US20140178986A1

US20140178986A1 - Soluble inhibitors of vascular endothelial growth factor and use thereof

Info

Publication number: US20140178986A1
Application number: US13/963,083
Authority: US
Inventors: Michael Klagsbrun; Shay Soker; Michael L. Gagnon
Original assignee: Childrens Medical Center Corp
Current assignee: Childrens Medical Center Corp
Priority date: 1997-12-09
Filing date: 2013-08-09
Publication date: 2014-06-26
Also published as: JP2014207904A; US20080261867A1; CA2313348A1; US8529905B2; JP2014193160A; US7736655B2; US20030104532A1; AU1633999A; JP2012044994A; WO1999029858A1; JP4404479B2; US20100267140A1; JP5480488B2; JP2002505078A; JP2009039133A; EP1037981A1; US7273612B2

Abstract

The present invention relates to cDNA encoding a soluble neuropilin protein (sNP) which is isolated from neuropilin (NP) producing cells or is recombinantly engineered from NP-encoding DNA. NP-1 and NP-2 are preferred NPs but any neuropilin or VEGF receptor (VEGFR), where the constituents share at least about 85% homology with either of the above VEGF₁₆₅R/NP-1 and NP-2. More preferably, such constituent shares at least 90% homology. Still more preferably, each constituent shares at least 95% homology.

Description

CROSS REFERENCE TO RELATED APPLICATIONS

This application is a continuation application that claims benefit under 35 USC §120 of co-pending U.S. patent application Ser. No. 12/756,737 filed on Apr. 8, 2010, which is a continuation application that claims benefit under 35 USC §120 of U.S. patent application Ser. No. 11/893,633 filed on Aug. 17, 2007 and issued as U.S. Pat. No. 7,736,655 on Jun. 15, 2010, which is a divisional application that claims benefit under 35 USC §121 of U.S. patent application Ser. No. 10/104,610 filed on Mar. 22, 2002 and issued as U.S. Pat. No. 7,273,612 on Sep. 25, 2007, which is a continuation application that claims benefit under 35 USC §120 of U.S. patent application Ser. No. 09/580,989 filed on May 30, 2000 and now abandoned, which is a continuation application of International Application No. PCT/US98/26138 filed on Dec. 9, 1998, which designated the U.S. and which claims benefit under 35 U.S.C. §119(e) of U.S. Provisional Application No. 60/069,155 filed on Dec. 9, 1997, U.S. Provisional Application No. 60/069,687 filed on Dec. 12, 1997, and U.S. Provisional Application No. 60/099,615 filed on Sep. 9, 1998, the contents of each of which are incorporated herein by reference in their entireties.

GOVERNMENT SUPPORT

The invention was made with Government Support under grant numbers CA37392 and CA45548 awarded by the National Institute of Health. The U.S. Government has certain rights in this invention.

SEQUENCE LISTING

The instant application contains a Sequence Listing which has been submitted in ASCII format via EFS-Web and is hereby incorporated by reference in its entirety. Said ASCII copy, created on Aug. 7, 2013, is named 701039-072096 SequenceListing.txt and is 49,152 bytes in size.

FIELD OF THE INVENTION

The present invention relates to vascular endothelial growth factor (VEGF). More particularly, the invention relates to soluble inhibitors of VEGF and use of those inhibitors in the treatment of disorders that are associated with VEGF.

BACKGROUND OF THE INVENTION

Blood vessels are the means by which oxygen and nutrients are supplied to living tissues and waste products are removed from living tissue. Angiogenesis refers to the process by which new blood vessels are formed. See, for example, the review by Folkman and Shing, J. Biol. Chem. 267, 10931-10934 (1992), Dvorak, et al., J. Exp. Med., 174, 1275-1278 (1991)). Thus, where appropriate, angiogenesis is a critical biological process. It is essential in reproduction, development and wound repair. However, inappropriate angiogenesis can have severe negative consequences. For example, it is only after many solid tumors are vascularized as a result of angiogenesis that the tumors have a sufficient supply of oxygen and nutrients that permit it to grow rapidly and metastasize. Because maintaining the rate of angiogenesis in its proper equilibrium is so critical to a range of functions, it must be carefully regulated in order to maintain health. The angiogenesis process is believed to begin with the degradation of the basement membrane by proteases secreted from endothelial cells (EC) activated by mitogens such as vascular endothelial growth factor (VEGF) and basic fibroblast growth factor (bFGF). The cells migrate and proliferate, leading to the formation of solid endothelial cell sprouts into the stromal space, then, vascular loops are formed and capillary tubes develop with formation of tight junctions and deposition of new basement membrane.
In adults, the proliferation rate of endothelial cells is typically low compared to other cell types in the body. The turnover time of these cells can exceed one thousand days. Physiological exceptions in which angiogenesis results in rapid proliferation typically occurs under tight regulation, such as found in the female reproduction system and during wound healing.
The rate of angiogenesis involves a change in the local equilibrium between positive and negative regulators of the growth of microvessels. The therapeutic implications of angiogenic growth factors were first described by Folkman and colleagues over two decades ago (Folkman, N. Engl. J. Med., 285:1182-1186 (1971)). Abnormal angiogenesis occurs when the body loses at least some control of angiogenesis, resulting in either excessive or insufficient blood vessel growth. For instance, conditions such as ulcers, strokes, and heart attacks may result from the absence of angiogenesis normally required for natural healing. In contrast, excessive blood vessel proliferation can result in tumor growth, tumor spread, blindness, psoriasis and rheumatoid arthritis.
Thus, there are instances where a greater degree of angiogenesis is desirable—increasing blood circulation, wound healing, and ulcer healing. For example, recent investigations have established the feasibility of using recombinant angiogenic growth factors, such as fibroblast growth factor (FGF) family (Yanagisawa-Miwa, et al., Science, 257:1401-1403 (1992) and Baffour, et al., J Vasc Surg, 16:181-91 (1992)), endothelial cell growth factor (ECGF)(Pu, et al., J Surg Res, 54:575-83 (1993)), and more recently, vascular endothelial growth factor (VEGF) to expedite and/or augment collateral artery development in animal models of myocardial and hindlimb ischemia (Takeshita, et al., Circulation, 90:228-234 (1994) and Takeshita, et al., J Clin Invest, 93:662-70 (1994)).
Conversely, there are instances, where inhibition of angiogenesis is desirable. For example, many diseases are driven by persistent unregulated angiogenesis, also sometimes referred to as “neovascularization.” In arthritis, new capillary blood vessels invade the joint and destroy cartilage. In diabetes, new capillaries invade the vitreous, bleed, and cause blindness. Ocular neovascularization is the most common cause of blindness. Tumor growth and metastasis are angiogenesis-dependent. A tumor must continuously stimulate the growth of new capillary blood vessels for the tumor itself to grow.
There is mounting evidence that VEGF may be a major regulator of angiogenesis (reviewed in Ferrara, et al., Endocr. Rev., 13, 18-32 (1992); Klagsbrun, et al., Curr. Biol., 3, 699-702 (1993); Ferrara, et al., Biochem. Biophjs. Res. Commun., 161, 851-858 (1989)). VEGF was initially purified from the conditioned media of folliculostellate cells (Ferrara, et al., Biochem. Biophjs. Res. Commun., 161, 851-858 (1989)) and from a variety of tumor cell lines (Myoken, et al., Proc. Natl. Acad. Sci. USA, 88:5819-5823 (1991); Plouet, et al., EMBO. J., 8:3801-3806 (1991)). VEGF was found to be identical to vascular permeability factor, a regulator of blood vessel permeability that was purified from the conditioned medium of U937 cells at the same time (Keck, et al., Science, 246:1309-1312 (1989)). VEGF is a specific mitogen for endothelial cells (EC) in vitro and a potent angiogenic factor in vivo. The expression of VEGF is up-regulated in tissue undergoing vascularization during embryogenesis and the female reproductive cycle (Brier, et al., Development, 114:521-532 (1992); Shweiki, et al., J. Clin. Invest., 91:2235-2243 (1993)). High levels of VEGF are expressed in various types of tumors, but not in normal tissue, in response to tumor-induced hypoxia (Shweiki, et al., Nature 359:843-846 (1992); Dvorak et al., J. Exp. Med., 174:1275-1278 (1991); Plate, et al., Cancer Res., 53:5822-5827; Ikea, et al., J. Biol. Chem., 270:19761-19766 (1986)). Treatment of tumors with monoclonal antibodies directed against VEGF resulted in a dramatic reduction in tumor mass due to the suppression of tumor angiogeneis (Kim, et al., Nature, 382:841-844 (1993)). VEGF appears to play a principle role in many pathological states and processes related to neovascularization. Regulation of VEGF expression in affected tissues could therefore be key in treatment or prevention of VEGF induced neovascularization/angiogenesis.
VEGF exists in a number of different isoforms that are produced by alternative splicing from a single gene containing eight exons (Ferrara, et al., Endocr. Rev., 13:18-32 (1992); Tischer, et al., J. Biol. Chem., 806:11947-11954 (1991); Ferrara, et al., Trends Cardio Med., 3:244-250 (1993); Polterak, et al., J. Biol. Chem., 272:7151-7158 (1997)). Human VEGF isoforms consists of monomers of 121, 145, 165, 189, and 206 amino acids, each capable of making an active homodimer (Polterak et al., J. Biol. Chem., 272:7151-7158 (1997); Houck, et al., Mol. Endocrinol., 8:1806-1814 (1991)). The VEGF₁₂₁and VEGF₁₆₅isoforms are the most abundant. VEGF₁₂₁is the only VEGF isoforms that does not bind to heparin and is totally secreted into the culture medium. VEGF₁₆₅is functionally different than VEGF₁₂₁in that it binds to heparin and cell surface heparin sulfate proteoglycans (HSPGs) and is only partially released into the culture medium (Houck, et al., J. Biol. Chem., 247:28031-28037 (1992); Park, et al., Mol. Biol. Chem., 4:1317-1326 (1993)). The remaining isoforms are entirely associated with cell surface and extracellular matrix HSPGs (Houck, et al., J. Biol. Chem., 247:28031-28037 (1992); Park, et al., Mol. Biol. Chem., 4:1317-1326 (1993)).
VEGF receptor tyrosine kinases, KDR/Flk-1 and/or Flt-1, are mostly expressed by EC (Terman, et al., Biochem. Biophys. Res. Commun., 187:1579-1586 (1992); Shibuya, et al., Oncogene, 5:519-524 (1990); De Vries, et al., Science, 265:989-991 (1992); Gitay-Goran, et al., J. Biol. Chem., 287:6003-6096 (1992); Jakeman, et al., J. Clin. Invest., 89:244-253 (1992)). It appears that VEGF activities such as mitogenicity, chemotaxis, and induction of morphological changes are mediated by KDR/Flk-1 but not Flt-1, even though both receptors undergo phosphorylation upon binding of VEGF (Millauer, et al., Cell, 72:835-846 (1993); Waltenberger, et al., J. Biol. Chem., 269:26988-26995 (1994); Seetharam, et al., Oncogene, 10:135-147 (1995); Yoshida, et al., Growth Factors, 7:131-138 (1996)). Recently, Soker et al., identified a new VEGF receptor which is expressed on EC and various tumor-derived cell lines such as breast cancer-derived MDA-MB-231 (231) cells (Soker, et al., J. Biol. Chem., 271:5761-5767 (1996)). This receptor requires the VEGF isoform to contain the portion encoded by exon 7. For example, although both VEGF₁₂₁and VEGF₁₆₅bind to KDR/Flk-1 and Flt-1, only VEGF₁₆₅binds to the new receptor. Thus, this is an isoform-specific receptor and has been named the VEGF₁₆₅receptor (VEGF₁₆₅R). It will also bind the 189 and 206 isoforms. VEGF₁₆₅R has a molecular mass of approximately 130 kDa, and it binds VEGF₁₆₅with a Kd of about 2×10⁻¹⁰M, compared with approximately 5×10⁻¹²M for KDR/Flk-1. In structure-function analysis, it was shown directly that VEGF₁₆₅binds to VEGF₁₆₅R via its exon 7-encoded domain which is absent in VEGF₁₂₁(Soker, et al., J. Biol. Chem., 271:5761-5767 (1996)). However, the function of the receptor was unclear.
The current treatment of angiogenic diseases is inadequate. Agents which prevent continued angiogenesis, e.g, drugs (TNP-470), monoclonal antibodies, antisense nucleic acids and proteins (angiostatin and endostatin) are currently being tested. See, Battegay, J. Mol. Med., 73, 333-346 (1995); Hanahan et al., Cell, 86, 353-364 (1996); Folkman, N. Engl. J. Med., 333, 1757-1763 (1995). Although preliminary results with the antiangiogenic proteins are promising, there is still a need for identifying genes encoding ligands and receptors involved in angiogenesis for the development of new antiangiogenic therapies.

