ES2371015T3 - TREATMENT OF IMAGES, DIAGNOSIS AND TREATMENT OF THE DISEASE. - Google Patents

Abstract

An antibody capable of selectively binding the human ECSM4 polypeptide, wherein said human ECSM4 is a naturally occurring polypeptide comprising a portion of 50 consecutive amino acid residues of the polypeptide sequence of Figure 4 or Figure 5, for use as a medicine..

Description

The present invention relates to a gene whose expression is selective for the endothelium and to the use of antibodies that bind to the product of this gene for use as a medicament.

The endothelium plays an essential role in many physiological and pathological processes and is known to be an exceptionally active transcription site. Approximately 1000 different genes are expressed in an endothelial cell. In contrast, it was found that red blood cells express 8, platelets 22 and smooth muscle 127 different genes (Adams et al, 1995). Some known specific endothelial genes attract a lot of attention from both basic research and clinical compilation. For example, the specific endothelial tyrosine kinases Tie, TIE2 / TEK, KDR and flt1 play crucial roles in the regulation of vascular integrity, inflammatory processes mediated by the endothelium and in angiogenesis (Sato et al, 1993, Sato et al, 1995, Fong et al, 1995, Shalaby et al, 1995, Alello et al, 1995). Angiogenesis is now widely recognized as a process that limits the growth of solid tumors. It is also involved in the formation of atherosclerosis and restenosis plaques. Finally, the endothelium plays an essential role in the complex and dynamic system that regulates coagulation and hemostasis.

Of the many different genes that are expressed in an endothelial cell, not all are completely selective of the endothelial cells, so that the genes and their products, and the molecules that bind to them, are usually not useful for imaging, the diagnosis and treatment of diseases. Therefore, there is still a need to find specific or selective molecules of endothelial cells.

In this document we report the identification of two highly selective endothelial genes that we have called: endothelial cell specific molecule 1 (ECSM1) and magic roundabout (endothelial cell specific molecule 4; ECSM4), although the present invention relates only to with human ECSM4. The term ECSM4 is used to indicate, as the context will make clear, the cDNA and polypeptides encoded by the gene. ECSM4 is surprisingly specific in its cell expression profile, and shows a selectivity for endothelial cells that is similar to that of the marker currently accepted in the art as the best endothelial cell marker (von Willibrand Factor). Obviously, such a high degree of specificity for endothelial cells is as unpublished as unexpected.

The human magic roundabout cDNA clone (ECSM4), with an ORF of more than 417 aa (GenBank Accession Number AK000805) and described in WO 99/46281 as a sequence of 3716 nucleotides was identified by BLAST searches against the contig 111518. This sequence is rich in prolines and has several regions of low amino acid complexity. The BLAST search against PRODOM (protein family database in HGMP, UK) identified a 120 bp region with homology to the conserved cytoplasmic domain of the family of transmembrane receptors involved in the axon repulsive guide (protein family ROBO1 DUTT1, E = 4e-07). The homology was extended to 468 aa (E = 1.3e-09) when a more rigorous analysis was carried out using ssearch (Smith and Waterman, 1981), but the similarity region was still contained even in the cytoplasmic domain. The ROBO1 DUTT1 family includes homolog 1 of the human roundabout (ROBO1), the mouse gene DUTT1 and the rat gene ROBO1 (Kidd et al, 1998, Brose et al, 1999). Due to this region of homology we call the gene represented by Hs. 111518 "magic roundabout" (ECSM4). In addition, the BLAST SBASE (database of protein domains in HGMP) suggested a region of similarity with the long domain of the precursor of the intracellular neural cell adhesion molecule (E = 2e-11). It should be noted that the true protein product of the magic roundabout is probably longer than the 417 aa encoded by clone AK000805, since apparently the ORF does not have an upstream limit (felt up), and the comparison of size with the human roundabout 1 (1651 aa) suggests that it is a much larger protein. This is confirmed by Figure 3, which shows that the translation product of the human ECSM4 gene has about 118 kDa. However, ECSM4 is smaller than other members of the roundabout family, with which it shares only two of the five Ig domains and two of the three fibronectin-like domains in the extracellular region. The probable proline-rich intracellular region that is homologous to those of roundabouts that are thought to bind c-abl. Figure 12 shows the complete amino acid sequence of human ECSM4 (1105 aa), and the mouse homologue sequence is shown in Figure 13. In WO 99/11293 and WO 99/53051 coding nucleotide sequences are presented, showing around 99% identity with the nucleotide sequence of ECSM4 given in Figure 12.

EP 1 074 617, WO 00/53756, WO 99/46281, WO 01/23523 and WO 99/11293 show other sequences that show homology to the ECSM4 polypeptide or nucleotide sequence. However, none of these publications reveals that these sequences are expressed selectively in the vascular endothelium, nor do they suggest that they can be expressed in that way.

Recently, interesting associations have been discovered between the genes of neuronal differentiation and those of endothelial cells. For example, a neuronal receptor for vascular endothelial growth factor (VEGF) neuropilin 1 was identified (Soker et al, 1998). The traditional belief was that VEGF was an exclusive growth factor of the endothelium. Similar processes to the guidance of neuronal axons are now involved in the guidance of endothelial cell migration during angiogenesis and capillary growth. Therefore, efrinB ligands and EphB receptors are involved in the demarcation of the arterial and venous domains (Adams et al, 1999). It is possible that the magic roundabout (ECSM4) is a specific homolog of the endothelium of the human roundabout 1 involved in the repulsive guidance of the endothelial cells, presumably with a different ligand, since the similarity is confined to the cytoplasmic region, that is, effector and it is known that guide receivers have a highly modular architecture (Barshaw and Goodman, 1999).

However, so far there has been no mention of the existence of an endothelial counterpart, nor that the expression pattern of the magic roundabout gene (ECSM4) is restricted to endothelial cells, especially angiogenic endothelial cells , or any function of the polypeptide encoded by said gene.

It should be noted that a surprising result of our RT-PCR analysis, described in Example 1, was that ECSM4 apparently shows a specificity for the endothelium (Fig. 1) comparable to the classic endothelial marker von Willebrand factor (vWF). The expression of known endothelium-specific genes is usually not 100% restricted to the endothelial cell. The data presented in this document show the very unexpected finding that ECSM4 is not expressed at detectable levels (at least by using the methods described in the examples) in cell types other than endothelial cells, given that the selectivity of the markers Known endothelial cells is less than 100%. Protection analysis with ribonuclease confirmed and extended this observation (Figure 14a). It was observed that the expression of ECSM4 was restricted to the endothelium (three different isolates) and was absent from fibroblasts, carcinoma cells and neurons. KDR and FLT1 are expressed in the male and female reproductive tract: in spermatogenic cells (Obermair et al, 1999), trophoblasts and in decidua (Clark et al, 1996). KDR has been shown to define hematopoietic stem cells (Ziegler et al, 1999). FLT1 is also present in monocytes. In addition to the endothelial cells vWF has a strong presence in megakaryocytes (Sporn et al, 1985, Nichols et al, 1985) and, consequently, is present in platelets. Similarly, multimerin is present in both endothelial cells (Hayward et al, 1993), and in platelets (Hayward et al, 1998).

Generally speaking, endothelial cells and hematopoietic cells descend from the same embryonic precursors: hemangioblasts, and these two cell lineages share many cell markers (for a review see Suda et al, 2000). Therefore, the finding that ECSM4 is not expressed in cells other than those of the vascular endothelium is very surprising.

The determination of genes whose expression is selective for the vascular endothelium allows selective targeting of these cells and, therefore, the specific delivery of molecules for imaging, diagnosis, prognosis, treatment, prevention and evaluation of therapies. for diseases associated with normal or aberrant vascular growth.

WO 99/426281 discloses the identification and characterization of new secreted and transmembrane polypeptides, and new nucleic acids encoding those polypeptides.

In the term "ECSM4 human polypeptide" we include a polypeptide whose sequence comprises or consists of the amino acid sequence given in Figure 4 or 5 or 12 or whose sequence is encoded by the nucleotide sequence given in Figure 4 between nucleotides 1 and 1395 or between nucleotides 2 and 948 of Figure 5 or between nucleotides 71 and 3442 of Figure 12. Preferably, the ECSM4 polypeptide is one whose amino acid sequence comprises the sequence given in Figure 4 or Figure

12.

In the term "ECSM4 human polypeptide" we include a polypeptide represented by SEQ ID No. 18085 of EP1074617, SEQ ID No. 211 of either WO 00/53756 or WO 99/46281, SEQ ID No. 24-27, 29, 30, 33, 34 , 38 or 39 of WO 01/23523, or SEQ ID No. 86 of WO 99/11293, or the polypeptide represented by SEQ ID No. 18084 or 5096 of EP1074617, SEQ ID No. 210 of either WO 00/53756 or WO 99/46281 , SEQ ID No. 24-27, 29, 30, 33, 34, 38 or 39 of WO 01/23523, or SEQ ID No. 86 of WO 99/11293

In the term "ECSM4 polypeptide A" we also include any polypeptide that occurs naturally and that contains a portion of 50 consecutive amino acid residues, or natural variants thereof, with the polypeptide sequence given in Figures 4, 5 or 12. Preferably , the polypeptide is a human polypeptide.

A first aspect of the invention provides an antibody capable of selectively binding the human ECSM4 polypeptide, wherein said human ECSM4 polypeptide is a naturally occurring polypeptide and comprising a portion of 50 consecutive amino acid residues of the polypeptide sequence of the Figure 4 or Figure 5 for use as a medicine.

The antibody can selectively bind a polypeptide comprising the amino acid sequence located between residues 1 and 467 of the polypeptide sequence of Figure 12.

The antibody can be a monoclonal antibody.

Embodiments of the features of this aspect of the invention are as described below in more detail.

A second aspect of the invention provides a compound comprising (I) an antibody as defined in the first aspect of the invention, and (II) an additional fraction, for use as a medicament.

Preferably the antibody and the additional fraction are covalently bound.

Preferably, an antibody that selectively binds to ECSM4 is one that selectively binds a polypeptide with the sequence GGDSLLGGRGSL, LLQPPARGHAHDGQALSTDL, EPQDYTEPVE, TAPGGQGAPWAEE or ERATQEPSEHGP, or a sequence located in the extracell portion of EC4. As described in more detail below, these sequences represent amino acid sequences that are only found in human ECSM4 and are not found in the mouse ECSM4 polypeptide sequence.

Preferably, the antibody that selectively binds to ECSM4 is the one that binds to a polypeptide whose amino acid sequence contains the sequence given in any of Figures 4, 5 or 12 but does not bind to the polypeptide represented by any of the SEC . ID. No. 18085 of EP 1 074 617, SEQ ID No. 211 of WO 00/53756 or W099146281, SEQ ID Nos. 24-27, 29, 30, 33, 34, 38 or 39 of WO 01/23523, or SEQ ID No. 86 of WO 99/11293, or encoded by any of the nucleotide sequences represented by the SEC. ID No. 18084 or 5096 of EP 1 074 617, SEQ ID No. 210 of WO 00 53756 or WO 99/46281, or SEQ ID Nos. 22, 23, 96 or 98 of WO 01/23523 and SEQ ID No. 31 of WO 99 / 11293.

In the term "antibody" we include not only complete immunoglobulin molecules, but also fragments thereof, such as Fab, F (ab ’) 2, Fv and other fragments thereof that maintain the antigen binding site. Similarly, the term "antibody" includes genetically modified derivatives of antibodies, such as single chain Fv molecules (scFv) and antibody domains (dAbs). The term also includes antibody-like molecules that can be produced by phage-display techniques or other random selection techniques of molecules that bind to ECSM4.

The heavy variable (VH) and light variable (VL) domains of the antibodies are involved in antigen recognition, a fact that was first recognized by early protease digestion experiments. A subsequent confirmation was obtained by the "humanization" of rodent antibodies. Variable domains of rodents can be fused to constant domains of human origin, so that the resulting antibody maintains the antigenic specificity of the original rodent antibody (Morrison et al (1984) Proc. Natl. Acad. Sci. USA 81, 6851 -6855).

That antigenic specificity is conferred by variable domains, and is independent of the constant domains. It is known from experiments that involve the expression in bacteria of antibody fragments, all containing one or more variable domains. These molecules include Fab type molecules (Better et al (1988) Science 240, 1041); Fv type molecules (Skerra et al (1988) Science 240, 1038); single chain Fv molecules (ScFv) in which the corresponding VH and VL domains are linked by a flexible oligopeptide (Bird et al (1988) Science 242, 423; Huston et al (1988) Proc. Natl. Acad. Sci. USA 85, 5879) and single domain antibodies (dAbs), which contain isolated V domains (Ward et al (1989) Nature 341, 544). A general review of the techniques involved in the synthesis of antibody fragments that maintain their specific binding sites can be found in Winter & Milstein (1991) Nature 349, 293-299.

By "ScFv molecules" we mean molecules in which the corresponding VH and VL domains are linked by means of a flexible oligopeptide.

The advantages of the use of antibody fragments, instead of complete antibodies, are of various levels. The smaller size of the fragments can lead to improved pharmacological properties, such as better penetration to the white site. The effector functions of the whole antibodies, such as complement binding, are eliminated. The Fab, Fv, ScFv and dAb fragments can be expressed and secreted in E. coli, thereby allowing easy production of large quantities of said fragments.

The whole antibodies and the F (ab ’) 2 fragments are" bivalent ". By "bivalent" we mean that said antibodies and F (ab ’) 2 fragments have two sites of combination with the antigen. On the contrary, the Fab, Fv, ScFv and dAb fragments are monovalent, having a single combination site to the antigen.

Although the antibody may be a polyclonal antibody, it is preferred to use a monoclonal antibody. In some circumstances, in particular if the antibody is to be administered repeatedly to a human patient, it is preferred that the monoclonal antibody be a human monoclonal antibody or a humanized monoclonal antibody.

Suitable monoclonal antibodies, which are reactive as such, can be prepared by employing known techniques, for example, those presented in "Monoclonal Antibodies; A manual of techniques ", H Zola (CRC Press, 1988) and in " Monoclonal Hybridoma Antibodies: Techniques and Application ", SGR Hurrell (CRC Press, 1982). Polyclonal antibodies that are polyspecific or monospecific can be produced. It is preferred that they be monospecific.

Chimeric antibodies are discussed in Neuberger et al (1998, 8th International Biotechnology Symposium Part 2, 792-799).

Suitably prepared non-human antibodies can be "humanized" by known methods, for example, by inserting the CDR regions of mouse antibodies into the framework of human antibodies.

The antibodies may be human antibodies in the sense that they possess the amino acid sequence of human anti-ECMS4 antibodies, but may be prepared using methods known in the art that do not require human immunization. For example, transgenic mice are available that contain, in essence, human immunoglobulin genes (see Vaughan et al (1998) Nature Biotechnol. 16, 535-539).

The additional fraction may be any fraction that gives the compound a useful property with respect to treatment or imaging or the diagnosis of diseases or other conditions or conditions that involve an undesirable formation of new vasculature. Such diseases or other conditions or conditions are described in more detail below. In particular, the additional fraction is some that is useful in the removal or imaging of new vasculature associated with the growth of a tumor. Preferably, the additional fraction is some that is capable of killing the endothelial cells to which the compound is directed.

In a preferred embodiment of the invention, the additional fraction is cytotoxic directly or indirectly. In particular, the additional fraction is preferably directly or indirectly toxic on cells of the new vasculature or cells that are in the vicinity of, and associated with the new vasculature.

In the term "directly cytotoxic" we include the sense that the fraction is some that is, by itself, cytotoxic. In the term "indirectly cytotoxic" we include the sense that the fraction is some that, although not itself cytotoxic, can induce cytotoxicity, for example, through its action on some other molecule or by another action on it.

In one embodiment, the cytotoxic fraction is a chemotherapeutic cytotoxic agent. Chemotherapeutic cytotoxic agents are well known in the art.

Chemotherapeutic cytotoxic agents, such as anticancer agents include: alkylating agents including nitrogen derivatives, such as mechloretamine (NH2), cyclophosphamide, ifosfamide, melphalan (L-sarcolysin) and chlorambucil; ethylenamines and methylenamines, such as hexamethylmelamine, thioepta; alkyl sulfonates, such as busulfan; nitrosureas, such as camustine (BCNU), lomustine (CCNU), semustine (methyl-CCNU) and streptozocin (streptozocin); and triazenos, such as decarbazine (DTIC; dimethyltriazenoimidazol-carboxamide); antimetabolites, including folic acid analogs, such as methotrexate (ametopterine); pyrimidine analogues, such as fluorouracil (5-fluorouracil; 5-FU), floxuridine (fluorodeoxyuridine; FUdR) and cytarabine (cytosine arabinoside); and purine analogs and related inhibitors, such as mercaptopurine (6-mercaptopurine; 6-MP), thioguanine (6-thioguanine; TG) and pentostatin (2’-deoxycoformycin). Natural products including vinca alkaloids, such as vinblastine (VLB) and vincristine; epipodophyllotoxins, such as etoposide and teniposide; antibiotics, such as dactinomycin (actinomycin D), daunorubicin (daunomycin; rubidomycin), doxorubicin, bleomycin, plicamycin (mitramycin) and mitomycin (mitomycin C); enzymes, such as L-asparaginase; and biological response modifiers, such as interferon alfenomas. Varied agents, including platinum coordination complexes, such as cisplatin (cis-DDP) and carboplatin; anthracenedione, such as mitoxantrone and anthracycline; substituted ureas, such as hydroxyurea; methyl hydrazine derivatives, such as procarbacin (N-methylhydrazine, MIH); and adrenocortical suppressors, such as mitotane (or, p’-DDD) and aminoglutethimide; taxol and analogues / derivatives; and hormone agonists / antagonists, such as flutamide and tamoxifen. Several of these agents have been previously bound to antibodies and other target-directed release agents, so that compounds of the invention that include such agents can be easily prepared by persons skilled in the art. For example, carbodiimide conjugation (Bauminger & Wilchek (1980) Methods Enzymol. 70, 151-159) can be used to conjugate a range of agents, including doxorubicin, to antibodies or peptides.

Carbodiimides include a group of compounds that have the general formula R-N = C = N-R ’, where R and R’ can be aliphatic or aromatic, and are used for the synthesis of peptide bonds. The preparatory procedure is simple, relatively fast and can be carried out under mild conditions. Carbodiimide-type compounds attack carboxylic groups to convert them into reactive sites for free amino groups.

