• E-mail
• Print
• Comment
• Font Size
• Digg
• del.icio.us
• Discuss article

Human Endosialin (Tumor Endothelial Marker 1) Is Abundantly Expressed in Highly Malignant and Invasive Brain Tumors

Posted on: Thursday, 9 December 2004, 03:00 CST

Abstract. Endosialin (tumor endothelial marker 1) is expressed preferentially by tumor endothelial cells but not by normal endothelium. Its protein domain architecture is homologous to that of CD93 and thrombomodulin (CD141), suggesting a similar function in mediating cell-cell interactions. The aim of this study was to investigate the expression pattern of endosialin in human brain tumors in a bid to decipher its contribution to tumor angiogenesis. We generated an antibody specifically recognizing human endosialin and used it to study endosialin expression in 30 human brain tumor specimens by immunoblotting and immunohistochemistry. Twenty of 30 tumors expressed endosialin in a heterogeneous manner. The largest proportion of endosialin-expressing tumors was found in highly invasive glioblasloma multiforme, anaplastic astrocytomas, and metastatic carcinomas. Hndosialin was localized to the endothelium of small and large vessels strongly stained for CD31 and was also expressed by Thy-1-positive libroblast-like cells close to the meninges and α-smooth muscle actin-positive cells in some vessels. Hndosialin colocalized with thrombomodulin, suggesting the proteins may have complementary functions in tumor progression.

Key Words: Angiogenesis; Endosialin; Glioblasloma multiforme: Thrombomodulin; Tumor endothelial marker 1.

INTRODUCTION

The search for novel molecular markers that distinguish neoplastic from normal vasculature is producing many attractive therapeutic targets (1). One such potential target is endosialin, which is a member of a family of proteins involved in cell-cell interactions, including CD93 (the AA4 antigen, stem cell marker (2- 4)) and thrombomodulin (CD141) (5). Endosialin was discovered as the antigen to a monoclonal antibody FB5 raised against human fetal fibroblasts during a search for novel molecular targets of angiogenesis (6). Immunohistochemical staining of a panel of normal and malignant human tissues with the FB5 antibody identified endosialin expression in a subset of carcinomas, sarcomas, and neuroectodermal tumors, but no expression was evident in normal tissues. Endosialin was localized predominantly on the endothelial cells of tumor blood vessels, but was also found on a small number of malignant cells and reactive stromal fibroblasts in a subset of sarcomas and epithelial tumors respectively (6). More specifically, endothelial cell expression was primarily in capillaries rather than large vessels and the proportion of positively stained vessels was remarkably variable within the vascular bed (6). This is consistent with a role for endosialin in vascular reorganization during which endothelial cells are proliferating, migrating, and remodeling to form new capillaries in regions of the tumor that have become oxygen starved due to active tumor growth. Subsequent studies by St. Croix et al using SAGH (serial analysis of gene expression) technology to dissect the gene expression profiles of endothelial cells in human colorectal carcinomas identified tumor cndothelial marker 1 (TEM1, later shown to have an identical cDNA sequence to endosialin) as being consistently upregulated in tumor endothelium, and confirmed its expression to be restricted to the vascular endothelium in colorectal and other tumor types (7). Furthermore, in situ hybridization analysis showed endosialin was expressed in the corpus luteum and in granulation (issue of healing wounds (7, 8).

Detailed molecular and bioinformatic characterization of endosialin FB5 immuno-isolated from a human neuroblastoma cell line branded it a type 1 transmembrane glycoprotein with a distinct functional domain organization not dissimilar to CD93 and thrombomodulin (CD141) (9). The 757 amino acid endosialin protein is encoded by a 2,274 bp intronlcss gene on chromosome 11q13 (6, 9). Each of the 3 receptors comprises a short N-terminal signal peptide, followed by globular C-type lectin-like domain (CTLD), a tandem array of epidermal growth factor-like (EGP) domains (3 in endosialin, whereas CD93 and thrombomodulin have 5 and 6, respectively), a serine/threonine-rich region containing a significant number of potential O-glycosylation sites, a single transmembrane region, and a short low complexity cytoplasmic tail (3- 5, 9, 10). Endosialin has an additional domain between the CTLD and EGF domain that has been designated a putative short consensus repeat or complement control protein (CCP) domain on the basis of several conserved amino acid motifs (9). The overall amino acid identity between the 3 proteins is relatively low; however, key regions of the protein sequences are highly conserved. In particular, there is a 39% and 33% sequence identity between the N- terminal region of endosialin and thrombomodulin and CD93, respectively, as well as several highly conserved cysteine residues implying a homologous three-dimensional structure (9). Finally, the 3 proteins display a conserved WIGL consensus motif within their CTLD, a unique feature found in several additional cell surface proteins that modulate endocytosis (3, 9).

