INHIBITORS AND USES

The present invention relates to tumour endothelium specific genes and polypeptides, to the use of antibodies that bind these polypeptides for imaging and targeting tumour vasculature, and to the use of inhibitors of these genes/polypeptides for inhibiting angiogenesis in solid tumours. In particular, the present invention relates to ECSM2, to the use of antibodies that bind ECSM2 for imaging and targeting tumour vasculature, and to the use of inhibitors of ECSM2 for inhibiting angiogenesis in solid tumours.

The endothelium plays a central role in many physiological and pathological processes and it is known to be an exceptionally active transcriptional site. Approximately 1 ,000 distinct genes are expressed in an endothelial cell, although many of them are not endothelial cell specific. In contrast red blood cells were found to express 8, platelets 22 and smooth muscle 127 separate genes (Adams et al (1995) Nature 377 (6547 Suppl): 3-174). Known endothelial specific genes attract much attention from both basic research and the clinical community. For example, the endothelial-specific tyrosine kinases Tie, TIE2/TEK, KDR, and flt1 are crucial players in the regulation of vascular integrity, endothelium-mediated inflammatory processes and angiogenesis.

We have previously used an in silico database screening approach to identify endothelial specific genes, and identified four new candidate endothelial specific genes (Huminiecki & Bicknell, 2000).

Ho et al (2003) used data mining and micro-array expression analysis to identify endothelial specific genes, and identified 64 genes that are either specific for endothelial cells or at least 3-fold preferentially expressed in endothelial cells.

Endothelial cells form a single cell layer that lines all blood vessels and regulates exchanges between the blood stream and the surrounding tissues. New blood vessels develop from the walls of existing small vessels by the outgrowth of endothelial cells in the process called angiogenesis. Endothelial cells even have the capacity to form hollow capillary tubes when isolated in culture. Once the vascular system is fully developed, endothelial cells of blood vessels normally remain quiescent with no new vessel formation, with the exception of the formation of new blood vessels in natural wound healing.

However, some tumours attract a new blood supply by secreting factors that stimulate nearby endothelial cells to construct new capillary sprouts. Angiogenesis plays a major role in the progression of solid tumours and is widely recognised as a rate-limiting process in the growth of solid tumours. Tumours that fail to attract a blood supply are severely limited in their growth. Thus the ability to inhibit inappropriate or undesirable angiogenesis may be useful in the treatment of solid tumours.

The development of new blood vessels is essential for both local tumour progression and the development of distant metastases. Tumour angiogenesis involves the degradation of the basement membrane by activated tissue or circulating endothelial precursors, proliferation and migration of endothelial cells, interaction with the extracellular matrix, morphological differentiation, cell adherence and vascular tube formation. Inhibition of tumour angiogenesis is thus a target for anti-tumour therapies, employing either angiogenesis inhibitors alone or in combination with standard cancer treatments. However, targeting anti-tumour agents to the site of angiogenesis depends upon the identification of specific markers of tumour angiogenesis. Indeed, it is now accepted that the growth of solid tumours is dependent on their capacity to acquire a blood supply, and much effort has been directed towards the development of anti-angiogenic agents that disrupt this process. It has also become apparent that targeted destruction of the established tumour vasculature is another avenue for exciting therapeutic opportunities (Neri & Bicknell, 2005). These therapeutic approaches depend upon the identification of specific tumour endothelial markers (TEMs).

In a screen for tumour-specific endothelial markers that might be candidates for anti- angiogenic tumour therapy, St Croix et al (2000) identified 79 genes that were differentially expressed between endothelial cells derived from tumour endothelium and normal colonic mucosa. The expression of 33 of these genes was elevated at least 10- fold in tumour endothelial cells, including 11 known and 14 as-then uncharacterised genes. In situ hybridization on tissue samples confirmed that the expression of eight of the nine uncharacterised genes that were studied in depth were specific for tumour endothelial cells. Moreover, these genes were also expressed on endothelial cells of other tumours including lung and brain tumours. Except for one gene, these genes were also expressed at elevated levels in other angiogenic states such as healing wounds.

Khodarev et al (2003) modelled tumour/endothelial-cell interactions by co-culturing U87 human glioma cells with human umbilical vein endothelial cells (HUVECs). U87 cells

induced an 'activated' phenotype in HUVECs, including an increase in proliferation, migration and net-like formation. Activation was observed in co-cultures where cells were either in direct contact or physically separated, suggesting an important role for soluble factor(s) in the phenotypic and genotypic changes observed. Expressional profiling of tumour-activated endothelial cells was evaluated using cDNA arrays and confirmed by quantitative PCR. Matching pairs of receptors/ligands were found to be coordinately expressed, including TGFβRII with TGF&3, FGFRII and cysteine-rich fibroblast growth factor receptor (CRF-1) with FGF7 and FGF12, CCR1 , CCR3, CCR5 with RANTES and calcitonin receptor-like gene (CALCRL) with adrenomedullin. (Khodarev et al (2003) "Tumour-endothelium interactions in co-culture: coordinated changes of gene expression profiles and phenotypic properties of endothelial cells", Journal of Cell Science 116: 1013-1022.)

Seaman et al (2007) compared gene expression patterns of endothelial cells derived from the blood vessels of eight normal resting tissues, five tumours, and regenerating liver. They identified organ-specific endothelial genes as well as 25 transcripts over- expressed in tumour versus normal endothelium, 13 of which were not found in the angiogenic endothelium of regenerating liver. Most of the shared angiogenesis genes were expected to have roles in cell-cycle control, but those specific for tumour endothelium were primarily cell surface molecules of uncertain function (Seaman et al (2007) "Genes that distinguish physiological and pathological angiogenesis." Cancer Cell. 11(6): 539-54).

Nevertheless, there is a need in the art for additional tumour endothelial markers.

We have now identified a number of genes (listed in Table 1) whose expression is highly specific to the tumour endothelium by data-mining public cDNA and SAGE libraries. We have thus identified these genes/polypeptides as novel Tumour Endothelial Markers (TEMs). TEMs are particularly good anticancer drug targets as they can be targeted directly via the blood supply.

Table 1. Newly Identified Tumour Endothelial Markers (TEMs)

Each of the genes/polypeptides listed in Table 1 are ones we have newly identified as having a high degree of tumour endothelial specificity. For ECSM2 and RHOJ (and for other genes identified in the screening), we have confirmed by RT-PCR that the expression of these genes is highly- or completely-specific for human umbilical vein endothelial cells (HUVECs) and adult human dermal microvascular endothelial cells (HDMECs).

By in-situ hybridisation we have also shown that ECSM2 is expressed solely in the endothelium of a number of different solid tumours, but not in non-tumour control samples. Furthermore, we have shown that inhibition of ECSM2 using siRNA technology significantly inhibits endothelial proliferation, which is an essential component of angiogenesis.

We have also found that RHOJ is specifically upregulated in endothelial cells, and that the downregulation of RHOJ using siRNA technology in HUVEC reduced cell growth/proliferation, significantly impaired tube formation on fibrin gels and Matrigel ^™ , and inhibited cell migration in scratch wound and chemotaxis assays. The powerful effect of RHOJ downregulation indicates that RHOJ may play an important role in tumour angiogenesis. We have also found that the downregulation of RHBDL6 using siRNA technology in HUVEC reduced cell growth/proliferation, caused defective tube formation on Matrigel ^™ , and inhibited cell migration in scratch wound assays. The powerful effect of RHBDL6 downregulation indicates that RHBDL6 may play an important role in tumour angiogenesis. By in-situ hybridisation we have also shown that LRRC8C is expressed

specifically in the endothelium of a squamous cell carcinoma and that PCHD12 is expressed specifically in the endothelium of breast cancer tissue and fibrous histiocytoma tissue, but not in non-tumour control samples. These data on RHOJ, RHBDL6, LRRC8C and PCHD12 are not shown herein, but are included in our co- pending PCT application filed on the same day as this application under Attorney Docket No. CRTBV/P41657PC, and which also claims priority from US provisional patent application No. 60/997,477 filed on 3 October 2007.

As a further confirmation, our screening also identified a number of genes, such as ROBO4, ANGPT2, VIM, SPARC, SPHK1 and MED28 that were previously known or putative Tumour Endothelial Markers.

Accordingly, we conclude that each of the nine genes listed in Table 1 genuinely encode TEMs. Therefore, we now consider that each of these genes/polypeptides will be valuable as markers of the tumour endothelium; that antibodies that selectively bind these polypeptides can be used to image and target the tumour endothelium; and that inhibitors of these genes/polypeptides would be therapeutically useful in the inhibition of tumour neoangiogenesis for the treatment of solid tumours.

A first aspect of the invention thus provides a method of inhibiting tumour angiogenesis in an individual in need thereof, the method comprising administering to the individual an inhibitor of ECSM2.

This aspect of the invention includes the use of an inhibitor of ECSM2 in the preparation of a medicament for inhibiting tumour angiogenesis in an individual. The invention further includes an inhibitor of ECSM2 for use in inhibiting tumour angiogenesis in an individual.

It is appreciated that inhibiting tumour angiogenesis is therapeutically beneficial in an individual having a solid tumour. Accordingly, a second aspect of the invention provides a method of combating a solid tumour in an individual, the method comprising administering to the individual an inhibitor of ECSM2.

This aspect of the invention includes the use of an inhibitor of ECSM2 in the preparation of a medicament for combating a solid tumour in an individual. The invention also includes an inhibitor of ECSM2 for use in combating a solid tumour in an individual.

By "combating" we include the meaning that the method or the medicament or the inhibitor can be used to alleviate symptoms of the tumour (i.e. the method is used palliatively), or to treat the tumour, e.g. to prevent the (further) growth of the tumour, to prevent the spread of the tumour (metastasis), or to reduce the size of the tumour. Typically, the tumour is associated with undesirable neovasculature formation and the treatment reduces this to a useful extent. The reduction of undesirable neovasculature formation may halt the progression of the tumour and can lead to a clinically useful reduction of tumour size and growth.

Typically, the individual has a solid tumour which can be treated by inhibiting angiogenesis, i.e. a solid tumour which is associated with new blood vessel production. The term "tumour" is to be understood as referring to all forms of neoplastic cell growth, including tumours of the lung, brain, colon, kidney, prostate and skin as well as tumours of the liver, pancreas, stomach, uterus, ovary, breast, lymph glands and bladder.

VEGF is a well-known TEM ₁ and the anti-VEGF monoclonal antibody Bevacizumab (Avastin™) by Genentech, Inc. was the first angiogenesis inhibitor approved by the FDA for the treatment of solid tumours. Bevacizumab has been investigated for efficacy in treatment in a number of cancers including metastatic or advanced colorectal cancer, breast cancer including recurrent or metastatic breast cancer, lung cancer including advanced non-squamous non-small cell lung cancer, advanced or metastatic renal cell carcinoma, pancreatic cancer, and ovarian cancer.

Accordingly, we consider that cancers including metastatic or advanced colorectal cancer, breast cancer including recurrent or metastatic breast cancer, lung cancer including advanced non-squamous non-small cell lung cancer, advanced or metastatic renal cell carcinoma, pancreatic cancer and ovarian cancer may be treatable by inhibiting angiogenesis using an inhibitor of ECSM2 as disclosed herein.

In addition, we have shown that ECSM2 is specifically expressed in the vascular endothelium of a range of solid tumours including metastatic adenocarcinoma, astrocytoma, ganglioma, ganglioglioma, glioblastoma, medulloblastoma, histiocytoma, and solid bladder, lung, oesphagus, stomach, breast and kidney tumours, but not in control tissues. Thus, in a further embodiment, the solid tumour to be combated by an inhibitor of ECSM2 may be a brain tumour, such as an astrocytoma, ganglioma,

ganglioglioma, metastatic adenocarcinoma, glioblastoma and medulloblastoma. Alternatively, the solid tumour to be combated may be a histiocytoma or a solid tumour of the bladder, lung, oesphagus, stomach, breast or kidney.

The therapy (treatment) may be on humans or animals. Preferably, the methods of the inventions are used to treat humans. It is appreciated that when the methods are for treatment of non-human mammals, it is preferred if the inhibitor is specific for the ECSM2 gene/polypeptide from the other species.

A third aspect of the invention provides an ex vivo method of inhibiting angiogenesis, the method comprising administering an inhibitor of ECSM2 to tissue or cells ex vivo. Typically, this is an ex vivo method of inhibiting angiogenesis in a model of tumour angiogenesis, such as those desribed below. Thus the cells may be established tumour cell lines or tumour cells that have been removed from an individual. The tissue or cells are preferably mammalian tissue or cells, and most preferably are human tissue or cells. Preferably, the tissue or cells comprise tumour endothelium, or are a model of tumour endothelium.

By an inhibitor of ECSM2 we include both inhibitors of the ECSM2 polypeptide and of the the ECSM2 gene/cDNA.

Suitable inhibitors of ECSM2 include antibodies that selectively bind to ECSM2. Other suitable inhibitors of ECSM2 include siRNA, antisense polynucleotides and ribozyme molecules that are specific for polynucleotides encoding the ECSM2 polypeptide, and which prevent its expression.

It is appreciated that polynucleotide inhibitors of ECSM2 may be administered directly, or may be administered in the form of a polynucleotide that encodes the inhibitor. Thus, as used herein, unless the context demands otherwise, by administering to the individual an inhibitor of ECSM2 which is a polynucleotide, we include the meanings of administering the inhibitor directly, or administering a polynucleotide that encodes the inhibitor. Similarly, as used herein, unless the context demands otherwise, by a medicament or a composition comprising an inhibitor of ECSM2 which is a polynucleotide, we include the meanings that the medicament or composition comprises the inhibitor itself, or comprises a polynucleotide that encodes the inhibitor.

Antibodies

Suitable antibodies which bind to ECSM2, or to specified portions thereof, can be made by the skilled person using technology long-established in the art. Methods of preparation of monoclonal antibodies and antibody fragments are well known in the art and include hybridoma technology (Kohler & Milstein (1975) "Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 256: 495-497); antibody phage display (Winter et al (1994) "Making antibodies by phage display technology." Annu. Rev. Immunol. 12: 433-455); ribosome display (Schaffitzel et al (1999) "Ribosome display: an in vitro method for selection and evolution of antibodies from libraries." J. Immunol. Methods 231 : 119-135); and iterative colony filter screening (Giovannoni et al (2001) "Isolation of anti-angiogenesis antibodies from a large combinatorial repertoire by colony filter screening." Nucleic Acids Res. 29: E27). Further, antibodies and antibody fragments suitable for use in the present invention are described, for example, in the following publications: "Monoclonal Hybridoma Antibodies: Techniques and Application", Hurrell (CRC Press, 1982); "Monoclonal Antibodies: A Manual of Techniques", H. Zola, CRC Press, 1987, ISBN: 0-84936-476-0; "Antibodies: A Laboratory Manuar 1 ^st Edition, Harlow & Lane, Eds, Cold Spring Harbor Laboratory Press, New York, 1988. ISBN 0- 87969-314-2; "Using Antibodies: A Laboratory Manuar 2 ^nd Edition, Harlow & Lane, Eds, Cold Spring Harbor Laboratory Press, New York, 1999. ISBN 0-87969-543-9; and "Handbook of Therapeutic Antibodies" Stefan Dϋbel, Ed., 1 ^st Edition, - Wiley-VCH, Weinheim, 2007. ISBN: 3-527-31453-9.

Antibodies that are especially active at inhibiting tumour angiogenesis are preferred for anti-cancer therapeutic agents, and they can be selected for this activity using methods well known in the art and described below.

By an antibody that selectively binds the ECSM2 polypeptide (i.e. the glycoprotein) we mean that the antibody molecule binds ECSM2 with a greater affinity than for an irrelevant polypeptide, such as human serum albumin (HSA). Preferably, the antibody binds the ECSM2 with at least 5, or at least 10 or at least 50 times greater affinity than for the irrelevant polypeptide. More preferably, the antibody molecule binds the ECSM2 with at least 100, or at least 1,000, or at least 10,000 times greater affinity than for the irrelevant polypeptide. Such binding may be determined by methods well known in the art, such as one of the Biacore ^® systems. Preferably, the antibody molecule selectively binds the ECSM2 without significantly binding other polypeptides in the body. It is preferred if the antibodies have an affinity for their target of at least 10 ^"7 M and more

preferably 10 ^'8 M, although antibodies with higher affinities, e.g. 10 ^"9 M, or higher, may be even more preferred.

By an antibody that selectively binds a specific portion of ECSM2 we mean that not only does the antibody selectively bind to the target as described above, the antibody molecule also binds the specified portion of the ECSM2 with a greater affinity than for any other portion of it. Preferably, the antibody binds the specified portion with at least 2, or at least 5, or at least 10 or at least 50 times greater affinity than for any other epitope on ECSM2. More preferably, the antibody molecule binds the specified portion with at least 100, or at least 1 ,000, or at least 10,000 times greater affinity than for than for any other epitope on the ECSM2. Such binding may be determined by methods well known in the art, such as one of the Biacore ^® systems. It is preferred if the antibodies have an affinity for their target epitope on the ECSM2 of at least 10 ^"7 M and more preferably 10 ^"8 M, although antibodies with higher affinities, e.g. 10 ^"9 M, or higher, may be even more preferred. Preferably, the antibody selectively binds the particular specified epitope within the ECSM2 and does not bind any other epitopes within it.

Preferably, when the antibody is administered to an individual, the antibody binds to the target ESM2 or to the specified portion thereof with a greater affinity than for any other molecule in the individual. Preferably, the antibody binds to (a specified portion of) the ECSM2 with at least 2, or at least 5, or at least 10 or at least 50 times greater affinity than for any other molecule in the individual. More preferably, the agent binds the ECSM2 (at the specific domain) with at least 100, or at least 1,000, or at least 10,000 times greater affinity than any other molecule in the individual.

The term "antibody" or "antibody molecule" as used herein includes but is not limited to polyclonal, monoclonal, chimaeric, single chain, Fab fragments and fragments produced by a Fab expression library. Such fragments include fragments of whole antibodies which retain their binding activity for a target substance, Fv, F(ab') and F(ab')2 fragments, as well as single chain antibodies (scFv), fusion proteins and other synthetic proteins which comprise the antigen-binding site of the antibody. The term also includes antibody-like molecules which may be produced using phage-display techniques or other random selection techniques for molecules which bind to the specified polypeptide or to particular regions of it. Thus, the term antibody includes all molecules which contain a structure, preferably a peptide structure, which is part of the recognition site (i.e. the part of the antibody that binds or combines with the epitope or antigen) of a natural antibody.

Furthermore, the antibodies and fragments thereof may be humanised antibodies, which are now well known in the art.

By "ScFv molecules" we mean molecules wherein the V _H and V _L partner domains are linked via a flexible oligopeptide. Engineered antibodies, such as ScFv antibodies, can be made using the techniques and approaches long known in the art. The advantages of using antibody fragments, rather than whole antibodies, are several-fold. The smaller size of the fragments may lead to improved pharmacological properties, such as better penetration to the target site. Effector functions of whole antibodies, such as complement binding, are removed. Fab, Fv, ScFv and dAb antibody fragments can all be expressed in and secreted from E. coli, thus allowing the facile production of large amounts of the fragments. Whole antibodies, and F(ab') ₂ fragments are "bivalent". By "bivalent" we mean that the antibodies and F(ab') ₂ fragments have two antigen combining sites. In contrast, Fab, Fv, ScFv and dAb fragments are monovalent, having only one antigen combining site.

It is appreciated, however, that ECSM2 is a glycoprotein (Armstrong et al 2008, Arterioscler Thromb Vase Biol. 28(9):1640-6. Epub 12 June 2008). Thus, the antibody that binds to ECSM2 may bind to any combination of the protein or carbohydrate components of ECSM2.

Antibodies may be produced by standard techniques, for example by immunisation with the appropriate (glyco)polypeptide or portion(s) thereof, or by using a phage display library.

If polyclonal antibodies are desired, a selected mammal (e.g., mouse, rabbit, goat, horse, etc) is immunised with an immunogenic polypeptide bearing a desired epitope(s), optionally haptenised to another polypeptide. Depending on the host species, various adjuvants may be used to increase immunological response. Such adjuvants include, but are not limited to, Freund's, mineral gels such as aluminium hydroxide, and surface active substances such as lysolecithin, pluronic polyols, polyanions, peptides, oil emulsions, keyhole limpet hemocyanin, and dinitrophenol. Serum from the immunised animal is collected and treated according to known procedures. If serum containing polyclonal antibodies to the desired epitope contains antibodies to other antigens, the polyclonal antibodies can be purified by immunoaffinity chromatography. Techniques for producing and processing polyclonal antisera are well known in the art.

Monoclonal antibodies directed against entire polypeptides or particular epitopes thereof can also be readily produced by one skilled in the art. The general methodology for making monoclonal antibodies by hybridomas is well known. Immortal antibody- producing cell lines can be created by cell fusion, and also by other techniques such as direct transformation of B lymphocytes with oncogenic DNA, or transfection with Epstein- Barr virus. Panels of monoclonal antibodies produced against the polypeptides listed above can be screened for various properties; i.e., for isotype and epitope affinity. Monoclonal antibodies may be prepared using any of the well known techniques which provides for the production of antibody molecules by continuous cell lines in culture.

It is preferred if the antibody is a monoclonal antibody. In some circumstance, particularly if the antibody is going to be administered repeatedly to a human patient, it is preferred if the monoclonal antibody is a human monoclonal antibody or a humanised monoclonal antibody, which are suitable for administration to humans without engendering an immune response by the human against the administered immunoglobulin. Suitably prepared non-human antibodies can be "humanised" in known ways, for example by inserting the CDR regions of mouse antibodies into the framework of human antibodies. Humanised antibodies can be made using the techniques and approaches described in Verhoeyen et al (1988) Science, 239, 1534-1536, and in Kettleborough et al, (1991) Protein Engineering, 14(7), 773-783. In some instances, Fv framework residues of the human immunoglobulin are replaced by corresponding non- human residues. In general, the humanised antibody will contain variable domains in which all or most of the CDR regions correspond to those of a non-human immunoglobulin, and framework regions which are substantially or completely those of a human immunoglobulin consensus sequence.

Completely human antibodies may be produced using recombinant technologies. Typically large libraries comprising billions of different antibodies are used. In contrast to the previous technologies employing chimerisation or humanisation of e.g. murine antibodies this technology does not rely on immunisation of animals to generate the specific antibody. Instead the recombinant libraries comprise a huge number of pre- made antibody variants wherein it is likely that the library will have at least one antibody specific for any antigen. Thus, using such libraries, an existing antibody having the desired binding characteristics can be identified. In order to find the good binder in a library in an efficient manner, various systems where phenotype i.e. the antibody or antibody fragment is linked to its genotype i.e. the encoding gene have been devised.