SUMMARY OF THE INVENTION

We have isolated a cDNA encoding the VEGF₁₆₅R gene (SEQ ID NO: 1) and have deduced the amino acid sequence of the receptor (SEQ ID NO:2). We have discovered that this novel VEGF receptor is structurally unrelated to Flt-1 or KDR/Flk-1 and is expressed not only by endothelial cells but by non-endothelial cells, including surprisingly tumor cells.
In ascertaining the function of the VEGF₁₆₅R we have further discovered that this receptor has been identified as a cell surface mediator of neuronal cell guidance and called neuropilin-1. Kolodkin et al., Cell 90:753-762 (1997). We refer to the receptor as VEGF₁₆₅R/NP-1 or NP-1.
In addition to the expression cloning of VEGF₁₆₅R/NP-1 cDNA, we isolated another human cDNA clone whose predicted amino acid sequence was 47% homologous to that of VEGF₁₆₅R/NP-1 and over 90% homologous to rat neuropilin-2 (NP-2) which was recently cloned (Kolodkin, et al., Cell 90, 753-762 (1997)).
Our results indicate that these neuropilins are expressed by both endothelial and tumor cells including breast, prostate and melanoma. (FIG. 18) We have shown that endothelial cells expressing both KDR and VEGF₁₆₅R/NP-1 respond with increased chemotaxis towards VEGF₁₆₅, not VEGF₁₂₁, when compared to endothelial cells expressing KDR alone. While not wishing to be bound by theory, we believe that VEGF₁₆₅R/NP-1 functions in endothelial cells to mediate cell motility as a co-receptor for KDR.
We have also shown in the Boyden chamber motility assay that VEGF₁₆₅stimulates 231 breast carcinoma cell motility in a dose-response manner (FIG. 15A). VEGF₁₂₁had no effect motility of these cells (FIG. 15B). Since tumor cells such as, 231 cells, do not express the VEGF receptors, KDR or Flt-1, while not wishing to be bound by theory, we believe that tumor cells are directly responsive to VEGF₁₆₅via VEGF₁₆₅R/NP-1.
We have also analyzed two variants of Dunning rat prostate carcinoma cells, AT2.1 cells, which are of low motility and low metastatic potential, and AT3.1 cells, which are highly motile, and metastatic. Cross-linking and Northern blot analysis show that AT3.1 cells express abundant VEGF₁₆₅R/NP-1, capable of binding VEGF₁₆₅, while AT2.1 cells don't express VEGF₁₆₅R/NP-1 (FIG. 18). Immunostaining of tumor sections confirmed the expression of VEGF₁₆₅R/NP-1 in AT3.1, but not AT2.1 tumors. Additionally, immunostaining showed that in subcutaneous AT3.1 and PC3 tumors, the tumor cells expressing VEGF₁₆₅R/NP-1 were found preferentially at the invading front of the tumor/dermis boundary. Furthermore, stable clones of AT2.1 cells overexpressing VEGF₁₆₅R/NP-1 had enhanced motility in the Boyden chamber assay. These results indicate that neuropilin expression is associated with angiogenesis and motile metastatic cancer cells, and thus is an important target for antiangiogenic and anticancer therapy.
We have now identified and cloned several neuropilin isoforms that are truncated in the C-terminal region to produce soluble neuropilin (sNP) ectodomains (FIG. 19). These isoforms were cloned after a Northern blot analysis revealed that some cell lines and tissues expressed smaller transcripts in addition to 7 kb neuropilin-1 (NP-1) and 7 kb neuropilin-2 (NP-2), that were apparently generated by alternative splicing. Intact neuropilins have a domains homologous to complement components, b domains homologous to coagulation factors, a c domain homologous to MAM, a transmembrane domain and a short 40 amino acid cytoplasmic domain (Kawakami A, et al., (1995) J. Neurobiol. 29: 1-17.) (FIG. 19). An isoform of neuropilin-1 was cloned that is C-terminally truncated right after the b domain. During transcription there is reading through a 5′ splice donor site so that part of an intron is expressed followed by termination, with the result that the c, transmembrane and cytoplasmic domains are replaced by three intron amino acids following the b domain. In addition, a neuropilin-2 isoform was cloned in which the C-terminal part of the b domain, the c domain, the transmembrane domain and the cytoplasmic domain are replaced by 8 intron amino acids. The truncated neuropilin-1 cDNA was expressed in COS cells and proteins in conditioned medium were analyzed by Western blot using specific anti-neuropilin-1 antibodies (FIG. 20). A 90 kDa protein produced by transfection of the truncated neuropilin-1 cDNA, but not of the vector control was found in conditioned medium but not in the lysate. Thus the neuropilin-1 isoform is a soluble form of neuropilin-1 (sNP1).
We have also expressed an engineered truncated soluble neuropilin-1 ectodomain receptor that contains the a, b and c domains (designated sNP1abc) by truncation at a site in the juxtamembrane domain.
sNPs are capable of binding to VEGF₁₆₅or any form of VEGF that contains exon 7 (SEQ ID NO: 15) and therefore are useful for inhibiting VEGF interaction not only with neuropilins but also with KDR/Flk-1 and Flt-1 as well. In addition, sNPs could also act as dominant negative receptors when expressed in cells by dimerizing with intact neuropilin receptors. Our results have shown that sNP1 protein preparations are excellent inhibitors of ¹²⁵I-VEGF₁₆₅binding to PAE/NP1 and of VEGF-mediated HUVEC proliferation (FIG. 21).
One preferred sNP for use in methods of the invention is an isolated soluble neurophilin-2 comprising amino acids 277 to 594 of SEQ ID NO:4, or a fragment or a homolog thereof, that reduces VEGF165 mediated HUVEC proliferation.
Accordingly, sNPs or nucleic acids, e.g., DNA or RNA, encoding sNPs are useful as inhibitors of VEGF and NP function and can be used to treat diseases, disorders or conditions associated with VEGF. sNPs can be used alone or in combination with other anti-VEGF strategies including, for example, those that antagonize VEGF directly (e.g. anti-VEGF antibodies, soluble VEGF receptor extracellular domains), or antagonize VEGF receptors (e.g. anti-KDR antibodies, KDR kinase inhibitors, dominant-negative VEGF receptors) (Presta L G, et al., Cancer Res. 57: 4593-4599 (1997), Kendall R L, et al., (1996) Biochem. Biophys. Res. Commun. 226: 324-328, Goldman C K, et al., (1998) Proc. Natl. Acad. Sci. USA 95: 8795-8800, Strawn L M, et al., (1996) Cancer Res. 56: 3540-3545, Zhu Z, et al., (1998). Cancer Res. 58: 3209-3214, Witte L, et al., (1998). Cancer Metastasis Rev. 17: 155-161.)
Diseases, disorders, or conditions, associated with VEGF, include, but are not limited to retinal neovascularization, hemagiomas, solid tumor growth, leukemia, metastasis, psoriasis, neovascular glaucoma, diabetic retinopathy, rheumatoid arthritis, osteoarthritis, endometriosis, mucular degeneration and retinopathy of prematurity (ROP).
In addition, the present invention relates to methods of screening for expression of a naturally occurring soluble neuropilins in selected tissues. Expression can be analyzed at the RNA level (in situ hybridization with specific probes corresponding to intron sequences), or at the protein level (Western blot detection of lower molecular masses). The relative distribution of intact and truncated neuropilin isoforms can then be determined. These techniques can be used to analyze sNP distribution in cells, tissues and biological fluids such as urine. sNP1 and sNP2 both contain C-terminal intron sequences that are absent in intact neuroplins. sNP 1 has 3 C-terminal intron amino acids (GIK) and 28 intron by in the cDNA. sNP-2 has 8 C-terminal intron amino acids (VGCSWRPL), residues 548-555 of SEQ ID NO:8) and 146 intron by in the cDNA. Thus, sNP specific probes can be prepared for in situ hybridization and to analyze for sNP distribution in tumors and normal tissue in a background of intact neuropilins.
Other aspects of the invention are disclosed infra.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1. Purification of VEGF₁₆₅R From 231 Cells. ¹²⁵I-VEGF₁₆₅(5 ng/ml) was bound and cross-linked to receptors on 231 cells and analyzed by SDS PAGE and autoradiography (lane 1). VEGF₁₆₅R was purified by CON A SEPHAROSE™ and VEGF₁₆₅affinity column chromatography and analyzed by SDS-PAGE and silver stain (lane 2). Two prominent bands were detected (arrows) and N-terminally sequenced separately. Their N-terminal 18 amino acid sequences are shown to the right of the arrows; SEQ ID NO: 25 (upper) and SEQ ID NO: 26 (lower). The published N-terminal sequences of human (SEQ ID NO: 9) and mouse neuropilin (SEQ ID NO: 10) (Kawakami et al., J. Neurobiol., 29, 1-17 (1995); He and Tessier-Lavigne, Cell 90, 739-751 1997) are shown below.

FIG. 2A and FIG. 2B. Isolation of VEGF₁₆₅R cDNA by Expression Cloning. Photomicrographs (dark field illumination) of COS 7 cells binding ¹²⁵I-VEGF₁₆₅. ¹²⁵I-VEGF₁₆₅was bound to transfected COS 7 cells which were then washed, fixed, and overlayed with photographic emulsion that was developed as described in the example. 2A. COS 7 cells were transfected with a primary plasmid pool (#55 of the 231 cell library) representing approximately 3×10³clones and one COS 7 cell binding ¹²⁵I-VEGF₁₆₅in the first round of screening is shown. 2B. Several COS 7 cells transfected with a single positive cDNA clone (A2) binding ¹²⁵I-VEGF₁₆₅after the third round of screening.

FIG. 3. The Deduced Amino Acid Sequence of Human VEGF₁₆₅R/NP-1 (SEQ ID NO:2). The deduced 923 amino acid sequence of the open reading frame of VEGF₁₆₅R/NP-1, clone A2 (full insert size of 6.5 kb) is shown. The putative signal peptide sequence (amino acids 1-21) and the putative transmembrane region (amino acids 860-883) are in boxes. The amino acid sequence obtained by N-terminal amino acid sequencing (FIG. 3, amino acids 22-39) is underlined. The arrow indicates where the signal peptide has been cleaved and removed, based on comparison of the N-terminal sequence of purified VEGF₁₆₅R/NP-1 and the cDNA sequence. The sequence of human VEGF₁₆₅R/NP-1 reported here differs from that reported by He et al. (He and Tessier-Lavigne, Cell 90, 739-751 (1997)) in that we find Lys₂₆rather than Glu₂₆, and Asp₈₅₅rather than Glu₈₅₅. Lys₂₆and Asp₈₅₅are found, however, in mouse and rat VEGF₁₆₅R/NP-1 (Kwakami et al., J. Neurobiol. 29, 1-17 (1995); He and Tessier-Lavigne, Cell 90, 739-751 1997).

FIG. 4 (FIG. 4A, and FIG. 4B) show the Comparison of the Deduced Amino Acid Sequence of Human VEGF₁₆₅R/NP-1 (SEQ ID NO:2) and NP-2 (SEQ ID NO:4). The deduced open reading frame amino acid sequences of VEGF₁₆₅R/NP-1 and NP-2 are aligned using the DNASIS program. Amino acids that are identical in both open reading frames are shaded. The overall homology between the two sequences is 43%.

FIG. 5. Northern Blot Analysis of VEGF₁₆₅R/NP-1 Expression in Human EC and Tumor-Derived Cell Lines. Total RNA samples prepared from HUVEC (lane 1) and tumor-derived breast carcinoma, prostate carcinoma and melanoma cell lines as indicated (lanes 2-8) were resolved on a 1% agarose gel and blotted onto a GeneScreen nylon membrane. The membrane was probed with ³²P-labeled VEGF₁₆₅R/NP-1 cDNA and exposed to X-ray film. Equal RNA loading was demonstrated by ethydium bromide staining of the gel prior to blotting. A major species of VEGF₁₆₅R/NP-1 mRNA of approximately 7 kb was detected in several of the cell lines.

FIG. 6. Northern Blot Analysis of VEGF₁₆₅R/NP-1 and KDR mRNA in Adult Human Tissues. A pre-made Northern blot membrane containing multiple samples of human mRNA (Clonetech) was probed with ³²P-labeled VEGF₁₆₅R/NP-1 cDNA (top) as described in FIG. 5, and then stripped and reprobed with ³²P-labeled KDR cDNA (bottom).

FIG. 7A and FIG. 7B. Scatchard Analysis of VEGF₁₆₅Binding to VEGF₁₆₅R/NP-1. 7A. Increasing amounts of ¹²⁵I-VEGF₁₆₅(0.1-50 ng/ml) were added to subconfluent cultures of PAE cells transfected with human VEGF₁₆₅R/NP-1 cDNA (PAE/NP-1 cells) in 48 well dishes. Non-specific binding was determined by competition with a 200-fold excess of unlabeled VEGF₁₆₅. After binding, the cells were washed, lysed and the cell-associated radioactivity was determined using a γ counter. 7B. The binding data shown in 7A were analyzed by the method of Scatchard, and a best fit plot was obtained with the LIGAND program (Munson and Rodbard, 1980). PAE/NP-1 cells express approximately 3×10⁵VEGF₁₆₅binding sites per cell and bind ¹²⁵I-VEGF₁₆₅with a Kd of 3.2×10⁻¹⁰M.

FIG. 8. Cross-linking of VEGF₁₆₅and VEGF₁₂₁to PAE cells Expressing VEGF₁₆₅R/NP-1 and/or KDR. ¹²⁵I-VEGF₁₆₅(5 ng/ml) (lanes 1-6) or ¹²⁵I-VEGF₁₂₁(10 ng/ml) (lanes 7-10) were bound to subconfluent cultures of HUVEC (lane 1), PC3 (lane 2), PAE (lanes 3 and 7), a clone of PAE cells transfected with human VEGF₁₆₅R/NP-1 cDNA (PAE/NP-1) (lanes 4 and 8), a clone of PAE cells transfected with KDR (PAE/KDR) (lanes 5 and 9), and a clone of PAE/KDR cells transfected with human VEGF₁₆₅R/NP-1 cDNA (PAE/KDR/NP-1) (lanes 6 and 10). The binding was carried out in the presence of 1 μg/ml heparin. At the end of a 2 hour incubation, each ¹²⁵I-VEGF isoform was chemically cross-linked to the cell surface. The cells were lysed and proteins were resolved by 6% SDS-PAGE. The polyacrylamide gel was dried and exposed to X-ray film. Solid arrows denote radiolabeled complexes containing ¹²⁵I-VEGF and KDR, open arrows denote radiolabeled complexes containing ¹²⁵I-VEGF and VEGF₁₆₅R/NP-1.

FIG. 9. Cross linking of VEGF₁₆₅to PAE/KDR Cells Co-expressing VEGF₁₆₅R/NP-1 Transiently. PAE/KDR cells were transfected with pCPhygro or pCPhyg-NP-1 plasmids as described in “Experimental Procedures”, and grown for 3 days in 6 cm dishes. ¹²⁵I-VEGF₁₆₅(5 ng/ml) was bound and cross linked to parental PAE/KDR cells (lane 1), to PAE/KDR cells transfected with pCPhygro vector control (V) (lane 2), to PAE/KDR cells transfected with pCPhyg-VEGF₁₆₅R/NP-1 plasmids (VEGF₁₆₅R/NP-1) (lane 3), and to HUVEC (lane 4).). The binding was carried out in the presence of 1 μg/ml heparin. The cells were lysed and proteins were resolved by 6% SDS-PAGE as in FIG. 8. Solid arrows denote radiolabeled complexes containing ¹²⁵I-VEGF₁₆₅and KDR. Open arrows denote radiolabeled complexes containing ¹²⁵I-VEGF₁₆₅and VEGF₁₆₅R/NP-1.