Water-soluble carbodiimide, 1-ethyl-3- (3-dimethylaminopropyl) carbodiimide (EDC) is particularly useful for the conjugation of a functional fraction to a binding fraction and can be used to conjugate doxorubicin to peptides that appear in tumors. The conjugation of doxorubicin and a binding fraction requires the presence of an amino group, which is provided by doxorubicin, and a carboxyl group, which is provided by the binding fraction, such as an antibody.

or a peptide.

In addition to the use of carbodiimides for direct peptide bond formation, EDC can also be used to prepare active esters, such as N-hydroxysuccinimide ester (NHS). The NHS ester, which binds only to amino groups, can in turn be used to induce the formation of an amide bond with the only amino group of doxorubicin. The combined use of EDC and NHS is frequently used for conjugation, with the aim of increasing the yield in conjugate formation (Bauminger & Wilchek, supra, 1980).

Other methods can also be used to conjugate a functional fraction to an antibody. For example, oxidation with sodium periodate can be employed, followed by reductive alkylation of appropriate reagents, as well as cross-linking with glutaraldehyde. However, it is recognized that, regardless of the method of production of a conjugate of the invention that is selected, the determination should be made that the antibody maintains its targeting ability and that the functional fraction maintains its relevant function.

In another embodiment, the cytotoxic fraction is a cytotoxic peptide or a polypeptide fraction, in which we include any fraction that leads to cell death. Cytotoxic peptides and polypeptide fractions are well known in the art and include, for example, ricin and abrin, Pseudomonas exotoxin, tissue factor, and the like. Methods for binding them to antibodies are also known in the art. The use of ricin as a cytotoxic agent is described in Burrows & Thorpe (1993) Proc. Natl Acad. Sci. USA 90, 8996-9000; and the use of tissue factor, which leads to localized blood coagulation and tumor infarction has been described in Ran et al (1998) Cancer Res. 58, 4646-4653 and Huang et al (1997) Science 275, 547 -550. Tsai et al (1995) Dis. Colon Rectum 38, 1067-1074 describes the A chain of the abrin, conjugated with a monoclonal antibody. Other inactivated ribosome proteins are described as cytotoxic agents in WO 96/06641. Pseudomonas exotoxin can also be used as a cytotoxic polypeptide fraction (see, for example, Aiello et al (1995) Proc. Natl. Acad. Sci. USA 92, 10457-10461)

Some cytokines, such as TNFa and IL-2, may also be useful as cytotoxic agents.

Some radioactive atoms can also be cytotoxic if given in sufficient doses. Therefore, the cytotoxic fraction may include a radioactive atom that, when used, releases a sufficient amount of radioactivity to the white site, such as being cytotoxic. Suitable radioactive atoms include phosphorus 32, iodine 125, iodine 131, indium 111, rhenium 186, rhenium 188

or ytrium 90, or any other isotope that emits enough energy to destroy neighboring cells, organelles or nucleic acids. Preferably, the isotopes and the density of the radioactive atoms in the compound of the invention are such that a dose of more than 4000 cGy (preferably at least 6000, 8000 or 10000 cGy) is delivered to the white site, preferably, to the cells in the white site and their organelles, in particular, the nucleus.

The radioactive atom can bind to the antibody in known ways. For example, EDTA or other chelating agent can be attached to the binding fraction and used to bind 111In or 90Y. Tyrosine residues can be labeled with 125I or 131I.

The cytotoxic fraction may be a suitable indirectly cytotoxic polypeptide. In a preferred embodiment, the indirectly cytotoxic polypeptide is a polypeptide that possesses enzymatic activity and can convert a relatively non-toxic prodrug into a cytotoxic drug. When the targeting fraction is an antibody, this type of system is often referred to as ADEPT (Enzyme therapy directed by antibody on prodrug). The system requires that the targeting fraction locate the enzyme portion at the desired site of the patient's body (that is, the site that expresses ECSM4, such as the new vasculation associated with a tumor) and then allow some time for the enzyme to be located at the site, administer a prodrug that is the enzyme substrate, the final product of which catalysis is a cytotoxic compound. The purpose of this strategy is to maximize the drug concentration at the desired site and minimize the drug concentration in normal tissues (see Senter, PD et al (1988) " Antitumor effects of antibody-alkaline phosphatase conjugates in combination with etoposide phosphate " Proc. Natl. Acad. Sci. USA 85, 4842-4846; Bagshawe (1987) Br. J. Cancer 56, 531-2; and Bagshawe,

K.D. et al (1988) " A cytotoxic agent can be generated selectively at cancer sites " Br. J. Cancer. 58, 700-703.)

The enzyme and the prodrug of the system that employs an enzyme directed to ECSM4 as described herein may be any of the proposals proposed above. The cytotoxic substance can be any existing anticancer drug, such as an alkylating agent; an agent that is sandwiched in the DNA; an agent that inhibits any key enzyme, such as dihydrofolate reductase, thymidine synthetase, ribonucleotide reductase, nucleoside kinase or topoisomerase; or an agent that causes cell death by interacting with any other cellular constituent. The etoposide is an example of a topoisomerase inhibitor.

Reported drug systems include: a phenol gas prodrug activated by an E. coli �-glucuronidase (Wang et al, 1992 and Roffler et al, 1991); a prodrug of doxorubicin activated by a human �-glucuronidase (Bosslet et al, 1994); other doxorubicin prodrugs activated by coffee seed a-galactosidase (Azoulay et al, 1995); daunorubicin prodrugs activated by coffee seed a-D-galactosidase (Gesson et al, 1994); a 5-fluorouridine prodrug activated by a �-D-galactosidase (Abraham et al, 1994); and methotrexate prodrugs (e.g. methotrexate-alanine) activated by carboxypeptidase A (Kuefner et al, 1990, Vitols et al, 1992 and Vitols et al, 1995). These and others are included in the following table.

Enzyme

Prodrug

Carboxypeptidase G2 Alkaline Phosphatase

L-glutamic acid and benzoic acid mustard derivatives, aniline mustards, phenol mustards and phenylenediamine mustards; fluorinated derivatives of these Etoposide phosphate Mitomycin phosphate

Beta glucuronidase

P-Hydroxyaniline mustard glucuronide Epirubicin glucuronide

Penicillin V amidase

Adriamicin-N phenoxyacetyl

Penicillin G amidase

N- (4’-hydroxyphenyl acetyl) palitoxin Doxorubicin and melphalan

Beta lactamase Beta glucosidase

nitrogen mustard p-phenylenediamine cephalosporin; doxorubicin derivatives; cephalosporin derived from vinblastine, cephalosporin mustard; Taxol derivative Cyanophenylmethyl-beta-D-gluco-pyranosiduronic acid

Nitroreductase

5- (Azaridine-1-yl -) - 2,4-dinitrobenzamide

Cytosine deaminase

5-Fluorocytosine

Carboxypeptidase A

Methotrexate-Alanine

(This table is adapted from Bagshawe (1995) Drug Dev. Res. 34, 220-230, where complete references can be obtained for these different systems; the taxol derivative is described in ML et al (1995) Chemistry & Biology 2 , 223).

Enzymes suitable for forming part of the enzymatic portion include: exopeptidases, such as carboxypeptidases G,

5 G1, and G2 (for glutamylated prodrugs mustards), carboxypeptidases A and B (for MTX-based prodrugs) and aminopeptidases (for 2-a-aminoacyl MTC prodrugs); endopeptidases, such as, e.g., thrombolysin (for thrombin prodrugs); hydrolases, such as phosphatases (e.g., alkaline phosphatase) or sulfatases (e.g., aryl sulfatases) (for phosphorylated or sulfated prodrugs); amidases, such as penicillin and arylacyl amidase amidases; lactamases, such as

�-lactamases; glycosidases, such as �-glucuronidase (for anthracyclines �-glucuronomide), a-galactosidase (for tonsillin) and

10 �-galactosidase (for anthracycline �-galactose); deaminase, such as cytosine deaminase (for 5FC); kinases, such as urokinase and thymidine kinase (for ganciclovir); reductases, such as nitroreductase (for CB1954 and the like), azoreductase (for azobenzene mustards) and DT-diaphorase (for CB1954); oxidases, such as glucose oxidase (for glucose), xanthine oxidase (for xanthine) and lactoperoxidase; DL-racemases, catalytic antibodies and cyclodextrins.

Prodrug is relatively non-toxic, compared to the cytotoxic drug. It usually has less than 10% of the

15 toxicity, preferably, less than 1% of the toxicity, according to the result of the measurement in a suitable in vitro cytotoxicity test.

It is likely that the fraction capable of converting a prodrug into a cytotoxic drug is active in isolation from the rest of the compound, but it only needs to be active when (a) it is in combination with the rest of the compound and (b) the compound is bound to, adjacent to, or internalized in white cells.

When the other fraction is a polypeptide, the antibody and the other fraction can be linked together by any of the conventional means of cross-linking polypeptides, such as those described generally in O'Sullivan et al (1979) Anal . Biochem 100, 100-108. For example, the antibody that selectively binds to ECSM4 may be enriched with thiol groups and the other fraction may react with a bifunctional agent capable of reacting with said thiol groups, for example, the iodoacetic acid ester of N-hydrosuccinimide (NHIA) or N-succinimidyl-3- (2-pyridylthio) propionate (SP-DP). The amide and thioester bonds, for example, achieved by m-maleimidobenzoyl-N-hydroxysuccinimide ester are generally more stable in vivo than the disulfide bridges.

Alternatively, the compound can be produced as a fusion compound, by recombinant DNA techniques, in which one segment of DNA contains the respective regions encoding the antibody and the other fraction, either adjacent to each other, or separated by a region that codes for a binding polypeptide that does not destroy the properties of the compound. Conceivably, the two portions of the compound can be totally or partially overlapped.

The DNA is then expressed in a suitable host to produce a polypeptide containing the compound of the invention.

The cytotoxic fraction can be a radiosensitizer. Radiosensitizers include fluoropyrimidines, thymidine analogs, hydroxyurea, gencitabine, fludarabine, nicotinamide, halogenated pyrimidines, 3-aminobenzamide, 3-aminobenzodiamide, ethanixadol, pimonidazole and misonidazole (see, for example , McGinn et al (1996) J. Natl. Cancer Inst. 88, 1193-11203; Shewach & Lawrence (1996) Invest. New Drugs 14, 257-263; Horsman (1995) Acta Oncol. 34, 571-587; Shenoy & Singh (1992) Clin. Invest. 10, .533-551; Mitchell et al (1989) Int. J. Radiat. Biol. 56, 827-836; Iliakis & Kurtzman (1989) Int. J. Radiat Oncol. Biol. Phys. 16, 1235-1241; Brown (1989) Int. J. Radiat. Oncol. Biol. Phys. 16, 987-993; Brown (1985) Cancer 55, 2222-2228).

Also, the delivery of genes to cells can radiosensitize them, for example, the insertion of the p53 or cyclin D gene (Lang et al (1998) J. Neurosurg. 89, 125-132; Coco Martin et al (1999) Cancer Res. 59, 1134-1140).

The additional fraction may be one that becomes cytotoxic, or releases a cytotoxic fraction, under irradiation. For example, boron isotope 10, when properly irradiated, releases particles that are cytotoxic (see, for example, US 4,348,376 to Goldenberg; Primus et al (1996) Bioconjug. Chem. 7, 532-535) .

Similarly, the cytotoxic fraction may be one that is useful in photodynamic therapy, such as photophrine (see, for example, Dougherty et al (1998) J. Natl. Cancer Inst. 90, 889-905).

The other fraction may contain a nucleic acid molecule that is directly or indirectly cytotoxic. For example, the nucleic acid molecule may be an antisense oligonucleotide which, being located in the white site, is capable of penetrating cells and causing death. The oligonucleotide, therefore, may be one that prevents the expression of an essential gene, or one that leads to a change in gene expression that causes apoptosis.

Examples of suitable oligonucleotides include those directed to bcl-2 (Ziegler et al (1997) J. Natl. Cancer Inst. 89, 1027-1036), and DNA polymerase a and topoisomerase IIa (Lee et al (1996) Anticancer Res. 16, 1805-1811.

Peptide nucleic acids may be useful instead of conventional nucleic acids (see Knudsen & Nielsen (1997) Anticancer Drugs 8, 113-118).

In another embodiment, the antibody may be contained in a delivery vehicle, for delivery of the nucleic acid to the blank. The delivery vehicle can be any appropriate delivery vehicle. It may, for example, be a liposome containing nucleic acid, or it may be a virus or a virus-like particle that is capable of delivering a nucleic acid. In these cases, the antibody that selectively binds to ECSM4 is usually present on the surface of the delivery vehicle. For example, the antibody that selectively binds to ECSM4, such as a suitable antibody fragment, may be present on the outer surface of a liposome and the nucleic acid to be delivered may be present inside the liposome. As another example, a viral vector, such as a retroviral or adenoviral vector, is constructed, so that the antibody that selectively binds to ECSM4 is bound to or located at, the surface of the particle, which allows The viral particle is directed to the desired site. Directed delivery systems are also known as modified adenovirus systems, described in WO 94/10323 in which, usually, the DNA is carried by the adenoviral or adenoviral type particle. Michael et al (1995) Gene Therapy 2, 660-668 describe a modification of adenoviruses to add a selective fraction of cells to a protein fiber. Dissected retroviruses are also available for use in the invention; for example, sequences that confer specific binding affinities can be constructed within pre-existing viral env genes (see Miller & Vile (1995) Faseb J. 9, 190-199 for a review of this and other dissected vectors for gene therapy).

Immunoliposomes (antibody-directed liposomes) that contain an antibody that selectively binds to ECSM4 can be employed. For the preparation of immunoliposomes MPB-PE (N- [4- (p-maleimidophenyl) butyryl] -phosphatidylethanolamine) is synthesized according to the method of Martin & Papahadjopoulos (1982) J. Biol. Chem. 257, 286-288. MBP-PE is incorporated into the liposomal bilayers to allow covalent binding of the anti-ECSM4 antibody, or a fragment thereof, to the liposomal surface. The liposome is conveniently loaded with DNA or other genetic construct, followed by sequential extrusion through polycarbonate membrane filters with pore sizes of 0.6 µm and 0.2 µm under nitrogen pressures of up to 0.8 MPa. After extrusion, the trapped DNA construct is separated from the free ADn construct by ultracentrifugation at 80,000 x g for 45 minutes. The MBP-PE liposomes just prepared in deoxygenated buffer are mixed with the freshly prepared antibody (or a fragment thereof) and the coupling reactions are carried out in a nitrogen atmosphere at 4 ° C under constant overnight rotation. Immunoliposomes are separated from unconjugated antibodies by ultracentrifugation at 80,000 x g for 45 min. Immunoliposomes can be injected intraperitoneally or directly into the tumor.

The nucleic acid delivered to the target site can be any DNA that leads, directly or indirectly, to cytotoxicity. For example, the nucleic acid can code for a ribozyme that is cytotoxic to the cell, or it can code for an enzyme that is capable of converting an essentially non-toxic prodrug into a cytotoxic drug (the latter system is sometimes called GDEPT: Enzymatic therapy of Prodroga directed by gene).

Ribozymes that may be encoded in the nucleic acid to be delivered to the target are described in Cech and Herschlag " Site-specific cleavage of single stranded DNA " US 5,180,818; Altman et al " Cleavage of targeted RNA by RNAse P " US 5,168,053, Cantin et al " Ribozyme cleavage of HIV-1 RNA " US 5,149,796; Cech et al. "RNA ribozyme restriction endoribonucleases and methods", US 5,116,742; Been et al " RNA ribozyme polymerases, dephosphorylases, restriction endonucleases and methods ", US 5,093,246; and Been et al. "RNA ribozyme polymerases, dephosphorylases, restriction endoribonucleases and methods; cleaves single-stranded RNA at specific site by transesterification ", US 4,987,071. Suitable targets for ribozymes include transcription factors, such as c-fos and c-myc, and bcl-2. Durai et al (1997) Anticancer Res. 17, 3307-3312 describe a hammerhead ribozyme against bcl-2.

EP 0 415 731 describes the GDEPT system. Similar considerations regarding enzyme selection and prodrug apply to the GDEPT system in the same way as to the ADEPT system described above.

The nucleic acid delivered to the blank site can code for a directly cytotoxic polypeptide.

Alternatively, the other portion may include a polypeptide or a polynucleotide that encodes a polypeptide that is not directly or indirectly cytotoxic, but that is of therapeutic benefit. Examples of such polypeptides include anti-proliferative or anti-inflammatory cytokines that may be beneficial for atherosclerosis, and immunomodulators.

or antiproliferative factors that influence blood clotting may be beneficial for cancer treatment.

The other fraction may be a useful angiogenesis inhibitor, such as angiostatin or endostatin peptides. The other fraction may also be a useful enzyme that converts a precursor polypeptide to angiostatin or endostatin. Human matrix metalloproteases such as macrophage elastase, gelatinase and stromolysin convert plasminogen to angiostatin (Cornelius et al (1998) J. Immunol. 161, 6845-6852). Plasminogen is a precursor of angiostatin.

In another embodiment of the invention, the other fraction is an easily detectable fraction.

In the term "easily detectable fraction" we include the sense that the fraction is one that, when located in the white site after administration of the compound to a patient, can be detected, usually, non-invasively from the outside the body and the white site can be detected. Therefore, the compounds are useful for imaging and diagnosis.

Typically, the easily detectable fraction is, or comprises, a readiactive atom that is useful in imaging. Suitable radioactive atoms include technetium 99, or iodine 123 for cylindrical studies. Other easily detectable fractions include, for example, spin marks for magnetic resonance imaging (MRI), such as iodine 123 again, iodine 131, indium 111, fluorine 19, carbon 13, nitrogen 15, oxygen 17, gadolinium, manganese or iron Obviously, the compound must have a sufficient amount of the appropriate atomic isotope so that the molecule can be easily detected.

The radioactive or other brand may be incorporated into the compound by known means. For example, the antibody that selectively binds to ECSM4 can be biosynthesized or can be synthesized by chemical synthesis of amino acids, using the appropriate amino acid precursors, which includes, for example, fluorine 19 instead of hydrogen. Marks such as 99Tc, 123I, 186Rh, 188Rh, and 111In can, for example, be linked by a cysteine residue to the binding fraction. Itrium 90 can be linked by a lysine residue. The IODOGEN method (Fraker er al (1978) Biochem. Biophys. Res. Comm. 80, 49-57) can be used to incorporate iodine 123. The reference (" Monoclonal Antibodies in Immunoscintigraphy " JF Chatal, CRC Press, 1989) Describe other methods in detail.

In another preferred embodiment of the invention, the other fraction is capable of selectively binding to a directly or indirectly cytotoxic fraction or an easily detectable fraction. Therefore, in this embodiment, the other fraction can be any fraction that binds to another compound or component that is cytotoxic or easily detectable.