The most commonly used such system is the so called phage display system where antibody fragments are expressed, displayed, as fusions with phage coat proteins on the surface of filamentous phage particles, while simultaneously carrying the genetic information encoding the displayed molecule (McCafferty et al, 1990, Nature 348: 552- 554). Phage displaying antibody fragments specific for a particular antigen may be selected through binding to the antigen in question. Isolated phage may then be amplified and the gene encoding the selected antibody variable domains may optionally be transferred to other antibody formats, such as e.g. full-length immunoglobulin, and expressed in high amounts using appropriate vectors and host cells well known in the art. Alternatively, the "human" antibodies can be made by immunising transgenic mice which contain, in essence, human immunoglobulin genes (Vaughan et al (1998) Nature Biotechnol. 16, 535-539).

It is appreciated that when the antibody is for administration to a non-human individual, the antibody may have been specifically designed/produced for the intended recipient species.

The format of displayed antibody specificities on phage particles may differ. The most commonly used formats are Fab (Griffiths et al, 1994. EMBO J. 13: 3245-3260) and single chain (scFv) (Hoogenboom et al, 1992, J MoI Biol. 227: 381-388) both comprising the variable antigen binding domains of antibodies. The single chain format is composed of a variable heavy domain (V _H ) linked to a variable light domain (V _L ) via a flexible linker (US 4,946,778). Before use as a therapeutic agent, the antibody may be transferred to a soluble format e.g. Fab or scFv and analysed as such. In later steps the antibody fragment identified to have desirable characteristics may be transferred into yet other formats such as full-length antibodies.

WO 98/32845 and Soderiind et al (2000) Nature BioTechnol. 18:852-856 describe technology for the generation of variability in antibody libraries. Antibody fragments derived from this library all have the same framework regions and only differ in their CDRs. Since the framework regions are of germline sequence the immunogenicity of antibodies derived from the library, or similar libraries produced using the same technology, are expected to be particularly low (Soderiind et al, 2000). This property is of great value for therapeutic antibodies, reducing the risk that the patient forms antibodies to the administered antibody, thereby reducing risks for allergic reactions, the occurrence of blocking antibodies, and allowing a long plasma half-life of the antibody.

Thus, when developing therapeutic antibodies to be used in humans, modern recombinant library technology (Soderlind et a/, 2001 , Comb. Chem. & High Throughput Screen. 4: 409-416) is now used in preference to the earlier hybridoma technology.

si RNA

Small interfering RNAs are described by Hannon ef a/. Nature, 418 (6894): 244-51 (2002); Brummelkamp et a/., Science 21 , 21 (2002); and Sui et a/., Proc. Natl Acad. Sci. USA 99, 5515-5520 (2002). RNA interference (RNAi) is the process of sequence- specific post-transcriptional gene silencing in animals initiated by double-stranded (dsRNA) that is homologous in sequence to the silenced gene. The mediators of sequence-specific mRNA degradation are typically 21- and 22-nucleotide small interfering RNAs (siRNAs) which, in vivo, may be generated by ribonuclease III cleavage from longer dsRNAs. 21 -nucleotide siRNA duplexes have been shown to specifically suppress expression of both endogenous and heterologous genes (Elbashir et at (2001) Nature 411 : 494-498). In mammalian cells it is considered that the siRNA has to be comprised of two complementary 21mers as described below since longer double- stranded (ds) RNAs will activate PKR (dsRNA-dependent protein kinase) and inhibit overall protein synthesis.

Duplex siRNA molecules selective for a polynucleotide encoding the ECSM2 polypeptide can readily be designed by reference to its cDNA sequence. For example, they can be designed by reference to the ECSM2 cDNA sequences in the Genbank Accession Nos. listed herein. Typically, the first 21-mer sequence that begins with an AA dinucleotide which is at least 120 nucleotides downstream from the initiator methionine codon is selected. The RNA sequence perfectly complementary to this becomes the first RNA oligonucleotide. The second RNA sequence should be perfectly complementary to the first 19 residues of the first, with an additional UU dinucleotide at its 3' end. Once designed, the synthetic RNA molecules can be synthesised using methods well known in the art.

siRNAs may be introduced into cells in the patient using any suitable method, such as those described herein. Typically, the RNA is protected from the extracellular environment, for example by being contained within a suitable carrier or vehicle. Liposome-mediated transfer, e.g. the oligofectamine method, may be used.

Antisense polynucleotides

Antisense nucleic acid molecules selective for a polynucleotide encoding the ECSM2 polypeptide can readily be designed by reference to its cDNA or gene sequence, as is known in the art. Antisense nucleic acids, such as oligonucleotides, are single-stranded nucleic acids, which can specifically bind to a complementary nucleic acid sequence. By binding to the appropriate target sequence, an RNA-RNA, a DNA-DNA, or RNA-DNA duplex is formed. These nucleic acids are often termed "antisense" because they are complementary to the sense or coding strand of the gene. Recently, formation of a triple helix has proven possible where the oligonucleotide is bound to a DNA duplex. It was found that oligonucleotides could recognise sequences in the major groove of the DNA double helix. A triple helix was formed thereby. This suggests that it is possible to synthesise a sequence-specific molecules which specifically bind double-stranded DNA via recognition of major groove hydrogen binding sites. By binding to the target nucleic acid, the above oligonucleotides can inhibit the function of the target nucleic acid. This could, for example, be a result of blocking the transcription, processing, poly(A) addition, replication, translation, or promoting inhibitory mechanisms of the cells, such as promoting RNA degradations.

Antisense oligonucleotides are prepared in the laboratory and then introduced into cells, for example by microinjection or uptake from the cell culture medium into the cells, or they are expressed in cells after transfection with plasmids or retroviruses or other vectors carrying an antisense gene. Antisense oligonucleotides were first discovered to inhibit viral replication or expression in cell culture for Rous sarcoma virus, vesicular stomatitis virus, herpes simplex virus type 1, simian virus and influenza virus. Since then, inhibition of mRNA translation by antisense oligonucleotides has been studied extensively in cell-free systems including rabbit reticulocyte lysates and wheat germ extracts. Inhibition of viral function by antisense oligonucleotides has been demonstrated ex vivo using oligonucleotides which were complementary to the AIDS HIV retrovirus RNA (Goodchild, J. 1988 "Inhibition of Human Immunodeficiency Virus Replication by Antisense Oligodeoxynucleotides", Proc. Natl. Acad. Sci. (USA) 85(15), 5507-11). The Goodchild study showed that oligonucleotides that were most effective were complementary to the poly(A) signal; also effective were those targeted at the 5' end of the RNA, particularly the cap and 5N untranslated region, next to the primer binding site and at the primer binding site. The cap, 5' untranslated region, and poly(A) signal lie within the sequence repeated at the ends of retrovirus RNA (R region) and the oligonucleotides complementary to these may bind twice to the RNA.

Typically, antisense oligonucleotides are 15 to 35 bases in length. For example, 20-mer oligonucleotides have been shown to inhibit the expression of the epidermal growth factor receptor mRNA (Witters et al., Breast Cancer Res Treat 53:41-50 (1999)) and 25- mer oligonucleotides have been shown to decrease the expression of adrenocorticotropic hormone by greater than 90% (Frankel et a/., J Neurosurg 91 :261-7 (1999)). However, it is appreciated that it may be desirable to use oligonucleotides with lengths outside this range, for example 10, 11 , 12, 13, or 14 bases, or 36, 37, 38, 39 or 40 bases.

Antisense polynucleotides may be administered systemically. Alternatively, and preferably, the inherent binding specificity of polynucleotides characteristic of base pairing is enhanced by limiting the availability of the polynucleotide to its intended locus in vivo, permitting lower dosages to be used and minimising systemic effects. Thus, polynucleotides may be applied locally to the solid tumour to achieve the desired effect. The concentration of the polynucleotides at the desired locus is much higher than if the polynucleotides were administered systemically, and the therapeutic effect can be achieved using a significantly lower total amount. The local high concentration of polynucleotides enhances penetration of the targeted cells and effectively blocks translation of the target nucleic acid sequences.

It will be appreciated that antisense agents may also include larger molecules which bind to polynucleotides (mRNA or genes) encoding the ECSM2 polypeptide and substantially prevent expression of the protein. Thus, antisense molecules which are substantially complementary to the respective mRNA are also envisaged.

The molecules may be expressed from any suitable genetic construct and delivered to the patient. Typically, the genetic construct which expresses the antisense molecule comprises at least a portion of the ECSM2 cDNA or gene operatively linked to a promoter which can express the antisense molecule in the cell. Preferably, the genetic construct is adapted for delivery to a human cell.

Ribozvmes

Ribozymes are RNA or RNA-protein complexes that cleave nucleic acids in a site- specific fashion. Ribozymes have specific catalytic domains that possess endonuclease activity. For example, a large number of ribozymes accelerate phosphoester transfer reactions with a high degree of specificity, often cleaving only one of several phosphoesters in an oligonucleotide substrate. This specificity has been attributed to the

requirement that the substrate bind via specific base-pairing interactions to the internal guide sequence ("IGS") of the ribozyme prior to chemical reaction.

Ribozyme catalysis has primarily been observed as part of sequence-specific cleavage/ligation reactions involving nucleic acids. For example, US Patent No

5,354,855 reports that certain ribozymes can act as endonucleases with a sequence specificity greater than that of known ribonucleases and approaching that of the DNA restriction enzymes. Thus, sequence-specific ribozyme-mediated inhibition of gene expression may be particularly suited to therapeutic applications, and ribozymes specific for a polynucleotide encoding the ECSM2 polypeptide may be designed by reference to the cDNA sequences listed in the Genbank Accession Nos. given herein.

Methods and routes of administering polynucleotide inhibitors, such as siRNA molecules, antisense molecules and ribozymes, to a patient, are described in more detail below.

Further agents that inhibit transcription of the genes encoding any of the above listed polypeptides can also be designed, for example using an engineered transcription repressor described in lsalan et al (Nat Biotechnol, 19(7): 656-60 (2001)) and in Urnov {Biochem Pharmacol, 64 (5-6): 919 (2002)). Additionally, they can be selected, for example using the screening methods described in later aspects of the invention.

ECSM2

The gene ECSM2 (endothelial cell-specific molecule 2) encodes a 205 amino acid residue polypeptide. ECSM2 corresponds to the cDNA from EP 0 682 113 A1 and EMBL Accession No E 10591.

By the ECSM2 polypeptide we include the meaning of a gene product of human ECSM2, including naturally occurring variants thereof. A cDNA sequence corresponding to a human ECSM2 mRNA is found in Genbank Accession No NM_001077693 and in EMBL Accession No E10591. Human ECSM2 polypeptide includes the amino acid sequence found in Genbank Accession No NP_001071161 and naturally occurring variants thereof. The ECSM2 polypeptide sequence from NP_001071161 is shown in Figure 5.

ECSM2 is a serine and proline rich 205 residue transmembrane protein, which has the characteristics of a cell adhesion molecule (EP 0 682 113). According to EP 0 682 113,

ECSM2 has a signal peptide at residues 1-24, and a transmembrane region at residues 119-146.

According to EP 0 682 113, ECSM2 may be useful in a large number of conditions including immune diseases, rheumatoid arthritis, allergy, arteriosclerosis, organ transplant rejection, myocardial infarction, brain infarction, and reperfusion failure. EP 0 682 113 does not mention that ECSM2 is associated with the tumour endothelium, or with solid tumours.

In our earlier publication (Huminiecki & Bicknell, 2000), we identified ECSM2 as being specifically expressed in endothelial cells. However, in this earlier report we did not consider whether or not it was specifically expressed in the tumour endothelium. Indeed, no function for ECSM2 was known or proposed, and no further corroborative experiments were undertaken.

We have now shown that ECSM2 is highly specific for the tumour endothelium. For example, in Examples 2 and 4 we show that it is specifically expressed in the vascular endothelium of a range of solid tumours including metastatic adenocarcinoma, ganglioma, ganglioglioma, glioblastoma, medulloblastoma, astrocytoma, histiocytoma, and solid bladder, lung, oesphagus, stomach, breast and kidney tumours. We also show in Examples 3 and 4 that knock-out of ECSM2 in cultured HUVECs using three different siRNA molecules resulted in the significant inhibition of the proliferation of these cells, which is a recognised precursor of angiogenesis. Thus we have shown that ECSM2 is a highly tumour-specific anti-angiogenesis target.

To the best of the inventors' knowledge, ECSM2 has not been previously been associated with the tumour endothelium, or with solid tumours. Indeed, to the best of the inventors' knowledge, neither inhibitors of the ECSM2 gene, such as siRNA, antisense molecules or ribozymes specific for ECSM2, nor inhibitors of the ECSM2 polypeptide, such as antibodies that selectively bind to ECSM2, have been associated with the inhibition of angiogenesis in the tumour endothelium, or with the treatment of solid tumours. Thus the invention includes a method of inhibiting tumour angiogenesis by administering an inhibitor of the gene/polypeptide ECSM2, a method of combating a solid tumour by administering an inhibitor of the gene/polypeptide ECSM2, and the use of an inhibitor of the gene/polypeptide ECSM2 in the manufacture of a medicament for inhibiting tumour angiogenesis or for combating a solid tumour.

In an embodiment, the solid tumour to be combated may be a brain tumour, such as an astrocytoma, ganglioma, ganglioglioma metastatic adenocarcinoma, glioblastoma and medulloblastoma. Alternatively, the solid tumour to be combated may be a histiocytoma or a solid tumour of the bladder, lung, oesphagus, stomach or kidney.

We have also observed endothelial restricted expression of ECSM2 in the skin in a psoriasis patient biopsy, which is a site of active angiogenesis. To the best of the inventors' knowledge, ECSM2 has not been previously been associated with psoriasis. Indeed, to the best of the inventors' knowledge, neither inhibitors of the ECSM2 gene, such as siRNA, antisense molecules or ribozymes specific for ECSM2, nor inhibitors of the ECSM2 polypeptide, such as antibodies that selectively bind to ECSM2, have been associated with the inhibition of angiogenesis for the treatment of psoriasis. Thus the invention includes a method of inhibiting psoriatic angiogenesis by administering an inhibitor of the gene/polypeptide ECSM2, a method of combating psoriasis by administering an inhibitor of the gene/polypeptide ECSM2, and the use of an inhibitor of the gene/polypeptide ECSM2 in the manufacture of a medicament for inhibiting psoriatic angiogenesis or for combating psoriasis. Preferences for the ECSM2 inhibitor in this embodiment are as described herein.

Typically, the antibody that selectively binds ECSM2 binds to the mature peptide (residues 25-205) and not to the signal peptide (residues 1-24).

In a preferred embodiment, the antibody that selectively binds ECSM2 binds to the extracellular region of ECSM2 (residues 1-122, as determined using the CBS TMHMM method for predicting transmembrane helices (http://www.cbs.dtu.dk/services/TMHMM- 2.0)). More preferably, the antibody binds to residues 25-122 of ECSM2. According to EP 0 682 113, the ECSM2 transmembrane region is at residues 1 19-146, and, according to our own studies, the transmembrane region is likely to be at residues 120-147. Thus, the extracellular region of ECSM2 may comprise residues 25-1 18 or 25-119, and the antibody may bind to this region.

Moreover, the inventors are not aware of any suggested therapeutic use of a polynucleotide inhibitor of ECSM2 such as siRNA, antisense molecules or ribozymes.

Thus the invention includes a polynucleotide inhibitor of the gene/polypeptide ECSM2, a polynucleotide inhibitor of the gene/polypeptide ECSM2 for use in medicine, and a

pharmaceutical composition comprising a polynucleotide inhibitor of the gene/polypeptide ECSM2 and a pharmaceutically acceptable carrier, diluent or excipient.

It is appreciated that polynucleotide inhibitor of ECSM2 may be administered directly, or may be administered in the form of a polynucleotide that encodes the inhibitor, for combating a soliud tumour. Thus the siRNA, antisense molecule or ribozyme specific for ECSM2 may be administered directly to the individual in need thereof, or they may be administered in the form of a polynucleotide that encodes the inhibitor.

Preferences for formulations and routes of administration are provided herein below. For example, when the solid tumour is a CNS or brain tumour such as astrocytoma, ganglioma, ganglioglioma, metastatic adenocarcinoma, glioblastoma and medulloblastoma, the inhibitor of ECSM2 may be administered directly to the site of the tumour in the CNS or brain using methods and apparatus well known in the art and described below. Additionally or alternatively, when a polynucleotide that encodes the inhibitor of ECSM2 is used, it may be in a vector under the control of a CNS or brain specific promoter, which are well known in the art and described below.

Formulations and routes of administration It is appreciated that the inhibitor of ECSM2 will typically be formulated for administration to an individual as a pharmaceutical composition, i.e. together with a pharmaceutically acceptable carrier, diluent or excipient.

By "pharmaceutically acceptable" is included that the formulation is sterile and pyrogen free. Suitable pharmaceutical carriers, diluents and excipients are well known in the art of pharmacy. The carrier(s) must be "acceptable" in the sense of being compatible with the inhibitor and not deleterious to the recipients thereof. Typically, the carriers will be water or saline which will be sterile and pyrogen free; however, other acceptable carriers may be used.

In an embodiment, the pharmaceutical compositions or formulations of the invention are for parenteral administration, more particularly for intravenous administration. In a preferred embodiment, the pharmaceutical composition is suitable for intravenous administration to a patient, for example by injection.

Formulations suitable for parenteral administration include aqueous and non-aqueous sterile injection solutions which may contain anti-oxidants, buffers, bacteriostats and solutes which render the formulation isotonic with the blood of the intended recipient; and aqueous and non-aqueous sterile suspensions which may include suspending agents and thickening agents.

In an alternative preferred embodiment, the pharmaceutical composition is suitable for topical administration to a patient.

Preferably, the formulation is a unit dosage containing a daily dose or unit, daily sub-dose or an appropriate fraction thereof, of the active ingredient.

The inhibitor may be administered orally or by any parenteral route, in the form of a pharmaceutical formulation comprising the active ingredient, optionally in the form of a non-toxic organic, or inorganic, acid, or base, addition salt, in a pharmaceutically acceptable dosage form. Depending upon the disorder and patient to be treated, as well as the route of administration, the compositions may be administered at varying doses.

In human therapy, the inhibitor will generally be administered in admixture with a suitable pharmaceutical excipient, diluent or carrier selected with regard to the intended route of administration and standard pharmaceutical practice.

For example, the inhibitor may be administered orally, buccally or sublingually in the form of tablets, capsules, ovules, elixirs, solutions or suspensions, which may contain flavouring or colouring agents, for immediate-, delayed- or controlled-release applications. The inhibitor may also be administered via intracavemosal injection.

Suitable tablets may contain excipients such as microcrystalline cellulose, lactose, sodium citrate, calcium carbonate, dibasic calcium phosphate and glycine, disintegrants such as starch (preferably corn, potato or tapioca starch), sodium starch glycollate, croscarmellose sodium and certain complex silicates, and granulation binders such as polyvinylpyrrolidone, hydroxypropylmethylcellulose (HPMC), hydroxy-propylcellulose (HPC), sucrose, gelatin and acacia. Additionally, lubricating agents such as magnesium stearate, stearic acid, glyceryl behenate and talc may be included.

Solid compositions of a similar type may also be employed as fillers in gelatin capsules. Preferred excipients in this regard include lactose, starch, a cellulose, milk sugar or high molecular weight polyethylene glycols. For aqueous suspensions and/or elixirs, the compounds of the invention may be combined with various sweetening or flavouring agents, colouring matter or dyes, with emulsifying and/or suspending agents and with diluents such as water, ethanol, propylene glycol and glycerin, and combinations thereof.

The inhibitor can also be administered parenterally, for example, intravenously, intra- arterially, intraperitoneally, intrathecally, intraventricularly, intrasternally, intracranially, intra-muscularly or subcutaneously, or they may be administered by infusion techniques. They are best used in the form of a sterile aqueous solution which may contain other substances, for example, enough salts or glucose to make the solution isotonic with blood. The aqueous solutions should be suitably buffered (preferably to a pH of from 3 to 9), if necessary. The preparation of suitable parenteral formulations under sterile conditions is readily accomplished by standard pharmaceutical techniques well-known to those skilled in the art.

The formulations may be presented in unit-dose or multi-dose containers, for example sealed ampoules and vials, and may be stored in a freeze-dried (lyophilised) condition requiring only the addition of the sterile liquid carrier, for example water for injections, immediately prior to use. Extemporaneous injection solutions and suspensions may be prepared from sterile powders, granules and tablets of the kind previously described.

For oral and parenteral administration to human patients, the daily dosage level of an inhibitor will usually be from 1 to 1 ,000 mg per adult (Ae. from about 0.015 to 15 mg/kg), administered in single or divided doses.

Thus, for example, the tablets or capsules of the inhibitor may contain from 1 mg to 1 ,000 mg of active agent for administration singly or two or more at a time, as appropriate. The physician in any event will determine the actual dosage which will be most suitable for any individual patient and it will vary with the age, weight and response of the particular patient. The above dosages are exemplary of the average case. There can, of course, be individual instances where higher or lower dosage ranges are merited and such are within the scope of this invention.

The inhibitor can also be administered intranasally or by inhalation and are conveniently delivered in the form of a dry powder inhaler or an aerosol spray presentation from a pressurised container, pump, spray or nebuliser with the use of a suitable propellant, e.g. dichlorodifluoromethane, trichlorofiuoromethane, dichlorotetrafluoro-ethane, a hydrofluoroalkane such as 1 ,1 ,1 ,2-tetrafluoroethane (HFA 134A3 or 1 ,1 ,1 ,2,3,3,3- heptafluoropropane (HFA 227EA3), carbon dioxide or other suitable gas. In the case of a pressurised aerosol, the dosage unit may be determined by providing a valve to deliver a metered amount. The pressurised container, pump, spray or nebuliser may contain a solution or suspension of the active compound, e.g. using a mixture of ethanol and the propellant as the solvent, which may additionally contain a lubricant, e.g. sorbitan trioleate. Capsules and cartridges (made, for example, from gelatin) for use in an inhaler or insufflator may be formulated to contain a powder mix of a antibody and a suitable powder base such as lactose or starch. Such formulations may be particularly useful for treating solid tumours of the lung.

Aerosol or dry powder formulations are preferably arranged so that each metered dose or "puff' contains at least 1 mg of the inhibitor for delivery to the patient. It will be appreciated that the overall daily dose with an aerosol will vary from patient to patient, and may be administered in a single dose or, more usually, in divided doses throughout the day.

Alternatively, the inhibitor can be administered in the form of a suppository or pessary, particularly for combating solid colorectal tumours or prostate tumours.

The inhibitor may also be administered by the ocular route. For ophthalmic use, the inhibitor can be formulated as, e.g., micronised suspensions in isotonic, pH adjusted, sterile saline, or, preferably, as solutions in isotonic, pH adjusted, sterile saline, optionally in combination with a preservative such as a benzylalkonium chloride. Alternatively, they may be formulated in an ointment such as petrolatum. Such formulations may be particularly useful for treating solid tumours of the eye, such as retinoblastoma.

