PLEXIN D1 AS A TARGET FOR TUMOR DIAGNOSIS AND THERAPY

Abstract

The present invention relates to plexin D1 for use as a targetable protein in the treatment or diagnosis of disorders that involve expression of plexin D1. Diagnosis is suitably effected by detecting the presence of plexin D1 in the body or a bodily tissue or fluid, whereas treatment is effected by targeting plexin D1 for delivery of therapeutics to the site where treatment is needed. The invention further relates to the use of molecules that bind plexin D1, a nucleic acid encoding plexin D1 or a ligand of plexin D1 for the preparation of a therapeutical composition for the treatment or diagnosis of disorders that involve expression of plexin D1. The disorders comprise disorders in which plexin D1 is expressed on tumor cells, tumor blood vessels or activated macrophages.

Claims

1.-38. (canceled)

39. A method for detecting the expression of plexin D1, the method comprising: providing a sample from a subject comprising tumor cells and/or activated macrophages; contacting the sample with an antibody or antibody fragment specific for plexin D1 in vitro; and detecting binding of the antibody or antibody fragment to the tumor cells and/or activated macrophages.

40. The method according to claim 39, wherein the tumor cells are from a tumor selected from the group consisting of brain tumors, astrocytomas, oligodendrogliomas, hemangioblastomas, colon carcinomas, ductal carcinomas of the colon, prostate carcinomas, renal cell carcinomas, renal clear cell carcinomas, ovary carcinomas, squamous cell carcinomas, melanomas, lung carcinomas, small-cell lung carcinomas, non-small cell lung carcinomas, and soft tissue sarcomas.

41. The method according to claim 39, wherein the tumor cells are from a tumor selected from the group consisting of prostate carcinomas, renal cell carcinomas, renal clear cell carcinomas, ovary carcinomas, squamous cell carcinomas, melanomas, lung carcinomas, small-cell lung carcinomas, non-small cell lung carcinomas, and soft tissue sarcomas.

42. The method according to claim 39, wherein the binding of the antibody or the antibody fragment to the tumor cells and/or activated macrophages indicates the presence of a disorder selected from the group consisting of brain tumors, astrocytomas, oligodendrogliomas, hemangioblastomas, colon carcinomas, ductal carcinomas of the colon, prostate carcinomas, renal cell carcinomas, renal clear cell carcinomas, ovary carcinomas, squamous cell carcinomas, melanomas, lung carcinomas, small-cell lung carcinomas, non-small cell lung carcinomas, and soft tissue sarcomas.

43. The method according to claim 39, wherein the binding of the antibody or the antibody fragment to the tumor cells and/or activated macrophages indicates the presence of a disorder selected from the group consisting of prostate carcinomas, renal cell carcinomas, renal clear cell carcinomas, ovary carcinomas, squamous cell carcinomas, melanomas, lung carcinomas, small-cell lung carcinomas, non-small cell lung carcinomas, and soft tissue sarcomas.

44. The method according to claim 39, wherein the antibody or antibody fragment is labeled with a detectable marker.

45. The method according to claim 44, wherein the detectable marker is selected from the group consisting of a radioactive label, a paramagnetic label, a fluorescent label, and a chemiluminescent label.

46. The method according to claim 39, wherein the sample is from bodily tissue or fluid.

47. The method according to claim 39, wherein the tumor cells are ovarian carcinoma cells.

Description

BRIEF DESCRIPTION OF THE FIGURES

[0087] The present invention will be further illustrated in the Examples that follow and that are in no way intended to limit the invention. In the Examples, reference is made to the following figures:

[0088] FIG. 1: Structural domains of plexin family members: four subfamilies have been identified, named PlexinA-PlexinD. Horizontally hatched boxes indicate Sema domains, diagonally hatched boxes indicate Met-related sequences (MRS) motifs, and the clear box indicates the atypical MRS motif of PLXND1. PlexinB subfamily members have a potential furin-like proteolytic site, marked by a grey ribbon. The transmembrane region is marked by a shaded box and is followed by two conserved intracellular domains, together comprising the SP-domain, marked by two ovals.

[0089] FIG. 2: Panel A) In situ hybridization analysis of cerebral Me157-VEGF-A.sub.165 lesions using a digoxigenin-labeled mouse-specific plxnd1 RNA probe. Tumor vessels are strongly positive (arrows) whereas brain capillaries, distant from the lesions, are negative (compare the ISH profile with the CD34 staining in FIG. 2, Panel B).

[0090] FIG. 3: Human PLXND1-specific ISH analyses of glioblastoma multiforme (Panel A), brain metastases of sarcoma (Panel B), melanoma (Panel C) and mammacarcinoma (Panel D). Insets show CD31 stainings of serial sections. Control ISH using sense probes were negative (not shown). Note that in these tumors, PLXND1 expression is not confined to the blood vessels. Also, in tumor cells, high levels of the PLXND1 transcript are found, t=tumor, V=vessel.

[0091] FIG. 4: ISH analysis using a human-specific digoxigenin-labeled RNA probe (Panel A) and immunohistochemical staining with CD31 (Panel B) of normal brain. Note that vessels are abundantly present but these do not express the plexin D1 transcript.

[0092] FIG. 5: Specificity of phages (Panel A) and corresponding single-domain antibodies (sdabs) (Panel B) A12 and F8 for peptide H.sub.2N-ALEIQRRFPSPTPTNC-CONH.sub.2 (SEQ ID NO:8). In Panel A, 10.sup.10 phages were allowed to bind to PLXND1-peptide, BSA, human IgG or irrelevant peptide as described in the text. After rigorous washing, bound phages were detected using an anti-M13 antibody. In Panel B, similar incubations were performed but now with soluble sdabs. After washing, bound sdabs were detected and semi-quantified via the VSV-G-tag.

[0093] FIG. 6: The dissociation constants (Kds) of the binding between single-domain antibodies A12 and F8 were determined using the Biacore 2000 (Uppsala, Sweden) biosensor. The sensor chip and protein coupling chemicals were purchased from Biacore AB. PLXND1-peptide-KLH conjugate (27 ?g/ml in Na-Acetate, pH 4.0) or BSA (1 ?g/ml in Na-Acetate, pH 5.0) was coupled to activated CM5 surfaces using N-ethyl-N-(dimethylaminopropyl)carbodiimide, N-hydroxysuccinimide, under conditions recommended by the manufacturer. Unreacted groups were inactivated by 1 M ethanolamine, pH 8.5.