The function of endosialin in tumor angiogenesis remains to be defined, although inferences can be made from its structure, with CTLD, EGF, and CCP domains being separately implicated in cell-cell interactions and cell adhesion events (11-15); thus, a similar function for endosialin in the context of angiogenesis is an attractive possibility.

To date, there has been no detailed analysis of endosialin expression in human brain tumor subtypes. Brain tumors are highly invasive and the most aggressive are the glioblastoma multiforme and anaplastic astrocytomas (16). Both these tumors are characterized by microvascular proliferation with proliferation of endothelial and smooth muscle cells. An association between microvascular proliferation as an indicator of tumor angiogenesis has been proposed, as has a relationship between tumor necrosis and microvascular changes. Therefore the localization of endosialin in tumor blood vessels could provide information about control of angiogenesis and a potential target for anti-angiogenesis therapy. Brain tumors are a particularly attractive target for anti- angiogenic therapy due to the inherent difficulties associated with standard surgical and therapeutic options and thus their notoriously poor prognosis (17).

In this study, we have examined endosialin expression in different types of human brain tumors by immunohistochemistry and immunoblotting of tumor specimens using an antibody generated in our laboratory. The objective was to study variations in expression between different tumor types in order to decipher the contribution of endosialin to the tumor pathology.

MATERIALS AND METHODS

Tissue Specimens

Fresh tumor tissue was collected from 30 patients (age range 28- 75 years) during surgical resection of brain tumors. None had received radiotherapy or chemotherapy. The Local Research Ethics Committee (Bro Taf Health Authority) granted ethical approval for this study (02/4648). Specimens were snapfrozen in isopentane within several hours of surgery or embedded in paraffin wax. Ten-m sections were cut and mounted on glass slides for immunohistochemical analysis. Hematoxylin and eosin (H&E) staining was performed on paraffin sections by standard techniques to determine the morphology and provide a histological classification of the specimen. The cases were categorized as astrocytoma WHO grade II (n = 5), astrocytoma WHO grade III (n = 3), glioblastoma multiforme (GBM) WHO grade IV (n = 6), oligodendroglioma WHO grade II (n = 2), ependymoma WHO grade II (n = 1), meningioma WHO grade I (n = 5), metastatic carcinoma (n = 7), or melanoma (n =1). Specimens were randomized and all subsequent immunostaining was carried out blind.

Generation of Recombinant Human Endosialin

Human endosialin was cloned from genomic DNA prepared from a THP1 monocyte cell line. The cDNA coding for the CTLD was amplified by PCR in a 50-L reaction volume containing 100 ng of genomic DNA, 200 nmol/L of each primer (P1 5' CCGTCTAGAATGCTGCTGCGCCTGTTGCTG 3' (forward) and P2 5' TTGGGATCCAGCTCGCAACTGCGCCCG TC 3' (reverse), Xbal and BamH1 restriction sites respectively are italic), 200 nM dNTPs and 1U of Platinum Pfx DNA polymerase (Invitrogen, Paisley, UK). Thermal cycling was at 94C for 4 min for initial denaturation, followed by 94C for 45 s, 58C for 1 min, and 72C for 2 min for 5 cycles, 94C for 30 s, 58C for 30 s, and 72C for l min for 15 cycles, 94C for 45 s, 60C for 1 min and 72C for 2 min for 10 cycles, and a final elongation step at 72C for 15 min. The required band was extracted from a 0.7% agarose gel using a gel extraction kit (Qiagen, Crawley, UK) according to the manufacturer's instructions. The fragment was cloned into a pDR2&945; expression vector containing a human IgG1 Fc tail sequence by means of the Xba1 and BamH1 restriction enzyme sites as previously described (4). The plasmid was sequenced on the ABI 3100 genetic analyzer (Applied Biosystems, Warrington, UK).

CHO cells were transfected with the plasmid using the LipofectAMINE Plus Reagent (Invitrogen). Stable transfectants were selected in 400 g/mL of hygromycin for 2 weeks. Fusion protein secretion was confirm\ed by immunoblotting of tissue culture supernatant (TCS) with an antibody against the human IgG1 Fc tail (Sigma, Poole, UK). The fusion protein was then purified from liter volumes of TCS by passage over a prosep A affinity column (Bioprocessing, Durham, UK). The column was washed with 0.1 M citrate buffer pH5 to remove bovine immunoglobulins before elution of the fusion protein in 0.2 M glycine. The protein was dialyzed into PBS and concentrated using an Amicon system to give a final concentration of 1.5 mg/mL.