The inhibitor may be applied topically in the form of a lotion, solution, cream, ointment or dusting powder, or may be transdermal^ administered, for example, by the use of a skin patch. For application topically to the skin, the inhibitor can be formulated as a suitable ointment containing the active compound suspended or dissolved in, for example, a mixture with one or more of the following: mineral oil, liquid petrolatum, white petrolatum,

propylene glycol, polyoxyethylene polyoxypropylene compound, emulsifying wax and water. Alternatively, they can be formulated as a suitable lotion or cream, suspended or dissolved in, for example, a mixture of one or more of the following: mineral oil, sorbitan monostearate, a polyethylene glycol, liquid paraffin, polysorbate 60, cetyl esters wax, cetearyl alcohol, 2-octyldodecanol, benzyl alcohol and water. Such formulations may be particularly useful for treating solid tumours of the skin.

For skin cancers, the inhibitors can also be delivered by electroincoϊporation (El). El occurs when small particles of up to 30 microns in diameter on the surface of the skin experience electrical pulses identical or similar to those used in electroporation. In El, these particles are driven through the stratum comeum and into deeper layers of the skin. The particles can be loaded or coated with inhibitor or can simply act as "bullets" that generate pores in the skin through which the inhibitor can enter.

Formulations suitable for topical administration in the mouth include lozenges comprising the active ingredient in a flavoured basis, usually sucrose and acacia or tragacanth; pastilles comprising the active ingredient in an inert basis such as gelatin and glycerin, or sucrose and acacia; and mouth-washes comprising the active ingredient in a suitable liquid carrier. Such formulations may be particularly useful for treating solid tumours of the mouth and throat.

In an embodiment, when the inhibitor is a polypeptide, such as an anti-ECSM2 antibody, it may be delivered using an injectable sustained-release drug delivery system. These are designed specifically to reduce the frequency of injections. An example of such a system is Nutropin Depot which encapsulates recombinant human growth hormone (rhGH) in biodegradable microspheres that, once injected, release rhGH slowly over a sustained period.

The antibody can be administered by a surgically implanted device that releases the drug directly to the required site, for example, into the eye to treat ocular tumours. Such direct application to the site of disease achieves effective therapy without significant systemic side-effects.

An alternative method of polypeptide delivery is the ReGeI injectable system that is thermo-sensitive. Below body temperature, ReGeI is an injectable liquid while at body temperature it immediately forms a gel reservoir that slowly erodes and dissolves into

known, safe, biodegradable polymers. The active drug is delivered over time as the biopolymers dissolve.

Polypeptide pharmaceuticals such as antibodies can also be delivered orally. The process employs a natural process for oral uptake of vitamin B ₁₂ in the body to co-deliver proteins and peptides. By riding the vitamin B ₁₂ uptake system, the protein or peptide can move through the intestinal wall. Complexes are synthesised between vitamin B ₁₂ analogues and the drug that retain both significant affinity for intrinsic factor (IF) in the vitamin B ₁₂ portion of the complex and significant bioactivity of the drug portion of the complex.

Polynucleotides may be administered by any effective method, for example, parenterally (e.g. intravenously, subcutaneously, intramuscularly) or by oral, nasal or other means which permit the polynucleotides to access and circulate in the patient's bloodstream. Polynucleotides administered systemically preferably are given in addition to locally administered polynucleotides, but also have utility in the absence of local administration. A dosage in the range of from about 0.1 to about 10 grams per administration to an adult human generally will be effective for this purpose.

The polynucleotide may be administered as a suitable genetic construct as is described below and delivered to the patient where it is expressed. Typically, the polynucleotide in the genetic construct is operatively linked to a promoter which can express the compound in the cell. The genetic constructs of the invention can be prepared using methods well known in the art, for example in Sambrook et a/ (2001).

Although genetic constructs for delivery of polynucleotides can be DNA or RNA, it is preferred if they are DNA.

Preferably, the genetic construct is adapted for delivery to a human cell.

Means and methods of introducing a genetic construct into a cell in an animal body are known in the art. For example, the constructs of the invention may be introduced into cells by any convenient method, for example methods involving retroviruses, so that the construct is inserted into the genome of the cell. For example, in Kuriyama et al (1991 , Cell Struc. and Func. 16, 503-510) purified retroviruses are administered. Retroviral DNA constructs comprising a polynucleotide as described above may be made using methods

well known in the art. To produce active retrovirus from such a construct it is usual to use an ecotropic psi2 packaging cell line grown in Dulbecco's modified Eagle's medium (DMEM) containing 10% foetal calf serum (FCS). Transfection of the cell line is conveniently by calcium phosphate co-precipitation, and stable transformants are selected by addition of G418 to a final concentration of 1 mg/ml (assuming the retroviral construct contains a neo ^R gene). Independent colonies are isolated and expanded and the culture supernatant removed, filtered through a 0.45 μm pore-size filter and stored at -70 ⁰ C. For the introduction of the retrovirus into tumour cells, for example, it is convenient to inject directly retroviral supernatant to which 10 μg/ml Polybrene has been added. For tumours exceeding 10 mm in diameter it is appropriate to inject between 0.1 ml and 1 ml of retroviral supernatant; preferably 0.5 ml.

Alternatively, as described in Culver et al (1992, Science 256, 1550-1552), cells which produce retroviruses may be injected. The retrovirus-producing cells so introduced are engineered to actively produce retroviral vector particles so that continuous productions of the vector occurred within the tumour mass in situ.

Targeted retroviruses are also available for use in the invention; for example, sequences conferring specific binding affinities may be engineered into pre-existing viral env genes (see Miller & Vile (1995) Faseb J. 9, 190-199, for a review of this and other targeted vectors for gene therapy).

Other methods involve simple delivery of the construct into the cell for expression therein either for a limited time or, following integration into the genome, for a longer time. An example of the latter approach includes liposomes (Nassander et al (1992) Cancer Res. 52, 646-653).

Other methods of delivery include adenoviruses carrying external DNA via an antibody- polylysine bridge (see Curiel (1993) Prog. Med. Virol. 40, 1-18) and transferrin-polycation conjugates as carriers (Wagner et al (1990) Proc. Natl. Acad. Sci. USA 87, 3410-3414). In the first of these methods a polycation-antibody complex is formed with the DNA construct or other genetic construct of the invention, wherein the antibody is specific for either wild-type adenovirus or a variant adenovirus in which a new epitope has been introduced which binds the antibody. The polycation moiety binds the DNA via electrostatic interactions with the phosphate backbone. The adenovirus, because it contains unaltered fibre and penton proteins, is internalised into the cell and carries into

the cell with it the DNA construct of the invention. It is preferred if the polycation is polylysine.

In an alternative method, a high-efficiency nucleic acid delivery system that uses receptor-mediated endocytosis to carry DNA macromolecules into cells is employed. This is accomplished by conjugating the iron-transport protein transferrin to polycations that bind nucleic acids. Human transferrin, or the chicken homologue conalbumin, or combinations thereof is covalently linked to the small DNA-binding protein protamine or to polylysines of various sizes through a disulphide linkage. These modified transferrin molecules maintain their ability to bind their cognate receptor and to mediate efficient iron transport into the cell. The transferrin-polycation molecules form electrophoretically stable complexes with DNA constructs or other genetic constructs of the invention independent of nucleic acid size (from short oligonucleotides to DNA of 21 kilobase pairs). When complexes of transferrin-polycation and the DNA constructs or other genetic constructs of the invention are supplied to the tumour cells, a high level of expression from the construct in the cells is expected.

High-efficiency receptor-mediated delivery of the DNA constructs or other genetic constructs of the invention using the endosome-disruption activity of defective or chemically inactivated adenovirus particles produced by the methods of Cotten et a/ (1992) Proc. Natl. Acad. ScL USA 89, 6094-6098 may also be used. This approach appears to rely on the fact that adenoviruses are adapted to allow release of their DNA from an endosome without passage through the lysosome, and in the presence of, for example transferrin linked to the DNA construct or other genetic construct of the invention, the construct is taken up by the cell by the same route as the adenovirus particle. This approach has the advantages that there is no need to use complex retroviral constructs; there is no permanent modification of the genome as occurs with retroviral infection; and the targeted expression system is coupled with a targeted delivery system, thus reducing toxicity to other cell types.

It will be appreciated that "naked DNA" and DNA complexed with cationic and neutral lipids may also be useful in introducing the DNA of the invention into cells of the individual to be treated. Non-viral approaches to gene therapy are described in Ledley (1995, Human Gene Therapy S, 1129-1144).

Although for solid tumours of specific tissues it may be useful to use tissue-specific promoters in the vectors encoding a polynucleotide inhibitor, this is not essential. This is because the targeted genes are only expressed, or selectively expressed, in the tumour endothelium. Accordingly, expression of ECSM2-specifιc inhibitors such as siRNA, antisense molecules and ribozymes in the body at locations other than the solid tumour would be expected to have no effect since ECSM2 is not expressed. Moreover, the risk of inappropriate expression of these inhibitors, in a cell that may express the target polypeptide at a low level, is miniscule compared to the therapeutic benefit to a patient suffering from a solid tumour.

Targeted delivery systems are also known, such as the modified adenovirus system described in WO 94/10323, wherein, typically, the DNA is carried within the adenovirus, or adenovirus-Iike, particle. Michael et al (1995) Gene Therapy 2, 660-668 describes modification of adenovirus to add a cell-selective moiety into a fibre protein. Mutant adenoviruses which replicate selectively in p53-deficient human tumour cells, such as those described in Bischoff et al (1996) Science 27 '4, 373-376 are also useful for delivering the genetic construct of the invention to a cell. Thus, it will be appreciated that a further aspect of the invention provides a virus or virus-like particle comprising a genetic construct of the invention. Other suitable viruses, viral vectors or virus-like particles include lentivirus and lentiviral vectors, HSV, adeno-assisted virus (AAV) and AAV-based vectors, vaccinia and parvovirus.

Methods of delivering polynucleotides to a patient are well known to a person of skill in the art and include the use of immunoliposomes, viral vectors (including vaccinia, modified vaccinia, adenovirus and adeno-associated viral (AAV) vectors), and by direct delivery of DNA, e.g. using a gene-gun and electroporation. Furthermore, methods of delivering polynucleotides to a target tissue of a patient for treatment are also well known in the art.

Methods of targeting and delivering therapeutic agents directly to specific regions of the body, including the brain, are well known to a person of skill in the art. For example, US Patent No 6,503,242 describes an implanted catheter apparatus for delivering therapeutic agents directly to the hippocampus. Methods of targeting and delivering agents to the brain can be used for the treatment of solid tumours of the brain, such as astrocytoma, ganglioma, ganglioglioma, metastatic adenocarcinoma, glioblastoma and medulloblastoma. In one embodiment, therapeutic agents including vectors can be

distributed throughout a wide region of the CNS by injection into the cerebrospinal fluid, e.g., by lumbar puncture (See e.g., Kapadia et al (1996) Neurosurg 10: 585-587). Alternatively, precise delivery of the therapeutic agent into specific sites of the brain can be conducted using stereotactic microinjection techniques. For example, the subject being treated can be placed within a stereotactic frame base (MRI-compatible) and then imaged using high resolution MRI to determine the three-dimensional positioning of the particular region to be treated. The MRI images can then be transferred to a computer having the appropriate stereotactic software, and a number of images are used to determine a target site and trajectory for microinjection of the therapeutic agent. The software translates the trajectory into three-dimensional coordinates that are precisely registered for the stereotactic frame. In the case of intracranial delivery, the skull will be exposed, burr holes will be drilled above the entry site, and the stereotactic apparatus used to position the needle and ensure implantation at a predetermined depth. The therapeutic agent can be delivered to regions of the CNS such as the hippocampus, cells of the spinal cord, brainstem, (medulla, pons, and midbrain), cerebellum, diencephalon (thalamus, hypothalamus), telencephalon (corpus stratium, cerebral cortex, or within the cortex, the occipital, temporal, parietal or frontal lobes), or combinations, thereof. In another embodiment, the therapeutic agent is delivered using other delivery methods suitable for localised delivery, such as localised permeation of the blood-brain barrier. US Patent Application No 2005/0025746 describes delivery systems for localised delivery of an adeno-associated virus vector (AAV) vector encoding a therapeutic agent to a specific region of the brain.

When a therapeutic agent for the treatment of a solid tumour of, for example, the brain, is enocoded by a polynucleotide, it may be preferable for its expression to be under the control of a suitable tissue-specific promoter. Central nervous system (CNS) specific promoters such as, neuron-specific promoters (e.g., the neurofilament promoter (Byrne and Ruddle, 1989) and glial specific promoters (Morii et al, 1991) are preferably used for directing expression of a polynucleotide preferentially in cells of the CNS. Preferably, the promoter is tissue specific and is essentially not active outside the central nervous system, or the activity of the promoter is higher in the central nervous system than in other cells or tissues. For example, the promoter may be specific for the spinal cord, brainstem, (medulla, pons, and midbrain), cerebellum, diencephalon (thalamus, hypothalamus), telencephalon (corpus stratium, cerebral cortex, or within the cortex, the occipital, temporal, parietal or frontal lobes), or combinations, thereof. The promoter may be specific for particular cell types, such as neurons or glial cells in the CNS. If it is

active in glial cells, it may be specific for astrocytes, oligodentrocytes, ependymal cells, Schwann cells, or microglia. If it is active in neurons, it may be specific for particular types of neurons, e.g., motor neurons, sensory neurons, or intemeurons. The promoter may be specific for cells in particular regions of the brain, for example, the cortex, stratium, nigra and hippocampus.

Suitable neuronal specific promoters include, but are not limited to, neuron specific enolase (NSE; Olivia et al (1991); GenBank Accession No: X51956), and human neurofilament light chain promoter (NEFL; Rogaev et al (1992); GenBank Accession No: L04147). Glial specific promoters include, but are not limited to, glial fibrillary acidic protein (GFAP) promoter (Morii et al (1991); GenBank Accession No:M65210), S100 promoter (Morii et al (1991); GenBank Accession No: M65210) and glutamine synthase promoter (Van den et al (1991); GenBank Accession No: X59834). In a preferred embodiment, the gene is flanked upstream (i.e., 5 ¹ ) by the neuron specific enolase (NSE) promoter. In another preferred embodiment, the gene of interest is flanked upstream (i.e., 5') by the elongation factor 1 alpha (EF) promoter. A hippocampus specific promoter that might be used is the hippocampus specific glucocorticoid receptor (GR) gene promoter.

Alternatively, for treatment of solid tumours of the heart, Svensson et al (1999) describes the delivery of recombinant genes to cardiomyocytes by intramyocardial injection or intracoronary-infusion of cardiotropic vectors, such as recombinant adeno-associated virus vectors, resulting in transgene expression in murine cardiomyocytes in vivo. MeIo et al (2004) review gene and cell-based therapies for heart disease. An alternative preferred route of administration is via a catheter or stent. Stents represent an attractive alternative for localized gene delivery, as they provide a platform for prolonged gene elution and efficient transduction of opposed arterial walls. This gene delivery strategy has the potential to decrease the systemic spread of the viral vectors and hence a reduced host immune response. Both synthetic and naturally occurring stent coatings have shown potential to allow prolonged gene elution with no significant adverse reaction. (Sharif et al, 2004).

It may be desirable to be able to temporally regulate expression of the polynucleotide inhibitor in the cell, although this is not essential for the reasons given above. Thus, it may be desirable that expression of the polynucleotide is directly or indirectly (see below) under the control of a promoter that may be regulated, for example by the concentration

of a small molecule that may be administered to the patient when it is desired to activate or, more likely, repress (depending upon whether the small molecule effects activation or repression of the said promoter) expression of the antibody from the polynucleotide. This may be of particular benefit if the expression construct is stable, i.e., capable of expressing the antibody (in the presence of any necessary regulatory molecules), in the cell for a period of at least one week, one, two, three, four, five, six, eight months or one or more years. Thus the polynucleotide may be operatively linked to a regulatable promoter. Examples of regulatable promoters include those referred to in the following papers: Rivera et al (1999) Proc Natl Acad Sci USA 96(15), 8657-62 (control by rapamycin, an orally bioavailable drug, using two separate adenovirus or adeno- associated virus (AAV) vectors, one encoding an inducible human growth hormone (hGH) target gene, and the other a bipartite rapamycin-regulated transcription factor); Magari et al (1997) J CHn Invest 100(11), 2865-72 (control by rapamycin); Bueler (1999) Biol Chem 380(6), 613-22 (review of adeno-associated viral vectors); Bohl et al (1998) Blood 92(5), 1512-7 (control by doxycycline in adeno-associated vector); Abruzzese et al (1996) J MoI Med 74(7), 379-92 (review of induction factors, e.g.. hormones, growth factors, cytokines, cytostatics, irradiation, heat shock and associated responsive elements).

For veterinary use, the inhibitor is typically administered as a suitably acceptable formulation in accordance with normal veterinary practice and the veterinary surgeon will determine the dosing regimen and route of administration which will be most appropriate for a particular animal.

Combination therapy

According to a National Cancer Institute Press Release dated 14 April 2005, updated 16 June 2005, ("Bevacizumab Combined With Chemotherapy Improves Progression-Free Survival for Patients With Advanced Breast Cancer"), the angiogenesis inhibitor anti- VEGF monoclonal antibody Bevacizumab improves the clinical outcome for a number of solid tumours when administered in combination with standard chemotherapy. Combinations that have been used include bevacizumab in combination with irinotecan, fluorouracil, and leucovorin; bevacizumab in combination with FOLFOX4 (a regimen of oxaliplatin, 5-fluorouracil and leucovorin); bevacizumab in combination with paclitaxel; and bevacizumab in combination with paclitaxel and carboplatin.

It is therefore appreciated that although the inhibitors of ECSM2 may be clinically effective in the absence of any other anti-cancer compound, it may be advantageous to administer these inhibitors in conjunction with a further anticancer agent.

Accordingly, a fourth aspect of the invention provides a pharmaceutical composition comprising: (i) an inhibitor of ECSM2, and (ii) at least one further anticancer agent, and a pharmaceutically acceptable carrier, diluent or excipient.

The further anticancer agent may be selected from alkylating agents including nitrogen mustards such as mechlorethamine (HN ₂ ), cyclophosphamide, ifosfamide, melphalan (L- sarcolysin) and chlorambucil; ethylenimines and methylmelamines such as hexamethylmelamine, thiotepa; alkyl sulphonates such as busulphan; nitrosoureas such as carmustine (BCNU), lomustine (CCNU), semustine (methyl-CCNU) and streptozocin (streptozotocin); and triazenes such as decarbazine (DTIC; dimethyltriazenoimidazole- carboxamide); antimetabolites including folic acid analogues such as methotrexate (amethopterin); pyrimidine analogues such as fluorouracil (5-fluorouracil; 5-FU), floxuridine (fluorodeoxyuridine; FUdR) and cytarabine (cytosine arabinoside); and purine analogues 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 (mithramycin) and mitomycin (mitomycin C); enzymes such as L-asparaginase; and biological response modifiers such as interferon alphenomes; miscellaneous agents including platinum coordination complexes such as cisplatin (c/s-DDP) and carboplatin; anthracenedione such as mitoxantrone and anthracycline; substituted urea such as hydroxyurea; methyl hydrazine derivative such as procarbazine (N-methylhydrazine, MIH); and adrenocortical suppressant such as mitotane (o,p'-DDD) and aminoglutethimide; taxol and analogues/derivatives; cell cycle inhibitors; proteosome inhibitors such as Bortezomib (Velcade ^® ); signal transductase (e.g. tyrosine kinase) inhibitors such as lmatinib (Glivec ^® ), COX-2 inhibitors, and hormone agonists/antagonists such as flutamide and tamoxifen.

The clinically used anticancer agents are typically grouped by mechanism of action:

Alkylating agents, Topoisomerase I inhibitors, Topoisomerase Il inhibitors, RNA/DNA antimetabolites, DNA antimetabolites and Antimitotic agents. The US NIH/National

Cancer Institute website lists 122 compounds (http://dtp.nci.nih.gov/docs/cancer/

searches/standard__mechanism.html), all of which may be used in conjunction with an inhibitor of ECSM2. They include Alkylating agents including Asaley, AZQ, BCNU, Busulfan, carboxyphthalatoplatinum, CBDCA, CCNU, CHIP, chlorambucil, chlorozotocin, c/s-platinum, clomesone, cyanomorpholino-doxorubicin, cyclodisone, dianhydrogalactitol, fluorodopan, hepsulfam, hycanthone, melphalan, methyl CCNU, mitomycin C, mitozolamide, nitrogen mustard, PCNU, piperazine, piperazinedione, pipobroman, porfiromycin, spirohydantoin mustard, teroxirone, tetraplatin, thio-tepa, triethylenemelamine, uracil nitrogen mustard, Yoshi-864; anitmitotic agents including allocolchicine, Halichondrin B, colchicine, colchicine derivative, dolastatin 10, maytansine, rhizoxin, taxol, taxol derivative, thiocolchicine, trityl cysteine, vinblastine sulfate, vincristine sulfate; Topoisomerase I Inhibitors including camptothecin, camptothecin, Na salt, aminocamptothecin, 20 camptothecin derivatives, morpholinodoxorubicin; Topoisomerase Il Inhibitors including doxorubicin, amonafide, m- AMSA, anthrapyrazole derivative, pyrazoloacridine, bisantrene HCL, daunorubicin, deoxydoxorubicin, mitoxantrone, menogaril, N,N-dibenzyl daunomycin, oxanthrazole, rubidazone, VM-26, VP-16; RNA/DNA antimetabolites including L-alanosine, 5- azacytidine, 5-fluorouracil, acivicin, 3 aminopterin derivatives, an antifol, Baker's soluble antifol, dichlorallyl lawsone, brequinar, ftorafur (pro-drug), 5,6-dihydro-5-azacytidine, methotrexate, methotrexate derivative, N-(phosphonoacetyl)-L-aspartate (PALA), pyrazofurin, trimetrexate; DNA antimetabolites including, 3-HP, 2'-deoxy-5-fluorouridine, 5-HP, alpha-TGDR, aphidicolin glycinate, ara-C, 5-aza-2'-deoxycytidine, beta-TGDR, cyclocytidine, guanazole, hydroxyurea, inosine glycodialdehyde, macbecin II, pyrazoloimidazole, thioguanine, thiopurine.

It is preferred if the further anticancer agent is selected from cisplatin, carboplatin, 5- flurouracil, paclitaxel, mitomycin C ₁ doxorubicin, gemcitabine, tomudex, pemetrexed, methotrexate, irinotecan, fluorouracil and leucovorin; or oxaliplatin, 5-fluorouracil and leucovorin; or paclitaxel and carboplatin.

A fifth aspect of the invention provides (i) an inhibitor of ECSM2, and (ii) at least one further anticancer agent as defined above in the fourth aspect of the invention, for use in medicine.

A sixth aspect of the invention provides a method of combating a solid tumour in an individual, the method comprising administering to the patient (i) an inhibitor of ECSM2 in

combination with (ii) at least one further anticancer agent as defined above in the fourth aspect of the invention.

Typically, the method comprises administering to the individual a pharmaceutical composition as defined above in the fourth aspect of the invention. However, it is appreciated that the inhibitor of ECSM2 and the further anticancer agent, may be administered separately, for instance by separate routes of administration. Thus it is appreciated that the inhibitor of ECSM2 and the at least one further anticancer agent can be administered sequentially or (substantially) simultaneously. They may be administered within the same pharmaceutical formulation or medicament or they may be formulated and administered separately.