[0094] Kinetic measurements were performed at 25? C. with a flow rate of 10 ml/minute in HBS-EP buffer (10 mM Hepes, pH 7.4, 150 mM NaCl, 3 mM EDTA, 0.005% surfactant P20). Six concentrations of Ni-affinity-purified sdabs (in the range of 1 mM to 50 ?M) were used to determine the dissociation constants (Kds) of the interaction with the PLXND1-peptide. After each experiment, regeneration of the sensor surface was performed with 10 mM NaOH.

[0095] Specific binding, defined by binding to a PLXND1-surface minus binding to a control BSA-surface, was analyzed using the BIAevaluation 4.1 software and a 1:1 Langmuir binding model. Affinities of single-domain antibodies A12 and F8 were 2.1?10.sup.?8 M and 3.5?10.sup.?8 M.

[0096] FIG. 7: Evaluation of plexin D1 specificity of single-domain antibodies. Panel A shows immunohistochemical staining of the growth plate of trabecular bone of a mouse embryo (E16.5) with single-domain antibody A12, utilizing the VSV-G tag for detection of the antibody. The inset shows an in situ hybridization of a similar embryonic structure using a mouse-specific digoxigenin-labeled plexin D1 probe. Note the overlap of plexin D1 in situ hybridization and immunostaining. Panel B is a representative example of a Me157-VEGF-A.sub.165 lesion in brain of a nude mouse. The vasculature, which is also positive in plexin D1 ISH (see also FIG. 2), is immunopositive with single-domain antibody F8.

[0097] FIG. 8: Immunostainings with single-domain antibody A12 on a selection of human brain tumors. Tumors shown are (Panel A) glioblastoma multiforme, metastases of (Panel B) melanoma (Panel C) mammacarcinoma and Panel D) renal cell carcinoma. The insets in Panels A and B consist of control stainings with anti-VSV antibody only, and show that the tumor staining is specific. Note that vessels and tumor cells are highly reactive with the antibody.

[0098] FIG. 9: Immunostainings with single-domain antibody A12 on a progression series of melanoma. Immunostainings were performed on a nevus, a dysplastic nevus and horizontal and vertical growth phases of melanoma. Note that only the neoplastic cells express plexin D1.

[0099] FIG. 10: Immunostainings with single-domain antibody A12 on sections of Me157-VEGF-A brain tumors in mice, treated with ZD6474. In untreated or placebo-treated mice, tumor vessels stain positive with this antibody. However, in ZD6474-treated mice, there is a dose-dependent decrease of plexin D1 expression. ZD6474 was given orally, once daily, in the dosage as indicated.

[0100] FIG. 11: Double immunostainings with the macrophage marker CD68 and single-domain antibody A12 on mammacarcinoma. A subpopulation of macrophages express plexin D1 as revealed by a staining protein.

[0101] FIG. 12: In vivo homing of phage A12, F8 or an irrelevant phage to Me157-VEGF.sub.165 brain lesions. Tumor-bearing mice were injected with 10.sup.12 phages in the tail vein and, after 5 minutes, mice were anesthetized and subjected to cardiac perfusion with 15 ml of phosphate-buffered saline. Mice were sacrificed, brains removed and frozen sections were analyzed for phage content and distribution. Panel A) M13 staining of a frozen section of brain Me 1 57-VEGF.sub.165 lesions. Phages are clearly vessel-associated, as evidenced by the anti-CD34 immunostaining on a serial section, shown in Panel B). The arrows point at a CD34-positive vessel, distant from the lesion, which is not highlighted by anti-M13 staining. The inset in Panel A shows a control experiment where an irrelevant phage was injected. Panel C) Distribution of sdab F8 after intravenous injection in tumor-bearing mice. Sdabs are visualized by immunohistochemistry using an anti-VSV antibody. Note that the sdab is detected in tumor vessels but not in normal brain capillaries. The inset shows the control experiment where an irrelevant sdab was injected. An interstitial localization was observed, consistent with the leaky nature of the vessels in these tumors. Panel D) Quantification of phage homing. Tumor tissue was dissected from 10 ?m frozen sections using laser capture dissection microscopy. Number of colony-forming phages (cfp) were counted after infection of TG1 cells. Twenty-fold more F8 phages were eluted from tumors than from comparable areas of unaffected brain tissue.

[0102] FIG. 13: Single-domain antibody homes to tumor vessels that are not per se newly formed. Nude mice were inoculated with a cell suspension of 1.5?10.sup.5 cells of the human glioma xenograft E98, which was obtained from a subcutaneous E98 tumor. After 3 weeks, phages carrying single-domain antibody F8 were injected in the tail vein, and after 5 minutes, mice were anesthetized and subjected to cardiac perfusion using 15 ml of phosphate-buffered saline. Mice were then sacrificed, brains removed and fixed in formalin. Serial sections were stained with antibodies against M13 p8 protein (Panel A), the endothelial marker CD34 (Panel B) and glut-1 (Panel C, a marker for pre-existent brain capillaries). Comparison of Panels A, B and C reveals that not only newly formed tumor vessels accumulate phage F8, but also non-dilated brain vessels that express glut-1, and that, therefore, are considered pre-existent brain vessels that had been incorporated in the tumor.

[0103] FIG. 14: Effects of extracellular domains of plexin D1 on the development of tumor vasculature. Double transfectants of the human melanoma cell line Me157, expressing VEGF-A.sub.165 and the extracellular domain of plexin D1 comprising amino acids 1-850, were injected in the right internal carotid artery of nude mice. After three weeks, the mice were subjected to Gadolinium-DTPA-enhanced magnetic resonance imaging. Panel A shows MR images of two control mice carrying Me157 brain tumors that express VEGF-A.sub.165 only, Panel B shows MR images of brains of two mice carrying the double transfectant. Vascular leakage, as assayed by Gd-DTPA extravasation, tends to be less in the double transtectants, suggesting that VEGF-A-induced vascular leakage is counteracted by the plexin D1 ectodomain. More importantly, blood vessels in the double-transfected tumors are activated, as indicated by upregulation of CD34, yet they express glut-1, strongly suggesting that these vessels are pre-existent vessels that are incorporated in the tumor by the phenomenon of co-option. Note that the blood vessels in the tumors expressing VEGF-A.sub.165 only, are negative for glut-1 and, therefore, can be considered to be newly formed.