The other fraction may, therefore, be an antibody that selectively binds to another compound or component, or it may be some other binding fraction, such as streptavidin, or biotin, or the like. The following examples illustrate the types of molecules that are included; The utility of other molecules can be easily deduced from the explanations provided in this document.

A bispecific antibody, in which a binding site includes the fraction that selectively binds to human ECSM4 and the second binding site includes a fraction that binds, for example, to an enzyme that is capable of converting a prodrug substantially Non-toxic in a cytotoxic drug.

A compound, such as an antibody that selectively binds to ECSM4, to which biotin, avidin or streptavidin has been bound, which have been labeled with an easily detectable probe can be used in conjunction with the biotin-labeled antibody in a system of Two-stage imaging, in which the labeled antibody is first located at the white site in the patient, and then, the labeled avidin or streptavidin are administered to the patient. The bispecific antibody and biotin / streptavidin (avidin) systems are reviewed in Rosebrough (1996) Q J Nucl. Med. 40,234-251.

In a preferred embodiment of the invention, the antibody that selectively binds to human ECSM4 and the other fraction are fused polypeptides.

The compounds of the second aspect of the invention are useful for the treatment, imaging or diagnosis of diseases, in particular diseases in which new undesirable vasculature may be formed, as described below in more detail.

It is appreciated that a nucleic acid molecule can encode a compound of the second aspect of the invention in which the antibody and the additional fraction are polypeptides that are fused.

Methods of binding polynucleotides are described below in greater detail.

The antibody or compound for use according to the invention can be formulated as a pharmaceutical composition comprising the antibody or compound and a pharmaceutically acceptable carrier.

The term "pharmaceutically acceptable" includes the formulation being sterile and pyrogen free. Pharmaceutically acceptable carriers are well known in the pharmaceutical art.

The carriers must be "acceptable" in the sense of being compatible with the antibody or compounds of the invention and not deleterious to the receptors thereof. Typically, the carriers will be sterile, pyrogen-free water or saline solution; however, other acceptable carriers may be employed.

Typically, pharmaceutical compositions or formulations are for parenteral administration, more particularly, for intravenous administration.

Formulations suitable for parenteral administration include sterile, aqueous and non-aqueous injection solutions, which may contain antioxidants, buffers, bacteriostatics and solutes that make the isotonic formulation with respect to the desired recipient's blood, and aqueous and non-aqueous sterile suspensions that They may contain suspending agents and agents that make them denser.

A compound according to the second aspect of the invention in which the additional fraction is an easily detectable fraction are useful can be used in an imaging method of the vascular endothelium in the body of an individual by administration to the individual of an effective amount of said compound. Usually, the vascular endothelium is related to angiogenesis.

Imaging of the vascular endothelium of an individual usually includes the detection of the location of the compound in said individual.

The detection of the compound or antibody can be carried out using methods well known in the art of imaging and clinical diagnosis. The specific method required will depend on the type of detectable label bound to the compound or antibody. For example, radioactive atoms can be detected using autoradiography or, in some cases, by magnetic resonance imaging (MRI), as described above.

The imaging of the vascular endothelium in the body is useful because it can provide information about the health of the body. It is particularly useful when the vascular endothelium is diseased, or is proliferating due to cancerous growth. Cancer imaging in a patient is especially useful, because it can be used to determine the size of a tumor and if it is responding to treatment. Since metastatic diseases involve the formation of new blood vessels, the method is useful in assessing whether a metastasis has occurred.

The vascular endothelium may be neovasculature, such as that produced in cancer.

A compound according to the second aspect of the invention in which the additional fraction is an easily detectable fraction can be used in a method for the diagnosis or prognosis, in an individual, of a condition involving the vascular endothelium, by administration to the individual of an effective amount of said compound ..

The medical condition may be one that includes aberrant or excessive growth of the vascular endothelium, such as cancer, atherosclerosis, restenosis, diabetic retinopathy, arthritis, psoriasis, endometriosis, menorrhagia, hemangiomas, and venous malformations.

The method can be one that is a diagnostic aid.

The diagnosis or support for the diagnosis of a medical condition that includes the vascular endothelium in an individual usually includes the detection of the location of the compound in the individual. Preferably, the endothelium is in a new vasculature, that is, angiogenic vasculature.

The function of ECSM4 may not be the promotion of proliferation of vascular endothelial cells. Therefore, the level of expression of ECSM4 polypeptides in an endothelial cell may not be informative regarding the health of the vascular endothelium. However, the location of the expression of ECSM4 polypeptides can be informative, since they represent the growth of blood vessels. Abnormal cell proliferation, such as cancer, can be diagnosed by detecting new vasculature.

A compound according to the second aspect of the invention in which the additional fraction is a cytotoxic or therapeutic fraction can be used in a method for treating an individual in need of treatment, by administering to the individual an effective amount of said compound.

The patient who requires treatment may have a proliferative disease or a condition that involves the vascular endothelium.

A number of diseases and ailments involve the unwanted formation of new vasculature. The formation of new vasculature is associated with cancer, psoriasis, atherosclerosis, menorrhagia, arthritis (both inflammatory and rheumatoid), macular degeneration, Pager's disease, retinopathy and its vascular complications (including proliferative and prematurity, and diabetics), benign vascular proliferations and fibrosis.

Cancer includes Kaposi's sarcoma, leukemia, lymphoma, myeloma, solid carcinomas (both primary and secondary (metastases), vascular tumors, including hemangioma (both capillary and juvenile (childhood)), hemangiomatosis and Hemagioblastoma

Thus, the compound of the second aspect of the invention can be used in a method of treating a patient who has a disease in which angiogenesis contributes to the pathology.

Usually, the disease is associated with the formation of new unwanted vasculature and treatment reduces it by an appropriate amount.

Tumors that can be treated by the methods of the invention include any tumors associated with the production of new blood vessels.

The term "tumor" should be understood as referring to all forms of neoplastic cell growth, including lung, liver, blood cell, skin, pancreas, stomach, colon, prostate, uterus tumors, breasts, lymph glands and bladder. Solid tumors are especially appropriate. However, it is currently estimated that blood cancers, including leukemia and lymphomas also involve the formation of new blood vessels and can be treated by the methods of the invention.

Normally for use in the aforementioned treatment methods, the other fraction is one that destroys or slows down, or reverses the growth of the new vasculature.

It will be readily appreciated that, depending on the particular compound used in the imaging, diagnosis or treatment, the timing of administration may vary, as well as the amount of other components used in the therapeutic systems described herein.

For example, in the case where the compound of the second aspect of the invention comprises an easily detectable fraction or a directly cytotoxic fraction, it may happen that only the compound, in a suitable formulation, is for administration to the patient, of course, Other agents, such as immunosuppressive agents and the like, may be administered to the patient.

With respect to the compounds that are detectably labeled, imaging takes place once the compound has located the white site.

However, if the compound is such that it requires another component to be useful for treatment, imaging and diagnosis, the compound can be administered and allowed to locate the white site and subsequently administer the other component after a period of time. suitable.

For example, with respect to the aforementioned ADEPT and GDEPT type systems, the enzyme-binding fraction compound is administered and located in the white site. Once this is done, the drug is administered.

Something similar happens, for example, with respect to the compounds in which the other fraction contained in the compound is such that it binds to another component, the compound can be administered first and allowed to be located in the white site and then Manage the other component.

Therefore, in one embodiment, a biotin-labeled anti-ECSM4 antibody is administered to the patient and, after a suitable period of time, labeled streptavidin is administered detectably. Once streptavidin has been located at the sites where the antibody has been located (i.e., the target sites), imaging takes place. It is believed that the compounds of the second aspect of the invention in which the other fraction is an easily detectable fraction, may be useful for the determination of the angiogenic state of tumors or other disease states in which angiogenesis contributes to the pathology. This may be an important factor that influences the nature and outcome of future therapy.

A compound according to the second aspect of the invention in which the additional fraction is an easily detectable label can be used in the manufacture of a diagnostic or prognostic agent for a condition involving the vascular endothelium.

A compound according to the second aspect of the invention in which the additional fraction is a cytotoxic or therapeutic fraction can be used in the manufacture of a diagnostic or prognostic agent for a condition that involves the vascular endothelium.

Conditions that involve the vascular endothelium are described below.

"Fragments" of the human ECSM4 polypeptide include polypeptides comprising at least five consecutive amino acids of the ECSM4 polypeptide. Preferably, a fragment of the polypeptide comprises an amino acid sequence that is useful as a peptide to produce an antibody that is specific for the human ECSM4 polypeptide. Normally, the fragments have at least 8 amino acids, preferably at least 10, more preferably at least 12 or 15 or 20 or 30

or 40 or 50 consecutive amino acids of the ECSM4 or ECSM1 polypeptide. Preferably, fragments of the ECSM4 polypeptide comprise but do not consist of any of the amino acid sequences represented by SEQ. ID. No. 18085 of EP 1 074 617, SEQ ID No. 211 of WO 00/53756 or W099146281, SEQ ID Nos. 24-27, 29, 30, 33, 34, 38 or 39 of WO 01/23523,

or SEQ ID No. 86 of WO 99/11293, or any of the amino acid sequences encoded by the SEC. ID No. 18084 or 5096 of EP 1 074 617, SEQ ID No. 210 of WO 00 53756 or WO 99/46281, or SEQ ID Nos. 22, 23, 96 or 98 of WO 01/23523 and SEQ ID No. 31 of WO 99 / 11293.

Typically, ECSM4 polypeptide fragments are ones that have portions of the amino acid sequence shown in Figure 4 or Figure 12.

The ECSM4 polypeptide fragment can be a fragment that has the sequence LSQSPGAVPQALVAWRA, DSVLTPEEVALCLEL, TYGYISVPTA, KGGVLLCPPRPCLTPT, WLADTW, WLADTWRSTRGSRD, SPPTTYGYIS, GSLANGWGSVEDV sequences of either sequence or any of these sequences. These peptides correspond to amino acids 165-181, 274-288, 311-320, 336-351, 8-13, 8-21, 307-316, 355-387, and 390-399 respectively of the human ECSM4 polypeptide shown in Figure 4 Peptides WLADTW, WLADTWRSTRGSRD, SPPTTYGYIS, GSLANGWGSASEDNAASRASLVSSSDGSFLAD and FARALAVAVD represent conserved regions between the mouse and human homologs of the ECSM4 polypeptide, and between the ECSM4 polypeptide and the dutt1 mouse protein. LSQSPGAVPQALVAWRA, DSVLTPEEVALCLEL, TYGYISVPTA and KGGVLLCPPRPCLTPT peptides may be useful for promoting antibodies.

Preferred peptides are peptides of at least 5 or 8 or 10 or 12 or 15 or 20 consecutive amino acid residues of these conserved sequences.

The ECSM4 polypeptide fragment may be a fragment having the sequence GGDSLLGGRGSL, LLQPPARGHAHDGQALSTDL, EPQDYTEPVE, TAPGGQGAPWAEE and ERATQEPSEHGP. These peptides correspond to regions of the human ECSM4 polypeptide (located at residues 4-16, 91-109, 227-236,288-300 and 444-455, respectively in Figure 12) that are not, or only poorly, conserved in the mouse counterpart (see Figure 14). As will be described later, such peptides may be particularly useful for activating antibodies against the human ECSM4 polypeptide.

According to the transmembrane domain prediction program called PRED-TMR (available on the website http://www.biophys.biol.uoa.gr) and an alignment of the amino acid sequence with the human protein Robo1 (whose transmembrane region is known), residues 1-467, as shown in Figure 12 have possibilities of being extracellular, and in addition to being extracellular and exposed, may include the binding site for the natural ligand. Therefore, fragments of ECSM4 that contain or consist of a sequence of the extracellular domain of residues 1-467 of Figure 12, may represent fragments useful for activating selective antibodies for cells expressing ECSM4 on their surface and which may also be useful in the modulation of the activity of the ECSM4 polypeptide.

Thus, the preferred fragments of the ECSM4 polypeptide are those fragments of the polypeptide sequence of Figure 12 that comprise at least 1, 3, or 5 amino acid residues that are not conserved when compared to mouse ECSM4 (as shown in Figure 13). More preferably, at least 7, 9, 11, or 13 amino acid residues in the fragment are not conserved between human ECSM4 and mouse ECSM4, and even more preferably, at least 15, 17, 19, or 21 residues of the fragment are not conserved between human ECSM4 and mouse ECSM4. The sequence of such fragments can be determined from the alignment of the human and mouse amino acid sequences shown in Figure 14.

A "variant" of the ECSM4 polypeptide includes natural variants, including allelic variations and mutant forms that occur in nature, and variants with insertions, deletions and substitutions, whether conservative or non-conservative, whose changes do not significantly alter the activity of said polypeptide. In the case of the ECSM4 polypeptide, as a specific endothelial homolog of human roundabout 1, it may very well be involved in the repulsive guidance of the endothelial cell.

By "conservative substitution" are meant combinations such as Gly, Ala; Val, Ile, Leu; Asp, Glu; Asn, Gln; Be, Thr; Lys, Arg; and Phe, Tyr.

In "non-conservative substitution" we include other substitutions, such as those in which the substituted residue mimics a particular modification of the replaced residue, for example a phosphorylated tyrosine or serine can be replaced by an aspartate or a glutamate, thanks to the similarity of the chain side of aspartate or glutamate with a phosphorylated residue (in this case, they carry a negative charge at neutral pH).

Other non-conservative substitutions that are included within the term "variants" are point mutations that alter one, sometimes two, and usually no more than three amino acids.

Variants of the human ECSM4 polypeptide include polypeptides comprising a sequence with at least 65% identity with the amino acid sequence given in Figure 4, or Figure 12, preferably at least 70% or 80% or 85% or 90% identity with said sequence, and more preferably at least 95%, or 98% identity with said amino acid sequence.

The percentage of identity can be determined using, for example, the LALIGN program (Huang and Miller, Adv. Appl. Math. (1991) 12: 337-357) available on the Expasy site (http: //www.ch.embnet .org / software / LALIGN form.html) using as parameters the global alignment option, the BLOSUM62 scoring matrix, -14 as a penalty for the opening of slits, -4 as a penalty for the extension of slits.

A further aspect of the invention provides an expression vector comprising a polynucleotide encoding a compound according to the second aspect of the invention in which the antibody and the additional fraction are polypeptides that are fused, for use as a medicament.

By "expression vector" we mean one that is capable, in an appropriate host, of expressing a polypeptide encoded by the polynucleotide.

Such vectors may be useful in medicine. Expression vectors comprising a polynucleotide encoding a compound according to the second aspect of the invention in which the antibody and the additional fraction are polypeptides that are fused, which are suitable for use in gene therapy are within the scope of the invention. Therefore, in a preferred embodiment, the vector is one that is suitable for use in gene therapy.

DNA (or in the case of retroviral vectors, APN) encoding the polypeptide can bind to a wide variety of other DNA sequences for introduction into an appropriate host. The companion DNA will depend on the nature of the host, the manner of introduction of the DNA into the host, and whether episomal maintenance or integration is desired.

Generally, the DNA is inserted into an expression vector, such as a plasmid in correct orientation and correct reading frame for expression. If necessary, the DNA may be linked to the appropriate transcriptional and transnational regulatory control nucleotide sequences recognized by the desired host, although such controls are generally available in the expression vector. Then the vector is introduced into the host by standard techniques.

Preferably, an antibody that selectively binds to ECSM4 is one that binds to a polypeptide whose amino acid sequence contains the sequence given in any of Figures 4, 5 or 12, but does not bind to the polypeptide represented by any of the IDs. SEQ No. 18085 of EP 1 074 617, ID. SEC. No 211, either WO 00/53756 or WO99 / 46281, ID. SEC. Nos. 24-27, 29, 30, 33, 34, 38 or 39 of WO 01/23523, or ID. SEC. No. 86 of WO 99/11293, or encoded by any of the nucleotide sequences represented by the ID. SEQ No. 18084 or 5096 of EP 1 074 617, ID. SEC. No. 210 of WO 00/53756

or WO 99/46281, or ID. SEC. Nos. 22, 23, 96 or 98 of WO 01/23523 and ID. SEC. No. 31 of WO 99/11293.

In the term "selectively bind" we include antibodies that bind at least ten times more strongly to ECSM4 than to another polypeptide; preferably at least 50 times with more force and, more preferably, at least 100 times more force. Such antibodies can be manufactured by methods well known in the art, using information related to differences in the amino acid sequences of ECSM4 and another polypeptide other than ECSM4.

Antibodies that selectively bind to ECSM4 can also modulate the function of the ECSM4 polypeptide. Antibodies that mimic the effect of true ligand binding by stimulation or activation of ECSM4, or that bind and thereby prevent subsequent binding and activation or stimulation of ECSM4 by the true ligand, and such modulating antibodies. The function are included in the scope of the invention. It will be appreciated that antibodies that modulate function are useful as a research instrument, for example, in the study of the effects of stimulation or activation of ECSM4, or subsequent processes triggered by such stimulation. Such antibodies are useful in medicine, for example, in the modulation of angiogenesis in patients. Specifically, modulation of angiogenesis by the administration of said antibodies may be useful for the treatment of a disease in a patient in which the modulation of angiogenesis is beneficial, such as cancer.

The following peptides may be useful as immunogens in the generation of antibodies, such as rabbit polyclonal serum: LSQSPGAVPQALVAWRA, DSVLTPEEVALCLEL, TYGYISVPTA and KGGVLLCPPRPCLTPT.

In a preferred embodiment, the antibody of the invention selectively binds to an amino acid sequence of the sequence GGDSLLGGRGSL, LLQPPARGHAHDGQALSTDL, EPQDYTEPVE, TAPGGQGAPWAEE or ERATQEPSEHGP. These sequences represent amino acid sequences that are not identical between the human and mouse ECSM4 polypeptide sequences. In general, human and mouse ECSM4 polypeptides show a high degree of identity, which makes the production of mouse antibodies against human ECSM4 particularly difficult, due to the lack of immunogenicity of much of the ECSM4 sequence. Human in mouse. Amino acid sequences that are absent from the mouse ECSM4 are more likely to be immunogenic in a mouse than the sequences that are present in the mouse ECSM4 (an alignment between the amino acid sequences of the human and mouse ECSM4 is found in Figure 14) . Therefore, polypeptide fragments containing sequences unique to human ECSM4 as described above are more useful than ECSM4 polypeptides whose sequences are found in both humans and mice, for the production of antibodies that selectively bind to the ECSM4 polypeptide. human.