This aspect of the invention includes the use of (i) an inhibitor of ECSM2 and (ii) at least one further anticancer agent as defined above in the fourth aspect of the invention, in the preparation of a medicament for combating a solid tumour in an individual.

The invention also includes the use of an inhibitor of ECSM2in the preparation of a medicament for combating a solid tumour in an individual who is administered at least one further anticancer agent as defined above in the fourth aspect of the invention. Typically the individual is administered the further anticancer agent at the same time as the medicament, although the individual may have been (or will be) administered the further anticancer agent before (or after) receiving the medicament containing the inhibitor.

The invention further includes the use of at least one further anticancer agent as defined above in the fourth aspect of the invention in the preparation of a medicament for combating a solid tumour in an individual who is administered an inhibitor of ECSM2. Typically the individual is administered the inhibitor at the same time as the medicament, although the patient may have been (or will be) administered the inhibitor before (or after) receiving the medicament containing the further anticancer agent.

The invention also includes (i) an inhibitor of ECSM2 and (ii) at least one further anticancer agent as defined above in the fourth aspect of the invention, for use in combating a solid tumour in an individual.

Suitable inhibitors of ECSM2 for the fourth, fifth and sixth aspects of the invention include

antibodies that selectively bind to the polypeptides, and siRNA, antisense polynucleotides and ribozyme molecules that are specific for the polynucleotides encoding these polypeptides, as discussed in detail above.

General preferences for the solid tumour and for the individual patient to be treated, and for the further anticancer agent, are as described above. However, when the further anticancer agent has been shown to be particularly effective for a specific tumour type, it is preferred if the inhibitor is used in combination with that further anticancer agent to treat that specific tumour type.

Tumour imaging, detection and diagnosis

In a further embodiment, the antibodies that selectively bind the ECSM2 polypeptide, when attached to a detectable moiety, may be useful in imaging, for example vascular imaging of tumours. Methods and compounds useful in vascular imaging of tumours are described in our earlier publication WO 02/36771 , incorporated herein by reference.

A seventh aspect of the invention provides compound comprising an antibody that selectively binds the ECSM2 polypeptide and a detectable moiety.

A compound comprising an anti-ECSM2 antibody as defined above and a detectable moiety can be used, in combination with an appropriate detection method, to detect the location of the compound in the individual, and hence to identify the sites and extent of tumour angiogenesis in the individual, as well as inhibiting the angiogenesis in the individual.

By a "detectable moiety" we include the meaning that the moiety is one which, when located at the target site following administration of the compound of the invention into a patient, may be detected, typically non-invasively from outside the body, and the site of the target located. Thus, the compounds of this aspect of the invention are useful in imaging and diagnosis, especially in the imaging and diagnosis of neovasculature of solid tumours.

Typically, the readily detectable moiety is or comprises a radioactive atom which is useful in imaging. Suitable radioactive atoms include technetium-99m or iodine-123 for scintigraphic studies. Other readily detectable moieties include, for example, spin labels 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. Clearly, the compound of the invention must have sufficient of the appropriate atomic isotopes in order for the molecule to be detectable.

The radio- or other label may be incorporated in the compound in known ways. For example, if the antibody may be biosynthesised or synthesised by chemical amino acid synthesis using suitable amino acid precursors involving, for example, fluorine-19 in place of hydrogen. Labels such as ^99m Tc, ¹²³ I, ¹⁸⁶ Rh, ¹⁸⁸ Rh and ¹¹¹ In can, for example, be attached via cysteine residues in the antibody. Yttrium-90 can be attached via a lysine residue. The IODOGEN method (Fraker et al (1978) Biochem. Biophys. Res. Comm. 80, 49-57) can be used to incorporate iodine-123. The reference ("Monoclonal Antibodies in Immunoscintigraphy", J. F. Chatal, CRC Press, 1989) describes other methods in detail.

The invention further includes a pharmaceutical composition comprising a compound according to this aspect of the invention and a pharmaceutically acceptable carrier, diluent or excipient. Preferences for the pharmaceutical composition are as described above.

An eighth aspect of the invention provides a method of imaging tumour neovasculature in an individual, the method comprising: administering to the individual a compound comprising an antibody that selectively binds the polypeptide ECSM2 and a detectable moiety, and detecting or imaging the location of the detectable moiety in the body.

Preferences for the antibody, the compound and the detectable moiety are as described above.

Typically, the individual has a solid tumour, and the neovasculature of the solid tumour is imaged. This method may be useful, for example, in determining the size of a previously diagnosed solid tumour, the effectiveness of a therapy against the solid tumour, or the extent of metastasis of the tumour. Methods for imaging a detectable moiety in the body are well known in the art, and include PET (positron emission tomography).

A ninth aspect of the invention provides a method of detecting, diagnosing or prognosing a solid tumour in an individual, the method comprising:

administering to the' individual a compound comprising an antibody that selectively binds the polypeptide ECSM2 and a detectable moiety, and detecting the presence and/or location of the detectable moiety in the body.

Preferences for the antibody, the compound and the detectable moiety are as described above. Thus, the localisation of the antibody at a particular organ in the body indicates that the individual may have or may be developing a solid tumour at that organ.

Targeted delivery of cytotoxic agents One avenue towards the development of more selective, better anticancer drugs is the targeted delivery of bioactive molecules to the tumour environment by means of binding molecules (for example, human antibodies) that are specific for tumour-endothelial markers. Due to their accessibility and to the therapeutic options that they allow (for example, intraluminal blood coagulation or recruitment of immune cells), vascular markers selectively expressed on tumour blood vessels seem to be ideally suited for ligand-based tumour-targeting strategies, opening new possibilities for the imaging and the therapy of cancer.

Accordingly, a tenth aspect of the invention provides a compound comprising an antibody that selectively binds the polypeptide ECSM2 and a cytotoxic moiety.

Typically the cytotoxic moiety is selected from a directly cytotoxic chemotherapeutic agent, a directly cytotoxic polypeptide, a moiety which is able to convert a prodrug into a cytotoxic drug, a radiosensitizer, a directly cytotoxic nucleic acid, a nucleic acid molecule that encodes a directly or indirectly cytotoxic polypeptide or a radioactive atom. Examples of such cytotoxic moieties, as well as methods of making the conjugates comprising the antibody and the cytotoxic moiety, are provided in our earlier publications WO 02/36771 and WO 2004/046191 , incorporated herein by reference.

The cytotoxic moiety may be directly or indirectly toxic to cells in neovasculature or cells which are in close proximity to and associated with neovasculature. By "directly cytotoxic" we include the meaning that the moiety is one which on its own is cytotoxic. By "indirectly cytotoxic" we include the meaning that the moiety is one which, although is not itself cytotoxic, can induce cytotoxicity, for example by its action on a further molecule or by further action on it.

In one embodiment the cytotoxic moiety is a cytotoxic chemotherapeutic agent. Cytotoxic chemotherapeutic agents are well known in the art. Cytotoxic chemotherapeutic agents, such as anticancer agents, include those listed above with respect to the seventh aspect of the invention.

Various of these cytotoxic moieties, such as cytotoxic chemotherapeutic agents, have previously been attached to antibodies and other targeting agents, and so compounds of the invention comprising these agents may readily be made by the person skilled in the art. For example, carbodiimide conjugation (Bauminger & Wilchek (1980) Methods Enzymol. 70, 151-159) may be used to conjugate a variety of agents, including doxorubicin, to antibodies. Other methods for conjugating a cytotoxic moiety to an antibody can also be used. For example, sodium periodate oxidation followed by reductive alkylation of appropriate reactants can be used, as can glutaraldehyde cross- linking. Methods of cross-linking polypeptides are known in the art and described in WO 2004/046191. However, it is recognised that, regardless of which method of producing a compound of the invention is selected, a determination must be made that the antibody maintains its targeting ability and that the attached moiety maintains its relevant function.

In a further embodiment of the invention, the cytotoxic moiety may be a cytotoxic peptide or polypeptide moiety by which we include any moiety which leads to cell death.

Cytotoxic peptide and polypeptide moieties are well known in the art and include, for example, ricin, abrin, Pseudomonas exotoxin, tissue factor and the like. Methods for linking them to targeting moieties such as 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 localised blood clotting and infarction of a tumour, has been described by Ran et al (1998) Cancer Res.

58, 4646^653 and Huang et al (1997) Science 275, 547-550. Tsai et al (1995) Dis.

Colon Rectum 38, 1067-1074 describes the abrin A chain conjugated to a monoclonal antibody. Other ribosome inactivating proteins are described as cytotoxic agents in WO 96/06641. Pseudomonas exotoxin may also be used as the cytotoxic polypeptide moiety

(see, for example, Aiello et a/ (1995) Proc. Natl. Acad. Sci. USA 92, 10457-10461).

Certain cytokines, such as TNFα and IL-2, may also be useful as cytotoxic agents.

Certain radioactive atoms may also be cytotoxic if delivered in sufficient doses. Thus, the cytotoxic moiety may comprise a radioactive atom which, in use, delivers a sufficient

quantity of radioactivity to the target site so as to be cytotoxic. Suitable radioactive atoms include phosphorus-32, iodine-125, iodine-131 , indium-111 , rhenium-186, rhenium-188 or yttrium-90, or any other isotope which emits enough energy to destroy neighbouring cells, organelles or nucleic acid. Preferably, the isotopes and density of 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 target site and, preferably, to the cells at the target site and their organelles, particularly the nucleus.

The radioactive atom may be attached to the antibody in known ways. For example EDTA or another chelating agent may be attached to the antibody and used to attach

11 ¹ 1Iιn or ⁹⁰ Y. Tyrosine residues may be labelled with ¹²⁵ I or ¹³¹ I.

The cytotoxic moiety may be a radiosensitizer. Radiosensitizers include fluoropyrimidines, thymidine analogues, hydroxyurea, gemcitabine, fludarabine, nicotinamide, halogenated pyrimidines, 3-aminobenzamide, 3-aminobenzodiamide, etanixadole, 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. Radial Biol. 56, 827-836; lliakis & Kurtzman

(1989) Int. J. Radial Oncol. Biol. Phys. 16, 1235-1241; Brown (1989) Int. J. Radial

Oncol. Biol. Phys. 16, 987-993; Brown (1985) Cancer 55, 2222-2228).

The cytotoxic moiety may be an indirectly cytotoxic polypeptide. In a particularly preferred embodiment, the indirectly cytotoxic polypeptide is a polypeptide which has enzymatic activity and can convert a relatively non-toxic prodrug into a cytotoxic drug.

When the targeting moiety is an antibody this type of system is often referred to as

ADEPT (Antibody-Directed Enzyme Prodrug Therapy). The system requires that the targeting moiety locates the enzymatic portion to the desired site in the body of the patient (e.g. the site of new vascular tissue associated with a tumour) and after allowing time for the enzyme to localise at the site, administering a prodrug which is a substrate for the enzyme, the end product of the catalysis being a cytotoxic compound. The object of the approach is to maximise the concentration of drug at the desired site and to minimise the concentration of drug in normal tissues (Senter et al (1988) "Anti-tumor 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, et al (1988) "A cytotoxic agent can be generated selectively at cancer sites" Br. J. Cancer. 58, 700-703.) Bagshawe (1995) Drug Dev. Res. 34, 220-230 and WO 2004/046191 , both of which are incorporated herein by reference, describe various enzyme/prodrug combinations which may be suitable in the context of this invention.

Typically, the prodrug is relatively non-toxic compared to the cytotoxic drug. Typically, it has less than 10% of the toxicity, preferably less than 1% of the toxicity as measured in a suitable in vitro cytotoxicity test.

It is likely that the moiety which is able to convert a prodrug to a cytotoxic drug will be active in isolation from the rest of the compound but it is necessary only for it to be active when (a) it is in combination with the rest of the compound and (b) the compound is attached to, adjacent to or internalised in target cells.

The further moiety may be one which becomes cytotoxic, or releases a cytotoxic moiety, upon irradiation. For example, the boron-10 isotope, when appropriately irradiated, releases α particles which are cytotoxic (see for example, US 4,348,376 to Goldenberg; Primus et al (1996) Bioconjug. Chem. 7, 532-535).

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

The invention further includes a compound according to the tenth aspect of the invention for use in medicine. The invention also includes a pharmaceutical composition comprising a compound according to the tenth aspect of the invention and a pharmaceutically acceptable carrier, diluent or excipient. Preferences for the formulation of pharmaceutical compositions are as described above.

It is appreciated that the compounds according to the tenth aspect of the invention can be used to inhibit tumour angiogenesis in an individual and to treat a solid tumour as discussed above with respect to the first and second aspects of the invention.

Thus, an eleventh aspect of the invention provides a method of inhibiting tumour angiogenesis in an individual, the method comprising administering to the individual a compound according to the tenth aspect of the invention.

A twelfth aspect of the invention provides method of combating a solid tumour in an individual, the method comprising administering to the individual a compound according to the tenth aspect of the invention.

The invention includes the use of a compound according to the tenth aspect of the invention in the preparation of a medicament for inhibiting tumour angiogenesis or for combating a solid tumour in an individual in an individual.

For the eleventh and twelfth aspects of the invention, preferences for the compound, the cytotoxic moiety, the individual to be treated, the types of solid tumour, the routes of administration, and so on are as defined above.

Screening A thirteenth aspect of the invention provides a method of identifying an agent that may be useful in the treatment of a solid tumour, or a lead compound for the identification of an agent that may be useful in the treatment of a solid tumour, the method comprising: providing a candidate compound that binds the polypeptide ECSM2, or a fragment thereof; and testing the candidate compound in an angiogenesis assay, wherein a candidate compound that inhibits angiogenesis in the assay may be an agent that is useful in the treatment of a solid tumour, or may be a lead compound for the identification of an agent that is useful in the treatment of a solid tumour.

A fourteenth aspect of the invention provides a method of identifying an agent that may be useful in the treatment of a solid tumour, or a lead compound for the identification of an agent that may be useful in the treatment of a solid tumour, the method comprising: providing a candidate compound; determining whether the candidate compound selectively binds to the polypeptide ECSM2, or a fragment thereof; and testing a candidate compound that selectively binds to the ECSM2 polypeptide or the fragment in an angiogenesis assay, wherein a candidate compound that selectively binds to the said ECSM2 polypeptide or fragment and which inhibits angiogenesis in the assay may be an agent that is useful in the treatment of a solid tumour, or may be a lead compound for the identification of an agent that is useful in the treatment of a solid tumour.

It is appreciated that these methods can be used to identify an anti-angiogenic factor, which may be an anti-cancer agent.

By an ECSM2 polypeptide we include a polypeptide having the sequence listed in Figure 5 (SEQ ID NO: 1), and naturally occurring variants thereof. It is appreciated that for the binding assay, it is not necessary to use a polypeptide having 100% sequence identity to the sequences listed in Figure 5 (whether over the full-length polypeptide or the fragment thereof). Accordingly, in this aspect of the invention it is possible to use a variant polypeptide having at least 80%, more preferably at least 85%, still more preferably at least 90%, yet more preferably at least 95%, or at least 96%, or at least 97%, or at least 98%, or at least 99% sequence identity with the sequences listed in Figure 5. It is preferred if the variant polypeptide has a consecutive region of at least 20 amino acid residues, more preferably at least 50 residues, of the sequence of the polypeptide listed in Figure 5. Such variants may be made, for example, using the methods of recombinant DNA technology, protein engineering and site-directed mutagenesis which are well known in the art.

The percent sequence identity between two polypeptides may be determined using suitable computer programs, for example the GAP program of the University of Wisconsin Genetic Computing Group and it will be appreciated that percent identity is calculated in relation to polypeptides whose sequence has been aligned optimally. The alignment may alternatively be carried out using the Clustal W program (Thompson et a/., (1994) Nucleic Acids Res 22, 4673-80). The parameters used may be as follows: Fast pairwise alignment parameters: K-tuple(word) size; 1 , window size; 5, gap penalty; 3, number of top diagonals; 5. Scoring method: x percent. Multiple alignment parameters: gap open penalty; 10, gap extension penalty; 0.05. Scoring matrix: BLOSUM.

It is also appreciated that in order to determine whether a candidate compound binds to a specified polypeptide, it is not necessary to use the entire full-length polypeptide in the binding assay, and fragments of the polypeptide may be usefully employed. Preferably, the fragment is at least 20 amino acid residues in length, and may be between 20 and 50 residues or between 50 and 100 residues or between 100 and 150 residues or between 150 and 200 residues in length, or more.

In an embodiment, the candidate compound may be an antibody that selectively binds the ECSM2 polypeptide, or a fragment thereof. Suitable antibodies are described above.

In another embodiment, the candidate compound may be a peptide. Suitable peptides that bind to the ECSM2 polypeptide, or a fragment thereof, may be identified by methods such as phage display of peptide libraries (Scott & Smith (1990) "Searching for peptide ligands with an epitope library." Science 249: 386-390; Felici ef al (1995) "Peptide and protein display on the surface of filamentous bacteriophage." Biotechnol. Annu. Rev. 1: 149-183); and Collins et al (2001) "Cosmix-plexing: a novel recombinatorial approach for evolutionary selection from combinatorial libraries." J. Biotechnol. 74: 317-338); including in vivo panning (Pasqualini et al (1997) "αv integrins as receptors for tumor targeting by circulating ligands. Nature Biotechnol. 15: 542-546), and solid-phase parallel synthesis (Frank (2002) "The SPOT-synthesis technique. Synthetic peptide arrays on membrane supports — principles and applications." J. Immunol. Methods 267: 13-26; and Pinilla et al (2003) "Advances in the use of synthetic combinatorial chemistry: mixture-based libraries." Nature Med. 9: 118-122). The dissociation constants of peptides are typically in the micromolar range, although avidity can be improved by multimerization (Terskikh ef al (1997) "Peptabody": a new type of high avidity binding protein. Proc. Natl Acad. Sci. USA 94, 1663-1668; and Wrighton ef al (1997) "Increased potency of an erythropoietin peptide mimetic through covalent dimerization. Nature Biotechnol. 15, 1261-1265).

In still another embodiment, the candidate compound may be an aptamer, i.e. a single- stranded DNA molecule that folds into a specific ligand-binding structure. Suitable aptamers that bind to the ECSM2 polypeptide, or a fragment thereof, may be identified by methods such as in vitro selection and amplification (Ellington & Szostak (1992) "Selection in vitro of single stranded DNA molecules that fold into specific ligand binding structures." Nature 355: 850-852; and Daniels ef al (2003) "A tenascin-C aptamer identified by tumor cell SELEX: systematic evolution of ligands by exponential enrichment." Proc. Natl Acad. Sci. USA 100, 15416-15421). The aptamer may be a nuclease-stable 'Spiegelmer ¹ (Helmling, S. ef al (2004) "Inhibition of ghrelin action in vitro and in vivo by an RNA-Spiegelmer." Proc. Natl Acad. Sci. USA 101 : 13174-13179). Aptamers typically have dissociation constants in the micromolar to the subnanomolar range.

In yet another embodiment, the candidate compound may be a small organic molecule. Suitable small molecule that bind to the ECSM2 polypeptide, or a fragment thereof, may be identified by methods such as screening large libraries of compounds (Beck-Sickinger & Weber (2001) Combinational Strategies in Biology and Chemistry (John Wiley & Sons, Chichester, Sussex); by structure-activity relationship by nuclear magnetic resonance (Shuker et al (1996) "Discovering high-affinity ligands for proteins: SAR by NMR. Science 27 '4: 1531-1534); encoded self-assembling chemical libraries Melkko et al (2004) "Encoded self-assembling chemical libraries." Nature Biotechnol. 22: 568-574); DNA- templated chemistry (Gartner et al (2004) "DNA-tem plated organic synthesis and selection of a library of macrocycles. Science 305: 1601-1605); dynamic combinatorial chemistry (Ramstrom & Lehn (2002) "Drug discovery by dynamic combinatorial libraries." Nature Rev. Drug Discov. 1 : 26-36); tethering (Arkin & Wells (2004) "Small-molecule inhibitors of protein-protein interactions: progressing towards the dream. Nature Rev. Drug Discov. 3: 301-317); and speed screen (Muckenschnabel et al (2004) "SpeedScreen: label-free liquid chromatography-mass spectrometry-based high- throughput screening for the discovery of orphan protein ligands." Anal. Biochem. 324: 241-249). Typically, small organic molecules will have a dissociation constant for the polypeptide in the nanomolar range, particularly for antigens with cavities. The benefits of most small organic molecule binders include their ease of manufacture, lack of immunogenicity, tissue distribution properties, chemical modification strategies and oral bioavailability.

The capability of a candidate compound to bind to or interact with the ECSM2 polypeptide or fragment thereof may be measured by any method of detecting/measuring a protein/protein interaction or other compound/protein interaction, as discussed further below. Suitable methods include methods such as, for example, yeast two-hybrid interactions, co-purification, ELISA, co-immunoprecipitation and surface plasmon resonance methods. Thus, the candidate compound may be considered capable of binding to the polypeptide or fragment thereof if an interaction may be detected between the candidate compound and the polypeptide or fragment thereof by ELISA, co- immunoprecipitation or surface plasmon resonance methods or by a yeast two-hybrid interaction or copurification method. It is preferred that the interaction can be detected using a surface plasmon resonance method. Surface plasmon resonance methods are well known to those skilled in the art. Techniques are described in, for example, O'Shannessy DJ (1994) "Determination of kinetic rate and equilibrium binding constants for macromolecular interactions: a critique of the surface plasmon resonance literature"

Curr Opin Biotechnol. 5(1):65-71 ; Fivash et a/ (1998) "BIAcore for macromolecular interaction." Curr Opin Biotechnol. 9(1):97-101 ; Malmqvist (1999) "BIACORE: an affinity biosensor system for characterization of biomolecuiar interactions." Biochem Soc Trans. 27(2):335-40.

It is appreciated that screening assays which are capable of high throughput operation are particularly preferred. Examples may include cell based assays and protein-protein binding assays. An SPA-based (Scintillation Proximity Assay; Amersham International) system may be used. For example, an assay for identifying a compound capable of modulating the activity of a protein kinase may be performed as follows. Beads comprising scintillant and a substrate polypeptide that may be phosphorylated may be prepared. The beads may be mixed with a sample comprising the protein kinase and 3 ² P-ATP or ³³ P-ATP and with the test compound. Conveniently this is done in a multi- well (e.g., 96 or 384) format. The plate is then counted using a suitable scintillation counter, using known parameters for ³² P or ³³ P SPA assays. Only ³² P or ³³ P that is in proximity to the scintillant, i.e. only that bound to the polypeptide, is detected. Variants of such an assay, for example in which the polypeptide is immobilised on the scintillant beads via binding to an antibody or antibody fragment, may also be used.

Other methods of detecting polypeptide/polypeptide interactions include ultrafiltration with ion spray mass spectroscopy/HPLC methods or other physical and analytical methods. Fluorescence Energy Resonance Transfer (FRET) methods, for example, well known to those skilled in the art, may be used, in which binding of two fluorescent labelled entities may be measured by measuring the interaction of the fluorescent labels when in close proximity to each other.