[0104] FIG. 15: Western blots were generated with recombinant plexin D1 ectodomains, expressed in E. coli and encompassing amino acids 47-506 (lanes 1) or 225-388 (lanes 2). Serum of mouse 25 was tested before (Panel A) and after (Panel B) immunization with plexin D1 region 47-506. As shown in Panel B, the mouse immune serum specifically recognized E. coli recombinant protein 47-506 (52 kDa, lane 1) and the protein encompassing plexin D1 residues 225-388 (a 18 kDa protein that lies completely within the sequence that was used for immunization, lane 2). The pre-immune serum did not show such a reactivity (Panel A). When tested in immunohistochemical stainings on a brain metastasis of an alveolar soft tissue sarcoma, the mouse immune serum (Panel D), but not the pre-immune serum (Panel C), showed positivity toward blood vessels and tumor cells, a staining pattern that was similar to that of single-domain antibody A12.

[0105] FIG. 16: Immunohistochemistry with monoclonal 1 ?M antibodies, obtained from B-lymphocytes from mouse 25. Antibodies 11F5H6 and 17E9C12 were selected based on reactivity against protein 47-506 in ELISA, and were analyzed for their potential to detect plexin D1 in frozen sections of human tumors. These antibodies showed strong positivity in brain metastases of sarcoma and melanoma, as illustrated in the figure. Of note, the insets in Panels C-F represent control stainings in which the primary antibody was omitted. Panels A and B show that these antibodies do not notably recognize vessel structures in normal brain tissue.

[0106] FIG. 17: Tumor homing of antibody 11F5H6. To further evaluate whether monoclonal antibody 11F5H6 is able to recognize tumor blood vessels, angiogenic Me157-VEGF-A tumors were grown in brains of nude mice, essentially as described in Example 10. Antibody 11F5H6 (1 mg) was injected in a lateral tail vein and allowed to circulate for 15 minutes. After this period, the mice were anesthetized with 1.3% isoflurane and the chest was opened, upon which a cardiac perfusion was performed with 20 ml phosphate-buffered saline. After this procedure, mice were decapitated, and brains removed and snap-frozen or fixed in formalin. Frozen sections of 4 ?m were stained with anti-IgM antibody. In FIG. 17, Panel A, it is shown that antibody 11F5H6 homes to and accumulates in tumor vessels but not in normal vessels (compare anti-IgM staining in Panel A with the anti-endothelial CD31 staining in Panel B). Such staining is not seen when performing anti-IgM staining on non-injected mice. Thus, 11F5H6 is a promising antibody that allows tumor targeting.

[0107] FIG. 18: Expression of Plexin D1 in macrophages in a mouse model of rheumatoid arthritis. Stainings were performed with single-domain antibody A12.

[0108] FIG. 19: expression of Plexin D1 in atherosclerosis. A subset of macrophages in human atherosclerotic plaques expresses plexin D1. Stainings were performed with single-domain antibody A12. A double staining was performed, displaying plexin D1 in red and the macrophage marker CD68 in blue. A purple color indicates co-expression.

[0109] The tables show the following:

[0110] Table 1: Analysis of different pathologies for plexin D1 expression.

[0111] Table 2: Plexin D1 expression in melanocytic lesions increases from benign to malignant lesions.

DETAILED DESCRIPTION

EXAMPLES

Example 1

Specific Expression of Plexin D1 on Tumor-Associated Blood Vessels

[0112] Plexin D1 is expressed on neurons but also endothelial cells in angiogenic vessels during embryogenesis.

[0113] The present invention demonstrates that plexin D1 is expressed on tumor-associated blood vessels but not on normal blood vessels. This has been shown by in situ hybridization of mouse brains, containing angiogenic human melanoma lesions (FIG. 2). The animal tumor model is described in (B. Kusters et al., Cancer Res. 63:5408-5413 (2003)). In short, tumor cells are injected via a microsurgical procedure in the right carotid artery, resulting in tumor growth in the parenchyma of the right brain hemisphere. After three weeks, at the onset of neurological symptoms, mice are sacrificed and brains removed and fixed in formalin.

[0114] Sections of 4 ?m were subjected to in situ hybridization with digoxigenin-labeled sense and antisense RNA fragments. RNA probes were generated by transcription using T3 and T7 RNA polymerase, respectively, from a PCR product, encompassing 600 bases in the 3-untranslated region, and which was flanked by T7 and T3 promoters (Van der Zwaag et al. (2002), supra).

[0115] In situ hybridizations using antisense RNA probes and sense RNA probes as negative controls, were performed using standard protocols. Sections were deparaffinated by melting paraffin at 60? C. and subsequent treatments with xylene and ethanol. After rehydration in phosphate-buffered saline (PBS), a proteinase K digestion was performed (10 ?g/ml PBS in 20 mM Tris-HCl pH7.4/5 mM EDTA) for 15 minutes at 37? C. Sections were post-fixed in 4% buffered formaldehyde for 10 minutes, and acetylated in 0.1 M acetic acid anhydrid. Slides were washed subsequently in 2?SSC (sodium Citrate/sodium chloride) and MILLIQ?. After drying, slides were hybridized with digoxigenin-labeled RNA probes overnight at 65? C. in 50% formamide/2?SSC.

[0116] High levels of plexin D1 RNA were observed in vessels of angiogenic Me157 tumors (FIG. 2) using a mouse-specific plexin D1 RNA probe. Tumor cells were also positive for the transcript. The non-perfect homology between mouse and human plexin D1 results in a weaker signal in the human tumor cells using the mouse probe.