Antibody Production

A rabbit was immunized subcutaneously with 400 g of endosialin CTLD fusion protein in complete Freund's adjuvant. Subsequent boosts were at 3 weekly intervals in incomplete Freund's adjuvant. Test bleeds were taken on which antibody titre was measured by ELISA. Once the titre had reached a plateau, the animal was killed and antiserum was harvested. Contaminating antibodies against human and bovine immunoglobulins were depleted by several passages over a CnBr column to which 10 mg each of human and bovine immunoglobulins had been coupled. Depletion was confirmed by ELISA. Specific antibodies against human endosialin CTLD were subsequently isolated by loading the antiserum onto a CnBr column to which 2 mg of the endosialin fusion protein had been coupled. Specific antibodies were eluted in 50 mM diethylamine, dialyzed into PBS, and concentrated to 0.7 mg/ mL using an Amicon system.

Flow Cytometry

CHO cells transfected with full-length human endosialin or the human C3aR (irrelevant control) were washed in PBS and harvested using PBS with 10 mM EDTA. Cells were suspended in PBS-1%BSA at a density of 10^sup 6^/mL and incubated with the affinity-purified antibody against human endosialin (1:200) for 1 hour at 4C. Cells were washed and incubated with an rPE-conjugated goat anti-rabbit antibody (Sigma). Cells were washed, suspended in PBS with 1%BSA, and analyzed on a FACSCalibur (BD Biosciences, Oxford, UK).

Western Blot Analysis

Lysates were prepared from frozen tumor tissues by mixing the tissue in Triton X100 lysis buffer (1% Triton X-100, 10 mM EDTA, 1 mM PMSF, 10 g/mL pepstatin, 10 g/mL leupeptin in PBS) (100 L/2mg of fresh tissue) for 1 hour followed by sonication of the tissue. Debris was cleared from the lysate by centrifugation for 5 min at 13,000 rpm. An aliquot of the lysate was mixed with an equal volume of Laemmli buffer and proteins were resolved by SDS-PAGE on a 7.5% gel under non-reducing conditions. Proteins were transferred onto a nitrocellulose membrane, blocked for 30 min in PBS-5% dried milk, and blotted for 1 hour at room temperature with the affinity- purified antibody against human endosialin (1:1,000). The blot was washed well in PBS-0.1%Tween 20 and blocked in PBS-5% milk before incubation with a peroxidase-conjugated goat anti-rabbit secondary antibody (1:1,000, Bio-Rad, Hemel Hempstead, UK). After a further 3 washes in PBS-0.1%Tween 20. the blot was developed using an enhanced chemiluminescent system (Pierce, Chester, UK) and exposed to x-ray film.

Immunohistochemistry

Immunohistochemistry was performed on transfected cells, frozen tumor sections and paraffin-embedded tumor sections. CHO cells transfected with the full-length human endosialin or the human C3aR were grown on sterile glass coverslips in Ham's F12 medium containing 10% FCS and 100 g/mL hygromycin at 37C with 5% CO2. Confluent cells were washed with saline and fixed for 10 s in ice- cold acetone. Frozen tissue sections were air-dried and fixed in ice- cold acetone for 10 min. Paraffin sections were dewaxed by three 5- minute incubations in xylene followed by 3 incubations in absolute ethanol. Endogenous peroxidases were blocked in 0.3% hydrogen peroxide for 30 min at room temperature, followed by antigen retrieval in 0.2% citrate buffer pH6 (endosialin staining) or 1 mM EDTA (CD31, CD34 and GFAP staining) with microwave heating for 30 min. Prior to immunostaining, all slides were washed in PBS and non- specific binding sites blocked by incubation with PBS-containing 1% BSA.