A further method of identifying a compound that is capable of binding to the ECSM2 polypeptide or fragment thereof is one where the polypeptide is exposed to the compound and any binding of the compound to the said polypeptide is detected and/or measured. The binding constant for the binding of the compound to the ECSM2 polypeptide may be determined. Suitable methods for detecting and/or measuring (quantifying) the binding of a compound to a polypeptide are well known to those skilled in the art and may be performed, for example, using a method capable of high throughput operation, for example a chip-based method. Technology, called VLSIPS™, has enabled the production of extremely small chips that contain hundreds of thousands or more of different molecular probes. These biological chips or arrays have probes

arranged in arrays, each probe assigned a specific location. Biological chips have been produced in which each location has a scale of, for example, ten microns. The chips can be used to determine whether target molecules interact with any of the probes on the chip. After exposing the array to target molecules under selected test conditions, scanning devices can examine each location in the array and determine whether a target molecule has interacted with the probe at that location.

It is appreciated that the identification of a candidate compound that binds to the polypeptide or fragment thereof may be an initial step in the drug screening pathway, and the identified compounds may be further selected e.g. for the ability to inhibit angiogenesis.

By "inhibiting angiogenesis" we include the meaning of reducing the rate or level of angiogenesis. The reduction can be a low level reduction of about 10%, or about 20%, or about 30%, or about 40% of the rate or level of angiogenesis. Preferably, the reduction is a medium level reduction of about 50%, or about 60%, or about 70%, or about 80% reduction of the rate or level of angiogenesis. More preferably, the reduction is a high level reduction of about 90%, or about 95%, or about 99%, or about 99.9%, or about 99.99% of the rate or level of angiogenesis. Most preferably, inhibition can also include the elimination of angiogenesis or its reduction to an undetectable level.

Methods and assays for determining the rate or level of angiogenesis, and hence for determining whether and to what extent a test compound inhibits angiogenesis, are known in the art. For example, US Patent No. 6,225,118 to Grant ef a/, incorporated herein by reference, describes a multicellular ex vivo assay for modelling the combined stages of angiogenesis namely the proliferation, migration and differentiation stages of cell development. The AngioKit, Catalogue No. ZHA-1000, by TCS CellWorks Ltd, Buckingham MK18 2LR, UK, is a suitable model of human angiogenesis for analysing the anti-angiogenic properties of compounds. The rate or level of angiogenesis can also be determined using the aortic ring assay and the sponge angiogenesis assay that are well known in the art.

Assays for endothelial cell proliferation, migration and invasion are also useful as angiogenesis assays. Suitable assays for endothelial cell proliferation and migration are known to a person of skill in the art and are described herein. Suitable assays for endothelial cell invasion are also known to a person of skill in the art and include the BD

BioCoat™ Angiogenesis System for Endothelial Cell Invasion which is available as Catalogue Nos. 354141 and 354142 from BD Biosciences, Bedford, MA, USA.

We also consider that a candidate compound that selectively binds to the ECSM2 polypeptide may inhibit migration of tumour endothelial cells, including bFGF- and VEGF- induced migration, inhibit proliferation of tumour endothelial cells, or invasion of tumour endothelial cells. Accordingly, candidate compounds that show inhibitory activity in the HUVEC migration assay, or that show anti-proliferative activity, or that show anti-invasive activity in an assay such as the BD BioCoat™ Angiogenesis System for Endothelial Cell Invasion (BD Biosciences, Bedford, MA, USA), may be therapeutically useful in combating solid tumours in which tumour endothelial cell migration, proliferation or invasion contributes to the angiogenesis of neovasculature and hence the pathology of solid tumours.

It is appreciated that these methods may be a drug screening methods, a term well known to those skilled in the art, and the candidate compound may be a drug-like compound or lead compound for the development of a drug-like compound.

The term "drug-like compound" is well known to those skilled in the art, and may include the meaning of a compound that has characteristics that may make it suitable for use in medicine, for example as the active ingredient in a medicament. Thus, for example, a drug-like compound may be a molecule that may be synthesised by the techniques of organic chemistry, less preferably by techniques of molecular biology or biochemistry, and is preferably a small molecule, which may be of less than 5000 Daltons and which may be water-soluble. A drug-like compound may additionally exhibit features of selective interaction with a particular protein or proteins and be bioavailable and/or able to penetrate target cellular membranes or the blood:brain barrier, but it will be appreciated that these features are not essential.

The term "lead compound" is similarly well known to those skilled in the art, and may include the meaning that the compound, whilst not itself suitable for use as a drug (for example because it is only weakly potent against its intended target, non-selective in its action, unstable, poorly soluble, difficult to synthesise or has poor bioavailability) may provide a starting-point for the design of other compounds that may have more desirable characteristics.

In an embodiment, the identified compound is modified, and the modified compound is tested for the ability to inhibit angiogenesis. Suitable assays for the inhibition of angiogenesis are described above.

It is appreciated that the screening methods can be used to identify agents that may be useful in combating solid tumours. Thus, the screening methods preferably also comprise the further step of testing the identified compound or the modified compound for efficacy in an animal model of cancer, particularly a solid tumour. Suitable models are known in the art and include Lewis lung carcinoma subcutaneous implants in mice (homograft in Black 57 mice) or HT29 xenografts subcutaneous implants in nude mice.

The invention may comprise the further step of synthesising, purifying and/or formulating the identified compound or the modified compound.

The invention may further comprise the step of formulating the compound identified into a pharmaceutically acceptable composition.

Compounds may also be subjected to other tests, for example toxicology or metabolism tests, as is well known to those skilled in the art.

Thus the invention includes a method for preparing an anticancer compound that may be useful in the treatment of a solid tumour, the method comprising identifying a compound using the screening methods described above and synthesising, purifying and/or formulating the identified compound.

Thus, the invention also includes a method of making a pharmaceutical composition comprising the step of mixing the compound identified using the methods described above with a pharmaceutically acceptable carrier.

All of the documents referred to herein are incorporated herein, in their entirety, by reference.

The listing or discussion of an apparently prior-published document in this specification should not necessarily be taken as an acknowledgement that the document is part of the state of the art or is common general knowledge

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

Figure 1 : A Venn diagram pictorial representation of the analyses carried out to predict TEMs. The results of an endothelial and tumour screen are combined to produce putative TEMs.

Figure 2: An overview of the EST-to-gene assignment process. Each EST sequence is BLAST searched against a Refseq mRNA database and the best mRNA is assigned that EST. In tandem, a mapping of all ESTs and Refseq mRNA to the human genome assigns ESTs to genes based on genome position. A decision tree makes the final assignment based on the quality of alignment and agreement between the two methods. If the genome position and BLAST result agree, the EST is assigned, if they don't agree but the BLAST result if of high quality (> 92% and > 100 bp alignment) the EST is also assigned. For any other result the EST is removed from the analysis.

Figure 3: Real time PCR was carried out on predicted endothelial genes and results show the power of the bioinformatics models as all genes examined were up-regulated or specific to HUVECs and/or HDMECs.

Figure 4: Real time PCR was carried out on predicted endothelial genes and results show the power of the bioinformatics models as all genes examined were up-regulated or specific to HUVECs and/or HDMECs.

Figure 5: Polypeptide sequence of human ECSM2 from Genbank Accession No. NP_001071161 (SEQ ID NO: 1) depicting the signal peptide (in bold) and the transmembrane region (underlined), and cDNA of human ECSM2 from Genbank Accession No. DQ462572 (SEQ ID NO: 2).

Figure 6: In situ hybridisation using an ECSM2 specific probe on sections of solid brain tumours and normal brain control. (A): ganglioma, (B): astrocytoma, (C): normal brain.

Figure 7: Effect of ECSM2-specific siRNA duplexes on HUVEC (Figure 7A) and MRC5 (Figure 7B) cell proliferation. Growth curves were measured by release of cells on exposure to trypsin followed by counting in a Coulter counter (n=4 ± SD). The experiment was repeated three times with similar results.

Figure 8: Quantitative PCR analysis of ECSM2 expression by cell lines in vitro. (A): The level of ECSM2 expression in four endothelial and 8 non-endothelial cell lines; HUVEC (human umbilical vein endothelial cells), HDMEC (human dermal microvascular endothelial cells), MEC (myometrial microvascular endothelial cells), NEC (normal endometrial microvascular endothelial cells), JURKAT (lymphoblast, acute T cell leukaemia), HELF (human embryonic lung fibroblast), MRC-5 (lung fibroblast), HL60 (peripheral blood leukocyte, acute promyelocytic leukaemia), MCF-7 (breast adenocarcinoma), U87-mg (glioblastoma-astrocytoma), SK-N-SH (neuroblastoma), SK23 (melanoma). Expression levels are shown relative to the level in HUVEC which was arbitrarily set at 1 , + SD. (B): the level of ECSM2 expression in human primary cell isolates including two endothelial (HUVEC and HDMEC) and five non-endothelial isolates: HBEC (human bronchial epithelial cells), hepatocytes, MRC-5 fibroblasts and peripheral blood lymphocytes. Data was analyzed using the method described by Pfaffl [48] and is shown + SE.

Figure 9: In situ hybridization analysis of ECSM2 expression in human tissues. Light-field (A, C, E, G, I and K) and dark-field (B, D, F, H, J and L). ECSM2 expression was restricted to the endothelium in all tissues analyzed. A and B, expression of ECSM2 in large and small vessels of breast carcinoma

(some small vessels are indicated by arrows in A).

C and D, ECSM2 expression restricted to the endothelium of ganglioglioma tissue (some small vessels are indicated by arrows in C).

E and F, ECSM2 expression in vessels of the skin from a psoriasis biopsy (arrows indicate vessels in E).

G and H, ECSM2 expression in the endothelium of placental tissue, (the line of corresponding arrows in G and H indicates ECSM2 expression by individual endothelial cells, the single arrow in G shows a transversely cut vessel).

I and J, endothelial expression of ECSM2 in fetal tissue (arrows in I indicate small vessels and vessel walls).

K and L, ECSM2 expression in breast carcinoma with vessels also stained for the endothelial marker CD34 (indicated by the brown staining in K), confirming the endothelial restricted expression of ECSM2.

Example 1 : A Method for Accurate Expressed Sequence Tag to Gene Assignment and a Novel Statistical Analysis of Differential Gene Expression Across Multiple

cDNA Libraries Applied to the Identification of Endothelial and Tumour Endothelial Genes.

Summary In this study, differential gene expression analysis using complementary DNA (cDNA) libraries has been improved by the introduction of an accurate method of assigning Expressed Sequence Tags (ESTs) to genes and a novel maximum likelihood statistical scoring of differential gene expression between two pools of cDNA libraries. These methods were applied to the latest available cell line and bulk tissue cDNA libraries in a two-step screen to predict novel tumour endothelial markers. Initially, endothelial cell lines were subtracted in silico from non-endothelial cell lines to identify endothelial genes. Subsequently, a second bulk tumour versus normal tissue subtraction was employed to predict tumour endothelial markers.

From an endothelial cDNA library analysis, 431 genes were found to be significantly up- regulated in endothelial cells with a False Discovery Rate adjusted q-value of 0.01 or less, and 104 of these were expressed only in endothelial cells. Combining the cDNA library data with the latest Serial Analysis of Gene Expression (SAGE) library data derived a complete list of 459 genes preferentially expressed in endothelium. 27 genes were predicted tumour endothelial markers in multiple tissues based on the second bulk tissue screen.

This ability to accurately assign an EST to a gene, statistically measure differential expression between two pools of cDNA libraries and predict putative tumour endothelial markers before entering the laboratory represents a significant advance.

Background

Study aim

The growth and survival of tumours is dependent on their ability to obtain a blood supply and damage inflicted on the tumour endothelium has been shown to effectively eradicate tumours [1]. It follows that the discovery of widely expressed tumour endothelial markers promises much clinical benefit [2]. The aim of this study was to apply novel bioinformatic methods to the latest public expression data repositories, with an emphasis on cDNA library analysis, to create an up-to-date list of putative endothelial genes and to predict tumour endothelial markers that are potential anti-cancer targets.

Previous work employing in silico analysis to identify tissue specific genes Previous studies [3-15] have employed cDNA or SAGE libraries to predict the transcriptional profiles of tissues of interest that were subsequently confirmed by experimental analysis. Our analysis [8] employed a cDNA subtractive Basic Local Alignment Search Tool (BLAST) [16] algorithm to predict endothelial specific genes. This approach required cross-referencing of the results to SAGE libraries to confidently predict endothelial expression due to a large number of false positives associated with the BLAST method of EST-to-gene assignment. In one study [7], Unigene's Digital Differential Display (DDD) tool was employed to predict endothelial genes, which is reliant on Unigene clusters. DDD requires at least 1 ,000 EST sequences from a cDNA library to be clustered into Unigene clusters for valid statistical analysis and can measure statistical significance accurately between only two libraries [15]. This 1 ,000 sequence limit of DDD can remove small, but often potentially relevant, cDNA libraries from an analysis.

Improving cDNA library analysis

We aimed to improve both the statistical analysis and EST-to-gene assignment methods used in subtractive in silico cDNA differential gene expression analyses. To eliminate the cDNA library false positive discovery rate of our previous study [8], EST-to-gene assignment was improved by combining human genome BLAST Like Alignment Tool (BLAT) [17], alignment data with a new BLAST protocol. Further, a new maximum likelihood ratio test was developed that is based upon the intrinsic variability of cDNA library counts and represents a maximally powerful approach to analyze this type of data. This method alleviated the need for the 1 ,000 Unigene cluster limit of the DDD tool, enabled any size cDNA library to be analysed and accurately determined differential expression across more than two cDNA libraries.

In silico Tumour Endothelial Marker prediction

A two-step analysis was performed to predict tumour endothelial markers (TEMs). The first stage identified endothelial genes by comparing the expression patterns of genes between endothelial and non-endothelial cell lines. The second stage involved a comparison of bulk tumour and bulk normal cDNA libraries to identify genes up-regulated in tumours. Putative TEMs are genes that were both endothelial and preferentially expressed in tumours. The Venn diagram in Figure 2 summarises the analysis.

Materials and Methods

Construction of databases

A large part of this study involved the collection and processing of data in the public domain with speed and accuracy, in particular the creation and use of a Relational Database Management System (RDBMS) MySQL database called dbestlibraries. The database was central to all processes in tandem with Perl scripts, which were written for the import of data, assignment of EST-to-gene symbols and the accurate calculation of the FDR-adjusted q-value results.

Data was collected from Genbank flat files (release 154) downloaded from the NCBI ftp://ftp.ncbi.nih.gov/genbank/ that supplied all cDNA library data imported into the database. 10,788 libraries containing 8,003,786 ESTs were imported into the database.

Information concerning 29,367 human reference sequence project mRNAs and gene predictions were downloaded from release 14 of the Reference sequence project at

(ftp://ftp.ncbi.nih.gov/refseq). Finally, all information relating to Refseq sequences was downloaded into the database from ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/.

Selection of EST library pools

The CGAP library finder: (http://cgap.nci.nih.gov/Tissues/LibraryFinder) was used as a tool for choosing which libraries to compare in tumour and endothelial screens. Additional endothelial cDNA libraries were discovered, using a Peri script to parse raw Genbank flat files, which identified libraries with keywords such as "cell lines" and "endothelial". Normalised or subtracted libraries were excluded from this analysis.

Normal versus tumour tissue screen Bulk tumour and normal cDNA libraries for six organs were chosen using the CGAP library browser. The combined algorithm was employed to perform virtual subtraction hybridisation between tumour and normal libraries of the same organ. All results were imported into the dbestlibraries database. Results with an FDR-adjusted q-value of 0.01 were significant.

cDNA library screen (EST-to-qene assignment)

To perform in silico virtual subtraction, two different protocols for assigning an EST to a gene were combined for greatest accuracy. The first protocol took advantage of the almost complete human genome by using genome address to assign an EST to a gene. A genome address of a gene or EST is the physical base pair position it occupies on a chromosome. Both cDNA pools and all Refseq mRNAs were aligned to the human

genome using BLAT to generate genome addresses. The BLAT alignment genome addresses were clustered using a Perl algorithm called the Jake cluster algorithm to identify EST sequences that overlapped with a gene and to assign them. To save processing time using BLAT, the human genome addresses of Refseq genes and ESTs were downloaded from the University of California Santa Cruz (UCSC) table browser page (http://genome.ucsc.edu/cgi-bin/hgTables). This file contained the, pre-processed BLAT output [17]. BLAT is designed to rapidly align DNA sequences that are 95% identical or more, over at least 40 base pairs.

For the second method of assigning an EST to a gene, each EST from both cDNA library pools was collected as a FASTA sequence and BLAST searched against a database of all Refseq mRNAs. An expectation cut-off of 1 was employed and the -v and -b BLAST options were set to 1. This ensured that only the best mRNA that matched the EST was returned in the BLAST results.

A Perl script algorithm (Jake cluster) was constructed to combine the results of the genome BLAT address with the BLAST search method. If the genome address assignment agreed with the BLAST result, then the EST was assigned to the gene; if they disagreed, only a high quality BLAST result allowed EST-to-gene assignment (>= 92% identity, >= 100 bp alignment).

Combining cDNA and SAGE library analysis

For Experiments 2 and 3 described in the results section, the cDNA analysis was combined with a SAGE library analysis for endothelial gene prediction. The SAGEmap xProfiler tool at NCBI was used for this: http://www.ncbi. nlm.nih.gov/projects/SAGE/index.cgi?cmd=expsetup. No SAGE analyses were carried out for the tumour screen, as there were insufficient bulk tumour or bulk normal libraries SAGE libraries available.

The SAGE experiments were performed using a fold difference factor of 10 and a 0% coefficient of variance cut off. Only genes with a p-value of 0.9 or more were considered significant. In pool "A" (endothelial cell line pool) there were 10 SAGE libraries containing 427,254 SAGE transcripts. In pool "B", the normal non-endothelial pool, there were 1 1 normal non-endothelial libraries with 329,470 transcripts. For the cancer cell line non-endothelial pool [8], there were 24 SAGE libraries consisting of 733,461 transcripts. As the cancer cell line non-endothelial pool was twice the size of the normal

non-endothelial pool, more genes were significantly up-regulated in the former due to pool size and statistics.

Statistical methods We now describe a statistical methodology for the comparison of two groups of cDNA libraries to enable the discovery of differentially expressed genes. The method combines a generalised maximum likelihood ratio test with a False Discovery Rate procedure (FDR) in order to provide a robust list of differentially expressed genes. The analyses extend our earlier work, which identified differentially expressed genes in a single group of cDNA libraries [15].

As described in [15], we consider the expression of gene j in a set of cDNA libraries. There are two groups of libraries: m libraries from non-endothelial cell lines, and n libraries from endothelial cell lines. We let Ni : 1< i < m be the number of ESTs sequenced in each non-endothelial cell line library, and Nm+i : 1< i < n be the number of ESTs sequenced in each of the endothelial cell line library. For each gene j, let xij be the number of copies of associated ESTs in library i.

For each gene, we compare two hypotheses concerning its frequency of expression in the libraries, using a generalised likelihood test. Under the null hypothesis, the gene is not differentially expressed and we would expect its frequency to be identical in both the non-endothelial and endothelial cell libraries. In contrast, under the alternative hypothesis, the gene is differentially expressed, and so we would expect the frequency to be different in the non-endothelial and endothelial cell lines.

In both cases, as long as the number of copies of ESTs from the gene is small relative to the total number of ESTs sequenced in the library, the distribution of the gene is well approximated by a Poisson distribution. Under the null hypothesis, the frequency is fj, then for library i, the number of ESTs is approximately distributed as a Poisson variable with parameter fiNi. Thus the likelihood of the observed data is

The maximum likelihood estimate of ^ under the null hypothesis can be found by solving:

^ = O

2)

(0)

And the solut fiiormn fj ^J is given by:

m+n γ- (O) _ ,=] ^lJ Jj m+n

3) σ '= ⁱ *

Equation 3 is simply the proportion of ESTs for the gene of interest among all ESTs in all of the libraries. Thus under the null hypothesis, the likelihood of the data ' ^} > is given by equation 1 with ^{~~ J} .

For the alternative hypothesis, the frequency of gene transcripts is different in the non- endothelial and endothelial cell line libraries. By a similar argument, we derive

/ Y- (IK frequencies for each gene J in the non-endothelial libraries ^Uj ' and the endothelial

libraries ( ^KJ f ^{j{2) J} ) which is given by:

n _ ⁽ =1 n

5) '-'

Equations 4 and 5 are very similar to equation 3, and simply represent the proportion of ESTs for the gene of interest among all ESTs in hypothesis, the likelihood of the data

(X ¹ ) is given by:

R, - bgif ^V /f∞j∑x,, +togσ; ²⁾ //; ⁰⁾ )∑ ^χ . _+IJ

7) ,=i ,=i

Equation 7 can be explained very simply: there are two terms, one for the non- endothelial libraries and one for the endothelial libraries. Each term is the log ratio of the frequency of the gene in the relevant libraries and the overall frequency of the gene, multiplied by the total number of ESTs for that gene in the relevant libraries. The equation is very similar to the R statistic derived in Stekel et al, 2000 [15].

Under Wilke's Theorem, ^J is distributed as a X distribution with a 1 degree of freedom. Therefore it is straightforward to compute a p-value for each gene. However, when analyzing all genes in the library in order to find those that are most differentially expressed, it is essential to combine the p-value with a False Discovery Rate Procedure [19]. Thus the results we present are the FDR-adjusted q-values.

A definition of terms: m = the number of non-endothelial cell line libraries n = the number of endothelial cell line libraries

^X '- ^J = the number of transcript copies of gene j in cDNA library i ' = the total number of clones sequenced in the cDNA library i x ^m+ ' ^j = the number of copies of gene j in the m + i'th cDNA library

N m ^+ι = the total number of clones sequenced in the m + i'th cDNA library

J ^J = the frequency of gene j

Computation of the statistics

To compute the FDR adjusted q-values for a given data set, we calculate the R-values for all genes. We then compute the p-values for every gene using the Chi squared distribution of 2R. The genes were then ordered according to the p-values, ranked from smallest to highest. Each p-value was adjusted by multiplying it by the number of genes in the analysis and dividing by its rank position (The smallest p-value is rank position 1). To derive the q-value, the list of ranked values was stepped through, comparing p-value and its adjusted value and always selecting the lowest.

Cell isolates and extraction of RNA Human umbilical vein endothelial cells (pooled HUVEC), adult human dermal microvascular endothelial cells (HDMEC), human bronchial epithelial cells (HBEC) and adult human Epidermal keratinocytes were obtained from TCS Cellworks (Botolph Claydon, UK). Cells were grown in their appropriate growth media and supplements) according to manufacturers instructions and RNA extracted at passage 2-3. Human lung fibroblasts (MRC-5) were obtained from American Type Culture Collection (Manassas, VA) and cultured in DMEM containing 10% FCS. All cells were grown at 37 ⁰ C in a humidified atmosphere of 5% CO ₂ in air.