Antibodies generated as a result of the use of amino acid sequences that are located in the extracellular portion of the ECSM4 polypeptide are likely to be useful as white molecules of the endothelial tissue. Therefore, it is particularly preferred that the antibody of the invention be raised against and, preferably, selectively binds to, an amino acid sequence that is unique to the human ECSM4 polypeptide, whose sequence is located toward the N-terminal end of the polypeptide. , and is found in the extracellular portion, located between residues 1 and 467 of the amino acid sequence shown in Figure 12.

Although amino acid sequences that are unique to human ECSM4 can be used to produce polyclonal antibodies, it is preferred to use them to produce monoclonal antibodies.

Peptides in which one or more of the amino acid residues are chemically modified, before or after peptide synthesis, can be used, provided that the function of the peptide, that is, the production of specific antibodies in vivo, remains substantially unchanged. . Such modifications include the formation of salts with acids or bases, especially physiologically acceptable organic or inorganic acids or bases, the formation of esters or amides from a terminal carboxyl group, and the binding of protective groups to amino acids, such as Nt- butoxycarbonyl Such modifications can protect the peptide from metabolism in vivo. Peptides may be present as single or multiple copies, for example, tandem repeats. These tandem or multiple repetitions may themselves be antigenic enough to obviate the use of carriers. It may be advantageous for the peptide to be synthesized as a loop, with the N-terminal and C-terminal ends attached, or to add one or more Cys residues to one end, to increase antigenicity, and / or to allow bond formation. disulfide If the peptide is covalently linked to a carrier, preferably, a polypeptide, then the arrangement is preferably such that the peptide of the invention forms a loop.

According to the immune theories in vogue, a carrier function must be present in any immunogenic formulation in order to stimulate, or increase the stimulation of, the immune system. It is thought that the best carrier contains (or forms, together with the antigen) a T-cell epitope. Peptides may be associated, for example, by cross-linking, with a separate carrier, such as serum albumins, myoglobin, bacterial toxoids. and hemocyanin. More recently, developed carriers that induce T-help in the immune response include hepatitis B core antigen (also called nucleocapsid protein), presumable T-cell epitopes, such as Thr-Ala-Ser-Gly-Val -Ala-Glu-Thr-Thr-Asn-Cys, -galactosidase and peptide 163-171 of interleukin-1. The last compound can be seen as a carrier or as an adjuvant, or both. Alternatively, several copies of the same or different peptides of the invention can cross-link, in this situation there is no separate carrier as such, but the carrier function can be provided by said cross-linking. Suitable crosslinking agents include those listed as such in the Sigma and Pierce catalogs, for example, glutaraldehyde, carbodiimide and succinimidyl 4- (N-maleimidomethyl) cyclohexane-1-carboxylate, whose last agent exploits the SH group of the residue of cysteine in the C-terminal (if present).

If the peptide is prepared by the expression of a sequence of a suitable nucleotide in a suitable host, then it may be advantageous to express the peptide as a fusion product with a peptide sequence that acts as a carrier. The Kabigen system "ECOSOC" is an example of such a rearrangement.

Peptides can be synthesized by the Fmoc-polyamide method of solid phase peptide synthesis, as described in Lu et al (1981) J. Org. Chem. 46, 3433, and references of said work. Temporary protection of the N-amino group is achieved by the 9-fluorenylmethyloxycarbonyl group (Fmoc). The repetitive rupture of this highly labile base protecting group is carried out using 20% piperidine in N, N-dimethylformamide. The functions of the side chains can be protected as their butyl ethers (in the case of serine, threonine and tyrosine), butyl esters (in the case of glutamic acid and aspartic acid), derived from butyloxycarbonyl (in the case of lysine and histidine ), trityl derivative (in the case of cysteine) and 4-methoxy-2,3,6-trimethyl-benzenesulfonyl derivative (in the case of arginine). In the case where glutamine or asparagine are C-terminal residues, the group 4,4’-dimethoxybenzidrile is used to protect the functionalities of the amido side chains. The solid phase support is based on a polymer of polydimethyl archilamide, consisting of the three dimethylacrylamide monomers (skeleton monomer), bisacrylethylene diamino (crosslinker) and acryloylsarcosine methyl ester (functionalizing agent). The cleavable bound peptide to resin agent employed is the acid labile 4-hydroxymethyl-phenoxyacetic acid derivative. All amino acid derivatives are added in the form of their preformed symmetric anhydride derivatives, with the exception of asparagine and glutamine, which are added using a reverse coupling procedure mediated by N, N-dicyclohexyl carbodiimide / 1-hydroxybenzotriazole . All coupling and deprotection reactions are monitored using ninhydrin, trinitrobecene sulfonic acid or isotine test procedures. Upon completion of the synthesis, the peptides are separated from the resin support with the concomitant removal of the side chain protecting groups, by treatment with 95% trifluoroacetic acid, containing a 50% cleaning mixture. The cleaners commonly used are ethanedithiol, phenol, anisole and water, of which the specific choice depends on the amino acids that make up the peptide being synthesized. Trifluoroacetic acid is removed by evaporation in vacuo, with subsequent trituration with diethyl ether that yields the crude peptide. The cleaners present are removed by a simple extraction procedure that, by lyophilization of the aqueous phase, yields the crude peptide, free of cleaners. Reagents for peptide synthesis are generally available from Calbiochem-Novabiochem (UK) Ltd, Nottingham NG7 2QJ, UK. Purification can be accomplished by any of, or a combination of, techniques such as size exclusion chromatography, ion exchange chromatography and (especially) reverse phase high performance liquid chromatography. Peptide analysis can be carried out using thin layer chromatography, reverse phase high resolution liquid chromatography, amino acid analysis after acid hydrolysis and by rapid atom bombardment mass spectrometry (FAB) analysis. .

Antibodies in an animal can be activated by immunization with a suitable peptide. Suitable peptides are described herein. Alternatively, with current technology, it is possible to make antibodies as defined herein without the need for animals. Such techniques include, for example, phage antibody sample technology, as is well known in the art. Suitable peptides, as described herein, can be used to select antibodies produced in this manner.

It will be appreciated that, with advances in antibody technologies, it may not be necessary to immunize an animal to produce an antibody. Synthetic systems, such as phage sample libraries, can be used. The use of these systems is included in the methods of the invention and the products of said systems are "antibodies" for the purposes of the invention.

It will be recognized that antibodies that recognize ECSM4 and variants or fragments thereof are useful research reagents and therapeutic agents, in particular when prepared as a compound of the invention as described above. Suitably, the antibodies of the invention are labeled for detection, for example, they can be labeled so that they can be detected directly or indirectly. Conveniently, the antibodies are labeled with a radioactive fraction or a colored fraction, or a fluorescent fraction, or they can bind to an enzyme. Typically, the enzyme is one that can convert a non-colored (or non-fluorescent) substrate into a colored (or fluorescent) product. The antibody may be labeled with biotin (or streptavidin) and then indirectly detected using streptavidin (or biotin) that has been labeled with a radioactive fraction or a colored fraction, or a fluorescent fraction, or the like,

or they may be bound to any enzyme of the type described above. Typically, the fragments have at least 8 consecutive amino acids, preferably at least 10, more preferably at least 12 or 15 or 20 or 30 or 40 or 50 consecutive amino acids of the ECSM4 or ECSM1 polypeptide.

Methods for detecting the presence of peptide fragments derived from major polypeptides are known in the art.

The antibody that selectively binds ECSM4 of the first aspect of the invention can be used in a method of modulating angiogenesis in an individual.

Preferably, the antibody is one that modulates the activity or function, directly or indirectly, of the individual ECSM4 polypeptide.

Preferred antibodies are those described in more detail above.

The production of antibodies that modulate the function of an exposed polypeptide on the cell surface is known in the art and discussed in more detail before. Such antibodies can modulate the function by mimicking the function of the natural ligand and stimulating the polypeptide for the activity or function, or it can modulate the function of the polypeptide by eliminating stimulation of the polypeptide by the ligand by steric obstruction of the site. of ligand binding, thereby preventing the binding of the natural ligand.

The invention will now be described in more detail with reference to the following Examples and Figures

Figure 1.

Experimental verification by PCR with reverse transcription. The candidates for specific endothelial genes predicted by combining the UniGene / EST search and xProfiler SAGE differential analysis (Table 8) were tested for expression in three endothelial and nine non-endothelial cell cultures. The endothelial cultures were the following: HMVEC (human microvascular endothelial cells), confluent culture of HUVEC (human umbilical venous endothelium cells) and proliferative culture of HUVEC. The non-endothelial cultures were the following: normal endometrial stromal cells (NES) grown in normoxia and NES cells grown in hypoxia, MDA 453 and MDA 468 breast carcinoma cell lines, HeLa, normoxia fibroblasts cultured in normoxia and FEK4 fibroblasts cultured in hypoxia, and SW480, HCT116, two cell lines of colorectal epithelium. ECSM1 showed complete endothelial specificity, while the magic roundabout / ECSM4 showed a marked preference for expression in the endothelium. It is interesting that apparently these two new genes have greater endothelial specificity than the specific marker gene of endothelial tissue: von Willebrand factor.

Figure 2

Contig sequence generated with Phrap for ECSM1 and the amino acid sequence of the product of its translation. The ESTs used to generate this contig are shown in Table 10.

Figure 3

In vitro transcription / translation of ECSM4. The cDNA encoding the full-size ECSM4 was cloned into a plasmid vector pBluescript. The circular and digested plasmids by Hindi were subjected to in vitro transcription / translation using a coupled in vitro TNT T7 Quick transcription / translation system (Promega Corporation) incorporating 35S-labeled Methionine according to the manufacturer's instructions. The reaction products were resolved by SDS PAGE and visualized by autoradiography. The luciferase plasmid was used as a positive reaction control. The numbers on the left indicate the position of the molecular weight markers, as a reference. The size of the band denoting ECSM4 is consistent with the calculated molecular weight of the 118 kDa polypeptide.

Figure 4

CDNA and computational translation of GenBank AK000805 (human ECSM4 / magic gazebo).

Figure 5

Contig sequence generated by human ECSM4 Phrap (magic roundabout) ESTs and translation of the encoded polypeptide. The DNA sequence is shown in the orientation in which the cDNA would be, which is opposite to that in which it was originally generated. The ESTs used to generate the contig are shown in Table 11. The beginning of the translation in this sequence is at position 2 of the sequence of the contig, and the end of the translation is at position 948.

Figure 6

An alignment of GenBank Access No. AK000805 ("magic sec.") And Phrap ("hs. 111518") generated the nucleic acid sequences of human ECSM4 presented in Figure 4 and Figure 5.

Figure 7

Nucleotide sequence and amino acid sequence of the mouse ECSM4 contig.

Figure 8

An alignment of the amino acid sequences of the mouse Robo1 protein ("T30805") and the human ECSM4 ("magic pep.").

Figure 9

An alignment of the amino acid sequences of the mouse Robo1 protein ("T30805") and mouse ECSM4 ("magic mouse pep.").

Figure 10

An alignment of the amino acid sequences of the ECSM4 proteins of human ("magic pep.") And mouse ("magic mouse pep."). Residues in bold indicate well preserved sequences. The mouse protein sequence is shown above and the human sequence is shown below.

Figure 11

Expression of the magic gazebo in vitro. (a) Ribonuclease protection analysis. Above, two probes for different regions (nucleotides 1 to 355 and 3333 to 3679) of the magic gazebo were used in the analysis (shown on the left and on the right). The RNAse protection assay was carried out with the small nuclear RNA U6 as a control (shown below) (Maxwell et al (1999) Nature 399: 271). Human cell lines and primary isolates: MRC-5, fibroblast cell line, MCF-7, breast carcinoma cell line. Neuro, SY-SH-5Y neuroblastoma cell line, HUVEC, umbilical venous endothelial isolation, HDMEC, dermal microvascular endothelium isolation and HMME2, mammary microvascular endothelium cell line. N, normoxia, H, hypoxia, P, proliferation, (b) Western analysis of cell lysates. A band at approximately 110 kD corresponds to MR and was stronger in cells exposed to hypoxia for 18 h. The experiment was repeated twice with similar results. Immunostaining was carried out as described in Brown et al (2000) Cancer Res. 60: 6298. Rabbit polyclonal serum was raised against the following peptides bound to hemocyanin: amino acids 165-181 (LSQSPGAVPQALVAWRA) and 274-288 (DSVLTPEEVALCLEL) (anti-serum 1) or peptides 311-320 (TYGYISVPTA) and 336-351 (KGGVLLCPPRPCLTPT) (anti-serum 2). Both anti sera gave identical results. For western analyzes, the anti serum was purified by affinity on a column "HP Hi-Trap activated by NHS" (Amersham) to which peptides used to activate anti serum 1 were attached.

Figure 12

Full length human ECSM4 cDNA and encoded protein sequence.

Figure 13

Full-length mouse ECSM4 cDNA (MuMR sec.) And encoded protein sequence.

Figure 14

Alignment of the amino acid sequences of the human ECSM4 (above) and mouse (below).

Figure 15

Alignment of the human ECSM4 cDNA sequences ("HuMR. Sec"; above) and mouse ECSM4 ("MuMR. Sec"; below).

Figure 16

In situ hybridization analysis of human placental tissue using ECSM4 as a probe. A bright field view with 10x magnification of a thin section of placental tissue. The arrow indicates a large blood vessel.

Figure 17

In situ hybridization analysis of human placental tissue using ECSM4 as a probe. A larger magnification of the bright field view of the thin section of placental tissue shown in Figure 16, centered on the blood vessel. The arrow points to endothelial cells around the lumen of the vessel.

Figure 18

In situ hybridization analysis of human placental tissue using ECSM4 as a probe. A larger magnification of the bright field view of the thin section of placental tissue shown in Figure 16, centered on the blood vessel and shown here in the dark field. The arrow shows positive staining of endothelial cells around the lumen of the vessel.

Figure 19

In situ hybridization analysis of colorectal and liver metastatic tissue, using ECSM4 as a probe. A bright field view of an enlarged section of metastatic colorectal and liver tissue with objectives of (A) 10x and (B) 20x. The area marked by the border (circulating * A) shows normal liver tissue. The arrow in (B) shows one of the blood vessels within the tissue of the metastatic tumor.

Figure 20

In situ hybridization analysis of colorectal and liver metastatic tissue, using ECSM4 as a probe. This is a dark field view of a section of metastatic colorectal and enlarged liver tissue with objectives of (A) 10x and (B) 20x. The area marked by the border (circulating *) shows normal liver tissue. The arrow in (B) shows one of the blood vessels within the tissue of the metastatic tumor corresponding to the vessel shown in Figure 19B. The expression of ECSM4 is restricted to the endothelial cells of the tumor blood vessels. Note that there is little expression in the normal surrounding tissue (*).

Figure 21

Western Blot using the mouse MGO-5 antibody as the primary antibody. Dilutions of the peptides derived from ECSM4, MR 165, MR 311, MR 366 and the Bovine Serum Albumin control polypeptide (BSA) were resolved by polyacrylamide gel electrophoresis with SDS and punctured on an Immobilon P membrane. an MGO-5 antibody probe for gout and was visualized using anti-rabbit antibody bound to alkaline phosphatase.

Figure 22

Immunostaining of a section of frozen placenta. A frozen thin section of human placenta was analyzed by immunohistochemistry without any primary antibody (negative control) and visualized using an anti-rabbit antibody coupled to alkaline phosphatase. Little background staining is observed.

Figure 23

Immunostaining of a section of frozen placenta. A frozen thin section of human placenta was analyzed by immunohistochemistry using a primary antibody that recognizes von Willibrand Factor (positive control), and visualized using a secondary anti-rabbit antibody bound to alkaline phosphatase. The arrows show high levels of vWF expression, restricted to vascular endothelial cells.

Figure 24

Immunostaining of a section of frozen placenta. A frozen thin section of human placenta was analyzed by immunohistochemistry using MGO-5 (a rabbit polyclonal antibody raised against the MR 165 peptide) as the primary antibody, and visualized using a secondary rabbit anti antibody bound to an alkaline phosphatase. The arrows show high levels of ECSM4 expression that are restricted to vascular endothelial cells. Note that the surrounding tissue shows little staining. The comparison with Figures 22 and 23 shows that the expression of ECSM4 is colocalized with that of vWF, a known marker of vascular endothelial cells.

Figure 25

Immunohistochemical analysis of HUVEC cells: von Willibrand factor (vWF). HUVEC cells were immobilized and analyzed by immunohistochemistry using an antibody that recognizes von Willibrand Factor (an endothelial cell marker) as the primary antibody and was visualized using a rabbit anthocyte antibody bound to alkaline phosphatase. The arrows show the expression of vWF in a subset of HUVEC cells.

Figure 26

Immunohistochemical analysis of HUVEC cells using the MGO-7 antibody. HUVEC cells were immobilized and analyzed by immunohistochemistry using the MGO-7 antibody (a polyclonal mouse antibody raised against MR 311 and MR 336 peptides) as the primary antibody, and visualized using an anti-rabbit antibody bound to alkaline phosphatase. The arrows show the expression of ECSM4 in a subset of the HUVEC cells. Note that the staining is located, mainly, on the cell surface.

Figure 27. Expression of the magic gazebo in vivo.

(TO)

MR expression detected by in situ hybridization in a placental arteriole (a) and venule (b) (left, bright field and right, dark field). (c) Immunohistochemical staining of the magic gazebo in a placental arteriole. Left, control of von Willibrand factor and right, magic gazebo. (B) Expression of MR in tumor endothelium. Ganglioglioma (a) x20 and (b) x50. Left, bright field; right, dark field. The arrows highlight a glass that runs diagonally across the section with an erythrocyte inside. Endothelial cells are strongly positive for MR expression. Papillary bladder carcinoma (c) x20 and (d) x50. The vascular nucleus of the papilla of the tumor is strongly positive, in particular the "flat" endothelial cells indicated by the arrows. A magic roundabout antisense probe was generated using T3 polymerase from the IMAGE EST clone 1912098 (GenBank acc. AI278949). The plasmid was linearized with EcoRI prior to probe synthesis. The in situ analysis was then carried out as described in Poulsom et al (1998) Eur.

J.

 Histochemistry 42: 121-132.

Example 1.

In silico cloning of new genes specific to the endothelium

We describe the use of two independent strategies for differential expression analysis combined with experimental verification for the identification of specific genes or that are preferentially expressed in the vascular endothelium.