Transfected cells were incubated with the affinity-purified antibody against human endosialin (1:200). Slides were washed in PBS, blocked in PBS-BSA before incubation with a FITC-conjugated donkey anti-rabbit secondary antibody (Jackson Immunoresearch Laboratories, West Grove, PA). Frozen tumor sections were double stained by incubation at room temperature with affinity-purified rabbit anti-endosialin and mouse monoclonal antibodies against either CD31 (1:200, Dako Cytomation Ltd, Ely, UK), CD141 (1:200, BD Biosciences), CD93 (BIIG-4, 1:200, [4]) CDw90 (Thy-1, 1:200, Serotec, Oxford, UK), GFAP (clone GA5, 1:200, Sigma), E-cadherin (1:200, R&D Systems, Abingdon, UK), or α-smooth muscle actin (α-SMA, 1:40, Serotec) (all human antigens). Tissues were washed in PBS, non-specific sites blocked as above, and incubated with rhodamine- or FITC-conjugated donkey anti-mouse and rabbit secondary antibodies (Jackson Immunoresearch Laboratories) at room temperature. DAPI (4,6, Diamidino-2-phenylindole) was used to stain the nuclei blue. All immunofluorescently labeled slides were washed well in PBS and mounted with Vectashield mounting medium (Vector Laboratories, Peterborough, UK). Paraffin sections were incubated with a single antibody against endosialin (rabbit polyclonal, 1:200), CD31 (as above), CD34 (mouse monoclonal, Qbend10 1:80, Skybio, Bedford, UK), or GFAP (mouse monoclonal, 1:400, Dako). Slides were washed in PBS, blocked as above, and incubated with a peroxidase-conjugated donkey anti-rabbit or anti-mouse secondary antibody (Jackson Immunoresearch Laboratories). Slides were washed in PBS and immunostaining was revealed by incubation with 0.05% DAB (3'3' diaminobenzidine) with 0.03% H^sub 2^O^sub 2^. Nuclei were counterstained with hematoxylin. Sections were washed and dried before mounting with Clerium mounting medium (Surgipath, Peterborough, UK). All image analysis was undertaken using a fluorescent/bright field microscope (DMLB, Leica, Milton Keynes, UK).

Semiquantitative analysis was earned out to determine the extent of tumor vasculature and of endosialin expression for each specimen. CD31 staining was used as the primary indicator of the extent of the vasculature and each specimen was scored according to the number of CD31-positive vessels. Specimens were considered positive if CD31 stained endothelial cells specifically. Scores ranged from (-) if they lacked CD31 expression to (+++) for specimens with a large number of CD31-positive vessels. Endosialin expression by endothelial cells was determined by double immunofluorescence staining for endosialin and CD31. Endosialin expression by endothelial cells was scored relative to the number of CD31- positive vessels in the specimen, with (-) indicating no endosialin expression, (+) indicating <10% of vessels positively stained, (++) indicating 10% to 50% of vessels positively stained, and (+++) indicating >50% of vessels positively stained.

RESULTS

Generation of Recombinant Endosialin

A plasmid containing a fusion protein of the endosialin CTLD and the human IgG1 Fc tail was generated. The integrity of the plasmid was determined by sequencing and compared with the Genbank sequence for endosialin (Accession number AF279142). The fusion protein sequence had a single base pair substitution at 717 bp, which is a conserved mutation in the endosialin CTLD (amino acid 239).

Characterization of Anti-Endosialin Antibody

We generated a polyclonal against an endosialin fusion protein. The specificity of the antibody for human endosialin was tested by immunostaining and flow cytometry of CHO cells transfected with full- length human endosialin or human C3aR, which acted as a negative control. The antibody strongly stained CHO cells transfected with endosialin by immunocytochemistry, but did not stain C3aR- transfected cells (Fig. 1A). Endosialin was expressed at the cell surface and in the cytoplasm as previously described (10). An irrelevant primary antibody against human CD93 did not stain CHO cells expressing endosialin (4). The specificity of the anti- endosialin antibody was confirmed by flow cytometry (Fig. IB). Immunoprecipitation of endosialin from CHO cells expressing the protein using the polyclonal antibody generated bands of 175 kDa, 120 kDa, and 90 kDa by Western blot of cell lysates representing the fully glycosylated protein, an intermediate form and the core protein respectively (data not shown).

Immunoblot Analysis of Tumor Lysates

We made lysates of 5 tumor specimens (carcinoma patients 3 and 6, GBM patient 5, astrocytoma II patient 3, and the melanoma patient) and a normal brain tissue specimen and immunoblotted for endosialin using the polyclonal antibody. All 5 specimens gave a broad band of approximately 120 kDa as expected for endosialin, with particularly strong expression in the carcinoma and GBM patients (Fig. 2). No expression was evident in the normal tissue. The quality of the lysate was confirmed by probing the blot with rabbit anti-MCP (CD46) (not shown). Due to the heterogeneous nature of tumor tissues, it is difficult to use the blot to compare levels of endosialin expression, despite normalization of the amount of tissue lysate loaded onto the gel. We therefore used immunohistochemistry to localize and provide a semiquantitative analysis of endosialin expression.