Example 2

Expression of Plexin D1 in Tumors

[0117] To investigate plexin D1 RNA expression in human tumor samples, in situ hybridizations were performed with a human-specific plexin D1 RNA probe. High plexin D1 RNA expression levels were found in a number of human tumors, including glioblastoma multiforme, brain metastases of sarcoma, renal cell carcinoma, adenocarcinoma of the colon and of the breast, both in tumor vasculature and tumor cells. A summary of plexin D1-expressing tumor types is given in Table 1. FIG. 3 shows some examples of in situ hybridizations, e.g., a glioblastoma, a brain metastasis of melanoma and a brain metastasis of colon carcinoma. Plexin D1 RNA was found not only on the tumor vasculature, but also excessively on the tumor cells themselves. Importantly as in FIG. 4, Panel A, no plexin D1 RNA expression is observed in normal brain vasculature. In FIG. 4, Panel B, a CD31 staining is shown, demonstrating that abundant vessels are present in these sections.

Example 3

Preparation on of Antibodies Against Plexin D1

[0118] To detect plexin D1 protein, antibodies were selected with affinity toward plexin D1. To this end, a M13 pHENIX phage library was constructed expressing Llama single-domain V-H antibodies, constructed by RT-PCR from Llama B-lymphocytes as described (S. van Koningsbruggen et at., J. Immunol. Methods 279:149-161 (2003)). The population of resulting cDNAs encoding V-H-single-domain antibody (sdab) fragments was ligated into phagemid vector pHENIXHis8VSV (results not shown), resulting in a fusion product with an 8*His-tag and a VSV-G-tag at the C-terminus. After electroporation in E. coli TG1 cells, ampicillin-resistant colonies were collected and pooled.

[0119] The resulting library had a complexity of 8?10.sup.8 clones. Eighty percent of plasmids contained full-length sdab insert as determined by PCR analysis and immunological dot-blot-detection of the VSV-G-tag in sdabs (see below). The phage library was propagated as phagemids in E. coli TG1 bacteria. Phage particles were rescued by infection with trypsin-sensitive helper phage M13K07 (50). Phages were purified and concentrated from the culture supernatant by precipitation with 20% Polyethyleneglycol/2.5 M NaCl via standard methodology.

[0120] To select for phages, displaying antibodies with affinity toward plexin D1, immunotubes (Nunc, Roskilde, Denmark) were coated overnight at 4? C. with 5 ?g/ml KLH-conjugated peptide (H.sub.2N-ALEIQRRFPSPTPTNC-CONH.sub.2 (SEQ ID NO:8), corresponding to amino acids 1-16 of the mature human PLXND1 protein (accession no. AY116661) in 50 mM NaHCO.sub.3 (pH 9.6). Of note, the glutamic acid on position 3 in this peptide is a lysine in the mouse sequence; the remaining amino acids are homologous to mouse plxnd1.

[0121] After rigorous washing with PBS/0.05% TWEEN? 20 (PBST), non-specific binding sites were blocked with 5% marvel in PBST (MPBST, 1 hour at room temperature (RT)) and 10.sup.13 phage particles from the library stock were incubated with the immobilized peptide for 90 minutes at RT. After rigorous washing with PBST and PBS, bound phages were eluted by trypsin treatment (10 mg/ml, 30 minutes RT).

[0122] After trypsin inactivation with 1% newborn calf serum, the eluate was used to infect log-phase TG1 cells to amplify PLXND1-binding phages and calculate number of binders.

[0123] To enrich for binding phages, four rounds of selection were performed. From the second round on, selections were performed against unconjugated peptides, immobilized on DNA-binding plates (Costar) to prevent selection of KLH-binders.

[0124] Individual PLXND1-binding phages with PCR-confirmed full-length sdab inserts were tested for specificity toward plexin D1. Wells of DNA-binding plates or immunoplates (Nunc) were coated overnight at 4? C. with PLXND1-peptide or an irrelevant peptide (1 ?g/well in PBS/0.5 M NaCl pH 9.0), Bovine serum albumin (1 ?g/well in 50 mM NaHCO.sub.3 pH 9.6) or human Immunoglobulin G (1 ?g/well in 50 mM NaHCO.sub.3 pH 9.6). After blocking non-specific binding sites with MPBST, wells were incubated with phages in MPBST for 1 hour at RT and non-bound phages removed by rigorous washing. Bound phages were detected using HRP-conjugated anti-M13 (Amersham Pharmacia Biotech, Piscataway, N.J., USA) and tetramethylbenzidine (TMB; bioMerieux B.V., Netherlands). The reaction was terminated with 2 M H.sub.2SO.sub.4 and enzymatic activity quantified by measuring absorbance at 450 nm using an ELISA reader.

[0125] Using this selection procedure, phages displaying V-H single-domain antibodies A12 and F8 on their surfaces were identified as specific binders. FIG. 5, Panel A, shows that M13 phage-associated antibodies A12 and F8 bind specifically to plexin D1 peptide, but not to bovine serum albumin, immunoglobulins or an irrelevant peptide.

[0126] Expression of soluble single-domain antibodies was induced in log-phase E. coli TG1 cells by culturing at 30? C. in 2?TYA medium/1 mM IPTG. Sdabs were collected by osmotic lysis using ice-cold TES buffer (200 mM TrisHCl, 0.5 mM EDTA, 500 mM sucrose) containing a protease inhibitor cocktail (Roche, Basel, Switzerland). Sdab concentrations were estimated via dot-blot analysis using the mouse monoclonal anti-VSV-G P5D4, alkaline phosphatase-conjugated rabbit anti-mouse immunoglobulin (Dako, Denmark) and NBT/BCIP staining. Sdabs were tested in ELISA for PLXND1-peptide specificity. Single-domain antibodies A12 and F8 did not bind to irrelevant peptide, not to bovine serum albumin, and not to human immunoglobin G (FIG. 5, Panel B). The dissociation constants (Kds) of the binding between single-domain antibodies A12 and F8 were determined using the Biacore 2000 (Uppsala, Sweden) biosensor. The sensor chip and protein coupling chemicals were purchased from Biacore AB. PLXND1-peptide-KLH conjugate (27 ?g/ml in Na-Acetate, pH 4.0) or BSA (1 ?g/ml in Na-Acetate, pH 5.0) was coupled to activated CM5 surfaces using N-ethyl-N-(dimethylaminopropyl) carbodiimide, N-hydroxysuccinimide, under conditions recommended by the manufacturer. Unreacted groups were inactivated by 1 M ethanolamine, pH 8.5.