The first strategy was based on the analysis of the expression in groups of ESTs in the human UniGene gene index (Schuler et al, 1997). Recurrent searches with slits were performed using BLAST (Altschul et al, 1997) with very high rigor against expressed sequence fragments (ESTs) divided into two groups. These two groups included endothelial and non-endothelial cell libraries derived from dbEST (Boguski et al, 1995). The second strategy used a second data mining tool: SAGEmap xProfiler. XProfiler is a free online tool that is part of the NCBI cancer genome anatomy project (CGAP) (Strausberg et al, 1997, Cole et al, 1995). Although these two strategies alone produced a daunting number of false positives, when both were combined, the predictions were exceptionally reliable and two new candidates for specific endothelial genes were identified. Full length cDNAs have been identified in sequence databases. Another gene (EST set) corresponds to a partial cDNA sequence of a large-scale cDNA sequencing project fr and contains a region similar to the intracellular domain of homolog 1 of the human roundabout (ROBO1).

UniGene / EST gene index screening

A group of endothelial sequences and a group of non-endothelial sequences were extracted using the sequence recovery system (SRS) version 5 from dbEST. The endothelial group consisted of 11117 ESTs from nine human endothelial libraries (Table 1). The non-endothelial group included 173137 ESTs from 108 libraries of human cell lines and microdissected tumors (Table 2). The ESTs were extracted from the April 2000 version of dbEST. Multiple FASTA files became a BLAST database that allowed searches using the pressdb program. Table 3 shows the expression status of five specific endothelial cell genes known in these two groups.

Subsequently, the longest representative sequence in each UniGene set (UniGene Build # 111 May 2000, multiple file FASTA hs.sec.uniq) was searched using a BLAST with very high rigor against these two groups. If no matches were found for these representative sequences, the rest of the sequences belonging to the set (UniGene multiple file-FASTA hs.sec) were used as BLAST search sequences. Finally, the sets with no matches in the non-endothelial group and at least one match in the endothelial group were selected.

The optimization of BLAST E value was crucial to the success of BLAST identity level searches. A too high E value would have produced gene paralogs. On the other hand, a very low E value (of high rigor) would have produced many false negatives, for example, true positives would not have been reported due to sequencing errors in the EST data: ESTs are sequences of a single sequencing pass , done on a large scale and low cost and have a high error rate (Aaronson et al, 1996). In this work, an E value of 10e-20 was used in searches against the non-endothelial EST group and an E value of 10e-30 of higher rigor in searches against the smaller endothelial group. These values were considered optimal after a series of test BLAST searches.

Differential analysis of SAGE and SAGEmap xProfiler data

The second data mining strategy that was used for the identification of specific or preferably new endothelial endothelial genes was the subtraction of the web-based SAGE library (SAGEmap xProfiler: http://www.ncbi.nlm.nih.gov/ SAGE / sagexpsetup.cgi). Two endothelial SAGE libraries (SAGE_Duke_HMVEC and SAGE_Duke_HMVEC + VEGF were compared with a total of 110790 sequences) to twenty-four libraries of non-endothelial cell lines, (the complete list in Table 4, a total of 733461 sequences). Table 5 shows the expression status of five known specific endothelial genes: von Willebrand factor (vWF), two endothelial vascular growth factor receptors: tyrosine kinase type fins 1 (flt1) and kinase domain receptor insert (KDR) , tie type tyrosine kinase receptor (TIE1) and tek type tyrosine kinase receptor (TIE2 / TEK), in these two SAGE sets.

Combined data provide highly accurate predictions.

Twenty known genes were selected in the UniGen / EST search (Table 6). These genes showed no matches in the non-endothelial set, and at least one match in the endothelial set. The list contained at least four specific genes of the endothelium: TIE1, TIE2 / TEK, LYVE1, and multimerin, indicating approximately 20% prediction accuracy. Other genes on the list, although they were, with certainty, preferentially expressed in endothelial cells, may not be specific to the endothelium. To increase the accuracy of the prediction we decided to combine UniGen / EST searches with the xPrifiler SAGE analysis. The output of the xProfiler consisted of a list of genes with ten times more number of fragements in the endothelium than in the non-endothelial group, ordered according to the certainty of the prediction. A threshold of 90% certainty was applied to this list. Table 7 shows how the data from the two strategies were combined: Identity level searches in the BLAST were carried out with mRNAs (known genes) or with you calculated using the Phrap (the EST groups that represented new genes) to investigate how these genes were represented in the endothelial and non-endothelial assemblies. The experimental verification that was carried out next, by RT-PCR (Figure 1) showed that the combined strategy was 100% certain, that is, the genes in the xProfiler list that did not have coincidences in the EST set non-endothelial and at least one coincidence in the endothelial group was really specific to the endothelium.

DISCUSSION

There have been several reports of computational analysis of tissue transcriptosomes. Generally, an expression profile is constructed, based on the number of fragments assigned to a given gene or class of genes (Bernstein et al, 1996, Welle et al, 1999, Bortoluzzi et al, 2000). An attempt can be made to identify specific tissue transcripts. For example, Vasmatzis et al, (1997) describes three new genes that are expressed exclusively in the prostate by in silico subtraction of libraries from the dbEST collection. You can also use cDNA libraries made for that purpose. Ten candidates for granulocyte-specific genes have been identified through extensive sequence analysis of cDNA libraries obtained from granulocytes and samples from 11 other tissues, specifically, a hepatocyte cell line, fetal liver, infantile liver, adult liver, subcutaneous fat, visceral fat, lung, colon mucosa, keratinocytes, cornea and retina (Itoh et al, 1998).

An analysis similar to the dbEST-based strategy, followed by Vasmatzis et al, is complicated by the fact that endothelial cells are present in all body tissues and endothelial ESTs contaminate all tissue libraries. To validate this, we use three well-known endothelium specific genes: KDR, FLT1, and TIE-2 as an interrogation for BLAST searches against dbEST. Transcripts were present in a wide range of tissues with multiple coincidences in highly vascularized tissues (e.g., placenta, retina), embryonic tissues (liver, spleen) or infant (brain). In addition, we found that the simple subtraction of endothelial EST libraries against all other dbEST libraries was not able to identify any specific gene (data not shown).

Two very different types of expression data were used in our data mining efforts. The UniGen / EST search was based on libraries of sequence fragments expressed from dbEST. There are 9 libraries of human endothelium in the current version of dbEST, with a relatively small total number of ESTs: 11117. Some specific well-known endothelial genes are not represented in this data set (table 3). This limitation raised our concerns that genes with low levels of expression would be ignored in our analysis. Therefore, we use another type of computable expression data: CGAP SAGE libraries. SAGE fragments are sometimes called small ESTs (usually 10-11 bp in length). Their main advantage is that they can be located within the cDNA unequivocally: they are immediately adjacent to the 3 ’NlaIII restriction site. Although there are only two endothelial CGAP SAGE libraries available so far, they contain an impressive total of ~ 111,000 fragments - a data set approximately 10 times larger than the ~ 11117 sequences in the endothelial EST set. The combined strategy proved to be very accurate (table 8, figure 1) when verified by RT-PCR.

We report here the identification of two new highly specific endothelial genes: endothelial cell specific molecule 1 (ECSM1 - UniGene entry Hs.13957) and magic roundabout (UniGene entry Hs. 111518). For a comprehensive summary of the available data of these genes see table 8.

Our combined data mining strategy in conjunction with experimental verification is a powerful functional genomics tool. This type of analysis can be applied to many cell types, not just endothelial cells. The challenge of identifying the function of the discovered genes remains, but bioinformatics tools such as structural genomics, or homology and motive searches can offer knowledge that can be experimentally verified.

In summary, this screening strategy has allowed the identification of new specific genes of endothelial cells and known genes whose expression was not known to be specific to endothelial cells. This identification allows us to increase our knowledge of the biology of the endothelial cell and also provides new pharmaceutical targets for the imaging, diagnosis and treatment of medical ailments related to the endothelium.

METHODS

PERL programs

A number of programs were generated in PERL to facilitate large-scale sequence retrieval, BLAST search submissions, and automatic BLAST output analysis.

Database Sequence Recovery

UniGene files stored locally (Build # 111, May 2000 version) were used to prepare this report. The UniGene website can be accessed at the URL: www.ncbi.nlm.nih.gov/UniGene/, and you can download the UniGene files from the ftp repository: ftp://ncbi.nlm.nih.gov/repository/unigene /. The representative sequences of the UniGene human subset (the longest EST within the set) are stored in the Hs.sec.uniq file, while all ESTs belonging to the set are stored in a separate file called Hs.sec.

The sequences were extracted from the dbEST database, which was accessed locally at the HGMP center using the getz command of the file recovery system (SRS version 5). This was done repeatedly using a PERL program for all libraries in the endothelial and non-endothelial subgroups, and the sequences were integrated into two multiple FASTA files.

Selection criteria for non-endothelial EST libraries

The selection of 108 non-endothelial dbEST libraries was primarily manual. Initially, we searched the list of all available dbEST libraries (http://www.ncbi.nlm.nih.gov/dbEST/libs_byorg.html) using the keyword "cells" and the phrase "cell line" . Although this search identified the largest number of libraries, it was necessary to add additional keywords for the list to be complete: "melanocyte", "macrophage", "HeLa", "fibroblast". In some cases, the detailed description of the library was consulted to confirm that the library had been obtained from a primary culture or cell line. We also add a number of CGAP microdissected tumor libraries. For this, the library browser (available at http://www.ncbi.nlm.nih.gov/CGAP/hTGI/lbrow/cgaplb.cgi) was used to search for the word “microdissected”.

Research of the UniGen gene index

The UniGen gene transcript index was screened against the GenBank EST division, dbEST. Both UniGen and dbEST were developed by the National Center for Biotechnology Information (NCBI). UniGen is a collection of EST groups that correspond to probable unique genes. It currently consists of four data sets: human, mouse, rat and zebrafish. The human data set consists of approximately 90,000 groups (UniGene Build # 111 May 2000). Through very rigorous identity searches using BLAST, we try to identify UniGen genes that have transcripts in the endothelium and do not have transcripts in the dbEST libraries of non-endothelial cells. In the course of the project the blast2 of the University of Washington was used as the BLAST implementation, which is a version of the BLAST that supports the introduction of slits. The E-value was set at 10e-20 in searches against the non-endothelial EST set and at 10e-30 in searches against the smaller endothelial set.

Although UniGen does not provide consensus sequences of its groups, the longest sequence within the group is identified. Therefore, this longer, representative sequence (multi file FASTA Hs.sec.uniq) was used to search for BLAST using very stringent conditions against endothelial and non-endothelial EST sets. If said representative sequence did not find any match, the rest of the group sequences (multi file FASTA UniGen Hs.sec) was then used for the BLAST search. Finally, the groups without coincidences in the non-endothelial group and with at least one coincidence in the endothelial group were selected using PERL programs that analyzed the text of the BLAST outputs.

Subtraction xProfiler SAGE

xProfiler allows an online user to carry out a differential comparison of any combination of forty-seven serial analyzes of gene expression libraries (SAGE) with a total of approximately 2,300,000 fragments of SAGE using a dedicated statistical algorithm (Chen et al , 1998). XProfiler can be accessed at http://www.ncbi.n1m.nih.gov/SAGE/sagexpsetup.cgi. SAGE is a quantitative expression technology in which genes are usually identified by a 10 or 11 bp sequence fragment adjacent to the NIaIII restriction site closest to the 3 'end of the cDNA (Velculescu et al, 1995 ).

The two available endothelial cell libraries (SAGE_Duke_HMVEC and SAGE_Duke_HMVEC + VEGF) defined set A and 24 non-endothelial libraries (see Table 4 for a list) defined set B. The strategy was verified by verifying the state of expression of the five Endothelium-specific genes that served as a reference in the two SAGE sets (Table 5) using a Gen-to-Fragment mapping (http://www.ncbi.nlm.nih.gov/SAGE/SAGEcid.cgi). Next, the xProfiler was used to select the genes with differential expression between the sets A and B. The output of the xProfiler consisted of a list of genes with a tenfold difference in the number of fragments found in the endothelial set with respect to the non-endothelial set, ordered according to the certainty of the prediction. A 90% certainty threshold was applied to this list.

The other CGAP online differential expression analysis tool, the Digital Differential Sample (DDD) is based on EST expression data (source library information) instead of using SAGE fragments. We tried to use this tool in a manner similar to the SAGE xProfiler mapping, but we have not been able to obtain useful results. Five of the nine endothelial libraries and sixty-four out of one hundred eight non-endothelial libraries used in our BLAST-oriented strategy were available for online analysis using DDD (http://www.ncbi.nlm.nih.gov/CGAP/info/ ddd.cgi). When this analysis was carried out, the fifteen genes with the highest scores were the following: annexin A2, gamma actin 1, large ribosomal protein P0, plasminogen activator type I inhibitor, thymosin beta 4, peptidyloprolyl isomerase A, ribosomal protein L13a, receptor of laminin 1 (ribosomal protein SA), translation elongation factor 1 alpha 1 in eukaryotes, vimentin, heavy ferritin polypeptide, L3 ribosomal protein, S18 ribosomal protein, L19 ribosomal protein, translationally controlled tumor protein1. This list was quite surprising; it did not include any of the well-known endothelium specific genes, did not show any overlap with the results of SAGE (Table 8), and contained many genes that, in the literature are reported as ubiquitous in their expression (ribosomal proteins, actin, vimetin, ferritin). An important advantage of our UniGen / EST search is that, instead of relying on data from source libraries and fall-back EST clustering algorithms, it actually performs identity comparisons through BLAST, in the search for transcripts that apply to a gene

Data mining in UniGen groups

To quickly access information about UniGen entries (eg, references to literature, STS sites, counterparts, function references), online resources are routinely used: the NCBI UniGene and LocusLink interfaces and the Online database Mendelian Inheritance in Man.

ESTs in UniGen groups are not assembled into you, so before sequence analysis they were created with you using the phrap assembler (for documentation about phrap see http://bozeman.mbt.washington.edu/phrap. does / phrap.html).

For the analysis of the genomic contig AC005795 (44 399 bp) containing ECSM1, the NIX Internet interface was used for analysis of multiple applications in large unknown nucleotide sequences. For more information regarding NIX see http://www.hgmp.mrc.ac.uk/NIX/. The alignment of ECSM1 against AC005795 was obtained using the NCBI interface to the Human Genome Interface: the NCBI Map Viewer. For more information on the NCBI Map Viewer see http://www.ncbi.nmi.nih.gov/genome/guide/.

To search for possible transmembrane domains and signal sequences in translated nucleotide sequences, three Internet-based applications were used: DAS http://www.biokemi.su.se/~server/DAS/ (Cserzo et al, 1997), TopPred2 http : //www.biokemi.su.se/-server/toppred2/ (Heijne 1992), and SignalP http://www.cbs.dtu.dk/services/SignalP/ (Nielsen et al, 1997).

PERL programs

A number of PERL programs were generated to facilitate large-scale recovery of sequences, the sending of the BLAST and the automatic analysis of the output of the BLAST.

Experimental verification

To experimentally verify the specificity of the expression we use the reverse transcription coupled to the polymerase chain reaction (RT-PCR). RNA was extracted from three endothelial and seven non-endothelial cell lines, grown in vitro. The endothelial cultures were the following: HMVEC (human microvascular endothelial cells), HUVEC (umbilical venous endothelial cells) confluent culture and proliferative culture of HUVEC. The non-endothelial cultures were the following: normal endometrial stromal cells (NES) grown in normoxia and NES grown in hypoxia, breast cancer cell lines MDA 453 and MDA 468, HeLa, fibroblasts FEK4 grown in normoxia and FEK4 fibroblasts cultured in hypoxia, and SW480, HCT116: two cell lines of colorectal epithelium.

If a fragmented sequence site (STS) was available, dbSTS PCR primers were employed and the PCR cycle conditions suggested at the dbSTS input were followed. Otherwise, primers were designed using the Primer3 program. The primers are listed in Table 9.

Tissue culture media, RNA extraction and cDNA synthesis

Cell lines were cultured in vitro, according to standard tissue culture protocols. In particular, the endothelial cell medium was supplemented with ECGS (endothelial cell growth supplement, Sigma), and heparin (Sigma) to stimulate growth. Total RNA was extracted using the RNeasy Minikit (Qiagen) and the cDNA was synthesized using the Reverse-IT (ABgene) first strand synthesis package.

REFERENCES

Aaronson J.S., B. Eckman, R.A. Blevins, J.A. Borkowski, J. Myerson, S. Imran, and K.O. Elliston 1996. Toward the development of a gene index to the human genome: an assessment of the nature of high-throughput EST sequence data. Genome Res. 6: 829-45.

Adams M.D., A.R. Kerlavage, R.D. Fleischmann, R.A. Fuldner, C.J. Bult, N.H. Lee, E.F. Kirkness, K.G. Weinstock, J.D. Gocayne, O. White, et al, 1995. Initial assessment of human gene diversity and expression patterns based upon 83 million nucleotides of ADNc sequence. Nature 377 (6547 Suppl): 3-174.

Adams R.H., G.A. Wilkinson, C. Weiss, F. Diella, N.W. Gale, U. Deutsch, W. Risau, and R. Klein. 1999. Roles of ephrinB ligands and EphB receptors in cardiovascular development: demarcation of arterial / venous domains, vascular morphogenesis, and sprouting angiogenesis. Genes Dev. 13 (3): 295-306.

Aiello L. P., E.A. Pierce, E.D. Foley, H. Takagi, H. Chen, L. Riddle, N. Ferrara, G.L. King, and L.E.H. Smith 1995. Suppression of retinal neovascularization in vivo by inhibition of vascular endothelial growth factor (VEGF) using soluble VEGF-receptor chimeric proteins. Proc. Natl Acad. Sci. USA. 92: 10457-10461.

Altschul S.F., T.L. Madden, A.A. Schaffer, J. Zhang, Z. Zhang, W. Miller, and D.J. Lipman 1997. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25: 3389-3402.

Banerji S., J. Ni, S.X. Wang, S. Clasper, J. Su, R. Tammi, M. Jones, and D.G. Jackson 1999. LYVE-1, a new homologue of the CD44 glycoprotein, is a lymph-specific receptor for hyaluronan. J Cell Biol. 144 (4): 789-801.

Bashaw G.J., and C.S. Goodman 1999. Chimeric axon guidance receptors: the cytoplasmic domains of slit and netrin receptors specify attraction versus repulsion. Cell 97 (7): 917-26.

Bates E.E., O. Ravel, M.C. Dieu, S. Ho, C. Guret, J.M. Bridon, S. Ait-OYahia, F. Briere, C. Caux, J. Banchereau, and S. Lebecque. 1997. Identification and analysis of a novel member of the ubiquitin family expressed in dendritic cells and mature B cells. Eur J Immunol. 27 (10): 2471-7.