Immunohistochemical Analysis of Tumor Vasculature

The extent of tumor vascularization was determined by immunostaining paraffin-embedded tissue sections for CD31 and CD34 (examples of each in a GBM (patient 5) and a carcinoma (patient 6) shown in Fig. 3b, c, j, k). All tumors were CD31-positive (Table). Endosialin expression in endothelial cells was determined by double immunofluorescence staining of frozen sections for endosialin and CD31. The extent of endosialin expression between tumors and within a tumor was highly variable. We observed express\ion in small capillary-like vessels as well as larger vessels. In no case were 100% of vessels positively stained for endosialin. The Table shows endosialin expression by endothelial cells for each tumor specimen. Two of 5 astrocytomas (WHO grade II), 2/3 anaplastic astrocytomas (WHO grade III), 6/6 GBM (WHO grade IV), 2/2 oligodendrogliomas (WHO grade II), 1/1 ependymoma (WHO grade II), 2/5 meningiomas (WHO grade I), 4/7 carcinomas, and 1/1 melanoma showed endothelial expression of endosialin. Figure 3e and 3m show colocalization of endosialin and CD31 in a GBM (patient 5) and a carcinoma (patient 7) specimen, respectively. CD31 was expressed at the cell surface facing the vessel lumen, whereas endosialin was distributed at the cell membrane and in intracellular stores. Endothelial expression of endosialin in these 2 cases was confirmed by single staining of paraffin wax sections with the polyclonal antibody against endosialin (not shown). Representative examples of an astrocytoma (WHO grade II, patient 1) and a metastatic carcinoma (patient 7) with CD31-positive endothelial cells, which are endosialin- negative, are shown in Figure 3q and 3r, respectively. Control staining using the secondary antibody only was negative.

Immunohistochemical Analysis of Specific Tumors

In order to further characterize the cellular localization of endosialin, 2 cases in which endosialin expression was considered not to be restricted to endothelial cells (GBM patient 5 and carcinoma patient 6) were analyzed in more detail by immunohistochemical staining. In each case, H&E staining confirmed the type of tumor (Fig. 3a, i). The GBM was strongly positive for GFAP confirming its astrocytic origins (Fig. 3d), although GFAP- positive astrocytes did not stain for endosialin (Fig. 3h). Interestingly, it is apparent that periendothelial cells also express endosialin (Fig. 3e). Moreover, cells positive for α- SMA, suggesting a pericyte or smooth muscle cell origin (Fig. 3s, t), were stained for endosialin. Interestingly, in analogy to endothelial cells, not all smooth muscle cells within a vessel or the tumor express endosialin. For instance, smooth muscle cells (or pericytes) expressing high levels of Thy-1 (CD90) did not display endosialin expression (Fig. 3f). These Thy-1-positive cells are negative for the NeuN neuronal marker (data not shown). In contrast to the GBM, tumor cells within the carcinoma were negative for GFAP (Fig. 31) and positive for E-cadherin (Fig. 3p). Of note, reactive stromal fibroblasts close to the pia membrane and surrounding the E- cadherin-positive cells were strongly stained for endosialin.

Expression of thrombomodulin (CD141) was examined in the same 2 cases. Thrombomodulin is expressed by endothelial cells in the GBM (colocalized with endosialin in Fig. 3g). Moreover, thrombomodulin was weakly expressed by the E-cadherin-positive tumor cells in the carcinoma patient, particularly at the interface with reactive fibroblasts strongly stained for endosialin (Fig. 3p) and endothelial cells (not shown).

Fig. 1. Characterization of polyclonal against human endosialin fusion protein. A: CHO cells transfected with human endosialin or human C3aR (negative control) were cultured on coverslips and stained with the affinity-purified antibody against endosialin by immunohistochemistry. A FITC-conjugated secondary antibody was used for detection and DAPI was used to counter-stain the nuclei. The antibody stained the endosialin-transfected cells (right panel), while they did not stain the control cells (left panel) (magnification 400). B: CHO cells transfected with human endosialin or human C3aR were stained with the polyclonal antibody by flow cytometry using an rPE-conjugated secondary antibody. The antibody stained endosialin-transfected cells strongly but did not stain C3aR- transfected cells.

Fig. 2. Western blot analysis of endosialin expression in representative brain tumor lysates and normal brain tissue lysate. No expression was evident in normal brain tissue in contrast to a robust expression in several brain tumors.