[0127] Kinetic measurements were performed at 25? C. with a flow rate of 10 ml/minute in HBS-EP buffer (10 mM Hepes, pH 7.4, 150 mM NaCl, 3 mM EDTA, 0.005% surfactant P20).

[0128] Six concentrations of Ni-affinity-purified sdabs (in the range of 1 mM to 50 ?M) were used to determine the dissociation constants (Kds) of the interaction with the PLXND1-peptide. After each experiment, regeneration of the sensor surface was preformed with 10 mM NaOH. Specific binding, defined by binding to a PLXND1-surface minus binding to a control BSA-surface, was analyzed using the BIAevaluation 4.1 software and a 1:1 Langmuir binding model.

[0129] Affinities of single-domain antibodies A12 and F8 were 2.1?10.sup.?8 M and 3.5?10.sup.?8 M, respectively (FIG. 6).

Example 4

Immunohistochemical Stainings with Single-domain Antibodies A12 and F8

[0130] The single-domain antibodies are tagged at the carboxyterminal end with a VSV-His-tag, enabling immunohistochemical stainings using an anti-VSV antibody. The following protocol was followed for immunohistochemical stainings with single-domain antibodies A12 and F8. Following deparaffinization, endogenous peroxidase activity was blocked by incubation with 0.03% H.sub.2O.sub.2. Antigen retrieval was performed by treatment with pronase according to standard protocols. Subsequently, slides were pre-incubated with normal horse or goat serum (to block non-specific binding sites in sections of human and mouse tissues, respectively), followed by incubation with sdabs for 1 hour. Sdabs were detected by sequential 1-hour incubations with a mouse or rabbit anti-VSV-G antiserum (Sigma-Aldrich Chemie B.V., Zwijndrecht, The Netherlands), biotinylated anti-mouse or anti-rabbit antibody as appropriate (Vector, Burlingame, Calif.), and avidin-biotin peroxidase complex (Vector, Burlingame, Calif.). Finally, peroxidase was visualized by the 3-amino-9-ethylcarbazole (ScyTek, Utah, USA) peroxidase reaction with hematoxylin as counterstain. All steps were performed at RT.

[0131] The specificity of the antibody A12 and F8 for plexin D1 in immunohistochemical stainings was first examined by staining mouse embryos in which expression patterns of plexin D1 on the RNA level were well characterized (Van der Zwaag et al. (2002), supra), and comparing profiles with immunostainings with anti-endothelial antibody anti-CD31 (DAKO, Glostrup, Denmark). In growth plate of trabecular bone of mice embryos at E16.5, immunostaining was observed on CD31-positive blood vessels. The staining profile correlated well to in situ hybridization for the plexin D1 transcript (FIG. 7, Panel A). The blood vessel origin of PLXND1 expression was further confirmed by performing stainings on serial sections with sdabs and anti-human anti-CD31 antibody (anti-human CD31).

Example 5

Staining of Tumor Cells with F8

[0132] Four ?m sections cerebral mouse xenografts of the human melanoma cell line Me157-VEGF-A (Kusters et al. (2003), supra) were stained with single-domain antibody F8, according to the protocol exemplified in Example 4. The antibody clearly recognized plexin D1 on tumor blood vessels (FIG. 7, Panel B). To further investigate plexin D1 protein expression on tumors, archival paraffin-embedded or tumor tissue of different origin (glioblastoma multiforme (FIG. 8, Panel A), brain metastases of melanoma (FIG. 8, Panel B) colon carcinoma (FIG. 8, Panel C) and renal cell carcinoma (FIG. 8, Panel D) were immunostained with anti-PLXND1 sdabs. Immunohistochemistry using antibody A12 and comparison with anti-human CD31 stainings on serial sections, showed expression on all tumors examined and confirmed plexin D1 expression on the protein level in tumor cells and in tumor blood vessels.

Example 6

Timing of Plexin Expression on Malignant Cells

[0133] To investigate whether expression of plexin D1 occurs on premalignant cells, a progression series of melanoma was stained, consisting of benign nevi, dysplastic nevi, radial growth phase melanoma, invasive melanoma and disseminated melanoma. Melanocytes in benign nevi and dysplastic nevi do not express the protein, whereas malignantly transformed cells, both in radial growth phase and vertical growth phase tumors are positive for the protein (FIG. 9 and Table 2).

Example 7

Activation State of Plexin D1-Expressing Cells

[0134] Plexin D1 expression is related to the activation state of the endothelial cells in tumor blood vessels. Treatment with ZD6474, an inhibitor of VEGFR2 and EGFR, was previously shown to block angiogenesis in a mouse brain tumor model, resulting in a phenotypic shift from an angiogenic to a non-angiogenic vessel co-opting phenotype (43). Treatment with ZD6474 resulted in a decrease of plexin D1 expression on tumor-associated blood vessels in a dose-dependent manner (FIG. 10). Thus, plexin D1 expression is a characteristic of activated endothelial cells.

Example 8

Immunohistochemistry with A12 on Normal Tissues

[0135] Expression of plexin D1 in normal brain, heart, skin, kidney, spleen, intestine, and endometrium was examined by immunohistochemistry using antibody A12. Vessels in proliferative myometrium-expressed plexin D1, showing that plexin D1 is associated not only with pathological angiogenesis, but also with physiological angiogenesis (not shown).

[0136] In some instances, co-immunostainings were performed with the CD68 macrophage marker. These stainings revealed that a subpopulation of macrophages expressed the protein (FIG. 11). Also, fibroblasts in skin and some proliferating intestinal epithelial cells were found to express plexin D1 (not shown).

Example 9

Staining of Macrophages in Inflammatory Diseases

[0137] To further examine the involvement of plexin D1 in diseases with prominent macrophage involvement, immunohistochemical stainings were preformed on atherosclerotic plaques, multiple sclerosis and rheumatoid arthritis. Macrophages express plexin D1.