FIG. 10. Inhibition of ¹²⁵I-VEGF₁₆₅Binding to VEGF₁₆₅R/NP-1 Interferes With Its Binding to KDR. ¹²⁵I-VEGF₁₆₅(5 ng/ml) was bound to subconfluent cultures of PAE transfected with human VEGF₁₆₅R/NP-1 cDNA (PAE/NP-1) (lanes 1 and 2), PAE/KDR cells (lanes 3 and 4), and PAE/KDR cells transfected with human VEGF₁₆₅R/NP-1 cDNA (PAE/KDR/NP-1) (lanes 5 and 16) in 35 mm dishes. The binding was carried out in the presence (

lanes

2, 4, and 6) or the absence (

lanes

1, 3, and 5) of 25 μg/ml GST-Ex 7+8. Heparin (1 μg/ml) was added to each dish. At the end of a 2 hour incubation, ¹²⁵I-VEGF₁₆₅was chemically cross-linked to the cell surface. The cells were lysed and proteins were resolved by 6% SDS-PAGE as in FIG. 9. Solid arrows denote radiolabeled complexes containing ¹²⁵I-VEGF₁₆₅and KDR, open arrows denote radiolabeled complexes containing ¹²⁵I-VEGF₁₆₅and VEGF₁₆₅R/NP-1.

FIG. 11A-FIG. 11C. A Model for VEGF₁₆₅R/NP-1 Modulation of VEGF₁₆₅Binding to KDR. 11A. Cells expressing KDR alone. 11B. Cells co-expressing KDR and VEGF₁₆₅R/NP-1. 11C. Cells co-expressing KDR and VEGF₁₆₅R/NP-1 in the presence of GST-Ex 7+8 fusion protein. A single KDR receptor or a KDR-VEGF₁₆₅R/NP-1 pair is shown in top panels. An expanded view showing several receptors is shown in the bottom panels. VEGF₁₆₅binds to KDR via exon 4 and to VEGF₁₆₅R/NP-1 via exon 7 (Keyt et al. J. Biol. Chem. 271, 5638-5646 (1996b); Soker et al., J. Biol. Chem. 271, 5761-5767 (1996)). A rectangular VEGF₁₆₅molecule represents a suboptimal conformation that doesn't bind to KDR efficiently while a rounded VEGF₁₆₅molecule represents one that fits better into a binding site. In cells expressing KDR alone, VEGF₁₆₅binds to KDR in a sub-optimal manner. In cells co-expressing KDR and VEGF₁₆₅R/NP-1, the binding efficiency of VEGF₁₆₅to KDR is enhanced. It may be that the presence of VEGF₁₆₅R/NP-1 increases the concentration of VEGF₁₆₅on the cell surface, thereby presenting more growth factor to KDR. Alternatively, VEGF₁₆₅R/NP-1 may induce a change in VEGF₁₆₅conformation that allows better binding to KDR, or both might occur. In the presence of GST-Ex 7+8, VEGF₁₆₅binding to VEGF₁₆₅R/NP-1 is competitively inhibited and its binding to KDR reverts to a sub-optimal manner.

FIG. 12. Human NP-2 amino acid sequence (SEQ ID NO:4).

FIG. 13, (FIG. 13A, FIG. 13B and FIG. 13C) show Human NP-2 amino acid sequence (SEQ ID NO:3).

FIG. 14, (FIG. 14A-FIG. 14F) show nucleotide (SEQ ID NO:1) and amino acid sequences (SEQ ID NO:2) of VEGF₁₆₅R/NP-1. The domains are indicated.

FIG. 15A and FIG. 15B. VEGF₁₆₅stimulation of MDA MB 231 cell motility. (15A) Dose response of VEGF₁₆₅motility activity. (15B) Both VEGF₁₆₅and bFGF stimulate motility but VEGF₁₂₁does not.

FIG. 16A, FIG. 16B and FIG. 16C show motility and neuropilin-1 expression of Dunning rat prostate carcinoma cell lines AT3-1 (high motility, high metastatic potential) and AT2.1 (low motility, low metastatic potential) cells. (16A) AT3.1 cells are more motile than AT2.1 cells in a Boyden chamber assay. (16B) 125I-VEGF₁₆₅cross-links neuropilin-1 on AT3.1 cells but does not cross-link to AT2.1 cells. (16C) AT3.1 but not AT2.1 cells express neuropilin-1, while both cell types express VEGF.

FIG. 17A and FIG. 17B. Overexpression of neuropilin-1 in AT2.1 cells. (17A) Western blot, (17B) motility activity. Three AT2.1 clones (

lanes

4,5,6) express higher amounts of neuropilin-1 protein and are more motile compared to parental AT2.1 cells or AT2.1 vector (AT2.1/V) controls and approach AT3.1 cell neuropilin-1 levels and migration activity.

FIG. 18 shows expression of NP-1, NP-2 and β-actin in cancer cell lines and endothelial cells using reverse transcriptase PCR with the following primers:

Human NP-1:
Forward (328-351):
5′ TTTCGCAACGATAAATGTGGCGAT 3′;	(SEQ ID NO: 11)

Reverse (738-719):
5′ TATCACTCCACTAGGTGTTG 3′.	(SEQ ID NO: 12)

Human NP-2:
Forward (513-532):
5′ CCAACCAGAAGATTGTCCTC 3′;	(SEQ ID NO: 13)

Reverse (1181-1162):
5′ GTAGGTAGATGAGGCACTGA 3′..	(SEQ ID NO: 14)

FIG. 19 is a schematic presentation of structures of (top) intact neuropilin (-1 and -2), of (middle) a newly cloned cDNA that encodes an ectodomain of neuropilin-1, and (bottom) of a newly cloned cDNA that encodes an ectodomain of neuropilin-2. These two new cDNAs represent alternative spliced isoforms.

FIG. 20 shows cDNA encoding the C-terminally truncated neuropilin-1 isoform was transfected into COS cells. A soluble 90 kDa protein (sNP1) was detected by Western blot in the conditioned medium of cells expressing sNP1 but not in the vector control. Intact 130 kDa neuropilin-1 expressed by MDA MB 231 cells is shown in the first lane.

FIG. 21A and FIG. 21B show Soluble neuropilin-1 protein preparations (21A) inhibit ¹²⁵I-VEGF₁₆₅binding to PAE/NP cells and (Right) inhibit VEGF₁₆₅mediated HUVEC proliferation. sABC is an engineered soluble neuropilin-1 truncated in the juxtamembrane region. sAB is a naturally occurring neuropilin-1 isoform missing c, TM and cytoplasmic domains. In this experiment sNP1 (21B) is sABC produced in Example 3.