DISCUSSION

Endosialin is a potentially important marker of neoplastic angiogenesis, which we have studied in human brain tumors. We have generated and validated a novel antibody that specifically recognizes human recombinant endosialin and used it to dissect the expression pattern of endosialin in different human brain tumor subtypes by immunohistochemistry and Western blot. We chose representative tumor samples for Western blot analysis and prepared tissue lysates in which endosialin expression was analyzed. This indicated that all 5 tumor specimens expressed endosialin, although it was not a quantitative assessment due to the heterogeneous nature of tumor specimens. To determine the exact cellular location of endosialin and provide a semiquantitative assessment of endosialin expression, we carried out immunohistochemical staining of tissue sections. Endosialin expression was established in endothelial cells by co-immunostaining with CD31-positive vessels in 20 of 30 (67%) tumor specimens examined. Of particular interest, 100% of the highly invasive GBM express endosialin whereas 67% of grade III anaplastic astrocytomas and 40% of the less invasive grade II astrocytomas express the protein. Furthermore, only 40% of the relatively benign meningiomas (WHO grade I) expressed endosialin. Endosialin is expressed by endothelial cells in a substantial number of carcinomas (62%) that have metastasized to the brain. All of the oligodendroglioma and ependymoma specimens expressed endosialin although it is difficult to extrapolate conclusions from this due to the small number of patients available. Normal tissue surrounding the tumor bed was negative for endosialin. Thus, endosialin expression in blood vessels correlates with high tumor grade and aggressive histological behavior.

Our data is consistent with previous work in that endosialin expression in endothelial cells was extremely heterogeneous with regard to the number and type of tumors as well as the number of positively stained blood vessels within a tumor (6). This pattern of expression is not surprising given the dynamic and irregular nature of tumor blood vessels compared with the normal vasculature (1, 18). However, contrary to previous findings, expression was not solely restricted to microvascular capillary-like vessels, but was also found in larger vessels (Fig. 3 e, h, n, s). Furthermore, we have detected in GBM, endosialin expression in a subset of α-SMA- positive cells in both small vessels with scant smooth muscle cell coverage and also in larger vessels with significant numbers of smooth muscle cells (Fig. 3s, t). GBM have been described as having numerous α-SMA-positive pericytes/vascular smooth muscle cells, which are thought to contribute to angiogenesis (19).

Given the de novo expression of endosialin by endothelial cells and some smooth muscle cells particularly in invasive tumors such as anaplastic astrocytomas, GBM, and metastatic carcinomas, we speculate that it may regulate tumor growth by controlling angiogenesis. Endosialin might serve one of several roles in the vasculature, including signaling from cell to cell or serving as a receptor for soluble ligands. In analogy to selectins, it has been hypothesized that the CTLD domain of endosialin facilitates homotypic interactions or heterotypic interactions with periendothelial cells (pericytes in small vessels) that lie adjacent to endothelial cells (10). Cell-cell interactions are an important component of cancer progression, whether it is the interaction between endothelial cells and smooth muscle cells during vascular remodeling, or between tumor cells and endothelial cells to facilitate metastatic spread via systemic blood vessels (20, 21).

Interestingly, studies on endosialin homologues, in particular thrombomodulin, have provided useful functional information that may be applicable to endosialin given their structural similarity and vascular colocalization. Thrombomodulin is classically a powerful cofactor in the natural anti-coagulation system, a function mediated by EGF domains 4-6 (22). The exact role of the lectin-like domain (CTLD) of thrombomodulin in tumor growth has not been clearly resolved; however, speculative suggestions include the control of tumor vessel wall hemostasis and downregulation of tumor cell metastasis by maintaining tumor cell adhesion (23, 24). The latter paradigm has been corroborated by Zhang et al and Olivot et al who found a negative correlation between thrombomodulin expression and tumor cell proliferation in vitro and in vivo (25, 26), whereas other studies implicate the CTLD as a key factor in regulation of proliferation and adhesion (27, 28). Interestingly, thrombomodulin promotes proliferation of vascular smooth muscle cells (29). Given the complexity of function associated with thrombomodulin, it is therefore likely that the function of endosialin is no less complex.

Fig. 3. Histopathological assessment of angiogenesis and endosialin expression in human brain tumors. Immunohistochemistry with single staining of paraffin-embedded and fluorescent double staining of frozen sections from surgically resected specimens from a human glioblastoma multiforme (patient 6) (a-h, s, t), carcinoma (patients 6 and 7) (i-p, r) and astrocytoma grade II (patient 1) (q) as described in Materials and Methods. a, i: H&E staining of paraffin-embedded sections showing tumor morphology. b, j: Staining of endothelial cells in paraffin-embedded sections for CD31. Both sections are strongly positive for CD31, indicating a highly vascularized tumor bed. c, k: Staining of endothelial cells in paraffin-embedded sections for CD34. d, 1: Staining of paraffin- embedded sections for GFAP. e: Double staining of frozen sections for endosialin (green) and CD31 (red). Note that the cytoplasm and membrane are strongly stained f\or endosialin. m: Double staining of frozen sections for endosialin (red) and CD31 (green). f: Double staining of a frozen section for endosialin (green) and Thy1 (red). g: Strong colocalization of endosialin (green) with thrombomodulin (CD141) (red) h: Double staining of a frozen section for endosialin (green) and GFAP (red). n: A carcinoma stained for endosialin (red) and CD93 (green). Thrombomodulin (red) (o) is expressed by tumor cells that are E-cadherin-positive (red) (p), however, there is no colocalization of either marker with endosialin, which is expressed by reactive fibroblasts at the edge of the tumor (q). r: Representative examples of CD31-positive vessels that did not express endosialin. s, t: Colocalization of endosialin with α- smooth muscle actin (α-SMA)-positive cells. Magnifications: a- d, i-1, 200; e-h, m-t, α400.