Example 10

Access to Plexin D1 in Tumor Vessels Via Intravenous Injection

[0138] The expression of plexin D1 protein on tumor blood vessels suggests that plexin D1 is accessible via intravenous injection. To test this, 2?10.sup.5 stably transfected Me157 cells expressing the VEGF-A.sub.165 isoform were microsurgically injected into the right internal carotid artery of BALB/C nude mice. After 18 days, when animals showed neurological symptoms (Kusters et al. (2003), supra), 10.sup.12 PLXND1-binding phages of clones A12, F8 or non-relevant phages were injected in the tail vein of nude mice, carrying established Me157-VEGF-A.sub.165 brain metastases (n=2 for A12, n=4 for F8, n=3 for control phage).

[0139] In two other groups of mice, 30 ?g sdab F8 or a control sdab (n=2 for each group) was intravenously injected. After 5 minutes, mice were anesthetized using isoflurane, the chests were opened, and non-bound phages were washed from the system by cardiac perfusion with 15 ml of phosphate-buffered saline (PBS). Then, mice were sacrificed by cervical dislocation, and parts of brains, hearts, lungs, livers, spleens and kidneys were snap frozen in liquid nitrogen.

[0140] Other parts were fixed in formalin to be paraffin-embedded. After short hematoxylin staining, tumors were dissected from 10 ?m brain sections using laser capture dissection microscopy (Leica laser dissection microscope). Equivalent areas were dissected from unaffected brain, contralateral to the tumor.

[0141] Subsequently, phages were eluted from dissected tissue samples using trypsin treatment and used to infect TG1 cells. Numbers of colony-forming phages were counted and used as a measure of tumor homing. To qualitatively assess tumor homing by phages or sdabs, 4 sections, serial to the sections used for laser dissection, were stained with anti-M13 p8 antibody (Abcam Limited, Cambridge, UK) to detect bound phages, or anti-VSV-G antibodies (Sigma-Aldrich) to detect single-domain antibodies.

[0142] Intravenous injection of M13 phages displaying anti-PLXND1 single-domain antibody F8, but not phages carrying irrelevant single-domain antibodies, in mice carrying angiogenic melanoma lesions resulted in accumulation of phages in tumor vessels but not to detectable specific presence of phages in normal brain vessels, nor blood vessels in liver, spleen, kidney (FIG. 12, Panels A and D, not shown). This indicates that plexin D1 is expressed at the luminal side of the endothelial cell specifically in tumor blood vessels and thus can be used as a targetable marker.

[0143] Injection of the partially purified single-domain antibody accordingly led to preferential tumor localization (FIG. 12, Panel C). In the latter situation, it must be considered that the small molecular weight of 20 kDa of the single-domain antibodies enable extravasation from the highly permeable tumor vessels and accumulation in the tumor interstitium. This latter effect is non-specific and is also observed with non-relevant single-domain antibodies. It is envisioned that antibodies of small molecular weight and relatively low affinities have higher penetrability through tumors and are more suitable for targeting the tumor cell compartment.

Example 11

Accumulation of F8 in Tumor Blood Vessels

[0144] Mice were injected transcranially with E98, a glioma xenograft line. E98 tumors are maintained as subcutaneous tumors. A Balbc/c nu/nu athymic mouse carrying a subcutaneous E98 tumor was killed and the tumor removed. The tumor was minced with a sterile scalpel and the homogenate was passed through a sterile 70 ?m mesh nylon filter. Twenty ?l of the resulting cell suspension, containing 150,000 cells, was injected transcranially in the brain of nude mice. After 3 weeks, M13 phages displaying single-domain antibody F8 were injected intravenously, and after five minutes the mouse was subjected to cardiac perfusion with 15 ml of phosphate-buffered saline.

[0145] The mice were killed, brains removed and fixed in formalin. Four ?m sections were subjected to immunohistochemistry with anti-M13 antibody, and serial sections were stained immunohistochemically with antibodies against CD34 (endothelial marker) and glut-1 (a marker for pre-existent brain endothelial cells (B. Kusters et al., Cancer Res. 62:341-345 (2002)).

[0146] Phages carrying anti-plexin D1 single-domain antibodies accumulated specifically in tumor-associated blood vessels, but not in normal vessels (FIG. 13). Importantly, phages also accumulated in tumor blood vessels that were positive for glut-1, and that, therefore, can be considered as pre-existent blood vessels, rather than newly formed blood vessels. This indicates that not only angiogenic blood vessels are subject to targeting with anti-plexin D1 antibodies, but also non-angiogenic, yet activated blood vessels in tumors.

Example 12

Recombinant Plexin D1 Ectodomains Inhibit Angiogenesis

[0147] Human melanoma Me157 cells were transfected with the VEGF-A.sub.165 coding sequence in vector pIREShyg. Stably transfected cells were selected by culturing in 200 ?g/ml hygromycin in Dulbecco's Modified Eagle medium (DMEM) supplemented with 10% fetal calf serum (FCS) and penicillin/streptomycin. Because expression of the hygromycin resistance gene is linked to that of the VEGF-A cDNA via the internal ribosomal entry site (IRES), all hygromycin-resistant cells will produce the VEGF-A protein also. Stably transfected Me157-VEGF cells were subsequently transfected with pIRESnco-PlexinD1 ED. The vector contains the cDNA encoding the extracellular domain from nucleotides 1-2745, linked via the IRES to expression of the neomycin resistance gene.

[0148] Double transfectants were injected in the right carotid artery of nude mice, and tumors were allowed to develop. At the onset of neurological symptoms (approximately 18 days) mice were subjected to Gadolinium-DTPA-enhanced magnetic resonance imaging. Subsequently, mice were sacrificed, brains fixed in formalin and subjected to immunohistochemical stainings to examine the tumor vasculature.