DETAILED DESCRIPTION OF THE INVENTION

The present invention relates to cDNA encoding a soluble neuropilin protein (sNP) which is isolated from neuropilin (NP) producing cells or is recombinantly engineered from NP-encoding DNA. NP-1 and NP-2 are preferred NPs but any neuropilin or VEGF receptor (VEGFR), where the constituents share at least about 85% homology with either of the above VEGF₁₆₅R/NP-1 and NP-2. More preferably, such constituent shares at least 90% homology. Still more preferably, each constituent shares at least 95% homology.
Homology is measured by means well known in the art. For example % homology can be determined by any standard algorithm used to compare homologies. These include, but are not limited to BLAST 2.0 such as BLAST 2.0.4 and i. 2.0.5 available from the NIH (See world wide web site: “ncbi-dot-nlm-dot-nkh-dot-gov/BLAST/newblast-dot-html”) (Altschul, S. F., et al. Nucleic Acids Res. 25: 3389-3402 (1997)) and DNASIS (Hitachi Software Engineering America, Ltd.). These programs should preferably be set to an automatic setting such as the standard default setting for homology comparisons. As explained by the NIH, the scoring of gapped results tends to be more biologically meaningful than ungapped results.
For ease of reference, this disclosure will generally talk about VEGF₁₆₅R/NP-1 and NP-2 and/or homologs thereof but all teaching are applicable to the above-described homologs.
The present invention further relates to isolated and purified sNP protein. sNP, as used herein, refers to a protein which can specifically bind to a vascular endothelial cell growth factor containing exon 7 (SEQ ID NO:15), e.g., VEGF₁₆₅, and has VEGF antagonist activity as determined, for example, by the human umbilical vein endothelial cell (HUVEC) proliferation assay using VEGF₁₆₅as set forth in Soker et al., J. Biol. Chem. 272, 31582-31588 (1997). Preferably, the sNP has at least a 25% reduction in HUVEC proliferation, more preferably a 50% reduction, even more preferably a 75% reduction, most preferably a 95% reduction.
VEGF antagonist activity of the sNPs may also be determined by inhibition of binding of labeled VEGF₁₆₅to VEGF₁₆₅R as disclosed in Soker et al., J. Biol. Chem. 271, 5761-5767 (1996)) or to PAE/NP cells as set forth in the Examples. Preferably, the portion inhibits binding by at least 25%, more preferably 50%, most preferably 75%.
The term “isolated” means that the polypeptide or polynucleotide, e.g., DNA, is removed from its original environment. For example, a naturally-occurring polynucleotides or polypeptides present in a living animal is not isolated, but the same polynucleotides or DNA or polypeptides, separated from some or all of the coexisting materials in the natural system, is isolated. Such polynucleotides could be part of a vector and/or such polynucleotides or polypeptides could be part of a composition, and still be isolated in that such vector or composition is not part of its natural environment.
The nucleotide and amino acid sequence of full length NP-1 is set forth in the Sequence listing as SEQ ID Nos: 1 and 2, respectively. The nucleotide and amino acid sequence of full length NP-2 is set forth in the Sequence listing as SEQ ID Nos: 3 and 4, respectively.
DNA encoding human VEGF₁₆₅R/NP-1 or NP-2 and recombinant human VEGF₁₆₅R/NP-1 or NP-2 may be produced according to the methods set forth in the Examples.
Mammalian cell lines which produce NP-1 or NP-2 include, but are not limited to, MDA-MB-231 cells (ATCC HTB-26), PC3 prostate carcinoma cells and human umbilical vein endothelial cells (HUVEC) (ATCC CRL 1730).
Other cells and cell lines may also be suitable for use to isolate sNP. Selection of suitable cells may be done by screening for sNP binding activity on cell surfaces, in cell extracts or conditioned medium or by screening for gene expression by PCR or hybridization. Methods for detecting soluble receptor activity are well known in the art (Duan, D-S. R. et al., (1991) J. Biol. Chem., 266, pp. 413-418).
Full length NP producing cells such as human HUVEC cells (American Type Culture Collection, ATCC CRL 1730) [Hoshi, H. and McKeehan, W. L., Proc. Natl. Acad. Sci. U.S.A., (1984) 81, pp. 6413-6417] are grown according to the recommended culture conditions of the ATCC. Intact NP as well as extracellular region (sNP-1 and sNP-2) are shown in FIG. 8. The intact receptors have a domains homologous to complement components, b domains homologous to coagulation factors, a c domain homologous to MAM, a transmembrane domain (TM) and a short 40 amino acid cytoplasmic domain (cyto). Two of the inhibitory forms of this receptor, which are the subject of the present invention, are also shown in FIG. 8 and set forth in the sequence listing as SEQ ID NOS:6 and 8 and lack all of the c domain, the transmembrane domain and the cytoplasmic domain. Preferred sNPs of the invention additionally lack the a domains.
Neuropilin-1 (SEQ ID NO:2) domains are as follows: a1 (amino acids 22-146), a2 (amino acids 147-273), b1 (amino acids 275-430), b2 (amino acids 431-587), c (amino acids 646-809), TM (amino acids 857-884), cyto (amino acids 885-923)
Neuropilin-2 (SEQ ID NO:4) domains are as follows: a1 (amino acids 24-148), a2 (amino acids 149-275), b1 (amino acids 277-433), b2 (amino acids 434-594), c (amino acids 642-800), TM (amino acids 865-893), cyto (amino acids 894-931).
Any of a variety of procedures may be used to molecularly clone sNP cDNA. These methods include, but are not limited to, direct functional expression of the sNP gene following the construction of an sNP containing cDNA library in an appropriate expression vector system.
Another method is to screen a sNP containing cDNA library constructed in a bacteriophage or plasmid shuttle vector with a labelled oligonucleotide probe designed from the predicted amino acid sequence of sNP. One method consists of screening a sNP containing cDNA library constructed in a bacteriophage or plasmid shuttle vector with a partial cDNA encoding at least part of the full length NP protein. This partial cDNA is obtained by the specific PCR amplification of sNP DNA fragments through the design of oligonucleotide primers from the known sequence of full length NP-encoding DNA.
It is readily apparent to those skilled in the art that other types of libraries, as well as libraries constructed from other cells or cell types, may be useful for isolating sNP-encoding DNA. Additionally, suitable cDNA libraries may be prepared from cells or cell lines which have sNP activity. The selection of cells or cell lines for use in preparing a cDNA library to isolate sNP cDNA may be done by first measuring secreted sNP activity using the methods described herein.
Preparation of cDNA libraries can be performed by standard techniques well known in the art. Well known cDNA library construction techniques can be found for example, in Molecular Cloning, A Laboratory Manual (2nd Ed., Sambrook, Fritsch and Maniatis, Cold Spring Harbor, N.Y. 1989).
It is also readily apparent to those skilled in the art that DNA encoding sNP may also be isolated from a suitable genomic DNA library. Construction of genomic DNA libraries can be performed by standard techniques well known in the art. Well known genomic DNA library construction techniques can be found in Sambrook, et. al., supra.
Another means of obtaining sNP molecules is to recombinantly engineer them from DNA encoding the partial or complete amino acid sequence of an NP, e.g., NP-1 or NP-2. Using recombinant DNA techniques, DNA molecules are constructed which encode at least a portion of the NP capable of binding VEGF containing exon 7 without stimulating mitogenesis. Standard recombinant DNA techniques are used such as those found in Sambrook, et al., supra.
Using one of the preferred methods of the present invention, cDNA clones encoding sNP are isolated in a two-stage approach employing polymerase chain reaction (PCR) based technology and cDNA library screening. In the first stage, DNA oligonucleotides derived from the extracellular domain sequence information from the known full length NP is used to design degenerate oligonucleotide primers for the amplification of sNP-specific DNA fragments. In the second stage, these fragments are cloned to serve as probes for the isolation of complete sNP cDNA from a commercially available lambda gt10 cDNA library (Clontech) derived from HUVEC cells (ATCC CRL 1730).
Using another method, DNA encoding sNP is constructed from a DNA sequence encoding an NP. For purposes of illustration, DNA encoding NP-1 is utilized. Using the receptor DNA sequence, a DNA molecule is constructed which encodes the extracellular domain of the receptor, or the VEGF binding domain only. Restriction endonuclease cleavage sites are identified within the receptor DNA and can be utilized directly to excise the extracellular-encoding portion. In addition, PCR techniques as described above may be utilized to produce the desired portion of DNA. It is readily apparent to those skilled in the art that other techniques, which are standard in the art, may be utilized to produce sNP molecules in a manner analagous to those described above. Such techniques are found, for example, in Sambrook et al., supra.
In a preferred method sNP cDNAs are tagged with a His domain in the N-terminus of the a domain and subcloned into the pcDNA3.1 mammalian expression plasmid. Each of the plasmids is transfected into CHO-K1 cells and G418 resistant clones are isolated. Conditioned medium is collected and applied to a CON A SEPHAROSE™ column, washed and Con A binding proteins are eluted. The eluate is applied to a Nickel column, washed and Ni⁺⁺ binding sNP proteins are eluted. Purified sNP is assayed for the ability to inhibit ¹²⁵I-VEGF₁₆₅binding to PAE/NP cells and VEGF₁₆₅stimulation of HUVEC proliferation and motility. Smaller fragments are produced by PCR.
Our results indicate that VEGF binds to the b domain of neuropilin and that the a and c domains are not needed. See, FIG. 19 Smaller portions of b domain lacking increasingly larger segments of the N- and C-termini can be prepared by PCR using appropriate oligonucleotide primers. The amplified cDNA is then ligated into an expression vector, expressed in COS cells and conditioned medium tested for the ability to inhibit ¹²⁵I -VEGF₁₆₅binding to PAE/NP1 cells as shown for sNPs in FIG. 21A.
Additional truncated forms of NP can be constructed which contain the transmembrane region. Retention of the transmembrane may facilitate orientation of the inhibitor molecule at the target cell surface. Construction of transmembrane region containing molecules is done by standard techniques known in the art including but not limited to utilizing convenient restriction endonuclease cleavage sites or PCR techniques as described herein.
The cloned sNP cDNA obtained through the methods described above may be recombinantly expressed by molecular cloning into an expression vector containing a suitable promoter and other appropriate transcription regulatory elements, and transferred into prokaryotic or eukaryotic host cells to produce recombinant sNP. Techniques for such manipulations are fully described in Sambrook, et al., supra, and are well known in the art.
Expression vectors are defined herein as DNA sequences that are required for the transcription of cloned copies of genes and the translation of their mRNAs in an appropriate host. Such vectors can be used to express eukaryotic genes in a variety of hosts such as bacteria, bluegreen algae, fungal cells, yeast cells, plant cells, insect cells and animal cells.
Specifically designed vectors allow the shuttling of DNA between hosts such as bacteria-yeast or bacteria-animal or bacteria-insect cells. An appropriately constructed expression vector should contain: an origin of replication for autonomous replication in host cells, selectable markers, a limited number of useful restriction enzyme sites, a potential for high copy number, and active promoters. A promoter is defined as a DNA sequence that directs RNA polymerase to bind to DNA and initiate RNA synthesis. A strong promoter is one which causes mRNAs to be initiated at high frequency. Expression vectors may include, but are not limited to, cloning vectors, modified cloning vectors, specifically designed plasmids or viruses.
A variety of mammalian expression vectors may be used to express recombinant sNP in mammalian cells. Commercially available mammalian expression vectors which may be suitable for recombinant sVEGF-R expression, include but are not limited to, pMC1neo (Stratagene), pXT1 (Stratagene), pSG5 (Stratagene), EBO-pSV2-neo (ATCC 37593) pBPV-I(8-2) (ATCC 37110), pdBPV-MMTneo(342-12) (ATCC 37224), pRSVgpt (ATCC 37199), pRSVneo (ATCC 37198), pSV2-dhfr (ATCC 37146), pUCTag (ATCC 37460), and gZD35 (ATCC 37565).
DNA encoding sNP may also be cloned into an expression vector for expression in a recombinant host cell. Recombinant host cells may be prokaryotic or eukaryotic, including but not limited to bacteria, yeast, mammalian cells including but not limited to cell lines of human, bovine, porcine, monkey and rodent origin, and insect cells including but not limited to drosophila, moth, mosquito and armyworm derived cell lines. Cell lines derived from mammalian species which may be suitable and which are commercially available, include but are not limited to, CV-1 (ATCC CCL 70), COS-1 (ATCC CRL 1650), COS-7 (ATCC CRL 1651), CHO-K1 (ATCC CCL 61), 3T3 (ATCC CCL 92), NIH/3T3 (ATCC CRL 1658), HeLa (ATCC CCL 2), C1271 (ATCC CRL 1616), BS-C-1 (ATCC CCL 26) and MRC-5 (ATCC CCL 171). Insect cell lines which may be suitable and are commercially available include but are not limited to 3M-S (ATCC CRL 8851) moth (ATCC CCL 80) mosquito (ATCC CCL 194 and 195; ATCC CRL 1660 and 1591) and armyworm (Sf9, ATCC CRL 1711).
The expression vector may be introduced into host cells via any one of a number of techniques including but not limited to transformation, transfection, liposome or protoplast fusion, and electroporation. The expression vector-containing cells are clonally propagated and individually analyzed to determine whether they produce sNP protein. Identification of sNP expressing host cell clones may be done by several means, including but not limited to immunological reactivity with anti-sNP antibodies, binding to radiolabelled VEGF, and the presence of host cell-secreted sNP activity.
Following expression of sNP in a recombinant host cell, sNP protein may be recovered to provide sNP in active form, capable of binding VEGF without stimulating mitogenesis. Several sNP purification procedures are suitable for use. sNP may be purified from cell lysates and extracts, or from conditioned culture medium, by various combinations of, or individual application of salt fractionation, ion exchange chromatography, size exclusion chromatography, hydroxylapatite adsorption chromatography, reversed phase chromatography, heparin sepharose chromatography, VEGF165 ligand affinity chromatography, and hydrophobic interaction chromatography.
In addition, recombinant sNP can be separated from other cellular proteins by use of animmuno-affinity column made with monoclonal or polyclonal antibodies specific for full length sNP, or polypeptide fragments of sNP.
Preferably, sNPs can be purified by transfecting sNP containing DNA constructs into COS cells (transient transfection) and CHO cells (stable transfectants). The constructs used can be double tagged near the N-termini of the neuropilin (in the a domain which is not needed for VEGF binding) with, for example, both His and myc tags. Lectin column chromatography, is useful as a first step in sNP purification. The second step in the purification is to use a nickel column to bind the His-tagged proteins, and if necessary, anti-myc antibodies. The present inventors have shown that tagged sNPs are fully active in inhibiting VEGF binding to cells (FIG. 21A). To purify non-tagged sNPs, a combination of lectin and VEGF affinity chromatography is sufficient as shown in the examples for purification of intact neuropilin-1.
Purified sNP proteins can then be tested for effects on VEGF-mediated endothelial cell (e.g. HUVEC) migration and proliferation and the migration of endothelial cells out of rat aortic rings (in vitro angiogenesis). sNP proteins can also be tested in vivo for inhibition of VEGF-mediated angiogenesis in chick CAM, and mouse cornea models. FGF-2, which should not interact with sNPs can be used as a control. Purified sNP protein and DNA encoding the protein can also be test mouse models, in particular PC3 tumors grown subcutaneously or orthotopically into nude mice, to look for inhibition of angiogenesis, tumor growth and metastases.
The inhibitor of the present invention can be used for the inhibition of VEGF mediated activity including angiogenesis and tumor cell motility. The inhibitor can be used either topically or intravascularly. For topical applications the formulation would be applied directly at a rate of about 10 ng to about 1 mg/cm2/day. For intravaneous applications, the inhibitor is used at a rate of about 1 mg to about 10 mg/kg/day of body weight. For internal use, the formulation may be released directly into the region to be treated either from implanted slow release polymeric material or from slow release pumps or repeated injections. The release rate in either case is about 100 ng to about 100 mg/day/cm3.
For non-topical application the inhibitor is administered in combination with pharmaceutically acceptable carders or diluents such as phosphate buffer, saline, phosphate buffered saline, Ringer's solution, and the like, in a pharmaceutical composition, according to standard pharmaceutical practice. For topical application, various pharmaceutical formulations are useful for the administration of the active compound of this invention. Such formulations include, but are not limited to, the following: ointments such as hydrophilic petrolatum or polyethylene glycol ointment; pastes which may contain gums such as xanthan gum; solutions such as alcoholic or aqueous solutions; gels such as aluminum hydroxide or sodium alginate gels; albumins such as human or animal albumins; collagens such as human or animal collagens; celluloses such as alkyl celluloses, hydroxy alkyl celluloses and alkylhydroxyalkyl celluloses, for example methylcellulose, hydroxyethyl cellulose, carboxymethyl cellulose, hydroxypropyl methylcellulose, and hydroxypropyl cellulose; polyoxamers such as PLURONIC™. Polyols exemplified by PLURONIC™ F-127; tetronics such as tetronic 1508; and alginates such as sodium alginate.
The sNPs of the invention can be combined with a therapeutically effective amount of another molecule which negatively regulates angiogenesis which may be, but is not limited to, TNP-470, platelet factor 4, thrombospondin-1, tissue inhibitors of metalloproteases (TIMP1 and TIMP2), prolactin (16-Kd fragment), angiostatin (38-Kd fragment of plasminogen), endostatin, bFGF soluble receptor, transforming growth factor beta, interferon alfa, soluble KDR and FLT-1 receptors and placental proliferin-related protein.
A sNP of the invention may also be combined with chemotherapeutic agents.
The DNA encoding a sNP of the invention can be used in the form of gene therapy and delivered to a host by any method known to those of skill in the art to treat disorders associated with VEGF.
A preferred embodiment of the present invention relates to methods of inhibiting angiogenesis of solid tumors to prevent further tumor growth and eventual metastasis. To this end, any solid tumor or the region surrounding the tumor accessible to gene transfer will be a target for the disclosed therapeutic applications. A DNA encoding an sNP, housed within a recombinant viral- or non-viral-based gene transfer system may be directed to target cells within proximity of the tumor by any number of procedures known in the art, including but not limited to (a) surgical procedures coupled with administration of an effective amount of the DNA to the site in and around the tumor (involving initial removal of a portion or the entire tumor, if possible); (b) injection of the gene transfer vehicle directly into or adjacent to the site of the tumor; and, (c) localized or systemic delivery of the gene transfer vector and/or gene product using techniques known in the art.
Any solid tumor that contains VEGF or neuropilin expressing cells will be a potential target for treatment. Examples, but by no means listed as a limitation, of solid tumors which will be particularly vulnerable to gene therapy applications are (a) neoplasms of the central nervous system such as, but again not necessarily limited to glioblastomas, astrocytomas, neuroblastomas, meningiomas, ependymomas; (b) cancers of hormone-dependent, tissues such as protstate, testicles, uterus, cervix, ovary, mammary carcinomas including but not limited to carcinoma in situ, medullary carcinoma, tubular carcinoma, invasive (infiltrating) carcinomas and mucinous carcinomas; (c) melanomas, including but not limited to cutaneous and ocular melanomas; (d) cancers of the lung which at least include squamous cell carcinoma, spindle carcinoma, small cell carcinoma, adenocarcinoma and large cell carcinoma; and (e) cancers of the gastrointestinal system such as esophageal, stomach, small intestine, colon, colorectal, rectal and anal region which at least include adenocarcinomas of the large bowel.
A DNA fragment encoding an sNP may be delivered either systemically or to target cells in the proximity of a solid tumor of the mammalian host by viral or non-viral based methods. Viral vector systems which may be utilized in the present invention include, but are not limited to, (a) adenovirus vectors; (b) retrovirus vectors; (c) adeno-associated virus vectors; (d) herpes simplex virus vectors; (e) SV 40 vectors; (f) polyoma virus vectors; (g) papilloma virus vectors; (h) picarnovirus vectors; and (i) vaccinia virus vectors.
The recombinant virus or vector containing the DNA encoding the sNP of the present invention is preferably administered to the host by direct injection into a solid tumor and/or quiescent tissue proximal to the solid tumor, such as adipose or muscle tissue. It will of course be useful to transfect tumor cells in the region of targeted adipose and muscle tissue. Transient expression of the sNPs in these surrounding cells will result in a local extracellular increase in these proteins and will promote binding with VEGF, thus inhibiting binding of VEGF to the receptors.
Non-viral vectors which are also suitable include DNA-lipid complexes, for example liposome-mediated or ligand/poly-L-Lysine conjugates, such as asialoglyco-protein-mediated delivery systems (see, e.g., Felgner et al., 1994, J. Biol. Chem. 269: 2550-2561; Derossi et al., 1995, Restor. Neurol. Neuros. 8: 7-10; and Abcallah et al., 1995, Biol. Cell 85:1-7). Direct injection of “naked” DNA may also be used.
The invention also provides a pharmaceutical pack or kit comprising one or more containers filled with one or more of the ingredients of the pharmaceutical compositions of the invention. Optionally associated with such container(s) can be a notice in the form prescribed by a governmental agency regulating the manufacture, use or sale of pharmaceuticals or biological products, which notice reflects approval by the agency of manufacture, use or sale for human administration.
All references cited above or below are herein incorporated by reference.
The present invention is further illustrated by the following Examples. These Examples are provided to aid in the understanding of the invention and are not construed as a limitation thereof.