Bernstein S.L., D.E. Borst, M.E. Neuder, and P. Wong. 1996. Characterization of the human fovea cDNA library and regional differential gene expression in the human retina. Genomics 32: 301-308.

Boguski M.S. 1999. Biosequence exegesis. Science 286: 453-5.

Boguski M.S. and G.D. Schuler 1995. Establishing a human transcript map. Nature Genetics: 10, 369-371.

Bortoluzzi S., F. d’Alessi, C. Romualdi, and G.A. Danieli 2000. The human adult skeletal muscle transcriptional profile reconstructed by a novel computational approach. Genome Research 10: 344-349.

Brose K., K.S. Bland, K.H. Wang, D. Arnott, W. Henzel, C.S. Goodman, M. Tessier-Lavigne, and T. Kidd. 1999. Slit proteins bind Robo receptors and have an evolutionarily conserved role in repulsive axon guidance. Cell 96 (6): 795-806.

Chen H., M. Centola, S.F. Altschul, and H. Metzger. 1998. Characterization of gene expression in resting and activated mast cells. J Exp Med. 188 (9): 1657-68.

Clark D.E., S.K. Smith, A.M. Sharkey, and D.S. Charnock-Jones. 1996 Localisation of VEGF and expression of its receptors flt and KDR in human placenta throughout pregnancy. Human Reproduction 11 (5): 1090-1098.

Cole K.A., D.B. Krizman, and M.R. Emmert-OBuck. 1999. The Genetics of Cancer - A 3D Model. Nat Genet 21 (1): 38-41.

Cserzo M., E. Wallin, I. Simon, G. von Heijne, and A. Elofsson. 1997. Prediction of transmembrane alpha-helices in prokaryotic membrane proteins: the Dense Alignment Surface method; Prot. Eng. 6: 673-676.

Dillon N., and P. Sabbattini. 2000. Functional gene expression domains: defining the functional unit of eukaryotic gene regulation. BioEssays 7: 657-665.

Felbor U., A. Gehrig, C.G. Sauer, A. Marquardt, M. Kohler, M. Schmid, and B.H.F. Weber 1998. Genomic organization and chromosomal localization of the interphotoreceptor matrix proteoglycan-1 (IMPG1) gene: a candidate for 6q-linked retinopathies. Cytogenet Cell Genet. 81: 12-17.

Fong G.H., J. Rossant, and M.L. Breitman 1995. Role of the Flt-1 receptor tyrosine kinase in regulating the assembly of vascular endothelium. Nature 376: 65-69.

Gerhold D., and C.T. Caskey 1996. It's the genes! EST access to human genome content. Bioessays 18: 973-81

Ginsburg, R.I. Handin, D.T. Bonthron, T.A. Donlon, G.A. Bruns, S.A. Latt, and S.H. Orkin 1985. Human von Willebrand factor (vWF): isolation of complementary DNA (cDNA) clones and chromosomal localization. Science 228: 1401-6.

Hayward C. P. M., G. E Rivard., W. H. Kane, J. Drouin, S. Zheng, J.C. Moore, and J.G. Kelton 1996. An autosomal dominant, qualitative platelet disorder associated with multimerin deficiency, abnormalities in platelet factor V, thrombospondin, von Willebrand factor, and fibrinogen and an epinephrine aggregation defect. Blood. 87: 4967-4978.

Hayward C.P., D.F. Bainton, J.W. Smith, P. Horsewood, R.H. Stead, T.J. Podor, T.E. Warkentin, and J.G. Kelton 1993. Multimerin is found in the alpha-Ogranules of resting platelets and is synthesized by a megakaryocytic cell line. J Clin Invest. 91 (6): 2630-9.

Hayward C.P., E.M. Cramer, Z. Song, S. Zheng, R. Fung, J.M. Masse, R.H. Stead, and T.J. Podor 1998. Studies of multimerin in human endothelial cells. Blood. 91 (4): 1304-17.

Heijne G. Membrane Protein Structure Prediction, Hydrophobicity Analysis and the Positive-inside Rule. 1992. J. Mol. Biol. 225: 487-494.

Itoh K., K. Okubo, H. Utiyama, T. Hirano, J. Yoshii, and K. Matsubara. (1998). Expression profile of active genes in granulocytes. Blood. 15: 1432-41

Kidd T., K. Brose, K.J. Mitchell, R.D. Fetter, M. Tessier-Lavigne, C.S. Goodman, and G. Tear. 1998. Roundabout controls axon crossing of the CNS midline and defines a novel subfamily of evolutionarily conserved guidance receptors. Cell 92 (2): 205-15.

Matthews W., C.T. Jordan, M. Gavin, N.A. Jenkins, N.G. Copeland, and I.R. Lemischka 1991. A receptor tyrosine kinase cDNA isolated from a population of enriched primitive hematopoietic cells and exhibiting close genetic linkage to c-kit. Proc Natl Acad Sci U S A. 88 (20): 9026-30.

Nichols W.L., D.A. Gastineau, L.A., Solberg, and K.G. Jr Mann. 1985. Identification of human megakaryocyte coagulation factor

V. Blood 65O (6): 1396-406.

Nielsen H., J. Engelbrecht, S. Brunak, and G. Heijne. 1997. Identification of prokaryotic and eukaryotic signal peptides and

prediction of their cleavage sites. Protein Engineering 10: 1-6. Obermair A., A. Obruca, M. Pohl, A. Kaider, A. Vales, S. Leodolter, J. Wojta, and W. Feichtinger. 1999. Vascular endothelial growth factor and its receptors in male fertility. Fert Ster. 72 (2): 269-275.

Partanen J., E. Armstrong, T.P. Makela, J. Korhonen, M. Sandberg, R. Renkonen, S. Knuutila, K. Huebner, K. and Alitalo. 1992. A novel endothelial cell surface receptor tyrosine kinase with extracellular epidermal growth factor homology domains. Mol Cell Biol. 12 (4): 1698-707.

Petrenko O., A. Beavis, M. Klaine, R. Kittappa, I. Godin, I.R. and Lemischka. 1999. The molecular characterization of the fetal

stem cell marker AA4. Immunity 10 (6): 691-700. Sato T. N, Y. Qin, C.A. Kozak, and K.L. Audus 1993. Tie-1 and tie-2 defines another class of putative receptor tyrosine kinase genes expressed in early embryonic vascular system. Proc. Nat. Acad. Sci. 90: 9355-9358.

Sato T. N., Y. Tozawa, U. Deutsch, K. Wolburg-OBuchholz, Y. Fujiwara, M. Gendron-Maguire, T. Gridley, H. Wolburg, W. Risau, and Y. Qin. 1995. Distinct roles of the receptor tyrosine kinases Tie-1 and Tie-2 in blood vessel formation. Nature 376: 70-74.

Schuler G.D. 1997. Pieces of the puzzle: expressed sequence tags and the catalog of human genes. J Mol Med. 75 (10): 694-8. Shalaby F., J. Rossant, T.P. Yamaguchi, M. Gertsenstein, X.F. Wu, M.L. Breitman, and A.C. Schuh 1995. Failure of blood-island formation and vasculogenesis in Flk-1-deficient mice. Nature 376: 62-65.

Shibayama S., J. Hirano, and H. Ono. 1997. cDNA encoding novel polypeptide from human umbilical vein endothelial cell.

European Patent Office Publication number: 0 682 113 A2. Shibuya M., S. Yamaguchi, A. Yamane, T. Ikeda, A. Tojo, H. Matsushime, and M. Sato. 1990. Nucleotide sequence and expression of a novel human receptor-type tyrosine kinase gene (flt) closely related to the fins family. Oncogene 5 (4): 519-24.

Smith T.F., and M.S. Waterman 1981. Identification of common molecular subsences: J Mol Biol. 147: 195-197.

Soker S., S. Takashima, H.Q. Miao, G. Neufeld, and M. Klagsbrun. 1998. Neuropilin-1 is expressed by endothelial and tumor cells as an isoform-specific receptor for vascular endothelial growth factor. Cell 92 (6): 735-45. Sporn L.A., S.I. Chavin, V.J. Marder, and D.D. Wagner 1985. Biosynthesis of von Willebrand protein by human

megakaryocytes. J Clin Invest. 76 (3): 1102-6

Strausberg R.L., C.A. Dahl, and R.D. Klausner 1997. New Opportunities for Uncovering the Molecular Basis of Cancer. Nat Genet 15: 415-6. Suda T., N. Takakura, and Y. Oike. 2000. Hematopoiesis and angiogenesis. Int J Hematol. 71 (2): 99-107 Tamura N., H. Itoh, Y. Ogawa, O. Nakagawa, M. Harada, T.H. Chun, T. Suga, T. Yoshimasa, and K. Nakao, 1996. cDNA

cloning and gene expression of human type I-alpha cGMP-Odependent protein kinase. Hypertension 27: 552-557.

Vasmatzis G., M. Essand, U. Brinkmann, L. Byungkook, and I. Pastan. 1997. Discovery of three genes specifically expressed in human prostate by expressed sequence tag database analysis. Proc. Natl Acad. Sci. USA. 95: 300-304. Velculescu V.E., L. Zhang, B. Vogelstein, and K.W. Kinzler 1995. Serial analysis of gene expression. Science 270: 484-7. Vikkula M., L.M. Boon, K.L. Carraway 3rd, J.T. Calvert, A.J. Diamonti, B. Gounmerov, K.A. Pasyk, D.A. Marchuk, M.L. Warman,

L.C. Cantley, J.B. Mulliken, and B.R. Olsen 1996. Vascular dysmorphogenesis caused by an activating mutation in the receptor tyrosine kinase TIE2. Cell 87 (7): 1181-90.

Walker M.G., and Volkmuth W. 2000. Matrix-remodeling associated genes identified by co-expression. Personal communication

Welle S., K. Bhatt, and C.A. Thornton 1999. Inventory of high-abundance mRNAs in skeletal muscle of normal men. Genome Res. May; 9 (5): 506-13.

Ziegler B. L., M. Valtieri, G.A. Porada, R. De Maria, R. Muller, B. Masella, M. Gabbianelli, I. Casella, E. Pelosi, T. Bock, E.D. Zanjani, and C. Peschle. 1999. KDR receptor: a key marker defining hematopoietic stem cells. Science 285: 1553-1558.

Table 1 Nine dbEST human endothelium libraries

Human aortic endothelium, 20 sequences, in vitro culture

Human endothelial cells, 346 sequences, primary isolation

Human endothelial cell (Y. Mitsui), 3 sequences, in vitro culture

Stratagene 937223 endothelial cell, 7171 sequences, primary isolation

Endothelial cells of the aorta, 1245 sequences, primary isolation

Endothelial cells of the aorta treated with TNF, 1908 sequences, primary isolation

Umbilical vein endothelial cells I, 9 sequences

HDMEC cDNA library, 11 sequences, in vitro culture

Umbilical vein endothelial cells II, 404 sequences

Table 2. Non-endothelial dbEST libraries.

1. Activated T cells I

2. Activated T cells II

3. Activated T cells III

4 activated T cells IV

5. Activated T cells IX

6. Activated T cells V

7. Activated T cells VI

8. Activated T cells VII

9. Activated T cells VIII

10. Activated T cells X

11. Activated T cells XI

12. Activated T cells XII

13. Activated T cells. XX

14. CAMA1Ee I cell line

15. CAMA1Ee II cell line

16. CCRF-CEM cells treated with cyclohexamide I

17. CdnA library of activated cell line B 3D5

18. HeLa chromosome 7 cDNA library

19. Colon carcinoma cell line (Caco-2) I

20. Colon carcinoma cell line (Caco-2) II

21. Colon carcinoma cell line (HCC)

22. Colon carcinoma cell line (HCC) II

23. HCC cell line (mouse liver metastasis)

24. HCC cell line (mouse liver metastasis) II

25. HeLa cDNA (T. Noma)

26. SRIG HeLa (synthetic retinoid-induced genes)

27. Monocyte-derived macrophages from Homo sapiens

28. HSC172 I cells

29. HSC172 II cells

30. Human gastric carcinoma cell line 23132

31. Bcap37 human breast cancer cell line

32. A431 human cell line subclone

33. AGZY-83a human cell line

34. Human cell line PCI-O6A

35. Human cell line PCI-O6B

36. Human cell line SK-N-MC

37. Human cell line TF-1 (D.L.Ma)

38. Human exocervical cells (CGLee)

39. HT1080 human fibrosarcoma cell line

40. HT1080-6TGc5 Human Fibrosarcoma Cell Line

41. SGC-7901 Human Gastric Cancer Cell Line

42. GM-CSF private human TF-1 cell line (Liu, Hongtao)

43. Human HeLa (Y. Wang)

44. Human HeLa cells (M.Lovett)

45. Jurkat human cell line mRNA (Thiele, K.)

46. Human erythroleukemic cells K5 62

47. A549.A549 human lung cancer cell line

48. HNE1 human nasopharyngeal carcinoma cell line

49. Human SK-ER3 neuroblastoma cells (M.Garnier)

50. Newborn human melanocytes (T. Vogt)

51. Patu 8988t human pancreas cancer cell line

52. mRNA of human primary melanocytes (I.M. Eisenbarth)

53. Human promyelocytic cell line HL60 (S. Herblot)

54. ARPE-19 human retina cell line

55. HSG human salivary gland cell line

56. Human leukocytes

57. Jurkat T I cells

58. Jurkat T II T cells

59. Jurkat T III T cells

60. Jurkat T V cells

61. Jurkat T VI cells

62. HepG2 liver cell line

63. LNCAP I cells

64. Macrophage I

65. Macrophage II

66. Macrophage, subtracted (total CdNA)

67. MCF7 cell line

68. Namalwa B cells I

69. Namalwa II B cells

70. NCI_CGAP_Br4

71. NCI_CGAP_Br5

72. NCI_CGAP_CLL1

73. NCI_CGAP_GCB0

74. NCI_CGAP_GCB1

75. NCI_CGAP_HN1

76. NCI_CGAP_HN3

77. NCI_CGAP_HN4

78. NCI_CGAP_HSC1

79 NCI_CGAP_Li1

80. NCI_CGA.P_Li2

81. NCI_CGAP_Ov5

82. NCI_CGAP_Ov6

83. NCI_CGAP_Pr1

84. NCI_CGAP_Pr10

85. NCI_CGAP_Pr11 1

86. NCI_CG.AP_Pr16

87. NCI_CGAP_Pr18

88. NCI_CGAP_Pr2

89. NCI_CGAP_Pr20

90. NCI_CGAP_Pr24

91. NCI_CGAP_Pr25

92. NCI_CGAP_Pr3

93. NCI_CGAP_Pr4

94. NCI_CGAP_Pr4.1

95. NCI_CGAP_Pr5

96. NCI_CGAP_Pr6

97. NCI_CGAP_Pr7

98. NCI_CGAP_Pr8

99. NCI_CGAP_Pr9

100. Normal human trabecular bone cells

101. Raji cells treated with cyclohexamide I

102. 0041 retinal pigmented epithelium cell line

103. HeLa cells treated with retinoid

104. Melanocito Soares 2NbHM

105. Soares Nb HSF senescent fibroblasts

106. HeLa Stratagene s3 937216 cells

107. Supt cells

108. Rhabdomyosarcoma T cell line, human adult

Table 3. Five genes that are known to be endothelial genes specific in the dbEST group.

The number of ESTs in the endothelial group is relatively small (~ 11,117) and not all known endothelial genes are represented.

Known specific endothelial gene

Endothelial matches the group no Coincidences in the endothelial group

von Willebrand factor (vWF)

one 27

flt1 VEGF receiver

- -

KDR VEGF receiver

one -

tyrosine kinase TIE1

- 5

tyrosine kinase TIE2 / TEK

- 2

Table 4 Twenty-four SAGE-CGAP libraries of non-endothelial cells.

SYMBOL

 DESCRIPTION

SAGE_HCT116

Colon, cell line derived from colorectal carcinoma

SAGE_Caco_2

Colon, colorectal carcinoma cell line

SAGE_Duke_H392

Brain, Duke glioblastoma multiforme cell line

SAGE_SW837

Colon, cancer cell line

SAGE_RKO

Colon, cancer cell line

SAGE_NHA (5th)

Brain, normal human astrocyte cells cultured in pass 5

SAGE_ES2-1

Clearly differentiated ovarian carcinoma ES-2 cell line Clear

SAGE_OVCA432-2

Ovary, OVCA432 carcinoma cell line

SAGE_OV1063-3

Ovary, OV1063 carcinoma cell line

SAGE_Duke_mhh-1

Brain, negative medulloblastoma cell line for c-myc mhh-1

SAGE_Duke_H341

Brain, medulloblastoma cell line positive for c-myc, H341

SAGE_HOSE_4

Ovary, normal surface epithelium

SAGE_OVP-5

Ovary, sampled cancer cell lines

SAGE_LNCaP

Prostate, cell line. Dependent of

Androgens

SAGE_HMEC-B41

HMEC-B41 normal human mammary epithelial cell cell culture

SAGE_MDA453

MDA-MB-453 human breast carcinoma cell line

SAGE_SKBR3

ATCC SK-BR-3 cell line. Human mammary adenocarcinoma

SAGE_A27 80-9

Ovary, A2780 ovarian cancer cell line

SAGE_Duke_H247_normal

Brain, glioblastoma multiforme H247 cell line

AGE_Duke_H247_Hipoxia

Brain, Duke glioblastoma multiforme cell line, H247, grown under conditions of 1.5% oxygen

SAGE_Duke_post_crisis_fibroblastos

Skin, postcrisis surviving fibroblast cell line

SAGE_Duke_premsis_ fibroblasts

Skin, clones of human fibroblasts transformed with T antigens

SAGE_A

Prostate, cancer cell line. Induced with synthetic androgen

SAGE_IOSE29-11

Ovary, surface epithelial cell line

Table 5. Five specific endothelial genes known in CGAP SAGE groups.

TIE1 and TIE2 / TEK have multiple coincidences in the non-endothelial group (mostly in normal or ovarian carcinoma cell lines). vWF is the most specific endothelial with 80 matches in the endothelial group and only one match in the non-endothelial group.

Known specific endothelial gene

Fragment in the non-endothelial SAGE bookshop Fragment in the endothelial SAGE bookshop

von Willebrand factor (vWF)

1 (carcinoma cell line) 80

flt1 VEGF receiver

- -

KDR VEGF receiver

1 (IOSE29 ovarian surface epithelial cell line) 6

tyrosine kinase TIE1

17 (cell lines of ovarian tumors and normal ovarian epithelia) 27

tyrosine kinase TIE2 / TEK

4 (ovarian carcinoma and glioblastoma multiforme cell lines) 2

Table 6.