[0149] When compared to controls, consisting of tumors expressing VEGF-A only, Gd-DTPA enhancement in T1-weighted magnetic resonance imaging (MRI) was less (compare FIG. 14, Panel A, representing two examples of Me157-VEGF-A.sub.165 tumors, with Panel B representing two examples of Me157-VEGF-A.sub.165/PLEXIND1-ED tumors. In the tumors expressing VEGF-A.sub.165 and Plexin D1 ectodomain, vasculature shows upregulation of the endothelial marker CD34 (a hallmark of endothelial activation by VEGF-A.sub.165). The vasculature in tumors expressing VEGF-A.sub.165 only is negative for the brain endothelial cell marker glut-1, which is consistent with the fact that these vessels are newly made and, therefore, lack brain-endothelial cell-specific markers. As can be seen in FIG. 12, Panel B, the vessels that are associated with tumors that express the plexin D1 ectodomain too, do express glut-1. This is a strong indication that these vessels are actually pre-existent. Thus, the plexin D1 ectodomain does not prevent activation of endothelial cells by VEGF-A.sub.165, but it does prevent the formation of neovasculature.

Example 13

High-Affinity Antibodies Against Plexin D1

[0150] A protein sequence, corresponding to amino acids 47-506 (the 459 most amino terminal amino acids of the mature protein), was expressed in E. coli M15 pREP4 cells, using the expression vecor pQE16 (Qiagen). The recombinant protein, which was produced in the bacterial cells as inclusion bodies, was dissolved in denaturing buffer, containing 4 M urea and 1 mM dithiothreitol (DTT) and afterwards gradually dialyzed against PBS. The protein was used to immunize BALB c/c mouse 25 according to standard procedures.

[0151] FIG. 15 shows the characteristics of the mouse serum. As shown in FIG. 15, Panel B, the mouse immune serum specifically recognized E. coli recombinant protein 47-506 (52 kDa, lane 1), and a second recombinant plexin D1 sequence of 18 kDa, comprising amino acids 225-388 (thus lying completely within the sequence that was used for immunization, lane 2). The pre-immune serum did not show such a reactivity (Panel A).

[0152] When tested in immunohistochemical stainings on a brain metastasis of an alveolar soft tissue sarcoma, the mouse immune serum (Panel D), but not the pre-immune serum (Panel C), showed positivity toward blood vessels and tumor cells, a staining pattern that was similar to that of single-domain antibody A12. Thus, the B-lymphocytes of this mouse were considered suitable to generate hybridomas of spleen B-lymphocytes with myeloma cell line SP2/0.

[0153] From these hybridomas, a number of antibody-producing cell lines were selected based on reactivity against protein 47-506 in ELISA, and were analyzed for their potential to detect plexin D1 in frozen sections of human tumors. Of these, 11F5 H6 and 17E9C12, both antibodies of the IgM subtype, showed strong positivity in brain metastases of sarcoma and melanoma, as illustrated in FIG. 16. The insets in Panels C-F represent control stainings in which the primary antibody was omitted. Panels A and B show that these antibodies do not notably recognize vessel structures in normal brain tissue.

Example 14

Monoclonal Antibody 11F5H6 is Able to Recognize Tumor Blood Vessels

[0154] To further evaluate whether monoclonal antibody 11F5H6 is able to recognize tumor blood vessels, angiogenic Me157-VEGF-A tumors were grown in brains of nude mice, essentially as described in Example 10. Antibody 11F5H6 (1 mg) was injected in a lateral tail vein and allowed to circulate for 15 minutes. After this period, the mice were anesthetized with 1.3% isoflurane and the chest was opened, upon which a cardiac perfusion was performed with 20 ml phosphate-buffered saline.

[0155] After this procedure, mice were decapitated, and brains removed and snap-frozen or fixed in formalin. Frozen sections of 4 ?m were stained with anti-IgM antibody. In FIG. 17, Panel A, it is shown that antibody 11F5H6 homes to and accumulates in tumor vessels but not in normal vessels (compare anti-IgM staining in FIG. 17, Panel A, with the anti-endothelial CD31 staining in FIG. 17, Panel B). Such staining is not seen when performing anti-IgM staining on non-injected mice. Thus, 11F5H6 is a promising antibody that allows tumor targeting.

Example 15

Plexin D1 Expression in Rheumatoid Arthritis

[0156] Plexin D1 is expressed in macrophages in mouse models of rheumatoid arthritis (FIG. 18). A subset of macrophages in human atherosclerotic plaques also expresses plexin D1 (FIG. 19). Stainings were performed with single-domain antibody A12. In FIG. 19, a double staining was performed, displaying plexin D1 in red and the macrophage marker CD68 in blue. A purple color indicates co-expression.