Example 1

Experimental procedures

Materials

Cell culture media, LIPOFECTIN® and LIPOFECTAMINE™ reagents for transfection were purchased from Life Technologies. Human recombinant VEGF₁₆₅and VEGF₁₂₁were produced in Sf-21 insect cells infected with recombinant baculovirus vectors encoding either human VEGF₁₆₅or VEGF₁₂₁as previously described (Cohen et al., Growth Factors, 7, 131-138 (1992); Cohen et al., J. Biol. Chem., 270, 11322-11326 (1995)). GST VEGF exons 7+8 fusion protein was prepared in E. Coli and purified as previously described (Soker et al., J. Biol. Chem., 271, 5761-5767 (1996)). Heparin, hygromycin B and protease inhibitors were purchased from Sigma (St. Louis, Mo.). ¹²⁵I-Sodium, ³²P-dCTP, and GeneScreen-Plus hybridization transfer membrane were purchased from DuPont NEN (Boston, Mass.). Disuccinimidyl suberate (DSS) and IODO-BEADS® were purchased from Pierce Chemical Co. (Rockford, Ill.). CON A SEPHAROSE™ was purchased from Pharmacia LKB Biotechnology Inc. (Piscataway, N.J.). RNAZOL-B™ was purchased from TEL-TEST Inc. (Friendswood, Tex.). Silver Stain kit and Trans-Blot PVDF membranes were purchased from Bio-Rad Laboratories (Hercules, Calif.). Multiple tissue northern blot membranes were purchased from Clontech (Palo Alto, Calif.). POLYATRACT® mRNA isolation kits were purchased from Promega (Madison, Wis.). REDIPRIME® DNA labeling kits and molecular weight markers were purchased from Amersham (Arlington Heights, Ill.). Plasmids: pcDNA3.1 was purchased from Invitrogen (Carlsbad, Calif.), and pCPhygro, containing the CMV promoter and encoding hygromycin B phosphorylase, was kindly provided by Dr. Urban Deutsch (Max Plank Institute, Bad Nauheim, Germany). Restriction endonucleases and Ligase were purchased from New England Biolabs, Inc (Beverly, Mass.). NT-B2 photographic emulsion and x-ray film were purchased from the Eastman Kodak company (Rochester N.Y.).

Cell Culture

Human umbilical vein EC (HUVEC) were obtained from American Type Culture Collection (ATCC) (Rockville, Md.), and grown on gelatin coated dishes in M-199 medium containing 20% fetal calf serum (FCS) and a mixture of glutamine, penicillin and streptomycin (GPS). Basic FGF (2 ng/ml) was added to the culture medium every other day. Parental porcine aortic endothelial (PAE) cells and PAE cells expressing KDR (PAE/KDR) (Waltenberger et al., J. Biol. Chem. 269, 26988-26995 (1994)) were kindly provided by Dr. Lena Claesson-Welsh and were grown in F12 medium containing 10% FCS and GPS. MDA-MB-231 cells and MDA-MB-453 cells were obtained from ATCC, and grown in DMEM containing 10% FCS and GPS. The human melanoma cell lines, RU-mel, EP-mel and WK-mel were kindly provided by Dr. Randolf Byer (Boston University Medical School, Boston, Mass.), and grown in DMEM containing 2% FCS, 8% calf serum and GPS. Human metastatic prostate adenocarcinoma, LNCaP and prostate carcinoma, PC3 cells were kindly provided by Dr. Michael Freeman (Children's Hospital, Boston, Mass.), and grown in RPMI 1640 containing 5% FCS and GPS.

Purification and Protein Sequencing

Approximately 5×10⁸MDA-MB-231 cells grown in 150 cm dishes were washed with PBS containing 5 mM EDTA, scraped and centrifuged for 5 min at 500 g. The cell pellet was lysed with 150 ml of 20 mM HEPES, pH 8.0, 0.5% triton X-100 and protease inhibitors including 1 mM AEBSF, 5 μg/ml leupeptin and 5 μg/ml aprotinin for 30 min on ice, and the lysate was centrifuged at 30,000×g for 30 min. MnCl₂and CaCl₂were added to the supernatant to obtain a final concentration of 1 mM each. The lysate was absorbed onto a CON A SEPHAROSE™ column (7 ml) and bound proteins were eluted with 15 ml 20 mM HEPES, pH 8.0, 0.2 M NaCl, 0.1% triton X-100 and 1 M methyl-α-D-mannopyranoside at 0.2 ml/min. The elution was repeated twice more at 30 minute intervals. The CON A SEPHAROSE™ column eluates were pooled and incubated for 12 h at 4° C. with 0.5 ml of VEGF₁₆₅-Sepharose beads, containing about 150 μg VEGF₁₆₅, prepared as described previously (Wilchek and Miron, Biochem. Int. 4, 629-635. (1982)). The VEGF₁₆₅-Sepharose beads were washed with 50 ml of 20 mM HEPES, pH 8.0, 0.2 M NaCl and 0.1% triton X-100 and then with 25 ml of 20 mM HEPES, pH 8.0. The beads were boiled in SDS-PAGE buffer and bound proteins were separated by 6% SDS-PAGE. Proteins were transferred to a TransBlot PVDF membrane using a semi-dry electric blotter (Hoeffer Scientific), and the PVDF membrane was stained with 0.1% Coomassie Brilliant Blue in 40% methanol. The two prominent proteins in a 130-140 kDa doublet were cut out separately and N-terminally sequenced using an Applied Biosystems model 477A microsequenator as a service provided by Dr. William Lane of the Harvard Microchemistry facility (Cambridge, Mass.).

Expression Cloning and DNA Sequencing

Complementary DNA (cDNA) was synthesized from 5 μg 231 mRNA. Double-stranded cDNA was ligated to EcoRI adaptors, and size-fractionated on a 5-20% potassium acetate gradient. DNA fragments larger than 2 kb were ligated to an eukaryotic expression plasmid, pcDNA3.1. The plasmid library was transfected into E. coli to yield a primary library of approximately 1×10⁷individual clones. A portion of the transformed bacteria was divided into 240 pools, each representing approximately 3×10³individual clones. DNA prepared from each pool was used to transfect COS-7 cells seeded in 12 well dishes, using the LIPOFECTIN® reagent according to the manufacturer's instructions. Three days after transfection, the cells were incubated on ice for 2 h with ¹²⁵I-VEGF₁₆₅(10 ng/ml) in the presence of 1 μg/ml heparin, washed and fixed with 4% paraformaldehyde in PBS. ¹²⁵I -VEGF₁₆₅binding to individual cells was detected by overlaying the monolayers with photographic emulsion, NT-B2, and developing the emulsion after two days as described (Gearing et al., 1989). Seven positive DNA pools were identified and DNA from one of the positive pools was used to transform E. Coli. The E. coli were sub-divided into 50 separate pools and plated onto 50 LB ampicillin dishes, with each pool representing approximately 100 clones. DNA made from these pools was transfected into COS-7 cells which were screened for ¹²⁵I-VEGF₁₆₅binding as described above. Twenty positive pools were detected at this step, and their corresponding DNAs were used to transform E. Coli. Each pool was plated onto separate LB ampicillin dishes and DNA was prepared from 96 individual colonies and screened in a 96-well two dimensional grid for ¹²⁵I-VEGF₁₆₅binding to tranfected COS-7 cells as described above. Seven single clones were identified as being positive at this step. The seven positive plasmid clones were amplified and their DNA was analyzed by restriction enzyme digestion. Six clones showed an identical digestion pattern of digest and one was different. One clone from each group was submitted for automated DNA sequencing.

Northern Analysis

Total RNA was prepared from cells in culture using RNAzol according to the manufacturer's instructions. Samples of 20 μg RNA were separated on a 1% formaldehide-agarose gel, and transferred to a GENESCREEN PLUS™ membrane. The membrane was hybridized with a ³²P labeled fragment of human VEGF₁₆₅R/NP-1 cDNA, corresponding to nucleotides 63-454 in the ORF, at 63° C. for 18 h. The membrane was washed and exposed to an x-ray film for 18 h. A commercially-obtained multiple human adult tissue mRNA blot (Clonetech, 2 μg/lane) was probed for human NP-1 in a similar manner. The multiple tissue blot was stripped by boiling in the presence of 0.5% SDS and re-probed with a ³²P labeled fragment of KDR cDNA corresponding to nucleotides 2841-3251 of the ORF (Terman et al., Oncogene 6, 1677-1683 (1991)).

Transfection of PAE Cells

Parental PAE cells and PAE cells expressing KDR (PAE/KDR) (Waltenberger et al., 1994) were obtained from Dr. Lena Claesson-Welsh. Human NP-1 cDNA was digested with XhoI and XbaI restriction enzymes and subcloned into the corresponding sites of pCPhygro, to yield pCPhyg-NP-1. PAE and PAE/KDR cells were grown in 6 cm dishes and transfected with 5 μg of pCPhyg-NP-1 using LIPOFECTAMINE™, according to the manufacturer's instructions. Cells were allowed to grow for an additional 48 h and the medium was replaced with fresh medium containing 200 μg/ml hygromycin B. After 2 weeks, isolated colonies (5-10×10³cell/colony) were transferred to separate wells of a 48 well dish and grown in the presence of 200 μg/ml hygromycin B. Stable PAE cell clones expressing VEGF₁₆₅R/NP-1 (PAE/NP-1) or co-expressing VEGF₁₆₅R/NP-1 and KDR (PAE/KDR/NP-1) were screened for VEGF₁₆₅receptor expression by binding and cross linking of ¹²⁵I-VEGF₁₆₅. For transient transfection, PAE/KDR cells were transfected with VEGF₁₆₅R/NP-1 as described above and after three days ¹²⁵I-VEGF₁₆₅cross-linking analysis was carried out.

Radio-Iodination of VEGF, Binding and Cross-Linking Experiments.

The radio-iodination of VEGF₁₆₅and VEGF₁₂₁using IODO-BEADS® was carried out as previously described (Soker et al., J. Biol. Chem. 272, 31582-31588 (1997)). The specific activity ranged from 40,000-100,000 cpm/ng protein. Binding and cross-linking experiments using ¹²⁵I-VEGF₁₆₅and ¹²⁵I-VEGF₁₂₁were performed as previously described (Gitay-Goren et al., J. Biol. Chem. 267, 6093-6098 (1992); Soker et al., J. Biol. Chem. 271, 5761-5767 (1996)). VEGF binding was quantitated by measuring the cell-associated radioactivity in a γ-counter (Beckman, Gamma 5500). The counts represent the average of three wells. All experiments were repeated at least three times and similar results were obtained. The results of the binding experiments were analyzed by the method of Scatchard using the LIGAND program (Munson and Rodbard, 1980). ¹²⁵I-VEGF₁₆₅and ¹²⁵I-VEGF₁₂₁cross linked complexes were resolved by 6% SDS/PAGE and the gels were exposed to an X-Ray film. X-ray films were subsequently scanned by using an IS-1000 digital imaging system (Alpha Innotech Corporation)

Purification of VEGF₁₆₅R

Cross-linking of ¹²⁵I-VEGF₁₆₅to cell surface receptors of 231 cells results in formation of a 165-175 kDa labeled complex (Soker et al., J. Biol. Chem. 271, 5761-5767 (1996)). These cells have about 1-2×10⁵VEGF₁₆₅binding sites/cell. In contrast to VEGF₁₆₅, VEGF₁₂₁does not bind to the 231 cells and does not form a ligand-receptor complex (Soker et al., J. Biol. Chem. 271, 5761-5767 (1996)). The relatively high VEGF₁₆₅R number and the lack of any detectable KDR or Flt-1 mRNA in 231 cells (not shown) suggested that these cells would be a useful source for VEGF₁₆₅R purification. Preliminary characterization indicated that VEGF₁₆₅R is a glycoprotein and accordingly, a 231 cell lysate prepared from approximately 5×10⁸cells was absorbed onto a CON A SEPHAROSE™ column. Bound proteins, eluted from the CON A SEPHAROSE™ column, were incubated with VEGF₁₆₅-Sepharose and the VEGF₁₆₅-affinity purified proteins were analyzed by SDS-PAGE and silver staining (FIG. 9, lane 2). A prominent doublet with a molecular mass of about 130-135 kDa was detected. This size is consistent with the formation of a 165-175 kDa complex of 40-45 kDa VEGF₁₆₅bound to receptors approximately 130-135 kDa in size (FIG. 9, lane 1). The two bands were excised separately and N-terminal amino acid sequencing was carried out (FIG. 1, right). Both the upper and lower bands had similar N-terminal amino acid sequences which showed high degrees of sequence homology to the predicted amino acid sequences in the N-terminal regions of mouse (Kawakami et al., J. Neurobiol, 29, 1-17 (1995)) and human neuroplilin-1 (NP-1) (He and Tessier-Lavigne, Cell 90739-751 (1997)).
Expression Cloning of VEGF₁₆₅R from 231 Cell-Derived mRNA
Concomitant with the purification, VEGF₁₆₅R was cloned by expression cloning (Aruffo and Seed, Proc. Natl. Acad. Sci. USA 84, 8573-8577 (1987a); Aruffo and Seed, EMBO J., 6, 3313-3316 (1987b); Gearing et al., EMBO J. 8, 3667-3676 (1989)). For expression cloning, 231 cell mRNA was used to prepare a cDNA library of approximately 10⁷clones in a eukaryotic expression plasmid. E. coli transformed with the plasmid library were divided into pools. The DNA prepared from each pool were transfected into COS-7 cells in separate wells and individual cells were screened for the ability to bind ¹²⁵I -VEGF₁₆₅as detected by autoradiography of monolayers overlayed with photographic emulsion (FIG. 2A). After three rounds of subpooling and screening, seven single positive cDNA clones were obtained. FIG. 2B shows binding of ¹²⁵I-VEGF₁₆₅to COS-7 cells transfected with one of these single positive clones (clone A2).
Restriction enzyme analysis revealed that six of the seven positive single clones had identical restriction digestion patterns but that one clone had a pattern that was different (not shown). Sequencing of one of these similar cDNA clones, clone A2 (FIG. 3), showed it to be identical to a sequence derived from a human-expressed sequence tag data bank (dbEST). This sequence also showed a high percentage of homology to the sequence of mouse neuropilin, NP-1 (Kawakami et al., J. Neurobiol 29, 1-17 (1995)). After we had cloned human VEGF₁₆₅R, two groups reported the cloning of rat and human receptors for semaphorin III and identified them to be NP-1 (He and Tessier-Lavigne, Cell 90, 739-751 (1997); Kolodkin et al., Cell 90, 753-762 (1997)). The 231 cell-derived VEGF₁₆₅R cDNA sequence is virtually identical (see figure legend 3 for exceptions) to the human NP-1 sequence (He and Tessier-Lavigne, Cell 90, 739-751 (1997)). Significantly, the predicted amino acid sequence obtained by expression cloning (FIG. 3) confirmed the identification of VEGF₁₆₅R as NP-1 that was determined by N-terminal sequencing (FIG. 1), and we have therefore named this VEGF receptor, VEGF₁₆₅R/NP-1.
The human VEGF₁₆₅R/NP-1 cDNA sequence predicts an open reading frame (ORF) of 923 amino acids with two hydrophobic regions representing putative signal peptide and transmembrane domains (FIG. 3). Overall, the sequence predicts ectodomain, transmembrane and cytoplasmic domains consistent with the structure of a cell surface receptor. The N-terminal sequence obtained via protein purification as shown in FIG. 1 is downstream of a 21 amino acid putative hydrophobic signal peptide domain, thereby indicating directly where the signal peptide domain is cleaved and removed. The short cytoplasmic tail of 40 amino acids is consistent with results demonstrating that soluble VEGF₁₆₅R/NP-1 released by partial trypsin digestion of 231 cells is similar in size to intact VEGF₁₆₅R/NP-1 (not shown).
Sequence analysis of the one clone obtained by expression cloning that had a different restriction enzyme profile predicted an open reading frame of 931 amino acids with about a 47% homology to VEGF₁₆₅R/NP-1 (FIG. 4). This human cDNA has a 93% sequence homology with rat neuropilin-2 (NP-2) and is identical to the recently cloned human NP-2 (Chen et al., Neuron, 19, 547-559 (1997)).