Cryopreserved human hepatocytes (TCS Cellworks) were thawed in Leibpvitch L15 medium (Invitrogen, Paisley, UK), centrifuged and resuspended in fresh media, RNA was extracted after 30 minutes incubation at 37 ⁰ C in 5% CO ₂ . Cryopreserved human peripheral blood lymphocytes were obtained from TCS Cellworks, after thawing they were washed in PBS and used immediately for RNA extraction.

Quantitative PCR

Total RNA was extracted from cells in culture using TRI reagent (Sigma, Dorset, UK) cDNA was prepared using a high capacity cDNA archive kit (Applied Biosystems, Cheshire, UK). The Universal ProbeLibrary system (Roche) was used for real time PCR analysis. Reactions were performed in triplicate using Absolute QPCR mix (ABgene, Epsom, UK) according to the manufacturer's instructions using 10ng cDNA.

Reactions were performed in a Rotor-GENE RG30000 thermocycler (Corbett Research,

UK) using the following cycling conditions; 95 ⁰ C for 10 minutes followed by 40 cycles of

95 ⁰ C for 15 seconds and 6O ⁰ C for 1 minute. The appropriate housekeeper genes were determined as described by Vandesompele et al [47] using the software geNorm. For

the cell type screen FLOT2, Ubiquitin C and B-Actin were used. The raw data was analysed using a method described by Pfaffl [48].

List of abbreviations BLASTBasic Local Alignment Search Tool

BLAT BLAST Like Alignment Tool cDNA Complementary DNA

DDD Digital Differential Display

EST Expressed Sequence Tag FDR False Discovery Rate

FDR-adjusted False Discovery Rate adjusted

GFP Green Fluorescent Protein

HBEC Human bronchial epithelial cells

HDMECs Human dermal micro-vascular endothelial cells HUVECs Human umbilical vein endothelial cells

MRC-5 Human lung fibroblasts

PCR Polymerase Chain Reaction

RDBMS Relational Database Management System

Refseq Reference Sequence Project SAGE Serial Analysis of Gene Expression

TEM Tumour Endothelial Marker

Results

Development of an algorithm for EST-to-gene assignment We have developed a new algorithm for assigning an EST to a gene that takes advantage of the almost complete human genome and combines it with a BLAST analysis to achieve an accurate result. Initially, two EST pools and all Reference sequence project (Refseq) mRNA sequences were aligned to the human genome using BLAT. Sequences occupying an ambiguous position in the genome were removed. The aligned sequences were then collected into Perl data structures and a custom-clustering algorithm (Jake cluster) assigned each EST to a gene or gene prediction based on their overlapping genome position. In the BLAST analysis, each EST was BLAST searched against a Refseq database of all mRNA and gene predictions. Only the best mRNA hit from the BLAST analysis was assigned to an EST. The BLAT and BLAST results were cross-referenced and accurate EST-to-gene assignments were made based on the following decision tree:

If genome BLAT mapping and BLAST results agreed, then that gene was assigned. If the results disagreed, then the BLAST result was accepted only if the alignment was of high quality that is greater than or equal to 100 base pairs with at least 92% identity.

A pictorial representation of the analysis is shown in Figure 3. The approach was able to assign ESTs to a gene even when the single pass cDNA sequencing of an EST was of low quality. Thus, first finding an unambiguous position in the genome that overlaps with a gene and then searching with BLAST to find the best gene, it was able to assign an EST to a gene. Further, using a high quality BLAST alignment alone for the assignment gives this approach the ability to also assign a gene that lies in a gap in the human genome sequence.

Validation of the EST-to-gene assignment algorithm The results of the BLAST subtraction method used in our previous work [8] were compared to those of the algorithm developed here. Using a custom relational database that we developed, cDNA libraries were collected and divided into 2 pools (endothelial and non-endothelial cells respectively) and formatted into BLAST databases. The same data were used in this experiment as used in the earlier study because the EST-to-gene BLAST protocol was dependent on an expectation value. The expectation value was optimised in the earlier work for predictive capacity by performing trial runs. Expectation values are dependent on the size of a BLAST database and so it was important to use the same data. This was possible for the endothelial cell lines as the exact 11 ,1 17 ESTs were collected. However, for the non-endothelial pool the EST count had increased from 173,137 to 178,653 as a result of further EST sequencing. The expectation value for the non-endothelial pool was kept at 10e-20 as the larger pool size made the expectation value more stringent and less likely to deliver false positive hits.

Although there was a good agreement between the old BLAST and combined method EST assignments for some genes, a problem with searching mRNA queries against an

EST database is that any EST is able to hit more than one gene using an expectation value cut-off. In reality this is not possible, as an EST is derived from a single transcript derived from a single gene. Table 2 shows the number of EST sequences that hit more than one gene for both EST-to-gene assignment methods. From the endothelial pool using the earlier [8] method there were 5,228 from the 1 1 ,117 ESTs that were assigned to more than one gene using an expectation cut-off of 10e-30. The ambiguous

assignment means that it is not possible to know which EST-to-gene assignment was correct without manual inspection and as such failed. Assuming the remaining 5,889 ESTs that hit only one gene were correctly assigned, this amounted to a 53% success rate. For the new, combined algorithm there were no EST sequences assigned to more than one gene and the success rate for EST-to-gene assignment from the endothelial pool was 91%.

Table 2: Comparison of the two EST-to-gene assignment methods. The new method of EST-to-gene assignment improved accuracy enabling a higher percentage of ESTs to be unambiguously assigned compared to the Huminiecki and Bicknell (2000) method [8].

Testing of statistical significance To measure statistical significance of differential gene expression using cDNA libraries there is a Digital Differential Display (DDD) tool available at Unigene. This tool employs the Fisher exact test [18] to measure statistical significance (P > 0.05) between two libraries. According to [15] the statistics used by DDD are not valid for measuring statistical significance across multiple cDNA libraries, but only between two cDNA libraries. A further requirement of DDD is that it is only valid for cDNA libraries that contain at least 1 ,000 sequences collected into Unigene clusters. Several endothelial cell line libraries used in this study contained less than 1 ,000 sequences and comparisons between more than two libraries were required. For these reasons, the DDD tool and Unigene clusters were of no use in this analysis.

The statistics in the analyses used here combine a generalised maximum likelihood ratio test with a False Discovery Rate (FDR) that accounts for the different size of the cDNA library pools. During cDNA library construction, bacterial colonies are picked at random from agar plates for single pass sequencing of the EST insert. This process is random and can be modelled by a Poisson distribution. To derive the appropriate statistical method, two hypotheses were compared with each other. The NULL hypothesis states there is no difference in gene expression between two cDNA library pools and any differences in gene expression are due to sampling errors from the picking of colonies.

Alternatively, the difference in gene expression could be due to a genuine biological effect. The maximum likelihood ratio statistic (R-statistic) is derived by dividing the likelihood of seeing the data under the null hypothesis into the likelihood of seeing the data under the alternative hypothesis.

A p-value can be derived from the R-statistic as 2R is Chi square distributed. It should be noted that multiple testing on all genes in the human genome and using a p-value would result in many false positives. To account for multiple testing errors, a False Discovery Rate adjusted (FDR-adjusted) procedure was employed [19]. A q-value of 0.01 represents 1% false discovery rate and means that 10 in 1000 significantly differentially expressed genes were false positives. A q-value of 0.01 was considered to be significant.

Application of the statistics to the analysis

Applying the new analysis and statistics to our previous data [8], 14 genes were predicted to be significantly endothelial specific and a further 160 were significantly up- regulated in endothelial cells. Table 3 lists the 14 predicted significantly endothelial specific genes.

Table 3: Endothelial predicted genes from the original cDNA libraries analysed with the new gene assignment tool and statistical methods. 14 genes were predicted as significantly endothelial specific. A further 160 genes were predicted as showing significantly upregulated endothelial expression (q-value <= 0.01) but were not endothelial specific (i.e. had EST hits in the non-endothelial pool). With the new analysis there was no longer a need to cross reference to SAGE libraries for accurate prediction.

It is of interest to compare the 16 predicted endothelial genes listed in Table 7 of our previous analysis [8] with those found here. Three of the original 16 genes were no longer predicted as significantly endothelial and Table 4 summarises the results. RAMP2 had no ESTs in either pool; COL4A1 was up-regulated in endothelial cells but not to significance with a q-value of 0.5. In contrast, RASIP1 was endothelial specific with a single EST found in the endothelial pool but absent from the non-endothelial pool. However, the q-value of 0.36 was again not statistically significant.

Table 4: Listing of the genes from Table 7 of Huminiecki and Bicknell (2000), and comparison with the new analysis. 13 of the 16 genes were significantly endothelial, however, non-endothelial hits to known endothelial genes showed that the choice of non- endothelial cell lines could be improved, q-values in bold denote a significance threshold of <= 0.01.

Three genes identified as endothelial specific in the original analysis were not found to be so here. ROBO4 hit the EST [GenBank:AA577940] from the library NCI_CGAP__HSC1 that is a flow-sorted and non-normalized bone marrow cDNA library. EST accession [GenBank:AI380234] hit CD93 that is from a B-cell, chronic lymphocytic leukaemia flow-sorted cell line (NCI_CGAP_CLL1), while vWF hit a non-endothelial EST from the NCI_CGAP_Br4 library [GenBank:AA721546]. The last library was prepared from micro-dissected normal breast duct tissue and in view of the extensive literature showing restriction of von Willibrand factor expression to endothelium, is presumably from endothelial contamination of the dissected tissue. In subsequent analyses the non- endothelial pool was refined to exclude such hits.

Current data with the new algorithm and statistics. Experiment 1.

Employing the new EST assignment algorithm and the novel statistical method, a similar subtractive screen to [8] was carried out but this time with the most recent publicly

available data. In our earlier 2000 study there were 11 ,117 endothelial EST sequences. This has now increased to 31 ,114 and 64% of the currently available endothelial cell data was new. Table 5 (below) lists the 30 endothelial cell libraries used.

In view of aberrant gene expression by carcinoma lines arising from genetic instability and endothelial contamination of libraries isolated by FACS sorting or micro-dissection, we constructed a non-endothelial pool with no carcinoma cell, flow sorted or micro- dissected lines of 136,336 ESTs from 208 Genbank normal, non-endothelial cDNA libraries (Experiment 1).

From a cDNA library endothelial subtraction analysis alone (Experiment 1 ), there were 431 genes that were significantly up-regulated in endothelial cell lines. Of these, 104 genes showed an endothelial specific profile (Table 6, below), as transcripts were absent in the non-endothelial pool. The gene with the most significant endothelial specific profile was the metallo-proteinase gene MMP1, a surprising result as literature suggests this gene is widely expressed [20, 21]. It is worthy of note however, that MMP1 is also up- regulated in endothelial cells according to SAGE library analysis in Experiments 2 and 3. This analysis predicted ROBO4, CD93 and VWF as endothelial specific genes.

cDNA and SAGE library analysis combined, Experiment 2

For the second experiment, the data from the cDNA analysis of Experiment 1 was combined with SAGE in the same way as the 2000 analysis. The SAGE analysis used the latest endothelial libraries against a pool of normal non-endothelial libraries. There were 10 endothelial cell line SAGE libraries containing 427,254 tags and 11 normal non- endothelial libraries of 329,470 tags. 74% of the SAGE library data was new since 2000 and submitted to SAGEmap [22]. The SAGE library screen was very similar to the cDNA library approach as it involved comparing two pools of SAGE library cell lines using the SAGEmap xProfiler tool. Only genes with at least 10x the number of transcripts per million tags was considered significant with a p-value of 0.9 or more. Results are presented in Table 7, below, that lists 27 endothelial genes.

cDNA and SAGE library analysis combined (including carcinoma cell line cDNA data). Experiment 3

Although cancer, micro-dissected and sorted libraries in non-endothelial cell lines were thought to contaminate and invalidate the analysis, there exist many of these libraries in the public domain. Thus, to maximise the chance of predicting a comprehensive set of

endothelial genes, a final experiment (Experiment 3) was performed using non- endothelial libraries that included cancer, micro-dissected and sorted libraries (178,653 ESTs and 733,461 SAGE tags). The SAGEmap xProfiler analysis was again combined with a cDNA library subtraction. 58 endothelial genes were predicted from this analysis (Table 8, below).

A comprehensive set of in silico predicted endothelial genes

Combining the results of all three analyses gave a non-redundant list of 459 genes preferentially expressed at a statistically significant level in endothelial cells.

Experimental validation of the endothelial gene prediction

Real time PCR was carried out on predicted endothelial genes to examine the predictive power of the in silico analyses. A random selection of 12 genes (ECSM2, MMP1 , SOX18, ERG, RHOJ, APLN ₁ MMRN2, STAB1 , LYL1 , ELTD1 , EFEMP1 and BMX) was PCR amplified from human umbilical vein endothelial cells (HUVECs), human dermal micro-vascular endothelial cells (HDMECs) and a selection of normal primary, non- endothelial isolates; human lung fibroblasts (MRC-5), human bronchial epithelial cells (HBEC), adult human epidermal keratinocytes, peripheral blood lymphocytes and hepatocytes. Total RNA was extracted and real-time PCR was performed to measure differential expression of these genes between the cell types. Figures 4 and 5 show the power of the bioinformatics models as all genes examined were either highly up- regulated or completely specific to HUVECs and/or HDMECs.

Tumour endothelial marker prediction Following the prediction of endothelial genes, a second screen was performed to identify genes up-regulated in tumours or foetal tissue. Bulk tissue cDNA libraries that contain endothelium were used. The subtraction procedure carried out compared bulk tumour with bulk normal cDNA libraries from the same organ or tissue. The analysis involved six tissues, namely, lung, brain, colon, kidney, prostate and skin. Three foetal tissues (lung, brain and kidney) were also screened since foetal tissues, like tumours, have active angiogenesis. Specifically, 237 brain tumour bulk tissue libraries containing 140,621 ESTs were used versus brain normal libraries; 24 brain foetal bulk tissue libraries containing 69,862 ESTs were used versus brain normal libraries; 302 brain normal bulk tissue libraries containing 100,554 ESTs were used versus brain tumour/foetal libraries; 178 lung tumour bulk tissue libraries containing 108,107 ESTs were used versus lung normal libraries; 10 lung foetal bulk tissue libraries containing 112,690 ESTs were used

versus lung normal libraries; 91 lung normal bulk tissue libraries containing 82,757 ESTs were used versus lung tumour/foetal libraries; 7 kidney bulk tumour tissue libraries containing 38,519 ESTs were used versus kidney normal libraries; 5 kidney bulk foetal tissue libraries containing 2,605 ESTs were used versus kidney normal libraries, 5 kidney bulk normal tissue libraries containing 72,476 ESTs were used versus kidney tumour/foetal libraries; 131 prostate bulk tumour tissue libraries containing 19,125 ESTs were used versus prostate normal libraries; 132 prostate bulk normal tissue libraries containing 68,480 ESTs were used versus prostate tumour libraries; 6 skin bulk tumour tissue libraries containing 12,484 ESTs were used versus skin normal libraries; 4 skin bulk normal tissue libraries containing 33,218 ESTs were used versus skin tumour libraries; 557 colon bulk tumour tissue libraries containing 143,025 ESTs were used versus colon normal libraries; and 134 colon bulk normal tissue libraries containing 37,269 ESTs were used versus colon tumour libraries to find differentially expressed genes.

By screening each tissue independently, the analysis was able to identify genes that were putatively up-regulated in a tissue specific fashion. Special attention was taken in choosing normal tissue libraries, to ensure that they contained no active angiogenesis (e.g. foetal libraries were avoided for the normal tissue pools). Genes that were both selectively or preferentially expressed in tumour or foetal tissues and preferentially expressed in endothelial cells constituted predicted TEMs. 27 genes were chosen as being potential TEMs based on the specific or/and significant up-regulation in multiple tissues (Table 9).

Discussion

Identifying genes of interest

There exists great interest in the identification of tissue specific genes as they often perform a unique function within that cell type. In the past, tissue specific genes were sought using a range of molecular subtraction techniques employing mRNA/cDNA from the cell type of interest and a putative 'control' cell. Examples of such techniques include subtractive hybridisation, PCR display and PCR select. These approaches have been highly successful but remain laborious and expensive. Recent approaches have included selective insertional gene trapping or FACs sorting of cell lineages labelled with GFP in e.g. zebrafish followed by gene chip analysis. Both techniques have been used to identify endothelial genes [23, 24], for example in zebrafish the endothelium and

precursors were labelled with FIi promoter GFP. Nevertheless, such techniques are still expensive and laborious.

In silico analysis An alternative is to analyse computationally the vast amount of expression data now available in the public domain. We performed such an analysis in 2000 [8] that identified several previously unknown endothelial genes including Robo4, the endothelial roundabout guidance gene. A critical finding in the earlier analysis was the need to cross reference a cDNA with a SAGE analysis to achieve accurate prediction of expression. A complementary approach by Ho [7] combined cDNA and SAGE library database mining with microarray analysis. Virtual subtraction was carried out on data in the public domain using available tools to identify putative endothelial genes. These genes were then micro-arrayed and probed with RNA samples from a selection of cultured endothelial and non-endothelial cell types. A comparison of results is made below.

Introduction of new data gave a better analysis

We became aware that these earlier computational techniques could be improved and there are several cogent reasons for repeating such an analysis now. Firstly, the vast increase in expression data available now compared to 2000, in particular the number of ESTs in the endothelial libraries has more than doubled. Secondly, following the publication of the human genome, we developed a new technique that combines a BLAST search with genome BLAT alignments to increase the accuracy of EST-to-gene assignment. This removes the ambiguity of EST-to-gene assignment and consequent inaccuracies of the results present in our earlier analysis. Finally, we developed a novel maximum likelihood statistic analysis that can identify differentially expressed genes across multiple cDNA libraries. Using these improvements to cDNA library analysis and the inclusion of the latest SAGE library data, we have derived what we consider to be a near definitive set of endothelial specific genes.

cDNA library analysis improvements

We previously showed that endothelial genes could not be reliably predicted by using cDNA library analysis alone. Two possible explanations for this were. 1) Computationally, the EST-to-gene assignment was inaccurate for some genes with the BLAST protocol chosen and 2) there was no statistical analysis applied to the EST counts in order to determine the significance of the differential gene expression.

Repeating our analysis on the cell lines used in our earlier [8] study validated the new approach. The new analysis proved the critical importance of accurate EST-to-gene assignment to enable a successful analysis using cDNA libraries alone. The new method produced a successful assignment of 91% of the ESTs compared with 53% for the earlier study. Using the 2007, as apposed to the 2000 data, gave 31,114 assigned endothelial ESTs and a success rate of 94%. It should be noted that this success rate is also dependent on the quality of cDNA libraries, but comparing like for like, the new algorithm improved the accuracy of assignment by 38%. In order to identify differentially expressed genes between two pools of cDNA libraries convincingly, it is essential to employ a rigorous statistical analysis. The method described here makes use of the intrinsic variability associated with cDNA library measurements and represents the most powerful statistical analysis possible associated with that model. We note that the test is more appropriate than a t-test, and more powerful than non-parametric statistics such as the Mann-Whitney test. Differential expression of cDNA libraries can be performed on line at the CGAP and Unigene. However, DDD was not used in these analyses as it does not employ the maximal statistics test and only performs differential expression between cDNA libraries that have at least 1000 EST sequences clustered into Unigene. In contrast, the maximum likelihood statistics used in these analyses can be applied to cDNA libraries of any size and the EST-to-gene assignment does not rely on Unigene clusters.

Comparison of endothelial genes with previous work

It is of interest to compare the results of this analysis with two previous bioinformatic analyses to identify endothelial genes, those of Huminiecki and Bicknell [8] and Ho et al. [7]. In our earlier (2000) study [8], 16 genes were predicted as endothelial by a combined SAGE and cDNA library analysis. From the 16 genes, 13 were also predicted as significantly endothelial in this study. The three genes that differed between the two analyses were COL4A1, RAMP and RASIP1. In the new analysis RASIP1 was endothelial specific but not to significance, COL4A1 was expressed in both cDNA library pools and RAMP was not expressed in either pool. It is interesting that ECSM2 was the most endothelial specific gene in both the Huminiecki and Bicknell [8] and Ho et al [7] studies and was predicted as endothelial here but it was not ranked first, ROBO4 and MMP1 ranked higher. Real time PCR (Figure 5) and in-situ hybridisation (data not shown) show extreme endothelial specificity for ECSM2 and its lower ranking is simply due to fewer ESTs, i.e. it is expressed at a lower level in the cDNA libraries. A comparison with the endothelial genes found in this study with that of Ho et al [7] reveals

30 of the 49 genes were predicted as significantly (q-value <=0.01) up-regulated in endothelial cells. A further 5 genes were endothelial specific but not to significance (q- value > 0.01). 14 genes failed to show significant or specific expression in endothelial cells. Interestingly, the second ranked endothelial gene from the [7] analysis, SHE, showed only a single endothelial EST in this analysis. We conclude that although tissue specific genes can be predicted by cDNA analysis alone, it is advisable to use as many data sources as possible in order to derive a comprehensive list of genes. Finally, our results show that it is better to use normal cell isolates than carcinoma cell lines or libraries derived from micro-dissected or FACS sorted cells for this type of analysis, since several characterised endothelial genes hit ESTs in these non-endothelial libraries (Table 4: VWF, ROBO4 and CDH5).

Extended analysis to find TEMs

We extended the analysis to identify which of the endothelial genes were expressed in tumours but not normal tissue. This was achieved by combining the endothelial screen with an analysis that compared gene expression between tumour and normal bulk tissue libraries from several organs. A gene was a predicted TEM if it was preferentially expressed in endothelial cells and tumour tissues but absent in normal tissues. A list of 27 promising new TEMs based on these analyses is given in Table 9. Each cell in Table 9 represents the result for a tumour/foetal screen for a particular organ. Ultimately, we wanted to find genes that showed tumour and foetal specific expression in all or most of the organs at a statistically significant level. Cells with bold, dotted-underlined text represent this type of result, having 0 ESTs in the normal pool and a q-value of less than 0.01. Genes showing significant and specific expression in multiple organs (brain, skin, kidney and foetal lung) were PLOD3 and THRAP4. However, these genes showed expression in normal tissue for several other organs. Likewise some genes were significantly up-regulated in tumours in multiple organs but were not specific to tumours (cells with double-underlined text). These genes, although putative TEMs, were considered of least therapeutic value, as some expression was evident in normal tissues and as such could not be used to specifically target tumours.