Results of the UniGene / EST research. Twenty known genes were selected in the UniGene / EST screening (no matches in the non-endothelial group and at least one match in the endothelial group). At least four of these genes are known to be specific endothelial genes: TIE1, TIE2 / TEK, LYVE1 and multimerin, indicating ~ 20% prediction accuracy. Other genes, although they are certainly expressed preferentially in endothelial cells, may not be specific endothelial genes.

Description

UniGene ID Endothelio matches

TIE1 endothelial tyrosine kinase receptor

Hs. 78824  5

Cytosolic phospholipase A2; Related to eicosanoid metabolism

HS 2111587 3

Branched Chain Alpha Keto Acid Dehydrogenase

Hs.1265 2

CGMP-dependent protein kinase, cloned from cDNA aortic cDNA, highly expressed in highly vascularized tissues such as aorta, heart, and uterus (Tamura et al, 1996)

Hs. 2,689 2

Hyaluronan Lymphatic Endothelium Hyaluronan Receptor 1 - LYVE1

HS177917  2

TRAP interaction protein: TNF signaling pathway

Hs. 2154  2

Multimerine: a large specific protein of the endothelium; binds to platelet factor V, is also found in platelets (Hayward et al, 1996)

HS.32934  2

Diubiquitin (a member of the ubicuitin family); reported in dendritic cells and B lymphocytes; Involved in antigen processing; This is the first evidence that it is also present in endothelial cells (Bates et al, 1997)

Hs.44532  2

Protein of the beta-transducin family; also homologous to the notchless gene of D. melanogaster: a new protein that contains WD40 type reflections, which modulates Notch signaling activity

Hs. 85570  2

Endothelial receptor tyrosine kinase with Notch signaling activity

Hs. 89640  2

Protein X associated with BCL2 (BAX)

Hs. 159428 2

Sepiapterin reductase mRNA

Hs. 160100 2

Retinoinco acid beta receptor (RARB)

Hs. 171495 2

ST2 receptor: an interleukin receptor homolog 1

Hs.66 one

Mitogen activated protein kinase 8 (MAPK8)

Hs. 859 one

ERG gene related to the ETS oncogene

Hs. 45514  one

PP35 similar to yhdg of E. coli and nifR3 of R. capsulatus

Hs. 97627  one

Matrix receptor proteoglycan; strongly expressed in retina and umbilical cord vein (Felbor et al, 1998)

H.129882 one

Semi-dehydrogenase methylmalonate gene

Hs. 170608 one

HTLV-1 related retroviral sequence

Hs.247963 one

Table 7.

The differential analysis of xProfiler was combined with data obtained from the UniGen / EST survey, achieving 100% prediction certainty. The output of xProfiler lists genes with 10 times more fragments in the endothelial set than in the non-endothelial set of the SAGE-CGAP libraries. The matches that correspond to these genes in the endothelial and non-endothelial EST sets were identified by identity level searches, using BLAST to find mRNA (known genes) or contig sequences assembled by phrap (EST groups representing new genes) . The genes are sorted according to the number of matches in the non-endothelial EST set. New known and predicted specific endothelial genes are in bold

UniGen ID

Gene description XProfiler prediction certainty Coincidences in the set of endothelial ESTs Coincidences in the set of non-endothelial ESTs

Hs.13957

 ESTs-ECSMI 97% 4 0

Hs. 111518

Magic gazebo, distant homology with human gazebo 1 100% 4 0

Hs. 268107

multimerine 92% 5 0

Hs. 155106

Protein 2 modifying the activity of the receptor type receptor 97% 0 0

calcitonin

HS 233355

ESTs 96% 0 0

Hs. 26530

 Response to serum impoverishment (phosphatidylserine binding protein) 94% 3 one

Hs. 83213

Fatty acid binding protein 4 100% 3 one

Hs. 110802

Von Willebrand factor 100% 25 one

Hs. 76206

 Caderina 5, VE-caderina (vascular endothelium) 100% 4 one

Hs.2271

Endothelin 1 98% 9 2

Hs. 119129

Collagen, type IV, alpha 1 100% 4 6

Hs. 78146

 Platelet cell adhesion molecule / endothelial cells (CD31 antigen) 99% 18 5

Hs. 76224

Fibulin-type extracellular matrix protein 1, containing EGF 100% 37 9

Hs. 75511

 Connective tissue growth factor 100% 3. 4 48

Table 8. Summary of the information available on the magic roundabout Table 9.

ID of

CDNA of ORF of Segments of Information of Description

group

size higher transmembrane mapping

from

full size a peptide

UniGen

signal Context

Y

genomic

size

Genomic clones

ECS

Hs. 1395 103 aa Neighbour

M1

7 confirmed genomic: tropomyosin

1100 bp

with 5 ’RACE dbSTS G26129 and G28043 Chromosome 19 Gene Map 98: Marker SGC33470.

Marker stSG3414, Interval D19S425-D19S4 18 AC005945, AC005795 (partial identity)

Glori

Hs. Partial cDNA 417 aa A domain of Neighbour Homology Region of

eta

111518 FLJ20798 transmembrane genomic: 486aa with portion

magic

fis clone predicted by protein 1 cytoplasmic of the

AC

TopPred2 and integral of roundabout,

YOU GIVE.

transmembrane

2067 bp

ADSU02031 (acc. No signal peptide (ITM1) dbSTS G14646 and of the family of the

AK000805)

detected in the G14937 guide protein

ORF

axonic: human

available from

ROBO1, rat ROBO1 and

417 aa

dutt1 mouse (E = 1,3e-09)

(PS)

1496 bp

however, the true protein product is very likely to be higher Chromosome 11, Gene Map 98: Marker SHGC-11739, Interval D11S1353-D11S 93 The ORF has no apparent upstream limit (upward direction). This, together with the size comparison with ROBO1 (1651 aa)

suggests that the

true protein has

high probability of being

much older

Possible polyA sites

alternatives: the clone of

CDNA from

adipose tissue seems

be polyadenylated in

a different position than

sequence

coming from the contig

from UniGen

List of primers used in RT-PCR reactions. DbSTS primers were used if a UniGen entry contained a fragmented sequence site (STS). Otherwise, primers were designed using the Primer3 program.

Gen

Primers (GenBank Sequence or Access for the STS)

ECSM1 Hs.13957

G26129

Magic gazebo Hs. 111518

G14937

Calcitonin receptor type receptor modifying activity 2

G26129

HS 233355

G21261

Fatty acid binding protein 4

5’-TGC AGC TTC CTT CTC ACC TT-3 ’5’-TCA CAT CCC CAT TCA CAC TG-3’

Von Willebrand factor

5’-TGT ACC ATG AGG TTC TCA ATG C-3 ’5’-TTA TTG TGG GCT CAG AAG GG-3’

Whey impoverishment response protein

G21528

Type IV collagen, alpha 1

G07125

Fibulin-type extracellular matrix protein 1, containing EGF

G06992

Connective tissue growth factor

5’-CAA ATG CTT CCA GGT GAA AAA-3 ’5’-CGT TCA AAG CAT GAA ATG GA-3’

Table 10

The ESTs that belong to the ECSM1 contig sequence are the following:

Sequences of EST (30) AI540508, IMAGE cDNA clone: 2209821, Uterus, 3 'reading, 2.1kb AI870175, IMAGE cDNA clone: 2.424998, Uterus, 3' reading, 1.7kb AI978643, IMAGE cDNA clone: 2491824 , Uterus, 3 'reading, 1.3kb AI473856, IMAGE cDNA clone: 2044374, Lymph node, reading 3' AI037900, IMAGE cDNA clone: 1657707, Embryo, 3 'reading, 1.2kb AI417620, IMAGE cDNA clone: 2115082, 3 'reading, 1.0kb AA147817, IMAGE cDNA clone: 590062, reading 3' AA968592, IMAGE cDNA clone: 1578323, 3 'reading, 0.7kb AW474729, IMAGE cDNA clone: 2853635, Uterus, reading 3 'R02352, IMAGE cDNA clone: 124282, reading 3', 0.7kb R01889, IMAGE cDNA clone: 124485, 5 'reading, 0.7kb AA446606, IMAGE cDNA clone: 783693, Embryo, reading 3' R02456, clone IMAGE cDNA: 124282, 5 'reading, 0.7kb T72705, IMAGE cDNA clone: 108686, 5' reading, 0.7kb R01890, IMAGE cDNA clone: 124485, reading 3 ', 0.7kb AA147925, cDNA clone IMAGE: 590014, reading 5 'AI131471; IMAGE cDNA clone: 1709098, Heart, reading 3 ', 0.6kb AA733177, cDNA clone 399421, Heart, reading 3' AI039489, IMAGE cDNA clone: 1658903, Embryo, reading 3 ', 0.6kb AI128585, cDNA clone IMAGE: 1691245, Heart, 3 'reading, 0.6kb AI540506, cDNA clone IMAGE: 2209817, Uterus, reading 3', 0.6kb AA894832, cDNAcloneIlVIAGE: 1502815, Kidney, reading 3 ', 0.5kb AW057578, clone IMAGE cDNA: 2553014, Sampled, 3 'reading, 0.3kb AA729975, IMAGE cDNA clone: 1257976, Germ cell, 0.3kb AI131016, IMAGE cDNA clone: 1706622, Heart, 3' reading, 0.2kb AA147965, clone IMAGE cDNA: 590062, reading 5 'AA446735, cDNA clone IMAGE: 783693, Embryo, reading 5' AA147867, cDNA clone IMAGE: 590014, reading 3 'AI497866, cDNA clone IMAGE: 2125892, Sampled, reading 3' T72636 , IMAGE cDNA clone: 108686, 3 'reading, 0.7kb

Table 11

ESTs within the magic roundabout sequence:

EST sequences in the magic roundabout (55): AI803963, IMAGE cDNA clone: 2069520, 3 'reading, 0.9kb W88669, IMAGE cDNA clone: 417844, 3' reading, 0.7kb AI184863, IMAGE cDNA clone: 1565500, Sampled, 3 'reading, 0.6kb AA011319, cDNA clone IMAGE: 359779, Heart, reading 3', 0.6kb AA302765, cDNAcloneATCC: 194652, Adipose tissue, reading 3 'AI278949, cDNA clone IMAGE: 1912098, Colon, reading 3 ', 0.7kb AI265775, cDNA clone IMAGE: 2006542, Ovary, reading 3' AA746200, cDNA clone IMAGE: 1324396, Kidney, 0.5kb N78762, cDNA clone IMAGE: 301290, Lung, read 3 'AI352263, cDNA clone IMAGE: 1940638, Embryo, 3 'reading, 0.6kb AA630260, cDNA clone IMAGE: 854855, Lung, 3' reading, 0.5kb C20950, cDNA clone (no name), reading 3 'W88875, cDNA clone IMAGE: 417844 , reading 5 ', 0.7kb AA156022, cDNA clone IMAGE: 590120, reading 3' N93972, cDNA clone: 309369, Lung, reading 3 ', 1.7kb AI217602, cDNA clone IMAGE: 1732380, Heart, reading 3' , 0.5kb

AW294276, IMAGE cDNA clone: 2726'347, reading 3 'AA010931, IMAGE cDNA clone: 359779, Heart, reading 5', 0.6kb AA3 03 624, cDNAcloneATCC: 115215, Aorta, reading 5 'AI366745, IMAGE cDNA clone : 1935056, reading 3 ', 0.5kb AA327257, cDNAcloneATCC: 127927, Colon, reading 5' C06489, cDNAclonehbc5849, Pancreas BE218677, cDNA clone IMAGE: 3176164, Lung, reading 3 'AA335675, cDNAcloneATCC: 137 reading, Testicles: 137 R84975, IMAGE cDNA clone: 180552, Brain, 3 'reading, 2.1kb AI926445, IMAGE cDNA clone: 2459442, Stomach, 3' reading, 1.9kb H61208, IMAGE cDNA clone: 236318, Ovary, reading 3 ', 1.9 kb AA335358, ATCC cDNA clone: 137019, Testis, reading 5 'AI129190, IMAGE cDNA clone: 1509564, Sampled, 3' reading, 0.8kb T59188, IMAGE cDNA clone: 74634, Spleen, 5 'reading, 0.8kb T59150 , IMAGE cDNA clone: 74634, Spleen, 3 'reading, 0.8kb R53174, IMAGE cDNA clone: 154350, Mama, 5' reading, 0.8kb AA156150, IMAGE cDNA clone: 590120, 5 'AA reading 302509, cDNA cloneATCC: 114727, Aorta, reading 5 'R99429, cDNA clone IMAGE: 201985, read 5', 2.4kb AI813787, cDNA clone IMAGE: 2421627, Pancreas, read 3 ', 1.2kb H62113, cDNA clone IMAGE: 236316, Ovary, 5 'reading, 1.0kb R16422, cDNA clone IMAGE: 129313, 5' reading, 0.7kb T48993 cDNA clone IMAGE: 70531, Placenta, reading 5 ', 0.6kb T05694, cDNA clone HFBDF13, Brain R84531, cDNA clone IMAGE: 180104, Brain, 5 'reading, 2.2kb AI903080, cDNA clone (no name), breast AI903083, cDNA clone (no name), breast AA302764, cDNA clone cATCC: 194652, Adipose tissue, reading 5 'AA341407, cDNA clone ATCC: 143064, Kidney, reading 5' W16503, cDNA clone IMAGE: 301194, Lung, reading 5 'AW801246, cDNA clone (unnamed), uterus AW959183, cDNA clone (unnamed) R85924, IMAGE cDNA clone: 180104, Brain, 3 'reading, 2.2kb

AA358843, cDNA clone cATCC: 162953, Lung, reading 5 ’ BE161769, cDNA clone (no name), head-neck W40341, IMAGE cDNA clone: 309369, Lung, 5 ’reading, 1.7kb AA876225, IMAGE cDNA clone: 1257188, Germ Cell, reading 3 ’ R99441, IMAGE cDNA clone: 202009, 5 ’reading, 2.3kb W76132, IMAGE cDNA clone: 344982, Heart, 5 ’reading, 1.4kb,

Table 12 110 ESTs in the magic mouse roundabout group (Mm. 2782)

AI427548, IMAGE cDNA clone: 521115, Muscle, reading 3 ’ AV022394, cDNA clone1190026N09, reading 3 ’ BB219221, cDNA clone530053H04, reading 3 ’ AI604803, IMAGE cDNA clone: 388336, Embryo, reading 3 ’ AI504730, IMAGE cDNA clone: 964027, Mammary Gland, reading 3 ’ AI430395, IMAGE cDNA clone: 388336, Embryo, reading 5 ’ AI181963, cDNA clone IMAGE: 1451626, Liver, reading 3 ’ AV020471, cDNA clone1190017N14, reading 3 ’ BB219225, cDNA clone530053H12, reading 3 ’ BB224304, cDNA clone530086A21, reading 3 ’ BB527740, cDNA clone 930042M18, reading 3 ’ W66614, IMAGE cDNA clone: 388336, Embryo, reading 5 ’ BB097630, cDNA clone 9430060E21, reading 3 ’ AI152731, cDNA clone IMAGE: 1478154, Uterus, reading 5 ’ AW742708, IMAGE cDNA clone: 2780289, inner ear, 170 sampled, 3 ’reading BB118169, cDNA clone9530064M17, reading 3 ’ AI839154, cDNA clone-M-AO0-ach-e-11-0-UI, reading 3 ’ BB206388, cDNA clone 430075J10, reading 3 ’ BB381670, cDNA clone C230015E01, reading 3 ’ BB 199721, cDNA clone 430017A19, reading 3 ’ AI593217, IMAGE cDNA clone: 1177959; Mammary gland, reading 3 ’ BB219411, cDNA clone530054L01, reading 3 ’ BB220744, cDNA clone530061M19, reading 3 ’ BB220944, cDNA clone530062022, reading 3 ’

BB390078, cDNA clone C230066L23, reading 3 ’ BB220730, cDNA clone530061L13, reading 3 ’ AI615527, IMAGE cDNA clone: 964027, Mammary gland, reading 5 ’ A1882477, IMAGE cDNA clone: 1396822, Mammary gland, reading 5 ’ AV025281, cDNA clone1200012D01, reading 3 ’ BB470462, cDNA clone CD230033L23, reading 3 ’ BB247620, cDNA clone 730020G03, reading 3 ’ BB555377, cDNA clone330019B13, reading 3 ’ BB512960, cDNA clone D730043I21 BB400157, cDNA clone C330017F17, reading 3 ’ BB320465, cDNA clone CB230385O10, reading 3 ’ BB105670, cDNA clone 9430096H10, reading 3 ’ BB441462, cDNA clone CD030027B11, reading 3 ’ BB137530, cDNA clone 9830142007, reading 3 ’ AA553155, IMAGE cDNA clone: 964027, Mammary gland, reading 5 ’ BB319763, cDNA clone CB230382G07, reading 3 ’ BB451051, cDNA clone CD130007I05, reading 3 ’ BB504672, cDNA clone636349J11, reading 3 ’ AI429453, IMAGE cDNA clone: 569122, Embryo, reading 3 ’ BB190585, cDNA clone330062J23, reading 3 ’ BB257082, cDNA clone 730076M18, reading 3 ’ BB386699, cDNA clone C230047P06, reading 3 ’ BB295814, cDNA clone CB130042A09, reading 3 ’ BB450972, cDNA clone D130007A22, reading 3 ’ AA718562, IMAGE cDNA clone: 1177959, Mammary gland, reading 5 ’ BB223775, cDNA clone530083K18, reading 3 ’ AV020555, cDNA clone1190018G05, reading 3 ’ BB226083, cDNA clone530095K11, reading 3 ’ BB482105, cDNA clone D430007019, reading 3 ’ BB381671, cDNA clone C230015E02, reading 3 ’ BB383758, cDNA clone C230030C02, reading 3 ’ BB257519, cDNA clone 730080D13, reading 3 ’ BB265667, cDNA clone830021117, reading 3 ’

BB254777, cDNA clone 730063K20, reading 3 ’ AV240775, cDNA clone 4732443F15, reading 3 ’ BB315010, cDNA clone CB230352H04, reading 3 ’ BB390074, cDNA clone C230066L16, reading 3 ’ BB517605, cDNA clone cD830025B17, reading 3 ’ BB484410, cDNA clone D430025H01, reading 3 ’ BB357583, cDNA clone C030022J01, reading 3 ’ AV225639, cDNA clone 3830431D12, reading 3 ’ BB554921, cDNA clone330016A12, reading 3 ’ BB161650, cDNA clone130061H21, reading 3 ’ BB106720, cDNA clone 9530002M22, reading 3 ’ BB535465, cDNA clone 030043P14, reading 3 ’ BB357738, cDNA clone C030024B10, reading 3 ’ AV285588, cDNA clone 5031411M12 BB188339, cDNA clone330048H22, reading 3 ’ AV337749, cDNA clone 6430404F19, reading 3 ’ BB065281, cDNA clone 8030443H10, reading 3 ’ BB148059, cDNA clone9930104N19, reading 3 ’ AV252251, cDNA clone 4833438P20, reading 3 ’ BB184506, cDNA clone330012J24, reading 3 ’ BB522445, cDNA clone 930007M08, reading 3 ’ BB520366, cDNA clone cD830041K23, reading 3 ’ AV127290, cDNA clone 2700047J01, reading 3 ’ BB248651, cDNA clone 730027F04, reading 3 ’ BB008452, cDNA clone 4732482M24, reading 3 ’ BB550719, cDNA clone 230024C07, reading 3 ’ BB182033, cDNA clone 230095N14, reading 3 ’ BB480258, cDNA clone33334SD17, read 3 ’ BB004855, cDNA clone 4732463E03, reading 3 ’ AV379748, cDNA clone 9230013A19, reading 3 ’ BB552137, cDNA clone 230035B12, reading 3 ’ BB288263, IMAGE cDNA clone: 3490042, Mama; 5 ’reading BB215681, cDNA clone530026M11, reading 3 ’

BB251356, cDNA clone 730046B16, reading 3 'BB503441, cDNA clone D630043F10, read 3' BB500571, cDNA clone D630029E03, read 3 'BB199833, cDNA clone A430017K13, read 3' BB533549, cDNA clone, reading 3'03303003 , reading 3 'BB213310, cDNA clone530009E09, reading 3' BB240699, cDNA clone A630083B 14, reading 3 'BB217106, cDNA clone A530040N24, reading 3' BB057432, cDNA clone 7120459H22, reading 3 'BB214645, reading clone' BB218254, cDNAcloncA530048K12, reading 3 'BB319841, cDNA cloneB230382006, reading 3' BB459759, cDNA cloneD130063G22, reading 3 'BB485618, cDNA cloneD430032M09, reading 3' BB517699, cDNA clone, cd9509 reading 3 'BBS36291, cDNA clone 030049D17, reading 3' BB552689, cDNA clone33331A16, reading 3 'BB552709, cDNA clone3333001C16, reading 3'

Example 2

The expression of ECSM4 is restricted to endothelial cells.