Expression of VEGF₁₆₅R/NP-1 in Adult Cell Lines and Tissues

Reports of NP-1 gene expression have been limited so far to the nervous system of the developing embryo (Takagi et al., Dev. Biol. 122, 90-100 (1987); Kawakami et al., J. Neurobiol. 29, 1-17 (1995); Takagi et al., Dev. Biol. 170, 207-222 (1995)). Cell surface VEGF₁₆₅R/NP-1, however, is associated with non-neuronal adult cell types such as EC and a variety of tumor-derived cells (Soker et al., J. Biol. Chem. 271, 5761-5767 (1996)). Northern blot analysis was carried out to determine whether cells that crossed-linked ¹²⁵I -VEGF₁₆₅also synthesized VEGF₁₆₅R/NP-1 mRNA. (FIG. 5). VEGF₁₆₅R/NP-1 mRNA levels were highest in 231 and PC3 cells. VEGF₁₆₅R/NP-1 mRNA was detected to a lesser degree in HUVEC, LNCaP, EP-mel and RU-mel cells. There was little if any expression in MDA-MB-453 and WK-mel cells. The VEGF₁₆₅R/NP-1 gene expression patterns were consistent with our previous results showing that HUVEC, 231, PC3, LNCaP, EP-mel and RU-mel cells bind ¹²⁵I-VEGF₁₆₅to cell surface VEGF₁₆₅R/NP-1 but that MDA-MB-453 and WK-mel cells do not (Soker et al., J. Biol. Chem. 271, 5761-5767 (1996)).
VEGF₁₆₅R/NP-1 gene expression was analyzed also by Northern blot in a variety of adult tissues in comparison to KDR gene expression (FIG. 6). VEGF₁₆₅R/NP-1 mRNA levels were relatively highly in adult heart and placenta and relatively moderate in lung, liver, skeletal muscle, kidney and pancreas. A relatively low level of VEGF₁₆₅R/NP-1 mRNA was detected in adult brain. Interestingly, previous analysis of NP-1 gene expression in mouse and chicken brain suggested that this gene was expressed primarily during embryonic development and was greatly diminished after birth (Kawakami et al., J. Neurobiol. 29, 1-17 (1995); Takagi et al., Dev. Biol. 170, 207-222 (1995)). The tissue distribution of KDR mRNA was similar to that of VEGF₁₆₅R/NP-1 with the exception that it was not expressed highly in the heart. These results indicate that VEGF₁₆₅R/NP-1 is expressed widely in adult non-neuronal tissue, including tissues in which angiogenesis occurs such as heart and placenta.

Characterization of VEGF₁₆₅Binding to VEGF₁₆₅R/NP-1

In order to characterize the binding properties of VEGF₁₆₅R/NP-1, porcine aortic endothelial (PAE) cells were transfected with the cDNA of VEGF₁₆₅R/NP-1. The PAE cells were chosen for these expression studies because they express neither KDR, Flt-1 (Waltenberger et al., J. Biol. Chem. 269, 26988-26995 (1994)) nor VEGF₁₆₅R. Stable cell lines synthesizing VEGF₁₆₅R/NP-1 (PAE/NP-1) were established and ¹²⁵I-VEGF₁₆₅binding experiments were carried out (FIG. 7). ¹²⁵I-VEGF₁₆₅binding to PAE/NP-1 cells increased in a dose-dependent manner and reached saturation at approximately 30 ng/ml demonstrating that VEGF₁₆₅R/NP-1 is a specific VEGF₁₆₅receptor (FIG. 7A). Scatchard analysis of VEGF₁₆₅binding revealed a single class of VEGF₁₆₅binding sites with a K_dof approximately 3.2×10⁻¹⁰M and approximately 3×10^{5 125}I-VEGF₁₆₅binding sites per cell (FIG. 7B). Similar K_dvalues were obtained for several independently-generated PAE/NP-1 clones, although the receptor number varied from clone to clone (not shown). The K_dof 3×10⁻¹⁰M for the PAE/NP-1 cell lines is consistent with the 2-2.8×10⁻¹⁰M K_dvalues obtained for VEGF₁₆₅R/NP-1 expressed naturally by HUVEC and 231 cells (Gitay-Goren et al., J. Biol. Chem. 267, 6093-6098 (1992); Soker et al., J. Biol. Chem. 271, 5761-5767 (1996)). The binding of ¹²⁵I-VEGF₁₆₅to PAE/NP-1 cells was enhanced by 1 μg/ml heparin (not shown), consistent with previous studies showing that heparin enhances ¹²⁵I-VEGF₁₆₅binding to VEGF₁₆₅R/NP-1 on HUVEC and 231 cells (Gitay-Goren et al., J. Biol. Chem. 267, 6093-6098 (1992); Soker et al., J. Biol. Chem. 271, 5761-5767 (1996)).

Isoform-Specific Binding of VEGF to Cells Expressing VEGF₁₆₅R/NP-1

VEGF₁₆₅, but not VEGF₁₂₁, binds to VEGF₁₆₅R/NP-1 on HUVEC and 231 cells (Gitay-Goren et al., J. Biol. Chem. 271, 5519-5523 (1992); Soker et al., J. Biol. Chem. 271, 5761-5767 (1996)). To ascertain whether cells transfected with VEGF₁₆₅R/NP-1 had the same binding specificity, PAE/NP-1 cells were incubated with ¹²⁵I-VEGF₁₆₅or ¹²⁵I -VEGF₁₂₁followed by cross-linking (FIG. 8). ¹²⁵I-VEGF₁₆₅did not bind to parental PAE cells (FIG. 8, lane 3) but did bind to PAE/NP-1 cells via VEGF₁₆₅R/NP-1 (FIG. 8, lane 4). The radiolabeled complexes formed with VEGF₁₆₅R/NP-1 were similar in size to those formed in HUVEC (FIG. 8, lane 1) and PC3 cells (FIG. 8, lane 2). On the other hand, ¹²⁵I-VEGF₁₂₁, did not bind to either parental PAE (FIG. 8, lane 7) or to PAE/NP-1 cells (FIG. 8, lane 8). These results demonstrate that the VEGF isoform-specific binding that occurs with cells expressing endogenous VEGF₁₆₅R/NP-1 such as HUVEC, 231 and PC3 cells, can be replicated in cells transfected with VEGF₁₆₅R/NP-1 cDNA and support the finding that VEGF₁₆₅R and NP-1 are identical.

Co-Expression of VEGF₁₆₅R/NP-1 and KDR Modulates VEGF₁₆₅Binding to KDR

To determine whether expression of VEGF₁₆₅R/NP-1 had any effect on VEGF₁₆₅interactions with KDR, PAE cells that were previously transfected with KDR cDNA to produce stable clones of PAE/KDR cells (Waltenberger et al., J. Biol. Chem. 269, 26988-26995 (1994)), were transfected with VEGF₁₆₅R/NP-1 cDNA and stable clones expressing both receptors (PAE/KDR/NP-1) were obtained. These cells bound ¹²⁵I-VEGF₁₆₅to KDR (FIG. 8, lane 6, upper complex) and to VEGF₁₆₅R/NP-1 (FIG. 8, lane 6, lower complex) to yield a cross-linking profile similar to HUVEC (FIG. 8, lane 1). On the other hand, the PAE/KDR/NP-1 cells bound ¹²⁵I-VEGF₁₂₁to form a complex only with KDR (FIG. 8, lanes 9 and 10), consistent with the inability of VEGF₁₂₁to bind VEGF₁₆₅R/NP-1.
It appeared that in cells co-expressing KDR and VEGF₁₆₅R/NP-1 (FIG. 8, lane 6), the degree of ¹²⁵I-VEGF₁₆₅-KDR 240 kDa complex formation was enhanced compared to the parental PAE/KDR cells (FIG. 8, lane 5). These results were reproducible and the degree of ¹²⁵I-VEGF₁₆₅-KDR 240 kDa complex formation in different clones was correlated positively with the levels of VEGF₁₆₅R/NP-1 expressed (not shown). However, it could not be ruled out definitively that these differential KDR binding results were possibly due to clonal selection post-transfection. Therefore, parental PAE/KDR cells were transfected with VEGF₁₆₅R/NP-1 cDNA and ¹²⁵I-VEGF₁₆₅was bound and cross-linked to the cells three days later in order to avoid any diversity of KDR expression among individual clones (FIG. 9). A labeled 240 kDa complex containing KDR was formed in parental PAE/KDR cells (FIG. 9, lane 1) and in PAE/KDR cells transfected with the expression vector (FIG. 9, lane 2). However, when ¹²⁵I-VEGF₁₆₅was cross-linked to PAE/KDR cells transiently expressing VEGF₁₆₅R/NP-1, a more intensely labeled 240 kDa complex, about 4 times greater, was observed (FIG. 9, lane 3), compared to parental PAE/KDR cells (FIG. 9, lane 1) and PAE/KDR cells transfected with expression vector (FIG. 9, lane 2). These results suggest that co-expression of KDR and VEGF₁₆₅R/NP-1 genes in the same cell enhances the ability of VEGF₁₆₅to bind to KDR.

A GST-VEGF Exon 7+8 Fusion Protein Inhibits VEGF₁₆₅Binding to VEGF₁₆₅R/NP-1 and KDR

We have shown that ¹²⁵I-VEGF₁₆₅binds to VEGF₁₆₅R/NP-1 through its exon 7-encoded domain (Soker et al., J. Biol. Chem. 271, 5761-5767 (1996)). In addition, a GST fusion protein containing the peptide encoded by VEGF exon 7+8 (GST-Ex 7+8), inhibits completely the binding of ¹²⁵I-VEGF₁₆₅to VEGF₁₆₅R/NP-1 associated with 231 cells and HUVEC (Soker et al., J. Biol. Chem. 271, 5761-5767 (1996); Soker et al., J. Biol. Chem. 272, 31582-31588 (1997)). When, added to PAE/NP-1 cells, the fusion protein completely inhibited binding to VEGF₁₆₅R/NP-1 (FIG. 10, lane 2 compared to lane 1). On the other hand, it did not inhibit ¹²⁵I-VEGF₁₆₅binding at all to KDR (FIG. 10, lane 4 compared to lane 3). Thus, these results demonstrate that GST-Ex 7+8 binds directly to VEGF₁₆₅R/NP-1 but does not bind to KDR. The effects of GST-Ex 7+8 are different, however, in cells co-expressing both VEGF₁₆₅R/NP-1 and KDR (PAE/KDR/NP-1). Consistent with the results in FIGS. 8 and 9, the degree of ¹²⁵I-VEGF₁₆₅binding to KDR in PAE/KDR/NP-1 cells (FIG. 10, lane 5) was greater than to the parental PAE/KDR cells (FIG. 10, lane 3). Interestingly, in PAE/KDR/NP-1 cells, GST-Ex 7+8 inhibited not only ¹²⁵I-VEGF₁₆₅binding to VEGF₁₆₅R/NP-1 completely as expected, but it also inhibited binding to KDR substantially which was unexpected (FIG. 10, lane 6 compared to lane 5). In the presence of GST-Ex 7+8, binding of ¹²⁵I-VEGF₁₆₅to KDR in these cells was reduced to the levels seen in parental PAE/KDR cells not expressing VEGF₁₆₅R/NP-1 (FIG. 10, lane 6 compared to lanes 3 and 4). Since the fusion protein does not bind directly to KDR, these results suggest that inhibiting the binding of ¹²⁵I-VEGF₁₆₅to VEGF₁₆₅R/NP-1 directly, inhibits its binding to KDR indirectly. Taken together, the results in FIGS. 8, 9 and 10 suggest that interactions of VEGF₁₆₅with VEGF₁₆₅R/NP-1 enhance VEGF interactions with KDR.