Investigation of a subset of predicted TEMs

As endothelium comprises less than 5% of tumour tissue, it was hypothesised that genes with a tumour specific although not statistically significantly different expression could still be a TEM. Such cells are shown in Table 9 with single-underlined text. The most promising TEMs from Table 9 were selected based on little or no expression in normal

tissue across all or multiple organs (cells with single-underlined and bold, dotted- underlined text, respectively). Of these angiopoietin 2 (ANGPT2), protocadherin 12 (PCDH 12) and leucine rich repeat containing 8 family, member C (LRRC8C) had expression profiles totally restricted to tumour or foetal tissues. ANGPT2, in these in silico results, was restricted to renal and colon tumour tissue in adults and lung in embryos that is supported by the current literature that indicates ANGPT2 is associated with tumour endothelium and tumour progression [25-28]. In contrast, leucine rich repeat containing 8 family member C was not found to be a TEM in the literature but a gene responsible for adipocyte differentiation [29]. A list of 9 putative novel TEMs with the best tumour profile is listed in Table 1, above. This table excludes genes that already have substantial literature (e.g. angiopoietin2) as possible or actual TEMs. Another gene with an interesting literature is mediator of RNA polymerase Il transcription subunit 28 homolog (S. cerevisiae, MED28): research has shown MED28 to be significantly up- regulated in tumours, its over expression is able to stimulate cellular proliferation and its expression is up-regulated by endothelial cells when exposed to tumour media [30, 31].

TEM experimental validation

To experimentally validate the nine newly identified TEMs (Table 1), in situ hybridisation and immunostaining are the most definitive direct methods. However, the sensitivity to optimisation of the first makes it tricky for high throughput analysis. The second requires antibodies and the time to prepare these slows progress. It is the possible that recently developed phage antibody technology may overcome the delay in TEM validation. Therefore to validate our approach at this time, the next section describes the analyses and literature search results for previously predicted tumour endothelial markers.

Validation of TEM prediction based on known TEMs

Delta4 has been cited to have endothelial specific expression [32-34] and to be up- regulated in tumour vessels [32, 35]. In this study, Delta4 (DLL4) was endothelial specific but was expressed at a very low level in endothelial cell cultures. Delta4 matched one EST from the endothelial pool and none from the non-endothelial pool, with an FDR-adjusted q-value of 0.28. Even though this gene is not statistically significantly up-regulated in endothelial cells, it shows some evidence of being endothelial specific as there are no ESTs found from the non-endothelial pool. DLL4 was found in brain and colon tumour tissues. However, the expression was not specific or significant in tumours. Thus, in our analysis DLL4 was not a predicted TEM. GPR124 (TEM5) was previously identified as a putative TEM using custom SAGE libraries analysis [14]. In the

current analysis, GPR124 failed to match any endothelial ESTs from the 31 ,114 EST in the endothelial pool. From the non-endothelial pool, GPR124 did match a single EST [GenBank: BF325872] from the AN0041 cDNA library derived from a normal amniotic fluid cell line. These results suggest that GPR124 is only expressed at a low level in normal tissue and is absent or at a very low level in cultured endothelial cells. In contrast, GPR124 was predicted as significantly and specifically up-regulated in multiple tumour tissues. Thus, GPR124 appears to be a tumour but not a tumour endothelial marker. TEM 1 (endosialin or CD 248) [14] has a count of 1 and 2 ESTs for the endothelial and non-endothelial pools respectively. The FDR-adjusted q-value for this gene was 0.61 , a non-significant value. One EST from the non-endothelial pool, accession [GenBank: CN484271] was from a primary human ocular pericyte cDNA library. This agrees with experimental findings of MacFadyen et a/ [36, 37] that have shown that endosialin is expressed by fibroblasts and a subset of pericytes associated with tumour vessels but not by tumour endothelium.

Several groups have independently reported ROBO4 as a TEM [38-40]. In this study ROBO4 was highly endothelial specific, both from the in silico and experimental analyses. In the tumour screen, ROBO4 was seen to be tumour specific in brain and kidney tumour tissues but not at a statistically significant level. Thus ROBO4 was predicted as a tumour endothelial marker, but not in all tumour types. In this case our analysis may be under predictive, as experimentally ROBO4 has been found to be a strong TEM [39, 40]. This also demonstrates the absolute need for experimental verification of bioinformatics predictions. Numerous studies have reported SPARC to be up-regulated in endothelial cells, to have a role in tissue remodelling and be linked to tumour progression [41-44]. Our analysis strongly predicted SPARC to be a TEM. SPARC was up-regulated in endothelial cells with a significant q-value of 8.4 x 10-10 and also significantly up-regulated in brain, colon, kidney and prostate tumour tissue. VIM was significantly up-regulated in multiple tumour tissues and endothelial cells. It is an abundant intermediate filament protein and a secreted form of VIM has been shown to be the antigen for the endothelial cell-specific antibody PAL-E [45]. The integrin receptor α2β1 interacts with intracellular endothelial VIM and plays a role in endothelial cell to collagen adhesion [46]. Therefore, a combination of experimental and in silico evidence predicts VIM as a TEM. As above angiopoietin 2 has tumour/foetal specific profile from the in silico analysis this is confirmed by current literature [25-28].

There is evidence both for and against the use of cDNA analyses for the prediction of TEMs. If TEM 1 , TEM5 and DLL4 are true TEMs then this technique is not 100% reliable. In contrast, the successful prediction of ROBO4, ANGPT2, VIM and SPARC shows that these methods do have the ability to predict a validated TEM. The novel predictions of this analysis await validation.

Conclusions

New cDNA library data is continually been submitted to Genbank and the amount of relevant information that can be mined is increasing. cDNA library analysis has been improved in this work by more accurate EST-to-gene assignment and the best possible statistics applied to the data. Using these tools on the latest data sets will lead to the prediction of new biologically and therapeutically important genes. This is enhanced by the statistics as they enable the inclusion of cDNA libraries of all sizes.

We have shown that these methods accurately predict the identity of endothelial and tumour endothelial genes by comparing our results with that of known genes. For example, ROBO4 has consistently been shown to be highly endothelial specific and was ranked second in this work. Known TEMs were also successfully predicted as shown by the identification of SPARC and Angiopoeitin 2.

References for Example 1

1. Burrows FJ, Thorpe PE: Eradication of large solid tumors in mice with an immunotoxin directed against tumor vasculature. Proc Natl Acad Sci U S A 1993, 90(19):8996-9000. 2. Neri D, Bicknell R: Tumour vascular targeting. Nat Rev Cancer 2005, 5(6):436- 446.

3. Bortoluzzi S, Bisognin A, Romualdi C, Danieli GA: Novel genes, possibly relevant for molecular diagnosis or therapy of human rhabdomyosarcoma, detected by genomic expression profiling. Gene 2005, 348:65-71. 4. Bortoluzzi S, Danieli GA: Towards an in silico analysis of transcription patterns. Trends Genet 1999, 15(3): 118-119.

5. Bortoluzzi S, Rampoldi L, Simionati B, Zimbello R, Barbon A, d'Alessi F, Tiso N, Pallavicini A, Toppo S, Cannata N et a A comprehensive, high-resolution genomic transcript map of human skeletal muscle. Genome Res 1998, 8(8):817-825. 6. Chen J, Sun M, Lee S ₁ Zhou G, Rowley JD, Wang SM: Identifying novel transcripts and novel genes in the human genome by using novel SAGE tags. Proc Natl Acad Sci U S A 2002, 99(19): 12257-12262.

7. Ho M, Yang E, Matcuk G, Deng D, Sampas N, Tsalenko A, Tabibiazar R, Zhang Y, Chen M, Talbi S et al: Identification of endothelial cell genes by combined database mining and microarray analysis. Physiol Genomics 2003, 13(3):249-262.

8. Huminiecki L, Bicknell R: In silico cloning of novel endothelial-specific genes. Genome Res 2000, 10(11): 1796-1806.

9. ltoh K, Kawasaki S, Kawamoto S, Seishima M, Chiba H, Michibata H, Wakimoto K, lmai Y, Minesaki Y, Otsuji M et al: Identification of differentially expressed genes in psoriasis using expression profiling approaches. Exp Dermatol 2005, 14(9):667-674.

10. ltoh K, Okubo K, Utiyama H, Hirano T, Yoshii J, Matsubara K: Expression profile of active genes in granulocytes. Blood 1998, 92(4): 1432-1441.

11. ltoh K, Okubo K, Yosii J, Yokouchi H, Matsubara K: An expression profile of active genes in human lung. DNA Res 1994, 1 (6):279-287. 12. Okubo K ₁ Hori N, Matoba R, Niiyama T, Fukushima A, Kojima Y, Matsubara K: Large scale cDNA sequencing for analysis of quantitative and qualitative aspects of gene expression. Nat Genet 1992, 2(3):173-179.

13. Rajkovic A, Yan MSC, Klysik M ₁ Matzuk M: Discovery of germ cell-specific transcripts by expressed sequence tag database analysis. Fertil Steril 2001 , 76(3):550- 554.

14. St Croix B, Rago C ₁ Velculescu V, Traverso G, Romans KE ₁ Montgomery E ₁ LaI A ₁ Riggins GJ ₁ Lengauer C, Vogelstein B et al: Genes expressed in human tumor endothelium. Science 2000, 289(5482): 1197-1202.

15. Stekel DJ, Git Y, Falciani F: The comparison of gene expression from multiple cDNA libraries. Genome Res 2000, 10(12):2055-2061.

16. Altschul SF, Gish W ₁ Miller W, Myers EW ₁ Lipman DJ: Basic local alignment search tool. J MoI Biol 1990, 215(3):403-410.

17. Kent WJ: BLAT-the BLAST-like alignment tool. Genome Res 2002, 12(4):656- 664. 18. Kanji GK (ed.): 100 Statistical Tests, 3 edn. London: Sage Publications; 1993.

19. Benjamini y, Hochberg Y: Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Statist Soc B 1995, 57(1):289-300.

20. Domeij H ₁ Modeer T ₁ Quezada HC, Yucel-Lindberg T: Cell expression of MMP-1 and TIMP-1 in co-cultures of human gingival fibroblasts and monocytes: the involvement of lCAM-1. Biochem Biophys Res Commun 2005, 338(4): 1825-1833.

21. Ghaffari A, Li Y, Karami A, Ghaffari M ₁ Tredget EE ₁ Ghahary A: Fibroblast extracellular matrix gene expression in response to keratinocyte-releasable stratifin. J Cell Biochem 2006, 98(2):383-393.

22. Lash AE, Tolstoshev CM, Wagner L, Schuler GD, Strausberg RL, Riggins GJ, Altschul SF: SAGEmap: a public gene expression resource. Genome Res 2000,

10(7): 1051 -1060.

23. Covassin L, Amigo JD, Suzuki K, Teplyuk V, Straubhaar J, Lawson ND: Global analysis of hematopoietic and vascular endothelial gene expression by tissue specific microarray profiling in zebrafish. Dev Biol 2006, 299(2):551-562. 24. Hirashima M, Bernstein A, Stanford WL, Rossant J: Gene-trap expression screening to identify endothelial-specific genes. Blood 2004, 104(3):711-718.

25. Gill KA, Brindle NP: Angiopoietin-2 stimulates migration of endothelial progenitors and their interaction with endothelium. Biochem Biophys Res Commun 2005, 336(2):392-396.

26. Li Q, Xu B, Fu L, Hao XS: Correlation of four vascular specific growth factors with carcinogenesis and portal vein tumor thrombus formation in human hepatocellular carcinoma. J Exp Clin Cancer Res 2006, 25(3):403-409.

27. Ochiumi T, Tanaka S, Oka S, Hiyama T, lto M, Kitadai Y, Haruma K, Chayama K: Clinical significance of angiopoietin-2 expression at the deepest invasive tumor site of advanced colorectal carcinoma, lnt J Oncol 2004, 24(3):539-547.

28. Oliner J, Min H, Leal J, Yu D, Rao S, You E, Tang X, Kim H, Meyer S, Han SJ et ah Suppression of angiogenesis and tumor growth by selective inhibition of angiopoietin- 2. Cancer Cell 2004, 6(5):507-516. 29. Tominaga K, Kondo C, Kagata T, Hishida T, Nishizuka M, Imagawa M: The novel gene fad158, having a transmembrane domain and leucine-rich repeat, stimulates adipocyte differentiation. J Biol Chem 2004, 279(33):34840-34848.

30. Liu C, Zhang L, Shao ZM, Beatty P, Sartippour M, Lane TF ₁ Barsky SH, Livingston E, Nguyen M: Identification of a novel endothelial-derived gene EG-1. Biochem Biophys Res Commun 2002, 290(1):602-612.

31. Lu M, Zhang L, Sartippour MR, Norris AJ, Brooks MN: EG-1 interacts with c-Src and activates its signaling pathway, lnt J Oncol 2006, 29(4):1013-1018.

32. Mailhos C, Modlich U, Lewis J, Harris A, Bicknell R, Ish-Horowicz D: Delta4, an endothelial specific notch ligand expressed at sites of physiological and tumor angiogenesis. Differentiation 2001 , 69(2-3):135-144.

33. Rao PK, Dorsch M, Chickering T, Zheng G, Jiang C, Goodearl A, Kadesch T, McCarthy S: Isolation and characterization of the notch ligand delta4. Exp Cell Res 2000, 260(2):379-386.

34. Shutter JR, Scully S, Fan W, Richards WG, Kitajewski J, Deblandre GA, Kintner CR, Stark KL: DII4, a novel Notch ligand expressed in arterial endothelium. Genes Dev

2000, 14(11):1313-1318.

35. Patel NS, Dobbie MS, Rochester M, Steers G, Poulsom R, Le Monnier K, Cranston DW, Li JL, Harris AL: Up-regulation of endothelial delta-like 4 expression correlates with vessel maturation in bladder cancer. Clin Cancer Res 2006, 12(16):4836- 4844.

36. MacFadyen J, Savage K, Wienke D, lsacke CM: Endosialin is expressed on stromal fibroblasts and CNS pericytes in mouse embryos and is downregulated during development. Gene Expr Patterns 2007, 7(3):363-369.

37. MacFadyen JR, Haworth O, Roberston D, Hardie D, Webster MT, Morris HR, Panico M, Sutton-Smith M, Dell A, van der Geer P et a Endosialin (TEM1 , CD248) is a marker of stromal fibroblasts and is not selectively expressed on tumour endothelium. FEBS Lett 2005, 579(12):2569-2575.

38. Bedell VM, Yeo SY, Park KW, Chung J, Seth P, Shivalingappa V ₁ Zhao J, Obara T, Sukhatme VP, Drummond IA et a roundabout4 is essential for angiogenesis in vivo. Proc Natl Acad Sci U S A 2005, 102(18):6373-6378.

39. Huminiecki L, Gorn M, Suchting S, Poulsom R, Bicknell R: Magic roundabout is a new member of the roundabout receptor family that is endothelial specific and expressed at sites of active angiogenesis. Genomics 2002, 79(4):547-552.

40. Seth P, Lin Y, Hanai J, Shivalingappa V, Duyao MP, Sukhatme VP: Magic, roundabout, a tumor endothelial marker: expression and signaling. Biochem Biophys

Res Commun 2005, 332(2):533-541.

41. Dalla-Torre CA, Yoshimoto M, Lee CH, Joshua AM, de Toledo SR, Petrilli AS, Andrade JA, Chilton-MacNeill S, Zielenska M, Squire JA: Effects of THBS3, SPARC and SPP1 expression on biological behavior and survival in patients with osteosarcoma. BMC Cancer 2006, 6:237. 42. lkuta Y, Nakatsura T, Kageshita T, Fukushima S, lto S, Wakamatsu K, Baba H, Nishimura Y: Highly sensitive detection of melanoma at an early stage based on the increased serum secreted protein acidic and rich in cysteine and glypican-3 levels. Clin Cancer Res 2005, 11 (22):8079-8088.

43. Kzhyshkowska J ₁ Workman G, Cardo-Vila M, Arap W, Pasqualini R, Gratchev A, Krusell L, Goerdt S, Sage EH: Novel function of alternatively activated macrophages: stabilin-1 -mediated clearance of SPARC. J Immunol 2006, 176(10):5825-5832.

44. Schittenhelm J, Mittelbronn M, Roser F, Tatagiba M, Mawrin C, Bornemann A: Patterns of SPARC expression and basement membrane intactness at the tumour-brain border of invasive meningiomas. Neuropathol Appl Neurobiol 2006, 32(5):525-531. 45. Xu B, deWaal RM, Mor-Vaknin N, Hibbard C, Markovitz DM ₁ Kahn ML: The endothelial cell-specific antibody PAL-E identifies a secreted form of vimentin in the blood vasculature. MoI Cell Biol 2004, 24(20):9198-9206.

46. Kreis S ₁ Schonfeld HJ, Melchior C, Steiner B, Kieffer N: The intermediate filament protein vimentin binds specifically to a recombinant integrin alpha2/beta1 cytoplasmic tail complex and co-localizes with native alpha2/beta1 in endothelial cell focal adhesions. Exp Cell Res 2005, 305(1 ): 110-121.

47. Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, De Paepe A, Speleman F: Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol 2002, 3(7):RESEARCH0034. 48. Pfaffl MW: A new mathematical model for relative quantification in real-time RT- PCR. Nucleic Acids Res 2001 , 29(9):e45.

Table 5: Endothelial cDNA libraries available at Genbank that were used in this study. 21 new libraries have been submitted since our previous analysis [8]. The 30 combined libraries incorporate 31 ,114 endothelial ESTs.

Table 6: The 104 genes in the human genome with the most endothelial-specific expression profile predicted by applying the new analysis to the latest cDNA libraries.

Table 7: Combining cDNA and SAGE library analysis for endothelial gene prediction (normal non-endothelial libraries). 27 genes were predicted to be endothelial specific using a combined SAGE and cDNA library analysis of the latest libraries. The genes are sorted in descending order according to the number of non-endothelial library hits. Experimentally well-characterised endothelial genes are hi hli hted in bold.

OO CO

OO

Table 8: Combining cDNA and SAGE library analysis for endothelial gene prediction (tumour, microdissected, sorted non-endothelial libraries)

00 en

OO CD

OO

OO CO

Table 9: Potential Tumour Endothelial Markers. 58 endothelial specific genes were predicted by SAGE-CGAP xProfiler. All SAGE and cDNA non-endothelial cell libraries, including those from transformed cell lines and those produced by tissue micro-dissection or cell sorting, were used in this analysis

Gene Brain colon kidney Lung skin prostate Foetal kidney Foetal lung Foetal brain

SPHK1 D 0.86 1 1 U 0.57 2 0 U 0.38 1 0 U 0.33 2 0 U 0.22 2 0 U 0.43 1 0 U 0.83 1 1

KCTD 15 U 0.33 2 0 U 0.62 1 0 D 0.42 0 2 U 0.08 6 0 U 0.45 3 3 D 0.79 0 2 U 0.01 13 0 U 0.21 2 0

LRRC8C U 0.33 2 0 U 0.62 1 0 - - U 0.41 1 0 U 0.33 2 0 - -

PCDH12 U 0.43 1 0 U 0.52 3 0 U 0.38 1 0 - - - - - - - - U 0.23 3 0 - -

SPARC U 0.00 155 31 U 0 01 82 6 U 0 00 37 8 D 0.18 21 29 U 0.71 12 27 U Q OQ 32 25 U 0 00 6 8 U 0.44 50 29 U 0.44 28 31

ANGPT2 U 0.52 3 0 U 0.13 3 0 - - - - U 0.43 1 0

CO VIM U 0 00 393 9 D 0.75 36 11 U 0 00 35 9 U 0.87 50 36 U 0.00 23 11 D 0.70 2 10 U 0.07 4 9 D 0.53 40 36 U 0.00 94 [ 9

CO

BGN U 0.00 61 4 U 0.95 26 7 U 0.87 2 3 U 0.14 21 6 U 0.00 7 0 U 0.55 1 1 D 0.79 0 3 D 0.85 7 6 D 0.81 2 4

C12orf11 U 0.11 5 0 U 0.46 4 0 D 0.63 1 4 D 0.37 0 1 U 0.31 1 0 U 0.41 1 0 D 0.79 0 4 U 0.84 2 1 U 0.21 2 0

C16orf30 U 0.07 6 0 U 0.40 5 0 D 0.20 1 12 D 0.37 0 1 - - D 0.54 0 2 U 0.79 1 12 U 0.04 13 1 - -

ECOP U 0.00 27 0 U 0.46 4 0 D 0.37 0 3 U 0.59 3 1 U 0.85 1 2 D 0.55 0 1 D 0.79 0 3 U 0.84 2 1

ECSM2 U 0.16 4 0 U 0.46 4 0 - - U 0.59 3 1 - - U 0.70 1 2 - - U 0.23 3 0 - -

ERG - - - - D 0.77 1 3 U 0.33 2 0 U 0.41 1 0 D 0.79 0 3 U 0.04 8 0 U 0.34 1 0

GBP4 U 0.43 1 0 U 0.62 1 0 U 0.03 5 0 U 0.37 7 2 U 0.31 1 0 D 0.54 0 2 D 0.23 0 2 - -

IKBKE U 0.07 6 0 U 0.57 2 0 U 0.38 1 0 D 0.59 2 3 - - U 0.71 6 3

LOC653949 U 0.07 6 0 U 0.57 10 1 U 0.00 32 0 D 0.37 3 6 D 0.53 0 1 D 0.55 1 8 D 0.43 4 6 U 0.34 1 0

CD O

27 predicted TEMs that were significantly endothelial were also up-regulated or specific to tumours. Three foetal tissues were also screened as they contain regions of active angiogenesis. The information held in each cell is as follows: e.g., U 0.21 3 0 (space delimited): U/D = gene was Up or Down-regulated in the tissue; 0.21 = FDR q-value; 3 0 = EST counts, 3 tumour bulk tissue ESTs and 0 normal bulk tissue 5 ESTs.

Single-underlined text denotes genes showing tumour or foetal specific expression (no counts in normal tissue) but not at a statistically significant level.

Double-underlined text denotes genes that showed statistically significant (q-value <= 0.01) differential expression but not specific (some expression seen in normal tissue).

10 Bold and dotted-underlined text is for genes that were both significantly (q-value <= 0.01) and specifically (no counts in normal tissue) up-regulated in tumour or foetal tissues.

Example 2: ECSM2 is specifically expressed in tumour vasculature

The expression of ECSM2 in sections of solid tumours and normal tissue was examined using ECSM2-specific probes. In situ hybridisation was performed on tissue sections using appropriate ECSM2-specific Exiqon ^® biotin-labelled LNA™ probes followed by visualisation with avidin coupled to FITC using standard procedures (Thomsen et al (2005) "Dramatically improved RNA in situ hybridization signals using LNA-modified probes" RNA 11: 1745-1748).

As shown in Figure 6, ECSM2 specific expression was seen in the blood vessels of the brain tumours ganglioma (A) and astrocytoma (B), but not in the control brain section (C). We have also found that ECSM2 was specifically expressed in the vasculature of other solid brain tumours including metastatic adenocarcinoma, glioblastoma and medulloblastoma (data not shown).

ECSM2 was also expressed in the neovasculature of a histiocytoma (data not shown).

We have also found by in situ hybridisation to a SuperBiochip tumour array that ECSM2 is specifically upregulated in the neovasculature of bladder, lung, oesphagus, stomach and kidney tumours but not expressed in control tissue sections (data not shown).

