Murine Endothelial Cell Lines As Models of Tumor Endothelial Cells

Vol. 10, 2179–2189, March 15, 2004 Clinical Cancer Research 2179

Murine Endothelial Cell Lines as Models of Tumor Endothelial Cells

Jennifer Walter-Yohrling, Sharon Morgenbesser, INTRODUCTION Cecile Rouleau, Rebecca Bagley, The mouse has become the critical host organism for the Michelle Callahan, William Weber, and discovery and development of modern cancer therapeutics. However, it is important that we understand the strengths and Beverly A. Teicher limitations of mouse models of malignant disease (1). The most Genzyme Corporation, Framingham, Massachusetts frequently used murine models are transplantable syngeneic or xenograft tumors and transgenic mice that develop autologous ABSTRACT tumor nodules during their lifetimes. These models allow the study of biological processes associated with tumor growth, Identification of appropriate models for in vivo and in including the tumor-host interactions with murine stromal cells. vitro preclinical testing of inhibitors of tumor angiogenesis The abnormality of tumor vasculature has been studied and progression is vital to the successful development of extensively in mouse models. Modzelewski et al. (2) isolated anticancer therapeutics. Although the focus is on human and characterized fresh tumor-derived endothelial cells from the molecular targets, most preclinical in vivo efficacy testing syngeneic RIF-1 fibrosarcoma. The dynamics of blood vessel occurs in mice. The goal of the current studies was to growth, structure, cellular composition and gene expression identify a murine endothelial cell line to model tumor endo- have been followed using the murine syngeneic mammary thelium for studying the antiangiogenic activity of therapeu- MCaIV adenocarcinoma, the murine syngeneic Lewis lung car- tic compounds in vitro. In situ hybridization was performed cinoma, and the murine transgenic RIP-Tag2 pancreatic islet on three s.c. grown syngeneic murine tumors (B16 mela- cell tumor (3–5). The mobilization of bone marrow progenitor noma, Lewis lung carcinoma, and CT26 colon carcinoma) to cells to circulation, the maturation of these cells into endothelial assess expression of murine homologs of human tumor en- precursor cells, and the incorporation of the endothelial precur- dothelial cell markers in the vasculature of these tumor sor cells from circulation into tumor vasculature has been elu- models. Seven murine endothelial cell lines were character- cidated largely in mice (6–14). These studies used murine ized for expression of the murine homologs of recognized syngeneic tumors, human tumor and bone marrow xenografts in endothelial cell surface markers as well as for tumor endo- immunodeficient mice and genetically engineered mice, and thelial cell surface markers. The seven murine endothelial have been used to recapitulate the dynamics of processes ob- cell lines had similar generation times and five of the seven served in human patients. lines were able to form tubes on Matrigel. Real-time-PCR Cancer therapeutics has moved away from general prolif- and flow cytometry analysis were used to evaluate relative eration-related targets such as DNA and tubulin toward more mRNA and protein expression of murine homologs of sev- nontraditional pathological processes including angiogenesis or eral recognized endothelial cell surface markers in the seven immunomodulation, as well as more selective molecular targets. cell lines. The expression of the mRNA for the murine In most cases, efficacy in mouse models remains the critical homologs of five tumor endothelial cell surface markers was determinant of whether a potential therapeutic moves into de- also evaluated. The 2H11 cell line expressed all five of the velopment in clinical trials. However, it has become evident that tumor endothelial cell surface markers as well as several the homology between the murine and human proteins of spe- well-recognized endothelial cells markers. The 2H11 cell line cific molecular targets is, frequently, not sufficient to depend on responds to known and novel antiangiogenic agents by in- efficacy of agents selected for the murine target to translate into hibition of proliferation and tube formation. These cells can highly effective therapy in the human clinic. This is most be used in in vitro angiogenesis assays for evaluating the evident in the selection of monoclonal antibodies where it is has potential antiangiogenic properties and interspecies cross- become necessary to develop monoclonal antibodies to the reactivity of novel compounds. murine homologue of the human target (11). Therefore, mouse and human agents need to be developed in parallel, so that the tumor-bearing mouse remains the critical efficacy hurdle. Like- wise, it is important to define cell-based models of murine tumor endothelial cells that can be used to evaluate potential therapeu- Received 7/9/03; revised 12/1/03; accepted 12/05/03. tics in cell culture to select those most appropriate for in vivo The costs of publication of this article were defrayed in part by the testing. payment of page charges. This article must therefore be hereby marked The isolation and maintenance of a pure population of advertisement in accordance with 18 U.S.C. Section 1734 solely to primary murine endothelial cells has proven to be difficult and indicate this fact. Requests for reprints: Beverly A. Teicher, Genzyme Corporation, 1 has restricted the use of these cells in angiogenesis assay sys- Mountain Road, Framingham, MA 01701, Phone: (508) 271-2843, tems. To overcome these barriers, several groups have generated FAX: (508) 620-1203, E-mail: [email protected]. murine endothelial cell lines from cells isolated from the axil-