Neuropilin-1 is an Isoform-Specific VEGF₁₆₅Receptor

Recently, we described a novel 130-135 kDa VEGF cell surface receptor that binds VEGF₁₆₅but not VEGF₁₂₁and that we named, accordingly, VEGF₁₆₅R (Soker et al., J. Biol. Chem. 271, 5761-5767 (1996)). We have now purified VEGF₁₆₅R, expression cloned its cDNA, and shown it to be identical to human neuropilin-1 (NP-1) (He and Tessier-Lavigne, Cell 90 739-751 (1997)). The evidence that VEGF₁₆₅R is identical to NP-1 and that NP-1 serves as a receptor for VEGF₁₆₅is as follows: i) purification of VEGF₁₆₅R protein from human MDA-MB-231 (231) cells using VEGF affinity, yielded a 130-140 kDa doublet upon SDS-PAGE and silver stain. N-terminal sequencing of both proteins yielded the same N-terminal sequence of 18 amino acids that demonstrated a high degree of homology to mouse NP-1 (Kawakami et al., J. Neurobiol. 29, 1-17 (1995)); ii) After we purified VEGF₁₆₅R from human 231 cells, the cloning of human NP-1 was reported (He and Tessier-Lavigne, Cell 90, 739-751 (1997)) and the N-terminal sequence of human VEGF₁₆₅R was found to be identical to a sequence in the N-terminal region of human NP-1; iii) Expression cloning using a 231 cell cDNA library resulted in isolation of several cDNA clones and their sequences were identical to the human NP-1 cDNA sequence (He and Tessier-Lavigne, Cell 90, 739-751 (1997)). The combination of purification and expression cloning has the advantage over previous studies where only expression cloning was used (He and Tessier-Lavigne, Cell 90, 739-751 (1997); Kolodkin et al., Cell 90, 753-762 (1997)), in allowing unambiguous identification of the NP-1 protein N-terminus; iv) Northern blot analysis of NP-1 gene expression was consistent with previous ¹²⁵I-VEGF₁₆₅cross-linking experiments (Soker et al., J. Biol. Chem. 271, 5761-5767 (1996)). Cells that bound VEGF₁₆₅to VEGF₁₆₅R synthesized relatively abundant NP-1 mRNA while cells that showed very little if any VEGF₁₆₅binding, did not synthesize much if any NP-1 mRNA; v) when NP-1 was expressed in PAE cells, the transfected, but not the parental cells, were able to bind VEGF₁₆₅but not VEGF₁₂₁, consistent with the isoform specificity of binding previously shown for HUVEC and 231 cells (Soker et al., J. Biol. Chem. 271, 5761-5767 (1996)). Furthermore, the K_dof ¹²⁵I-VEGF₁₆₅binding of to PAE expressing NP-1 was about 3×10⁻¹⁰M, consistent with previous K_dbinding values of 2-2.8×10⁻¹° M for 231 cells and HUVEC (Soker et al., J. Biol. Chem. 271, 5761-5767 (1996)); and vi) The binding of VEGF₁₆₅to cells expressing NP-1 post-transfection was more efficient in the presence of heparin as was the binding of this ligand to HUVEC and 231 cells (Gitay-Goren et al., J. Biol. Chem. 267, 6093-6098 (1992); Soker et al., J. Biol. Chem. 271, 5761-5767 (1996)). Taken together, these results show not only that VEGF₁₆₅R is identical to NP-1 but that it is a functional receptor that binds VEGF₁₆₅in an isoform-specific manner. Accordingly, we have named this VEGF receptor VEGF₁₆₅R/NP-1.
In addition to the expression cloning of VEGF₁₆₅R/NP-1 cDNA, another human cDNA clone was isolated whose predicted amino acid sequence was 47% homologous to that of VEGF₁₆₅R/NP-1 and over 90% homologous to rat neuropilin-2 (NP-2) which was recently cloned (Kolodkin et al., Cell 90, 753-762 (1997)). NP-2 binds members of the collapsin/semaphorin family selectively (Chen et al., Neuron 19, 547-559 (1997)).
The discovery that NP-1 serves as a receptor for VEGF₁₆₅was a surprise since NP-1 had previously been shown to be associated solely with the nervous system during embryonic development (Kawakami et al., J. Neurobiol. 29, 1-17 (1995); Takagi et al., Dev. Biol. 170, 207-222 (1995)) and more recently as a receptor for members of the collapsin/semaphorin family (He and Tessier-Lavigne, Cell 90739-751 (1997); Kolodkin et al., Cell 90, 753-762 (1997)). NP-1 is a 130-140 kDa transmembrane glycoprotein first identified in the developing Xenopus optic system (Takagi et al., Dev. Biol. 122, 90-100 (1987); Takagi et al., Neuron 7, 295-307 (1991)). NP-1 expression in the nervous system is highly regulated spatially and temporally during development and in particular is associated with those developmental stages when axons are actively growing to form neuronal connections. (Fujisawa et al., Dev. Neurosci. 17, 343-349 (1995); Kawakami et al., J. Neurobiol 29, 1-17 (1995); Takagi et al., Dev. Biol. 170, 207-222 (1995)). The NP-1 protein is associated with neuronal axons but not the stomata (Kawakami et al., J. Neurobiol 29, 1-17 (1995)). Functionally, neuropilin has been shown to promote neurite outgrowth of optic nerve fibers in vitro (Hirata et al., Neurosci. Res. 17, 159-169 (1993)) and to promote cell adhesiveness (Tagaki et al., Dev. Biol. 170, 207-222 (1995)). Targeted disruption of NP-1 results in severe abnormalities in the trajectory of efferent fibers of the peripheral nervous system (Kitsukawa et al., Neuron 19, 995-1005 (1997)). Based on the these studies, it has been suggested that NP-1 is a neuronal cell recognition molecule that plays a role in axon growth and guidance (Kawakami et al., J. Neurobiol. 29, 1-17 (1995); He and Tessier-Lavigne, Cell 90, 739-751 (1997); Kitsukawa et al., Neuron 19, 995-1005 1997; Kolodkin et al., Cell 90, 753-762 (1997)).
Our results are the first to show that VEGF₁₆₅R/NP-1 is also expressed in adult tissues, in contrast to the earlier studies that have shown that NP-1 expression in Xenopus, chicken and mouse is limited to the developmental and early post-natal stages (Fujisawa et al., Dev. Neurosci. 17, 343-349 (1995); Kawakami et al., J. Neurobiol. 29, 1-17 (1995); Takagi et al., Dev. Biol. 170, 207-222 (1995)). For example, in mice, NP-1 is expressed in the developing nervous system starting in the dorsal root ganglia at day 9 and ceases at day 15 (Kawakami et al., J. Neurobiol. 29, 1-17 (1995). Our Northern blot analysis of human adult tissue demonstrates relatively high levels of VEGF₁₆₅R/NP-1 mRNA transcripts in heart, placenta, lung, liver, skeletal muscle, kidney and pancreas. Interestingly, there is very little relative expression in adult brain, consistent with the mouse nervous system expression studies (Kawakami et al., J. Neurobiol. 29, 1-17 (1995)). VEGF₁₆₅R/NP-1 is also expressed in a number of cultured non-neuronal cell lines including EC and a variety of tumor-derived cells. A possible function of VEGF₁₆₅R/NP-1 in these cells is to mediate angiogenesis as will be discussed below.
In addition, NP-1 has been identified as a receptor for the collapsin/semaphorin family by expression cloning of a cDNA library obtained from rat E 14 spinal cord and dorsal root ganglion (DRG) tissue (He and Tessier-Lavigne, Cell 90, 739-751 (1997); Kolodkin et al., Cell 90, 753-762 (1997)). The collapsin/semaphorins (collapsin-D-1/Sema III/Sem D) comprise a large family of transmembrane and secreted glycoproteins that function in repulsive growth cone and axon guidance (Kolodkin et al., Cell 75, 1389-1399 (1993)). The repulsive effect of sema III for DRG cells was blocked by anti-NP-1 antibodies (He and Tessier-Lavigne, Cell 90, 739-751 (1997); Kolodkin et al., Cell 90, 753-762 (1997)). The K_dof sema III binding to NP-1, 0.15-3.25×10⁻¹° M (He and Tessier-Lavigne, Cell 90, 739-751 (1997); Kolodkin et al., Cell 90, 753-762 (1997)) is similar to that of VEGF₁₆₅binding VEGF₁₆₅/NP-1, which is about 3×10⁻¹⁰M. These results indicate that two structurally different ligands with markedly different biological activities, VEGF-induced stimulation of EC migration and proliferation on one hand, and sema III-induced chemorepulsion of neuronal cells, on the other hand, bind to the same receptor and with similar affinity. An interesting question is whether the two ligands bind to the same site on VEGF₁₆₅R/NP-1 or to different sites. VEGF₁₆₅R/NP-1 has five discrete domains in its ectodomain, and it has been suggested that this diversity of protein modules in NP-1 is consistent with the possibility of multiple binding ligands for NP-1 (Takagi et al., Neuron 7, 295-307 (1991); Feiner et al., Neuron 19 539-545 (1997); He and Tessier-Lavigne, Cell 90 739-751 (1997). Preliminary analysis does not indicate any large degree of sequence homology between sema III and VEGF exon 7 which is responsible for VEGF binding to VEGF₁₆₅R/NP-1 (Soker et al., J. Biol. Chem. 271, 5761-5767 (1996)). However there may be some 3-dimensional structural similarities between the two ligands. Since both neurons and blood vessels display branching and directional migration, the question also arises as to whether VEGF₁₆₅displays any neuronal guidance activity and whether sema III has any EC growth factor activity. These possibilities have not been examined yet. However, it may be that VEGF requires two receptors, KDR and NP-1 for optimal EC growth factor activity (Soker et al., J. Biol. Chem. 272, 31582-31588 (1997)) and that sema III requires NP-1 and an as yet undetermined high affinity receptor for optimal chemorepulsive activity (Feiner et al., Neuron 19, 539-545 (1997;) He and Tessier-Lavigne, Cell 90, 739-751 (1997); Kitsukawa et al., Neuron 19, 995-1005 (1997)), so that the presence of NP-1 alone might not be sufficient for these ligands to display novel biological activities. Future studies will determine whether there are any connections between the mechanisms that regulate neurogenesis and angiogenesis.

VEGF₁₆₅R/NP-1 Role in Angiogenesis

VEGF₁₆₅R/NP-1 modulates the binding of VEGF₁₆₅to KDR, a high affinity RTK that is an important regulator of angiogenesis as evidenced by KDR knock out experiments in mice (Shalaby et al., Nature 376, 62-66 (1995). The affinity of KDR for VEGF₁₆₅is about 50 times greater than for VEGF₁₆₅R/NP-1 (Gitay-Goren et al., J. Biol. Chem. 287, 6003-6096 (1992); Waltenberger et al., J. Biol. Chem. 269, 26988-26995 (1994)). When VEGF₁₆₅R/NP-1 and KDR are co-expressed, the binding of ¹²⁵I-VEGF₁₆₅to KDR is enhanced by about 4-fold compared to cells expressing KDR alone. The enhanced binding can be demonstrated in stable clones co-expressing VEGF₁₆₅R/NP-1 and KDR (PAE/KDR/NP-1 cells), and also in PAE/KDR cells transfected transiently with VEGF₁₆₅R/NP-1 cDNA where clonal selection does not take place. Conversely, when the binding of ¹²⁵I-VEGF₁₆₅to VEGF₁₆₅R/NP-1 in PAE/KDR/NP-1 cells is inhibited completely by a GST fusion protein containing VEGF exons 7+8 (GST-Ex 7+8), the binding to KDR is inhibited substantially, down to the levels observed in cells expressing KDR alone. The fusion protein binds to VEGF₁₆₅R/NP-1 directly but is incapable of binding to KDR directly (Soker et al., J. Biol. Chem. 272, 31582-31588 (1997)). Although, not wishing to be bound bytheory, we believe that VEGF₁₆₅binds to VEGF₁₆₅R/NP-1 via the exon 7-encoded domain and facilitates VEGF₁₆₅binding to KDR via the exon 4-encoded domain (FIG. 11). VEGF₁₆₅R/NP-1, with its relatively high receptor/cell number, about 0.2-2×10⁵(Gitay-Goren et al., J. Biol. Chem. 287, 6003-6096 (1992); Soker et al., J. Biol. Chem. 271, 5761-5767 (1996)), appears to serve to concentrate VEGF₁₆₅on the cell surface, thereby providing greater access of VEGF₁₆₅to KDR. Alternatively, binding to VEGF₁₆₅R/NP-1, VEGF₁₆₅undergoes a conformational change that enhances its binding to KDR. The end result would be elevated KDR signaling and increased VEGF activity. Although we can demonstrate enhanced binding to KDR, to date we have not been able to demonstrate enhanced VEGF mitogenicity for PAE/KDR/NP-1 cells compared to PAE/KDR cells. One reason is that these cell lines do not proliferate readily in response to VEGF as do HUVEC (Waltenberger et al., J. Biol. Chem. 269, 26988-26995 (1994). Nevertheless, we have shown that VEGF₁₆₅, which binds to both KDR and VEGF₁₆₅R/NP-1, is a better mitogen for HUVEC than is VEGF₁₂₁, which binds only to KDR (Keyt et al., J. Biol. Chem. 271, 5638-5646 (1996b); Soker et al., J. Biol. Chem. 272, 31582-31588 (1997). Furthermore, inhibiting VEGF₁₆₅binding to VEGF₁₆₅R/NP-1 on HUVEC by GST-EX 7+8, inhibits binding to KDR and also inhibits VEGF₁₆₅-induced HUVEC proliferation, down to the level induced by VEGF₁₂₁(Soker et al., J. Biol. Chem. 272, 31582-31588 (1997)). Taken together, these results suggest a role for VEGF₁₆₅R/NP-1 in mediating VEGF₁₆₅, but not VEGF₁₂₁mitogenic activity. The concept that dual receptors regulate growth factor binding and activity has been previously demonstrated for TGF-13, bFGF and NGF (Lopez-Casillas et al., Cell 67, 785-795 (1991); Yayon et al., Cell 64, 841-848 (1991; Barbacid, Curr. Opin. Cell Biol. 7, 148-155 (1995)).
Another connection between VEGF₁₆₅R/NP-1 and angiogenesis comes from studies in which NP-1 was overexpressed ectopically in transgenic mice (Kitsuskawa et al., Develop. 121, 4309-4318 (1995)). NP-1 overexpression resulted in embryonic lethality and the mice died in utero no later than on embryonic day 15.5 and those that survived the best had lower levels of NP-1 expression. Mice overexpressing NP-1 displayed morphologic abnormalities in a limited number of non-neural tissues such as blood vessels, the heart and the limbs. NP-1 was expressed in both the EC and in the mesenchymal cells surrounding the EC. The embryos possessed excess and abnormal capillaries and blood vessels compared to normal counterparts and in some cases dilated blood vessels as well. Some of the chimeric mice showed hemorrhaging, mainly in the head and neck. These results are consistent with the possibility that ectopic overexpression of VEGF₁₆₅R/NP-1 results in inappropriate VEGF₁₆₅activity, thereby mediating enhanced and/or aberrant angiogenesis. Another piece of evidence for a link between NP-1 and angiogenesis comes from a recent report showing that in mice targeted for disruption of the NP-1 gene, the embryos have severe abnormalities in the peripheral nervous system but that their death in utero at days 10.5-12.5 is most probably due to anomalies in the cardiovascular system (Kitsukawa et al., Neuron 19, 995-1005 (1997)).
VEGF₁₆₅R/NP-1 is Associated with Tumor-Derived Cells
The greatest degree of VEGF₁₆₅R/NP-1 expression that we have detected so far occurs in tumor-derived cells such as 231 breast carcinoma cells and PC3 prostate carcinoma cells, far more than occurs in HUVEC. The tumor cells express abundant levels of VEGF₁₆₅R/NP-1 mRNA and about 200,000 VEGF₁₆₅receptors/cell (Soker et al., J. Biol. Chem. 271, 5761-5767 (1996)). On the other hand, these tumor cells do not express KDR or Flt-1 so that VEGF₁₆₅R/NP-1 is the only VEGF receptor associated with these cells. The tumor cells are therefore useful for testing whether VEGF₁₆₅R/NP-1 is a functional receptor for VEGF₁₆₅in the absence of a KDR background. To date, we have not been able to show that VEGF₁₆₅R/NP-1 mediates a VEGF₁₆₅signal in tumor-derived cells as measured by receptor tyrosine phopshorylation. Nevertheless, VEGF₁₆₅might have an effect on tumor cells by inducing some, as yet undetermined activity such as enhanced survival, differentiation, or motility. A recent report has demonstrated that glioma cells express a 190 kDa protein that binds VEGF₁₆₅but not VEGF₁₂₁efficiently (Omura et al., J. Biol. Chem. 272, 23317-23322 (1997)). No stimulation of tyrosine phosphorylation could be demonstrated upon binding of VEGF₁₆₅to this receptor. Whether the 190 kDa isoform-specific receptor is related to VEGF₁₆₅R/NP-1 is not known presently.
VEGF₁₆₅R/NP-1 may have a storage and sequestration function for VEGF₁₆₅. One might envision that VEGF₁₆₅is produced by a tumor cell and binds to VEGF₁₆₅R/NP-1 on that cell via the exon 7-encoded domain (Soker et al., J. Biol. Chem. 271, 5761-5767 (1996)). The stored VEGF₁₆₅could be then released to stimulate tumor angiogenesis in a paracrine manner. Alternatively, VEGF₁₆₅R/NP-1 may mediate a juxtacrine effect in which VEGF₁₆₅is bound to VEGF₁₆₅R/NP-1 on a tumor cell via the exon 7-encoded domain and is also bound to KDR on a neighboring EC via the exon 4-encoded domain (Keyt et al., J. Biol. Chem. 271, 5638-5646 (1996b)). Such a mechanism could result in a more efficient way for tumor cells to attract EC, thereby enhancing tumor angiogenesis.
In summary, we have demonstrated by independent purification and expression cloning methods that the VEGF isoform specific receptor, VEGF₁₆₅R, is identical to NP-1, a cell surface protein previously identified as playing a role in embryonic development of the nervous system and as being a receptor for the collapsins/semaphorins. Furthermore, binding to VEGF₁₆₅R/NP-1 enhances the binding of VEGF₁₆₅to KDR on EC and tumor cells.