Example 3: ECSM2 siRNA inhibits proliferation of HUVECs in culture

Methods

The effect of ECSM2-specific siRNA on the proliferation of HUVEC cells was assessed based on the protocol described by Nishiwaki et al (2003, "Introduction of short interfering RNA to silence endogenous E-selectin in vascular endothelium leads to successful inhibition of leukocyte adhesion". Biochem Biophys Res Commun 310: 1062- 6). 5 x 10 ⁵ cells in a 10 cm tissue culture plate were treated with 0.2% lipofectamine and 12.5 nM ECSM2-specific siRNA in a total volume of 6 ml of optimem for 4 h at 37 ⁰ C. After 4 h the transfection media was removed and replaced with fresh tissue culture media (DMEM, 10% foetal calf serum, 1 ng/ml FGF). MRC5 human fibroblasts, that do not express ECSM2, were used as controls.

The target sequences of the three different ECSM2-specific siRNA duplexes used in this experiment are (5'- 3'): ECSM2 siRNA 1 : AGACAGCAUCACCCUUAUC (SEQ ID NO: 3)

ECSM2 SiRNA 2: AGAGGUGGUGACAGAGAGA (SEQ ID NO: 4) ECSM2 siRNA 3: AUACUGCCUGCUUCUCCCA (SEQ ID NO: 5)

Results By cell counting, we show that ECSM2-specific siRNA significantly inhibits HUVEC proliferation (Figure 7A). Similar results were seen with three different siRNAs. Since endothelial proliferation is an essential component of angiogenesis, inhibitors of endothelial proliferation are usually anti-angiogenic.

The MRC5 cells do not express ECSM2, and the ECSM2-specific siRNAs had no effect of SiRNA on these cells (Figure 7B).

Example 4: Further Studies on the role of ECSM2

Summary

We aimed to characterize the expression and function of ECSM2 that our bioinformatics analysis in Example 1 predicted to be tumour endothelium specific. A full-length cDNA was isolated and confirmed the predicted ECSM2 polypeptide to be a putative 205 amino acid transmembrane protein that bears no homology to any known protein. Quantitative PCR analysis in vitro and in situ hybridization analysis in vivo confirmed ECSM2 expression to be exclusively endothelial. ECSM2's predicted plasma membrane expression was confirmed by cell surface expression of an ECSM2-GFP fusion protein in endothelial cells (data not shown). A yeast two-hybrid analysis using the ECSM2 intracellular domain identified filamin A as an interacting protein (data not shown). Filamin A anchors transmembrane proteins to the actin cytoskeleton acting as a scaffold for various signalling proteins [1] and is known to interact with and regulate the function of several endothelial transmembrane molecules [2-4]. The interaction was confirmed by precipitation of filamin-A from endothelial cell lysates by a GST-tagged intracellular domain of ECSM2 (data not shown). Functional studies with siRNA knockdown of ECSM2 markedly inhibited endothelial cell proliferation, but not apoptosis, adhesion, migration or tube formation.

We have reported this Example in Armstrong et al (2008, Arterioscler Thromb Vase Biol. 28(9):1640-6. Epub 12 June 2008), which is incorporated herein by reference in its entirety.

Methods

Identification of full length cDNA sequence and cloning of ECSM2 cDNA 5' and 3' rapid amplification of cDNA ends (RACE) was used to determine the full sequence of the ECSM2 transcript, using the SMART RACE cDNA Amplification kit (BD Biosciences, Clontech, UK). The ECSM2 transcript was amplified by PCR on total HUVEC cDNA using the upstream primer 5'-

TACTCGAGATGGACAGAGCCTCCACTGA-3' (SEQ ID NO: 6) designed to include the Xho1 site and the downstream primer 5'- TACCGCGGCACCTCATCACTTTCCTTGC -3' (SEQ ID NO: 7) designed to include the Sac Il site and inserted into the pBluescript vector (Stratagene, UK).

Quantitative PCR cDNA was prepared using total RNA and the random priming High-Capacity cDNA Archive kit (Applied Biosystems, UK). For the larger cell line screen, standard curve analysis was performed to obtain relative expression levels for ECSM2 and the housekeeping gene β-2- microglobulin to which ECSM2 expression was normalized. To perform the primary cell type screen, the housekeeper genes flotillin-2, ubiquitin C and β- actin were chosen using the method described by Vandesompele et al. (2002) with the software geNorm [5]. Data was analysed using a method described by Pfaffl [6].

In situ hybridization

In situ hybridization analysis was performed using radioactively labelled probes as described by Poulsom et al. 1998 [7]. In some cases slides were also stained with CD34 antibody using a method described by Jeffery et al. 2003 [8]. The ECSM2 transcript specific probe (nucleotides 15-980) was used to generate the in situ hybridization probe.

Detection of the ECSM2 protein

Polyclonal ECSM2 antibodies were generated in rabbits against the peptide TQTSSSQGGLGGLSLTTEP (SEQ ID NO: ) and were used at 1 :400 to detect ECSM2 protein by immunoblotting. Prior to immunoblotting for ECSM2, cell lysates were treated with N-glycosidase to remove N-linked carbohydrates.

Identification of interacting proteins

A yeast two-hybrid screen using the ECSM2 intracellular domain against a placental cDNA library was performed using the BD Matchmaker Library Construction and Screening kit (Clontech, BD Biosciences, UK). Briefly, AH109 yeast were cotransformed with the ECSM2-pGBKT7 (GAL4 binding domain) bait vector, the Smal linearized pGADT7 (GAL-4 activation domain) library vector and a placental library of cDNA fragments. The co-transformation reaction was initially plated onto triple drop out medium (SD/-His/-Trp/-Leu) and yeast colonies with a positive phenotype were then subjected to a more stringent screen by testing for growth on quadruple drop out medium (SD/-Ade/-His/-Trp/-I_eu). The filamin A interaction was confirmed by a pull down assay using a bacterially expressed ECSM2 intracellular domain-GST fusion protein bound to glutathione agarose beads and HUVEC lysate. To detect precipitation of filamin A, samples were separated using SDS-PAGE and either Coomassie stained for visualization of the GST fusion proteins or immunoblotted using anti-filamin A monoclonal antibody (Chemicon).

Transfection with siRNA and functional assays

2.5 to 5 x 10 ⁵ HUVEC were seeded into 10 cm plates the day before transfection. Transfection was carried out with lipofectamine 2000 in optimem for 20- 30 minutes. For cell growth assays, the cells were counted in a Coulter Counter. Migration, adhesion and 'Matrigel' tube forming assays were carried out using standard methods [9].

Results

Identification, Cloning and Sequence Analysis of ECSM2

To obtain the full length ECSM2 transcript sequence, the sequence of an ECSM2 EST (BI823114) was used to design primers for 5' and 3 ^' rapid amplification of cDNA ends (RACE). RACE was performed using human umbilical vein endothelial cell (HUVEC) RNA. Sequencing of RACE experimental products revealed the ECSM2 transcript to be 1030 bp in length. Subsequent northern blot analysis on human breast tissue RNA, confirmed the RACE data with expression of a major transcript at approximately 1 kb (data not shown). Primers were subsequently designed to the 5 ^' and 3 ^' ends of the full length ECSM2 sequence, which enabled PCR cloning of the transcript from HUVEC cDNA. Translation of the entire cDNA sequence revealed ECSM2 to contain an open reading frame (ORF) of 618 bp with a 67 bp 5 ^' untranslated region (UTR) and 345 bp 3' UTR. The complete ECSM2 nucleotide sequence did not match any database sequences other than ESTs (GenBankTM/EBI). The ECSM2 nucleotide sequence has

been submitted to GenBank ^™ /EBI Data Bank with accession number DQ462572 and is listed in Figure 5 (SEQ ID NO: 2).

Database searches and fluorescence in situ hybridization analysis (data not shown) revealed that the ECSM2 gene mapped to human chromosome 5 at position 5q31 , and spans 10.3 kb.

The conceptual ECSM2 protein was comprised of 205 amino acids (Figure 5; SEQ ID NO: 1) and extensive database searches revealed that ECSM2 had no known homologues and contained no functional domains. Analysis of the predicted amino acid sequence revealed that ECSM2 contained a putative 24 amino acid signal sequence and 27 amino acid sequence transmembrane domain (residues 120-147).

ECSM2 was predicted to contain a single N-linked glycosylation site at residue 96 by the presence of the universal acceptor sequence Asn-X-(Ser/Thr). ESTs from several other species showed homology to ECSM2 (data not shown). ClustalW analysis of the hypothetical proteins showed extensive conservation of the predicted intracellular domain of the ECSM2 protein, but much less in the extracellular domain (data not shown). While the transmembrane and intracellular domains showed 75% similarity between zebrafish and man, conservation of the extracellular domain was poor even within mammals.

Quantitative analysis of ECSM2 expression

Quantitative PCR analysis was performed on a selection of human cells to examine the pattern of ECSM2 expression. ECSM2 was expressed in all four endothelial cell types investigated but absent in non-endothelial cell lines or primary isolates (Figure 8). This supports the data obtained in Example 1 and shown in Figure 4. After endothelial cells, the highest expression was seen in aortic smooth muscle cells but was only 4% of that in HUVEC. There are two vascular smooth muscle cDNA libraries in the public databases, the HCASM2 and Sugana coronary artery smooth muscle cell libraries that between them encode 16,254 EST's. A BLAST search of the ECSM2 nucleotide sequence (NM_001077693) against the two libraries identified no hits. It is possible that the 4% signal detected in the real time PCR analysis of the vascular smooth muscle cells could be due to trace contamination of the isolate with endothelium.

In situ hybridization analysis of ECSM2 expression

In situ hybridization studies of ECSM2 expression were performed in order to determine expression of ECSM2 in human tissues in vivo. This supports and supplements the data obtained in Example 2 and shown in Figure 6. Endothelial restricted expression of ECSM2 was observed in human breast carcinoma (Figure 9A-B), human ganglioglioma

(Figure 9C-D), the skin in a psoriasis patient biopsy (Figure 9E-F), placenta (Figure 9G-

H) and foetal tissue (Figure 9/-J). To confirm that expression was endothelial, human breast carcinoma sections were also immunostained for the endothelial specific marker

CD34. Colocalization of the signals for ECSM2 and CD34 confirmed endothelial specific expression of ECSM2 in vivo (Figure 9K-L).

Detection of the ECSM2 protein in endothelial cells, analysis of the ECSM2 qlvcosylation pattern and sub-cellular localization

To confirm that the ECSM2 transcript was translated to produce the conceptual ECSM2 protein in endothelial cells, we detected the ECSM2 protein by immunoblotting with polyclonal rabbit anti-sera. The deglycosylated protein was of the size (18.7 kDa) expected from the predicted sequence (data not shown).

We predicted that ECSM2 had a single N linked glycosylation site at residue N96, and we found that ECSM2 was only glycosylated at this residue (data not shown).

To examine the cellular localization of ECSM2, the full-length protein was expressed with GFP fused at its C terminus and transfected into 293 cells or HUVEC. In 293 cells ECSM2-GFP showed membrane expression (data not shown), the cells were also stained with the nuclear marker 4',6-diamidino _: 2-phenylindole (DAPI), which confirmed that the expression seen was at the plasma membrane and not the nuclear membrane. To examine ECSM2 localization in endothelial cells, ECSM2-GFP was expressed in HUVEC. In live cells, fusion protein expression was localized at the cell surface and particularly on cellular protrusions such as filopodia (data not shown). ECSM2-GFP expressing HUVEC were also fixed and stained with phalloidin to visualize F-actin which is known to associate around the plasma membrane. Membrane expression of ECSM2 mostly co-localized with F-actin expression, particularly where many filopodia were present (data not shown).

Identification of filamin A as an ECSM2 binding protein

To identify proteins that interact with ECSM2, the intracellular domain of ECSM2 (residues 147-205) was used as bait in a yeast two hybrid analysis of a human placental library. Three clones isolated encoded the C-terminal region of filamin A. The 3 clones encoded amino acids 1658-1846 and 2092-2310, which are in the 15-16 and 19-21 beta repeat sheets regions respectively. To confirm this interaction, the intracellular domain of ECSM2 fused to GST was used to pull down endogenous filamin A from a HUVEC lysate, revealing a filamin A band at 280 kDa which was absent in the GST-fusion control (data not shown).

Functional studies with si RNA

These are the same studies conducted in Example 3 and shown in Figure 7. Real time PCR identified three siRNA's that efficiently (at a concentration of 12.5 nM) knockdown ECSM2 in HUVEC, and Western blotting with rabbit anti-sera confirmed protein knockdown (data not shown). ECSM2 knockdown was found to have no effect on HUVEC apoptosis, adhesion, migration (in a scratch wound or Boyden chamber assay) or tube formation on Matrigel (data not shown) but markedly inhibited endothelial cell proliferation determined by counting cells (Figure 7A) or MTT assay (data not shown). There was no effect on growth of non-ECSM2 expressing MRC5 fibroblasts (Figure 7B). Introduction of siRNA into cells can give artefactual results due to induction of interferon secretion from the transfected cell (the so-called interferon response). Real time PCR quantitation of the interferon sensitive genes (IG20 and OAS1) showed no difference between transfectants and controls (data not shown) and confirmed that these duplexes do not induce the interferon response at the 12.5 nM concentration used in the cell proliferation experiments.

Discussion

Endothelial cells are one of the most transcriptionally active cell types [10], they are involved in many physiological processes and play a role in the development of diseases such as atherosclerosis and cancer. As a result, we sought to identify novel endothelial specific genes. There is a particular interest in the identification of membrane spanning proteins that respond to environmental signals at the cell surface and deliver that information to the cell cytoskeleton. In this study, we describe the identification and characterization of the transmembrane protein ECSM2, a previously uncharacterized filamin A binding protein.

We originally identified ECSM2 (Genbank DQ462572) by bioinformatics [11], and it was named Endothelial Cell Specific Gene-2 (ECSM2) due to its putative endothelial restricted expression but has otherwise remained a hypothetical protein. This work has confirmed the ECSM2 mRNA as a 1030bp transcript and validated its highly restricted endothelial expression experimentally. Quantitative PCR analysis of primary cell isolates and cell lines cultured in vitro showed that ECSM2 was expressed only in endothelial cells. In situ hybridization analysis of ECSM2 expression in a range of human tissues showed that the ECSM2 transcript displayed highly restricted endothelial expression, with little or no expression observed in other cell types. Endothelial expression of ECSM2 was confirmed by immunostaining of adjacent sections for the endothelial marker CD34. Our findings are supported by the work of Ho et a/ (2003), who reported that ECSM2 was the most endothelial specific of 64 endothelial genes identified in a combined bioinformatics and microarray screen [12]. Furthermore, Pelossi et al (2002) have shown that ECSM2 is a marker of primary hemangioblasts and endothelial progenitors but not other hematopoietic progenitor cells [13]. Thus, ECSM2 shows potential as an endothelial marker and the ECSM2 transcript has already been used as a measure of endothelial cell contamination [14] and differentiation [13].

The ECSM2 gene is expressed from a 10.3 kb locus on human chromosome 5q31 and encodes a 205 amino acid protein, which sequence analysis predicted to contain a signal sequence and transmembrane region. Extensive database searches revealed the ECSM2 to be unique in that it bore no homology to any known protein and contained no functional domains. Putative ECSM2 orthologs were identified in other mammalian species and ClustalW alignment of the orthologs revealed that they all contained a highly conserved intracellular domain and a variable extracellular domain. Without wishing to be bound by theory, the conservation could indicate that there is a strong selective pressure to maintain the integrity of the intracellular domain sequence, or alternatively, interaction with another factor (e.g. a virus) could have driven rapid evolution of the extracellular domain.

Expression of ECSM2 protein in endothelial cells was confirmed by immunoblotting with anti-ECSM2 polyclonal sera. Due to a large N-linked carbohydrate moiety, we deglycosylated the ECSM2 protein prior to detection with polyclonal sera. Following deglycosylation, we found that the molecular weight of ECSM2 matched the predicted 18.7 kDa. Site directed mutagenesis studies identified this glycosylation site as

asparagine 96, suggesting this region of ECSM2 to be extracellular, supported by the presence of a later transmembrane domain and the preceding signal peptide. Cell surface expression of ECSM2 was subsequently confirmed by using an ECSM2-GFP fusion protein in 293 cells and HUVEC. ECSM2 expression was uniform across the entire membrane in 293 cells. By contrast, in HUVEC expression was concentrated at certain points such as the filopodia, which may be a feature of ECSM2 expression in the presence of its interacting proteins.

A mechanism for controlling the localization of ECSM2 within the plasma membrane is through the intracellular interactions of ECSM2 with other proteins. To investigate

ECSM2 protein interactions, yeast two hybrid analysis was performed using the intracellular domain of ECSM2 as bait to screen a human placental cDNA library. The screen identified a putative interaction between filamin A and ECSM2 and this was subsequently confirmed by using an ECSM2 fusion protein to precipitate endogenous filamin A from HUVEC lysate. Filamin A belongs to the filamin family of actin binding proteins that link actin filaments at the cell membrane and help maintain cell structure.

Filamins are considered to be key players in mammalian cell locomotion [15].

Mammalian filamins exhibit marked promiscuity in their protein interactions and have been shown to bind to more than 30 different proteins [16]. Despite the diversity of the interacting proteins the regions of the filamin that they bind to are principally specified by their function. Thus, a few smaller proteins that participate in signalling processes recognise repeats 1 to 15. In contrast, receptor proteins that comprise the largest group of interactors, all recognise repeats 16 to 24 and it is this interaction that mediates cross- talk between the extracellular environment and the actin matrix. The interaction of the intracellular domain of ECSM2 with repeats 19 to 21 (and to a lesser extent 15 and 16) are consistent with ECSM2 bearing such a role in endothelial cells that are known to express filamin A [17].

Our finding that siRNA knockdown of ECSM2 in HUVEC inhibited cell proliferation, but not migration or tube formation, was somewhat unexpected in view of the known functions of filamin A described above. Recently, a 90 kDa fragment of filamin A has been implicated in the androgen dependant growth of prostate cells [18]. The study showed that the 90 kDa filamin fragment underwent nuclear translocation and when present in the nucleus decreased AKT phosphorylation and blocked cell proliferation.

The 90 kDa fragment incorporates the sites of filamin A that bind ECSM2 (repeats 19-21) and it is possible that ECSM2 promotes endothelial proliferation by sequestering the 90 kDa filamin A fragment away from the nucleus. It is noteworthy in this context that the only tissues in which ECSM2 was detected by in situ hybridisation (tumours, placenta, psoriasis and embryonic tissue) were all sites of active angiogenesis. According, we suggest that ECSM2 could mediate a novel growth regulatory pathway in endothelium.

References for Example 4

1. Feng Y, Walsh CA. The many faces of filamin: a versatile molecular scaffold for cell motility and signalling. Nat Cell Biol. 2004;6: 1034-1038.

2. Calderwood DA, Huttenlocher A, Kiosses WB, Rose DM, Woodside DG, Schwartz MA, Ginsberg MH. Increased filamin binding to beta-integrin cytoplasmic domains inhibits cell migration. Nat Cell Biol. 2001 ;3: 1060-1068.

3. Feng S, Resendiz JC, Lu X, Kroll MH. Filamin A binding to the cytoplasmic tail of glycoprotein lbalpha regulates von Willebrand factor-induced platelet activation. Blood.

2003;102:2122-2129.

4. Ott I, Fischer EG, Miyagi Y, Mueller BM, Ruf W. A role for tissue factor in cell adhesion and migration mediated by interaction with actin-binding protein 280. J Cell Biol. 1998;140:1241-1253. 5. Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, De Paepe A, Speleman F. Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol. 2002;3:RESEARCH0034.

6. Pfaffl MW. A new mathematical model for relative quantification in real-time RT- PCR. Nucleic Acids Res. 2001 ;29:e45. 7. Poulsom R ₁ Longcroft JM, Jeffery RE, Rogers LA, Steel JH. A robust method for isotopic riboprobe in situ hybridisation to localise mRNAs in routine pathology specimens. Eur J Histochem. 1998;42:121-132.

8. Jeffery R, Hunt T, Poulsom R. In Situ hybridisation combined with immunohistochemistry to localise gene expression. In: Harris A and Brookes S. Eds. Breast Cancer Research Protocols: Human Press Inc; 2003.

9. Angiogenesis assays: critical appraisal of current techniques Staton CA, Lewis CE, Bicknell R, eds John Wiley and Sons Ltd; 2006.

10. Adams MD, Kerlavage AR, F.leischmann RD, Fuldner RA, BuIt CJ, Lee NH, Kirkness EF, Weinstock KG, Gocayne JD, White O, et al. Initial assessment of human gene diversity and expression patterns based upon 83 million nucleotides of cDNA sequence. Nature. 1995;377:3-174.

11. Huminiecki L, Bicknell R. In silico cloning of novel endothelial-specific genes. Genome Res. 2000; 10: 1796-1806.

12. Ho M, Yang E, Matcuk G, Deng D, Sampas N, Tsalenko A, Tabibiazar R, Zhang Y, Chen M, Talbi S, Ho YD, Wang J, Tsao PS, Ben-Dor A, Yakhini Z, Bruhn L,

Quertermous T. Identification of endothelial cell genes by combined database mining and microarray analysis. Physiol Genomics. 2003; 13:249-262.

13. Pelosi E, Valtieri M, Coppola S, Botta R, Gabbianelli M ₁ LuIIi V, Marziali G, Masella B, Muller R, Sgadari C, Testa U, Bonanno G, Peschle C. Identification of the hemangioblast in postnatal life. Blood. 2002; 100:3203-3208.

14. Permana PA, Nair S, Lee YH, Luczy-Bachman G ₁ Vozarova De Courten B, Tataranni PA. Subcutaneous abdominal preadipocyte differentiation in vitro inversely correlates with central obesity. Am J Physiol Endocrinol Metab. 2004;286:E958-962.

15. Stossel TP, Condeelis J, Cooley L ₁ Hartwig JH, Noegel A, Schleicher M, Shapiro SS. Filamins as integrators of cell mechanics and signalling. Nat Rev MoI Cell Biol. 2001 ;2:138-145.

16. Popowicz GM, Schleicher M, Noegel AA, Holak TA. Filamins: promiscuous organizers of the cytoskeleton. Trends Biochem Sci. 2006;31 :411-419.

17. Gorlin JB ₁ Yamin R, Egan S, Stewart M, Stossel TP, Kwiatkowski DJ, Hartwig JH. Human endothelial actin-binding protein (ABP-280, nonmuscle filamin): a molecular leaf spring. J Cell Biol. 1990; 111:1089-1105.

18. Wang Y, Kreisberg Jl, Bedolla RG, Mikhailova M, deVere White RW, Ghosh PM. A 90 kDa fragment of filamin A promotes Casodex-induced growth inhibition in Casodex- resistant androgen receptor positive C4-2 prostate cancer cells. Oncogene. 2007;26:6061-6070.

Example 5: Treatment of a solid tumour in an animal model

A mouse model of a solid tumour (e.g. either a Lewis lung carcinoma subcutaneous homograft implant in Black 57 mice or an HT29 subcutaneous xenograft implant in nude mice) is treated with intravenous infusions of saline solutions of a pharmaceutical composition comprising antibodies that selectively bind to the ECSM2 polypeptide. The infusions are administered weekly for a time of 2 to 4 months. The tumour regresses in the animal model compared to the controls.