TABLE

Endosialin and CD31 Expression in Human Brain Tumor and Normal Human Brain Specimens by Immunohistochemical Analysis

In several carcinomas we have also observed endosialin expression in reactive stromal fibroblasts. It has been suggested that Thy-1- positive fibroblasts are capable of differentiating into myofibroblasts (30, 31). This finding correlates with the published observations by Rettig et al that a small proportion of epithelial cancers contained a population of fibroblasts strongly stained for endosialin (6). The function of endosialin in fibroblasts is unknown.

In summary, while endosialin expression is unlikely to be useful in the histological differentiation of brain tumor types, its expression may be viewed as a poor prognostic indicator because it appears to be restricted to a subset of vessels undergoing angiogenesis. We have provided for the first time an in-depth analysis of endosialin expression in primary and metastatic human brain tumors. Endosialin expression is associated with high-grade primary and metastatic tumors and it is absent in normal blood vessels. The results have enabled us to speculate that endosialin may play a role in facilitating cell-cell interactions that ultimately may contribute to tumor angiogenesis, metastasis and proliferation. Interestingly, Rettig et al reported that the radiolabeled FB5 anti-endosialin antibody was rapidly internalized into endothelial cells (6). Hence, endosialin expressing blood vessels may be an attractive target for anti-cancer immunotherapy. Experiments along these lines arc now highly warranted.

ACKNOWLEDGMENTS

We thank Dr. I. Anegon (Institut National de la Santo et de la Recherche Mdicale Unit 437, Nantes, France) for the kind gift of the pDR2&945; expression vector.

REFERENCES

1. Ruoslahti E. Specialization of tumor vasculature. Nat Rev Cancer 2002;2:83-90

2. MeKearn JP, Baum C, Davie JM. Cell surface antigens expressed by subsets of pre-B cells and B cells. J Immunol 1984;132:332-39

3. Dean YD, McGreal EP, Akatsu H, Casque P. Molecular and cellular propertics of the rat AA4 antigen, a C-type lectin-like receptor with structural homology to thrombomodulin. J Biol Chem 2000;275:34382-92

4. McGreal HR Ikewaki N, Akatsu H. Morgan BP, Casque P. Human CIqRp is identical with CD93 and the mNI-11 antigen but does not bind CMq. J Immunol 2002;168:5222-32

5. Conway EM, Van de Wouwer M, Pollefeyt S, et al. The lectin- like domain of thrombomodulin confers protection from neutrophil- mediated tissue damage by suppressing adhesion molecule expression via nuclear factor kappaB and mitogen-activated protein kinase pathways. J Exp Med 2002;196:565-77

6. Rettig WJ, Garin-Chesa P, Healey JH, Su SL, Jatte EA, Old LJ. Identification of endosialin, a cell surface glycoprotein of vascular endothelial cells in human cancer. Proc Natl Acad Sci USA 1992;89:10832-36

7. St Croix BS, Rago C, Velculescu V, et al. Genes expressed in human tumor endothelium. Science 2000;289:1197-202

8. Carson-Walter EB, Watkins DN, Nanda A, Vogelstein B, Kinzler KW, St Croix BS. Cell surface tumor endothelial markers are conserved in mice and humans. Cancer Res 2001;61:6649-55

9. Christian S, Ahorn H, Koehler A, et al. Molecular cloning and characterisation of endosialin, a C-type lectin-like cell surface receptor of tumor endothelium. J Biol Chem 2001;276:7408-14

10. Opavsky R, Haviernik P, Jurkovicova D, et al. Molecular characterisation of the mouse TEM1/endosialin gene regulated by cell density in vitro and expressed in normal tissues in vivo. J Biol Chem 2001;276:38795-807

11. Weis WI, Taylor ME, Drickamer K. The C-type lectin superfamily in the immune system. Immunol Rev 1998;163:19-34

12. Ni Y, Tizard I. Lectin-carbohydrate interaction in the immune system. Vet Immunol Immunopathol 1996;55:205-23

13. Appella E, Weber IT, Blasi F. Structure and function of epidermal growth factor-like regions in proteins. FEES Lett 1988;231:1-4