In situ hybridization (ISH) of normal and tumor tissues showed that ECSM4 expression is restricted to cells

5 of the vascular endothelium in angiogenic vessels of adults only. Normal tissue analysis showed that ECSM4 expression is detected in human placenta and fetal umbilical cord tissue of 10.8 weeks of menstrual age. As shown in Figure 16, the expression of ECSM4 is highly specific for vascular endothelial cells of placental blood vessels. In addition, expression was absent through a large number of other normal tissues that were analyzed, including the adult liver, brain and large vessels, prostate, colon, intestine.

10 thin, the heart, the eye (choroid and sclera), the ovary, the stomach, the breasts, and the fetal bladder, the testicles, the kidneys (15.8 weeks) and the fetal heart, the kidneys, the adrenal gland , the intestine (11.3 weeks), fetal brain (10.6 weeks) and fetal eye (16.5 weeks) (data not shown).

ISH analysis of liver and colorectal tissue metastatic biopsies showed that ECSM4 expression was restricted to vascular endothelial cells of tumor vessels only (Figure 17 and 18). Not detected

15 expression in normal surrounding tissue. In addition, the increased expression in the vicinity of necrotic tissues (Figure 18, the necrotic tissue is indicated by the bright signal marked *) is indicative and consistent with the induction of ECSM4 expression by hypoxia. As such, ECSM4 may be a new hypoxia regulated gene.

The highly restricted pattern of ECSM4 expression in angiogenic vessels in normal and tumor tissues in adults is completely consistent with the pattern of selective expression to endothelial cells through in-house analysis.

Silico described in Example 1.

Methods

Blocks of tumors and tissues embedded in paraffin and fixed with formalin were obtained from the archives of the group of Breast Pathologies of the Imperial Cancer Research Fund of the Guys Hospital, London, United Kingdom. An antisense ribosonde for the ECSM4 cDNA was prepared for specific location of the ECSM4 mRNA by in situ hybridization. The methods for pretreatment, hybridization, washing and rinsing of the sheets in Ilford K5 for autoradiography have been previously described (Poulsom, R., Longcroft, JM, Jeffrey, RE, Rogers, L., and Steel, JH (1998) Eur J. Histochem. 42, 121-132). The films were exhibited 7 to 15 days before the development in Kodak D19 and the opposite staining with Giemsa. Sections were examined under the dark field conditions of reflected or conventional light (Olympus BH2 with epi-lighting) under an objective of x5, x10 or x20 that allowed to see the grains of individual autoradiographic plates as bright objects on a dark background.

Example 3

The ECSM4 polypeptide is detected only in endothelial cells.

Antibodies capable of selectively binding the ECSM4 polypeptide were generated and used in immunohistochemistry to demonstrate the presence of the ECSM4 polypeptide in a group of cell types (Figures 21 through 26). Tissue samples were prepared through standard techniques in immunohistochemical techniques.

Generation of antibodies that recognize ECSM4.

Peptides MR 165, MR 311 and MR 336 were fused to Keyhole Limpet hemocyanin (KLH) before rabbit immunization for polyclonal antibody production. The MGO-5 antibody was derived from rabbits immunized with the MR 165 peptide, while the MGO-7 was derived from rabbits immunized with a mixture of MR 311 and MR 336. The sequence of the peptides used to generate the polyclonal antibodies is shown. below, with the reference within the amino acid sequence of the full-length human ECSM4, as shown in Figure 12.

MR 165 = LSQSPGAVPQALVAWRA (681-697)

MR 274 = DSVLTPEEVALCLEL (790-804)

MR 311 = TYGYISVPTA (827-836)

MR 336 = KGGVLLCPPRPCLTPT (852-867)

Example 4

The EST sequence of the magic gazebo identified in the bioinformatics search for specific endothelial transcripts was used to isolate a cDNA 3800 base pairs in length from a human heart cDNA library. A search using gene-specific primers showed that it was present in libraries of the heart, fetal and adult brain, liver, lung, kidney, muscle, placenta and small intestine, but absent in leukocytes of peripheral blood, spleen and testicles. The biggest expression was in the placenta library. Comparison of the sequence of the magic roundabout to that of the roundabout revealed a transmembrane protein with homology throughout the entire protein sequence but absence of some extracellular domains. Thus, MR has two immunoglobulin domains and two fibronectin domains in the extracellular portion compared to five immunoglobulin domains and two fibronectin domains in the extracellular portion of specific neuronal roundabouts. A transmembrane domain was identified through (i) the use of a PRED-TMR transmembrane prediction program and (ii) the use of an alignment between the sequences of the human MR and the human ROBO1 peptide. Both methods identified the same residues corresponding to the transmembrane region of human MR, amino acids 468-490. Therefore, aa 1-467 are extracellular and aa 491-1007 are intracellular. The intracellular domain contains a region rich in putative proline that is homologous to those in the roundabout that are thought to be coupled to c-abl (Bashaw et al (2000) Cell 101: 703-715).

The fragmented sequence site (STS) of the human SHGC-11739 (GenBank acc. G14646) was mapped into the mRNA of the magic roundabout in a BLAST dbSTS search. This STS was located on chromosome 11 on the physical map of Stanford G3 (region 5647.00 cR10000 LOD 1.09 bin 129). However, a large amount of the sequence is missing and the genomic structure is unknown. The search in the RIKEN database identified the murine magic roundabout. The predicted molecular weight of the peptide nucleus of human MR was 107457 kDa. This was confirmed by in vitro translation (figure 3).

Example 5

Expression is detectable in tumors

In situ hybridization was used to characterize the expression of ECSM4 in vivo. It was found that the expression of ECSM4 is very restricted (Table 13), with a signal that is not detectable in many tissues, including neuronal tissue. In contrast, strong expression was detected in placenta and in a range of tumors, including brain, bladder and liver cancer metastases (Figure 27). Expression within the tumors was restricted to the tumor vasculature. Immunohistochemical staining of the placenta confirmed the specific endothelial expression of the protein.

5 A search for ECSM4 in CGAP SAGE libraries was detected only in endothelial and tumor libraries (Table 14). This was consistent with the results of in situ hybridization in adults, which showed that the expression was restricted to tumor vessels (metastasis of colon cancer in the liver, ganglioglioma and carcinomas of the bladder and lung).

Table 13 Expression of the magic gazebo in human tissue in vivo.

Expression detected Placenta and fetal umbilical cord tissue (10.8 weeks of menstrual age)

10 Metastatic vessels in the liver of colorectal tumor, ganglioglioma, bladder and mammary carcinoma. Undetected expression Adult liver, brain and large caliber vessels, prostate, solon, small intestine, heart, choroid and sclera of the eye,

ovary, stomach, breast

Table 14 CGAP SAGE libraries in which the magic roundabout was found based on the mapping of fragments 15 to genes

library

Fragments / million fragments

HDMEC

171

HDMEC + VEGF

224

Medulloblastoma

102

Glioblastoma multiforme

85

Ovary, serous adenocarcinoma

59

Glioblastoma multiforme, sampled

48

HMDEC, human dermal microvascular endothelial cells; VEGF, vascular endothelial growth factor

Example 6

Induction of ECSM4 in hypoxic endothelial cells

20 Initial RT-PCR detected the expression of ECSM4 in endothelial cell lines, but not in other cell lines, such as fibroblasts (normal endometrium and FEK4), colon carcinoma (SW480 and HCT116), breast carcinoma (MDA453 and MDA468) and HeLa cells. The ribonuclease protection analysis confirmed and extended this result (Figure 11a). It was found that ECSM4 expression is restricted to the endothelium (three different isolates) and absent in fibroblasts, carcinoma cells and neurons. The induction of ECSM4 in hypoxia in endothelial cells (but not in non-endothelial cells) is

25 observed when the expression of ECSM4 was analyzed using two different RNAse protection probes. The expression was, on average 5.5 and 2.6 times higher in hypoxia for HUVEC and HDMEC respectively. Western analysis identified a weak band of 110 kD in human dermal microvascular endothelial cells (HDMEC), but which was absent from non-endothelial cell types (Figure 11b). The band was more intense when HDMEC cells were exposed to 18 h of hypoxia, consistent with the hypothesis that ECSM4 is a gene regulated by hypoxia.

Claims (24)

one.

An antibody capable of selectively binding the human ECSM4 polypeptide, wherein said human ECSM4 is a naturally occurring polypeptide comprising a portion of 50 consecutive amino acid residues of the polypeptide sequence of Figure 4 or Figure 5, for use as a medicine..

2.

An antibody according to Claim 1, which selectively binds a polypeptide comprising the amino acid sequence located between residues 1 and 467 of the polypeptide sequence of Figure 12.

3.

 An antibody according to Claim 1 or 2 that selectively binds any of the amino acid sequences GGDSLLGGRGSL, LLQPPARGHAHDGQALSTDL, EPQDYTEPVE, TAPGGQGAPWAEE, or ERATQEEPSEHGP ,.

Four.

An antibody according to any one of Claims 1 to 3 which is a monoclonal antibody.

5.

A compound comprising (i) an antibody as defined in any one of claims 1 to 4, and (ii) an additional fraction, for use as a medicament.

6.

A compound according to Claim 5 wherein the additional fraction is an easily detectable fraction.

7.

A compound according to claim 6 wherein the easily detectable fraction comprises a radioactive atom.

8.

A compound according to claim 7 wherein the radioactive atom is technetium-99m or iodine-123.

9.

A compound according to Claim 6 wherein the easily detectable fraction comprises an appropriate amount of any one of iodine-123, iodine-131, indium-111, fluorine-19, carbon-13, nitrogen-15, oxygen-17, gadolinium , manganese or iron.

10.

 A compound according to claim according to claim 5, wherein the additional fraction is a direct fraction

or indirectly cytotoxic.

eleven.

 A compound according to claim 10 wherein the cytotoxic fraction is a directly cytotoxic chemotherapeutic agent.

12.

 A compound according to claim 11, wherein the cytotoxic fraction includes a radioactive atom.

13.

 A compound according to claim 12 wherein the radioactive atom is any one of phosphorus-32, iodine-125, iodine-131, indium-111, rhenium-186, rhenium-188 or yttrium-90.

14.

 A compound according to claim 10 wherein the cytotoxic fraction is a directly cytotoxic polypeptide.

fifteen.

A compound according to Claim 10 wherein the cytotoxic fraction is a fraction that is capable of converting a relatively non-toxic prodrug into a cytotoxic drug.

16.

 A compound according to claim 10 wherein the cytotoxic fraction is a radiosensitizer.

17.

 A compound according to claim 5 wherein the additional fraction comprises a nucleic acid molecule.

18.

 A compound according to claim 17 wherein the nucleic acid molecule is a directly cytotoxic nucleic acid molecule.

19.

 A compound according to claim 17 wherein the nucleic acid molecule encodes a polypeptide directly or indirectly cytotoxic.

twenty.

 A compound according to claim 17 wherein the nucleic acid encodes a therapeutic polypeptide.

twenty-one.

 A compound according to Claim 5 wherein the additional fraction selectively binds a directly or indirectly cytotoxic fraction.

22

A compound according to Claim 5 wherein the additional fraction selectively binds to an easily detectable fraction.

2. 3.

 A compound according to Claim 5 wherein the antibody and the additional fraction are polypeptides that are fused.

24.

 An expression vector comprising a polynucleotide encoding a compound according to claim 23, for use as a medicament.

 REFERENCES CITED IN THE DESCRIPTION This list of references cited by the applicant is solely for the convenience of the reader. It is not part of the European Patent document. Although great care has been taken in the compilation of references, errors or omissions cannot be excluded and the EPO rejects any responsibility in this regard. Patent documents cited in the description

•

WO 9946281 A [0005] [0006] [0014] [0023] [0118] [0136]

•

WO 9911293 A [0005] [0006] [0014] [0023] [0118] [0136]

•

WO 9953051 A [0005]

•

EP 1074617 A [0006] [0014] [0023] [0118] [0136]

•

WO 0053756 A [0006] [0014] [0023] [0118] [0136]

•

WO 0123523 A [0006] [0014] [0023] [0118] [0136]

•

WO 99426281 A [0012]

•

WO 9606641 A [0044]

•

US 4348376 A, Goldenberg [0059]

•

WO 9410323 A [0064]

•

US 5180818 A [0067]

•

US 5168053 A [0067]

•

US 5149796 A [0067]

•

US 5116742 A [0067]

•

US 5093246 A [0067]

•

US 4957071 A [0067]

•

EP 0415731 A [0068]

•

EP 0682113 A2 [0186] Non-patent related literature cited in the description

•

Morrison et al. Proc. Natl Acad. Sci. USA, 1984, vol. 81, 6851-6855 [0025]

•

 Better et al. Science, 1988, vol. 240, 1041 [0026]

•

 Skerra et al. Science, 1988, vol. 240, 1038 [0026]

•

Bird et al. Science, 1988, vol. 242, 423 [0026]

•

Huston et al. Proc. Natl Acad. Sci. USA, vol. 85, 5879 [0026]

•

Ward et al. Nature, 1989, vol. 341, 544 [0026]

•

 Winter; Milstein Nature, 1991, vol. 349, 293-299 [0026]

•

 H Zola. Monoclonal Antibodies; A manual of techniques. CRC Press, 1988 [0031]

•

 SGR Hurrell. Monoclonal Hybridoma Antibodies: Techniques and Application. CRC Press, 1982

[0031]

•

Neuberger et al. 8th International Biotechnology Symposium, 1998, 792-799 [0032]

•

Vaughan et al. Nature Biotechnol., 1998, vol. 16, 535-539 [0034]

•

Bauminger; Wilchek Methods Enzymol., 1980, vol. 70, 151-159 [0039]

•

 Burrows; Thorpe Proc. Natl Acad. Sci. USA, 1993, vol. 90, 8996-9000 [0044]

•

Ran et al. Cancer Res., 1998, vol. 58, 4646-4653 [0044]

•

 Huang et al. Science, 1997, vol. 275, 547-550 [0044]

•

Tsai et al. Dis. Colon Rectum, 1995, vol. 38, 1067-1074 [0044]

•

 Aiello et al. Proc. Natl Acad. Sci. USA, 1995, vol. 92, 10457-10461 [0044]

•

Senter, P.D. et al. Antitumor effects of antibody-alkaline phosphatase conjugates in combination with etoposide phosphate. Proc. Natl Acad. Sci. USA, 1988, vol. 85, 4842-4846 [0048]

•

Bagshawe Br. J. Cancer, 1987, vol. 56, 531-2 [0048]

•

Bagshawe, K.D. et al. A cytotoxic agent can be generated selectively at cancer sites. Br. J. Cancer., 1988, vol. 58, 700-703 [0048]

•

Bagshawe Drug Dev. Reps., 1995, vol. 34, 220-230 [0050]

•

Rodrigues, M.L. et al. Chemistry & Biology, 1995, vol. 2, 223 [0050]

•

 O'Sullivan et al. Anal. Biochem., 1979, vol. 100, 100-108 [0054]

•

McGinn et al. J. Natl. Cancer Inst., 1996, vol. 88, 1193-11203 [0057]

•

Shewach; Lawrence Invest. New Drugs, 1996, vol. 14, 257-263 [0057]

•

 Horsman Acta Oncol., 1995, vol. 34, 571-587 [0057]

•

 Shenoy; Singh. Clin. Invest., 1992, vol. 10, 533-551 [0057]

•

Mitchell et al. Int. J. Radiat. Biol., 1989, vol. 56, 827-836 [0057]

•

Iliakis; Kurtzman Int. J. Radiat. Oncol. Biol. Phys., 1989, vol. 16, 1235-1241 [0057]

•

 Brown Int. J. Radiat. Oncol. Biol. Phys., 1989, vol. 16, 987-993 [0057]

•

 Brown Cancer, 1985, vol. 55, 2222-2228 [0057]

•

 Lang et al. J. Neurosurg, 1998, vol. 89, 125-132 [0058]

•

 Coco Martin et al. Cancer Res, 1999, vol. 59, 1134-1140 [0058]

•

 Primus et al. Bioconjug Chem., 1996, vol. 7, 532-535 [0059]

•

Dougherty et al. J. Natl. Cancer Inst., 1998, vol. 90, 889-905 [0060]