Experimental Rationale

We have discovered that tumor cell neuropilin-1 mediates tumor cell motility and thereby metastasis. In a Boyden chamber motility assay, VEGF₁₆₅(50 ng/ml) stimulates 231 breast carcinoma cell motility in a dose-response manner, with a maximal 2-fold stimulation (FIG. 15A). On the other hand, VEGF₁₂₁has no effect on motility of these cells (FIG. 15B). Since 231 cells do not express KDR or Flt-1, these results suggest that tumor cells are directly responsive to VEGF₁₆₅and that VEGF₁₆₅might signal tumor cells via neuropilin-1. Possible candidates for mediating VEGF₁₆₅-induced motility of carcinoma cells are PI3-kinase (PI3-K) (Carpenter, et al. (1996) Curr. Opin. Cell Biol. 8: 153-158.). Since 231 cells do not express KDR or Flt-1, these results suggest that tumor cells are directly responsive to VEGF₁₆₅and that VEGF₁₆₅might signal tumor cells via neuropilin-1.
The other type of evidence is that neuropilin-1 expression might be associated with tumor cell motility. We have analyzed two variants of Dunning rat prostate carcinoma cells, AT2.1 cells, which are of low motility and low metastatic potential, and AT3.1 cells, which are highly motile, and metastatic. Cross-linking and Northern blot analysis show that AT3.1 cells express abundant neuropilin-1, capable of binding VEGF₁₆₅, while AT2.1 cells don't express neuropilin-1 (FIG. 16). Immunostaining of tumor sections confirms the expression of neuropilin-1 in AT3.1, but not AT2.1 tumors. Furthermore, the immunostaining shows that in subcutaneous AT3.1 and PC3 tumors, the tumor cells expressing neuropilin-1 are found preferentially at the invading front of the tumor/dermis boundary. To determine more directly whether neuropilin-1 expression is correlated with enhanced motility, neuropilin-1 was overexpressed in AT2.1 cells (FIG. 17). Three stable clones of AT2.1 cells overexpressing neuropilin-1 had enhanced motility in the Boyden chamber assay. These results indicate that expression of neuropilin-1 in AT2.1 cells enhances their motility. Taken together, it appears that neuropilin-1 expression on tumor cells is associated with the motile, metastatic phenotype.

Example 2

Construction of sNP-1 and sNP-2

The cDNAs encoding the soluble forms of neuropilin-1 and neuropilin-2 were cloned from an oligo dT-primed cDNA library which was synthesized from PC3 cell mRNA.
Soluble Neuropilin-1 (sNP-1) cDNA Cloning:
The sNP-1 cDNA deviates from the full length NP-1 cDNA between the b2
and c domains after amino acid 641, at the position of an exon-exon boundary. The 3′ end of the sNP-1 clone possesses 28 bp of intron sequence, encoding three novel amino acids and a translation stop codon.
An oligonucleotide (GAAGTATACGGTTGCAAGATA SEQ ID NO:16) designed from within the b1 domain was used in 3′RACE (rapid amplification of cDNA ends) to clone the 3′ end of the sNP-1 cDNA. The full length sNP-1 cDNA was subsequently cloned from the PC3 library by RT-PCR using primers at the 5′ (GCGTTCCTCTCGGATCCAGGC SEQ ID NO:17) and 3′ (CAGGTATCAAATAAAATAC SEQ ID NO:18) ends of the sNP-1 open reading frame (ORF). The sNP-1cDNA was tagged with His and c-myc domains (amino acids HHHHHHQQKLISQQNL SEQ ID NO:19) in the N-terminus of the a1 domain between amino acids 43 and 44 of sNP-1. The complete tagged sNP-1 cDNA was subcloned into the pcDNA3.1 mammalian expression plasmid. The nucleotide and amino acid sequence of the sNP-1 are set forth in the sequence listing as SEQ ID NOS:5 and 6, respectively.
Soluble Neuropilin-2 (sNP-2) cDNA Cloning:
The sNP-2 cDNA deviates from the full length NP-2 cDNA within the b2 domain after amino acid 547, at the position of an exon-exon boundary.
The 3′ end of the sNP-2 clone possesses 146 bp of intron sequence, encoding 8 novel amino acids and a translation stop codon.
An oligonucleotide GGCTGCCGGGTAACAGATGC SEQ ID NO:20) designed from within the b1 domain was used in 3′RACE (rapid amplification of cDNA ends) to clone the 3′ end of the sNP-2 cDNA. The full length sNP-2 cDNA was subsequently cloned from the PC3 library by RT-PCR using primers at the 5′ (ATGGATATGTTTCCTCTC SEQ ID NO:21) and 3′ (GTTCTTGGAGGCCTCTGTAA SEQ ID NO:22) ends of the sNP-2 open reading frame (ORF). The sNP-2 cDNA was tagged with His and c-myc domains (amino acids HHHHHHQQKLISQQNL SEQ ID NO:19) in the N-terminus of the a1 domain between amino acids 31 and 32 of sNP-2. The complete tagged sNP-2 cDNA was subcloned into the pcDNA3.1 mammalian expression plasmid. The nucleotide and amino acid sequence of sNP-2 are set forth in the sequence listing as SEQ ID NOS:7 and 8 respectively.

Example 3

Preparation of Soluble NP-1 (Domains AB and C)

1. The sequence of NP-1 between the BamHI site (base 100) and the XbaI site (base 4687) was subcloned between the BamHI and XbaI site in pBluscript II KS (+) (Stratagene, La Jola Calif.) to yield pBS-NP1.
2. PCR was performed on NP-1 sequence with the following primers:
Primer 1 (Forward): NdeI site (bold and underlined) at NP-1 base 2200) GGAATTC
GTTTTAACTGTGAA (SEQ ID NO:23); Primer 2 (Reverse): Outside the transmembrane membrane domain at NP-1 base 2823 including 6 histidine (his-tag) and an XbaI site (bold and italics) GC
TTAATGATGATGATGATGATGGGTCTTCAACACATTGCC (SEQ ID NO:24) The PCR DNA product (approx. 600 bp) was digested with NdeI and XbaI and purified from an agarose gel. The plasmid pBS-NP1 was digested with NdeI and XbaI and the large fragment containing the extracellular portion of NP-1 was purified from an agarose gel and was served as the vector. Ligation of the above PCR product and the vector was performed and the resulting plasmid was named pBS-sNPhis.
3. The plasmid pBS-sNPhis was digested with BamHI and XbaI and the fragment containing the extracellular part of NP-1 (including the his-tag) was subcloned in the BamHI and XbaI sites of pCPhygro (described in the above examples and in Soker et al., Cell 92:735 (1998) to yield pCPhyg-sNPhis.
4. The plasmid pCPhyg-sNPhis was transfected to CHO cells and hygromicine resistant clones were selected and tested for expression of soluble NP-1. soluble NP-1 was purified from the medium by using nickel Sepharose beads.
5. Clones were tested for sNP-1 expression in the following manner. Medium was conditioned for 24 hours and the conditioned medium was incubated with the lectin ConA for 24 hours. ConA bound material was analyzed by SDS-PAGE and Western blotting using an antibody against the A domain of neuropilin-1.
The references cited throughout the specification are incorporated herein by reference.
The present invention has been described with reference to specific embodiments. However, this application is intended to cover those changes and substitutions which may be made by those skilled in the art without departing from the spirit and the scope of the appended claims.

Claims

1.-26. (canceled)

27. An isolated cDNA encoding a human soluble neuropilin polypeptide which specifically binds VEGF₁₆₅and reduces VEGF-mediated HUVEC proliferation.

28. The isolated cDNA of claim 27 wherein the encoded soluble neuropilin polypeptide is a soluble neuropilin-2 polypeptide.

29. The isolated cDNA of claim 28, wherein the encoded soluble neuropilin-2 polypeptide comprises amino acids 277-594 of SEQ ID NO: 4, or a fragment thereof that reduces VEGF₁₆₅mediated HUVEC proliferation and binds to a VEGF protein encoded by SEQ ID NO: 15.

30. The isolated cDNA of claim 28, wherein the encoded soluble neuropilin-2 polypeptide comprises a neuropilin-2 fragment that binds to VEGF₁₆₅and inhibits VEGF₁₆₅-mediated HUVEC proliferation, the fragment lacking the c domain (amino acids 642-800 of SEQ ID NO: 4), transmembrane domain (amino acids 865-893 of SEQ ID NO: 4) and cytoplasmic domain (amino acids 885-923 of SEQ ID NO: 4) and comprising the b1 domain (amino acids 277-433 of SEQ ID NO: 4) and b2 domain (amino acids 434-594 of SEQ ID NO: 4) of human neuropilin-2.

31. The isolated cDNA of claim 30, wherein the encoded neuropilin-2 polypeptide consists of amino acids 277-594 of SEQ ID NO: 4.

32. The isolated cDNA of claim 28, wherein the encoded soluble neuropilin-2 consists of the amino acid sequence of SEQ ID NO:8.

33. The isolated cDNA of claim 28, having the nucleotide sequence of SEQ ID NO:7.

34. A vector comprising the cDNA of claim 28.

35. The vector of claim 34 that is an expression vector.

36. A host cell comprising the vector of claim 34.

37. The host cell of claim 36 that expresses the cDNA.

38. The host cell of claim 36, wherein the host cell is selected from the group consisting of a bacteria cell, a yeast cell, a mammalian cell and an insect cell.

39. The isolated cDNA of claim 27 wherein the encoded soluble neuropilin polypeptide is a soluble neuropilin-1 polypeptide.

40. The isolated cDNA of claim 39, wherein the encoded soluble neuropilin-1 polypeptide comprises amino acids 275-587 of SEQ ID NO: 2, or a fragment thereof that reduces VEGF₁₆₅mediated HUVEC proliferation and binds to a VEGF protein encoded by SEQ ID NO: 15.

41. The isolated cDNA of claim 39, wherein the encoded soluble neuropilin-1 polypeptide comprises a neuropilin-1 fragment that binds to VEGF₁₆₅and inhibits VEGF₁₆₅-mediated HUVEC proliferation, the fragment lacking the c domain (amino acids 646-809 of SEQ ID NO: 2), transmembrane domain (amino acids 857-884 of SEQ ID NO: 2) and cytoplasmic domain (amino acids 885-923 of SEQ ID NO: 2) and comprising the b1 domain (amino acids 275-430 of SEQ ID NO: 2) and b2 domain (amino acids 431-587 of SEQ ID NO: 2) of human neuropilin-1.

42. The isolated cDNA of claim 41, wherein the encoded neuropilin-1 polypeptide consists of amino acids 275-587 of SEQ ID NO: 4.

43. The isolated cDNA of claim 39, wherein the encoded soluble neuropilin-1 consists of the amino acid sequence of SEQ ID NO:6.

44. The isolated cDNA of claim 39, having the nucleotide sequence of SEQ ID NO:5.