14. Pigott R, Needham LA, Edwards RM, Walker C, Power C. Structural and functional studies of the endothelial activation antigen endothelial leucocyte adhesion molecule-1 using a panel of monoclonal antibodies. J Immunol 1991;147:130-35

15. Wei X-Q, Orchardson M, Gracie JA, et al. The sushi domain of soluble IL-15 receptor α is essential for binding IL-15 and inhibiting inflammatory and allogenic responses in vitro and in vivo. J Immunol 2001;167:277-82

16. Kirsch M, Santarius T, Black PM, Schackert G. Therapeutic anti-angiogenesis for malignant brain tumors. Onkologie 2001;24:423- 30

17. Behin A, Hoang-Xuan K, Carpentier AF, Delattre J-Y. Primary brain tumors in adults. Lancet 2003;61:323-31

18. Dvorak HF. How tumors make bad blood vessels and stroma. Am J Pathol 2003;162:1747-57

19. Wesseling P, Schlingemann RO, Rietveld EJ, Link M, Burger PC, Ruiter DJ. Early and extensive contribution of pericytes/vascular smooth muscle cells to microvascular proliferation in glioblastoma multiforme: An immuno-light and immuno-electron microscopic study. J Neuropathol Exp Neurol 1995;54:304-10

20. Bischoff J. Cell Adhesion in vascular biology. J Clin Invest 1997;100:837-39

21. Chavakis E. Dimmeler S. Regulation of endothelial cell survival and apoptosis during angiogenesis. Arterioscler Thromb Vasc Biol 2002;22:887-93

22. Sadler JE, Lentz SR, Sheehan JP, Tsaing M, Wu Q. Structure- function relationships of the thrombin-thrombomodulin interaction. Haemostasis 1993;23:183-93

23. Tamura A, Matsubara O, Hirokawa K, Aoki N. Detection of thrombomodulin in human lung cancer cells. Am J Pathol 1993;142:79- 85

24. Tezuka Y, Yonezawa S, Maruyama I, et al. Expression of thrombomodulin in esophageal squamous cell carcinoma and its relationship to lymph node metastasis. Cancer Res 1995;55:4196-200

25. Zhang Y, Weiler-Guettler H, Chen J, et al. Thrombomodulin modulates growth of tumor cells independent of its anticoagulant activity. J Clin Invest 1998;101:1301-9

26. Olivot J-M, Estebanell E, Lafay M, Brohard B, Aiach M, Rendu F. Thrombomodulin prolongs thrombin-induced extracellular signalregulated phosphorylation and nuclear retention in endothelial cells. Circ Res 2001;88:681-87

27. Weiler H, Isermann BH. Thrombomodulin. J Thromb Haemost 2003;1:1515-24

28. Huang H-C, Shi G-Y, Jiang S-J, et al. Thrombomodulin- mediated cell adhesion. J Biol Chem 2003;278:46750-59

29. Tohda G, Oida K, Okada Y, et al. Expression of thrombomodulin in atherosclerotic lesions and mitogenic activity of recombinant thrombomodulin in vascular smooth muscle cells. Arterioscler Thromb Vasc Biol 1998;18:1861-69

30. Steiniger B, Barth P, Hellinger A. The perifollicular and marginal zones of the human splenic white pulp; do fibroblasts guide lymphocyte migration? Am J Pathol 2001;159:501-12

31. Koumas L, Smith TJ, Feldon S, Blumberg N, Phipps RP. Thy-1 expression in human fibroblast subsets defines myofibroblastic or lipofibroblastic phenotypes. Am J Pathol 2003;163:1291-300

Received April 30, 2004

Revision received August 6, 2004

Accepted September 3, 2004

JENNIFER BRADY, MSC, JAMES NBAL, FRCPATH, NAGARAJAN SADAKAR, FRCS, AND PHILIPPE GASQUE, PHD

From Brain Inflammation and Immunity Group (JB, PG), Department of Medical Biochemistry and Immunology, Department of Histopathology (JN), University of Wales College of Medicine. Department of Neurosurgery (NS), University Hospital of Wales, Cardiff. United Kingdom.

Correspondence to: Dr. Philippe Casque. Brain Inflammation and Immunity Group, Department of Medical Biochemistry and immunology. University of Wales College of Medicine, Heath Park, Cardiff CF14 4XN, United Kingdom. E-mail: gasque@cardiff.ac.uk

Supported by The Wales Office of Research and Development for Health and Social Care (WORD), WSDHSR S01/020.

Copyright American Association of Neuropathologists, Inc. Dec 2004

Source: Journal of Neuropathology and Experimental Neurology

More News in this Category