TABLE-US-00001 Sequences A12(SEQIDNO:1): ATGAAATACCTATTGCCTACGGCAGCCGCTGGATTGTTATTACTCGCGGC CCAGCCGGCCATGGCCCAGGTGCAGCTGCAGGAGTCTGGGGGAGGCTTGG TGCAGCCTGGGGGGTCTCTGAGACTCTCCTGTGCAGCCTCTGGAAGCAGT ATCAGTATCAATAACTGGGGCTGGTACCGCCAGGCTCCAGGAAAACAGCG CGAGCGGGTCGCAGCTATATCTGGTGGTAAAACAGTCTATGCGGACTCCG TGAAGGGCCGATTCACCATCTCCAGAGACAACGCCAAGAACACGGTGTAT CTGCAAATGAACAGCCTGAAACCTGAGGATACGGCCGTCTATTACTGTAG AGCAGTCCGGAAAAGTACGGGTTGGCTTAGGGGGCTTGACGTCTGGGGCC AGGGGACCCAGGTCACCGTCTCCGCAGAACCCAAGACACCAAAACCACAA CCAGCGGCCGCACATCATCACCATCATCACCATCATTATACAGACATAGA GATGAACCGACTTGGAAAGGGGGCCGCATAG A12proteinsequence(SEQIDNO:2) MKYLLPTAAAGLLLLAAQPAMAQVQLQESGGGLVQPGGSLRLSCAASGSS ISINNWGWYRQAPGKQRERVAAISGGGKTVYADSVKGRITISRDNAKNTV YLQMNSLKPEDTAVYYCRAVRKSTGWLRGLDVWGQGTQVTVSAEPKTPKP QPAAAHHHHHHHHYTDIEMNRLGKGAA@ F8(SEQIDNO:3): ATGAAATACCTATTGCCTACGGCAGCCGCTGGATTGTTATTACTCGCGGC CCAGCCGGCCATGGCCCAGGTGCAGCTGCAGGAGTCTGGGGGAGGATTGG TGCAGGCTGGAGACTCTCTGAGACTCTCCTGTGCAGCCTCTGGACGCACC TTCAGTACTTTGATTATGGCCTGGTTCCGCCAGGCTCCAGGGAAGGAGCG TGAATTTGTAGCGGCGATTAGCCGGGGTGGCGGTAGCACAAGCTATGCAG ACTCCGTGAAGGGCCGATTCACCATCTCCAGAGACAATTCCAAGAACGCG GTGTATCTACAAATGAACAGCCTGAAACCTGATGACACGGCCGTCTATTA CTGTAATGCCCGGTACGGTAGCCGAATTTACTGGGGCCAGGGGACCCAGG TCACCGTCTCCTCAGAACCCAAGACACCAAAACCACAACCAGCGGCCGCA CATGATCACCATCATCACCATCATTATACAGACATAGAGATGAACCGACT TGGAAAGGGGGCCGCATAG F8proteinsequence(SEQIDNO:4) MKYLLPTAAAGLLLLAAQPAMAQVQLQESGGGLVQAGDSLRLSCAASGRT FSTLIMAWFRQAPGKEREFVAAISRGGGSTSYADSVKGRFTISRDNSKNA VYLQMNSLKPDDTAVYYCNARYGSRIYWGQGTQVTVSSEPKTPKPQPAAA HHHHHHHHYTDIEMNRLGKGAA@ Sequencesinglechainantibody,derivedfrom antibody11F5H6(SEQIDNO:5) MKYLLPTAAAGLLLLAAQPAMADYKDIVMTQTPLSLPVSLGDQASISCRS SQSIVHSNGNTYLEWYLQKPGQSPKLLIYKVFNRLSGVPDRFSGSGSGTD FTLKISRVEAEDLGVYYCFQGSHVPLITGAGTKLELKRGGGGSGGGGSGG GGRAPGGGGSEVQLQQSGPELVKPGASMKISCKASGYSFTGYTMNWVKQS HGKNLEWIGLINPYNGGTSYNQKFKGKATLTVDKSSSTAYMELLSLTSED SAVYYCARAITTDGWFAYWGQGTLVTVSAAAAHHHHHHHHYTDIEMNRLG KGAA Sequencesinglechainantibody,derivedfrom antibody17EC12(SEQIDNO:6) MKYLLPTAAAGLLLLAAQPAMADYKDIQMTQTPSSLAVSAGEKVTMSCKS SQSVLYSSNQKNYLAWYQQKPGQSPKLLIYWASTRESGVPDRFFGSGSGT DFTLTISSVQAEDLAVYYCHQYLSSWTFGGGTKLEIKRGGGGSGGGGSGG GGSGGGGSQVQLQQSGAEINKPGASVKLSCTASGFNIKDTYMHWVKQRPE QGLEWIGRIDPANGNTKYDPKFQGKATITADTSSNTAYLQLSSLTSEDTA VYYCAMDYWGQGTSVTVSSAAAHHHHHHHHYTDIEMNRLGKGAA

TABLE-US-00002 TABLE 1 PLXND1 expression in human tissues Tissue PLXND1 expression Malignant Adenocarcinoma of oesophagus (n = 1) Tumor vessels and tumor cells Adenocarcinoma of rectum (n = 5) Tumor vessels, tumor cells and macrophages Adenocarcinoma of prostate (n = 1) Tumor vessels and tumor cells Alveolar soft part sarcoma of femur (n = 1) Tumor vessels and tumor cells Astrocytoma (n = 1) Tumor vessels Carcinoid tumor of lung (n = 1) Tumor vessels, tumor cells and macrophages Ductal carcinoma in situ of mamma (n = 5) Tumor vessel, tumor cells, macrophages, fibroblasts Follicular lymphoma (n = 8) Tumor vessels Glioblastoma Multiforme (n = 3) Tumor vessels and tumor cells Brain metastasis of adenocarcinoma (n = 4) (mamma, lung, rectum) Tumor vessels and tumor cells Brain metastasis of alveolar soft part sarcoma (n = 1) Tumor vessels and tumor cells Brain metastasis of renal cell carcinoma (n = 1) Tumor vessels and tumor cells Liver metastasis of adenocarcinoma colon (n = 2) Tumor vessels, tumor cells and macrophages Lobular carcinoma in situ of mamma (n = 3) Tumor vessels and tumor cells weakly positive, macrophages and fibroblasts Lymph node metastasis ductal mamma carcinoma (n = 1) Tumor cells and some tumor vessels Ovary metastasis of adenocarcinoma colon (n = 1) Tumor cells and myofibroblasts Renal cell carcinoma (n = 1) Tumor vasculature and tumor cells Urothelial cell carcinoma of prostate (n = 2) Tumor vessel, tumor cells and macrophages Non-malignant Bladder (n = 1) Macrophages Blood vessel, atherosclerosis (n = 6) Macrophages Bone marrow (n = 2) Brain cortex (n = 1) Some neurons perinuclear Brain, Alzheimer + CAA (n = 1) Endometrium Proliferation phase (n = 5) Macrophages Secretion phase (n = 4) Macrophages Secretion/menstruation phase (n = 1) Macrophages Endometriosis interna (n = 1) Macrophages Heart (n = 1) Some muscle cells perinuclear Large intestine (n = 1) Some luminal staining of epithelium, macrophages, fibroblasts Liver (n = 1) Liver cells perinuclear granular, macrophages Lung (n = 2) Macrophages Mamma (n = 2) Some epithelial cells perinuclear Mamma, ductal hyperplasia (n = 1) Focal epithelial cells perinuclear, macrophages Oesophagus (n = 1) Macrophages Small intestine (n = 1) Some luminal staining of epithelium macrophages, fibroblasts Spleen (n = 1) Macrophages

TABLE-US-00003 TABLE 2 PLXND1 expression in melanoma progression series Absent Moderate Abundant Naevocellular naevi (n = 18) 18 Atypical naevi (n = 14) 14 Melanomas in situ (n = 5) 5 Primary melanomas (n = 26) 4 2 20 Melanoma metastases Lymph node (n = 9) 1 2 6 Skin (n = 5) 1 1 3 Brain (n = 5) 5 Lung (n = 1) 1