Table 1 List of endothelial cell markers studied, the antibody used for FACSa analysis, and primer sequences for mRNA expression analysis EC marker Antibody Forward primer Reverse primer CD34/sialoucin Pharmingen cat. 553731 gaagacccttattacacgga gctgaatggccgtttct CD36/GPIIIb Chemicon cat. MAB1258 N/A N/A CD105/endoglin Pharmingen cat. 550546 cagcaagcgggagcccgtggt ggtgctctgggtgctcccgat ENDRB N/A gcagaggactggccatttgga gcaacagctcgatatctgtcaatact P1H12/CD146 Chemicon cat. MAB16985 N/A N/A Tie1 N/A tggagatagtgagccttgga cagtttcgaggctgctccatgcg Tie2 Chemicon cat. MAB1148 gagattgttagcttaggaggcac gtctcattagatcatacacctcatcat VCAM1/CD106 Pharmingen cat. 553330 gacctgttccagcgagggtcta cttccatcctcatagcaattaaggtgg VEGFR1 N/A ctgggagcctgcacgaagcaa gtcacgtttgctcttgaggtagt VEGFR2 Pharmingen cat. 555308 gcagacagaaatacgtttgagttgg agtgattgccccatgtgga FACS, fluorescence-activated cell sorter; EC, endothelial cell; ENDRB, endothelin-B receptor; VCAM, vascular cellular adhesion molecule; VEGFR, vascular endothelial growth factor receptor; cat., catalogue number.

lary lymph node (15–17), embryonic yolk saks (6), and aorta, sue-type and vasculature-specific expression patterns. Recently, brain, and heart capillaries (18). In selecting the appropriate several novel markers have been identified that are specific for murine in vitro model of tumor endothelial cells, it is ideal to human tumor endothelial cells (TEM; 19). The TEM expression identify a murine endothelial cell line that maintains the expres- pattern for the murine homologue of several of these markers sion of traditional endothelial cell markers as well as markers was determined in murine B16 melanoma and human HCT116 found to be expressed by human tumor endothelial cells (19, colon carcinoma xenografts by in situ hybridization (20). These 20). Thus, in the process of targeted antiangiogenic drug dis- markers provide criteria for identification of appropriate murine covery, having identified cell-based models expressing the hu- endothelial cells for studying tumor endothelial properties. man target and cell-based models expressing the murine target The goal of these studies was to identify by the behavior, allows comparison of the sensitivity of the human and mouse as well as the mRNA and protein expression of normal and targets to therapeutic candidates and allows an informed deci- tumor endothelial cell markers, the murine endothelial cell sion to be made regarding selection of therapeutic candidates for line(s) that is the most appropriate model of tumor endothelium testing in in vivo models with murine vasculature. for in vitro study of the biology of tumor angiogenesis. O’Connell et al. (15–17) generated several murine endothelial cell lines from endothelial cells isolated from axillary and inguinal lymph nodes of adult C3H/HeJ mice and transformed MATERIALS AND METHODS the cells using SV40. When implanted in mice, these immor- Chemicals. Murine endothelial cell lines, SVEC4–10, talized endothelial cells produced nodules that displayed the SVEC4–10EE2, SVEC4–10EHR1, 2F2B, 2H11, IP1B, and characteristics of Kaposi’s sarcoma, a multifocal malignant neo- IP2-E4, were purchased from American Type Culture Collec- plasm consisting of spindle cells and believed to be derived tion (Manassas, VA). Human dermal neonatal microvascular from endothelial cells. The SVEC4–10EE2 cell line was de- endothelial cells (HMVEC) and EGM2-MV were obtained from rived from the parent SVEC4–10 cell line by the extraction of Cambrex (East Rutherford, NJ). Cell culture media, versene, tumors generated after the injection of the parent cell line in and fetal bovine serum (FBS) were purchased from Invitrogen nude mice. These cells were then sub-cloned using conditioned Corp. (Carlsbad, CA). Primary antibodies against endothelial medium from SVEC4–10EE2 to isolate a slower growing mu- cell markers were purchased from PharMingen (San Diego, CA) rine endothelial cell line designated SVEC4–10EHR1. IP2E4 and Chemicon (Temecula, CA; Table 1). Phycoerythrin and and 3B11 cell lines were sub-cloned from ascites fluid generated fluorescein-conjugated secondary antibodies were obtained in mice that received implantation with SVEC4–10EHR1 cells. from Jackson ImmunoResearch Laboratories (West Grove, PA). The 2H11 cell line was sub-cloned from a nodular spindle-like PCR primers that recognize the human and mouse gene were tumor derived from the SVEC4–10EHR1 cell line. The 2F2B purchased from Integrated DNA Technologies (Coralville, IA). cell line was developed from a lobular cutaneous tumor result- Trizol was purchased from Sigma (St. Louis, MO), and the RNA ing from the implantation of SVEC4–10EE2 cells. Each of the Extraction Kit was obtained from Qiagen (Valencia, CA). The seven murine endothelial cell lines has distinctive cellular mor- High-Capacity cDNA Archive Kit, Taqman rRNA Control Re- phology as well as different endothelial cell characteristics, agents, Taqman Universal PCR Master Mix, and Sybr Green including the expression of von Willebrand factor/Factor VIII, PCR Master Mix were purchased from Applied Biosystems the incorporation of acetylated low-density lipoprotein, the up- (Foster City, CA). Cell Titer Glo was purchased from Promega regulation of class II MHC expression in response to IFN␥, and (Madison, WI). the formation of tubular networks on Matrigel (15–17). Cell Culture. The murine endothelial cell lines 2F2B, There are several markers that have proven useful for 2H11, 3B11, IP2E4, SVEC4–10, SVEC4–10EE2, and SVEC4– identifying endothelial cells in situ and in vitro, including plate- 10EHR1 were maintained in DMEM plus 10% fetal bovine

