ATCC under accession number 40397.

15. A method for treating a human patient suffering from a medical condition characterized by the presence in said patient of cells associated with said medical condition, which cells have CALLA on their surfaces, said method comprising administering to said patient the soluble CALLA fragment of Claim 3, in an amount sufficient to inhibit proliferation of said cells.

16. A method of reducing leukemic cell growth, comprising administering to an individual in whom leukemic cell growth is present an inhibitor of CALLA in an amount sufficient to inhibit leukemic cell growth.

17. A method of Claim 16 wherein the inhibitor of CALLA comprises a substance capable of binding to CALLA and of reducing enzymatic activity of CALLA.

18. A method of Claim 16 wherein the inhibitor of CALLA comprises a substance capable of binding to CALLA and preventing, thereby, binding of CALLA and naturally-occurring ligands which bind CALLA.

19. A method of producing an analgesic effect in an individual, compris ing administering to an individual in whom an analgesic effect is desired a substance which interferes with the ability of CALLA to cleave opioid pentapeptides.

20. A composition for treatment of leukemic cell growth in an Individual, comprising an inhibitor of CALLA.

21. A composition of Claim 20, in which the inhibitor comprises a substance capable of binding to CALLA and of reducing enzymatic activity of CALLA.

22. A composition of Claim 20 in which the inhibitor comprises a substance capable of binding to CALLA and preventing binding of CALLA and naturally-occurring ligands which bind CALLA.

23. An analgesic composition comprising a substance which interferes with the ability of CALLA to cleave opioid pentapeptides.