let endothelial cell adhesion molecule-1/CD31, von Willebrand serum in a humidified 10% CO2 environment. HMVECs were factor, P1H12, and vascular endothelial growth factor receptor 2 maintained in EGM2-MV that includes 5% FBS, vascular en- (VEGFR2; Refs. 21–24). These cell surface proteins have tis- dothelial growth factor (VEGF), basic fibroblast growth factor,

epidermal growth factor, and insulin-like growth factor, in a vested and prepared as formalin-fixed, paraffin-embedded 3–5 ␮ humidified 5% CO2 atmosphere. m sections on slides. All treatments were carried out at room Generation Time and Proliferation. To determine cel- temperature, unless otherwise stated. Sections were deparaf- lular growth rates, cells from each of the seven murine endo- finized in xylene, washed in 100% ethanol, then hydrated in 85, thelial cell lines (2 ϫ 103) were placed in each well of a 96-well 75, and 50% ethanol in distilled water. After incubation in plate in DMEM supplemented with 0, 2, 5 or 10% FBS. The diethyl pyrocarbonate-treated water (Quality Biological, Inc., cells were collected after 24, 48, 72, and 96 h using Cell Titer Gaitherburg, MD), sections were permeabilized by treatment Glo and extrapolating cell number from a standard curve for with pepsin in 0.2 N hydrochloric acid, washed briefly in PBS, each cell line. Generation time was calculated by applying an then fixed in 4% paraformaldehyde. Sections were acetylated in exponential curve fit to the cell number values and calculating acetic anhydride/0.1 M triethanolamine (pH 8.0), equilibrated for the time required for the cell number to double. For proliferation 10 min in 5ϫ SSC [3 M sodium chloride, 0.3 M sodium citrate experiments, cells were exposed to each agent for 48 h, and cell (pH 7.0); Invitrogen], and prehybridized for 1 to2hat55°Cin numbers were determined as described above. mRNA hybridization buffer (DAKO, Carpinteria, CA). Sections In Vitro Tube Formation. Matrigel (BD Biosciences) were hybridized with digoxigenin riboprobes (100–200 ng/ml) was added to the wells of a 48-well plate in a volume of 120 ␮l in mRNA hybridization buffer (DAKO) overnight at 55°C. and allowed to solidify at 37°C for 30 min. After the Matrigel After removing unbound riboprobes by washing, sections were solidified, cells from each of the seven murine endothelial cell incubated with RNase (Ambion, Austin, TX) to remove any 4 lines (2 ϫ 10 cells) were added in 200 ␮l of DMEM supple- nonspecific-bound riboprobe. Sections were treated with perox- mented with 10% FBS. The cells were incubated at 37°C with idase block (DAKO) to eliminate any endogenous peroxidase, humidified 95% air/5% CO2 for5hinthepresence or absence then blocked with a 1% blocking reagent (DIG nucleic acid of a potential inhibitor. detection kit; Roche), containing rabbit immunoglobulin frac- Quantitative PCR. A confluent T75 flask of cells was tion (DAKO), in Tris buffered saline. Rabbit antidigoxigenin- harvested and lysed using Trizol. RNA was isolated by phenol horseradish peroxidase (DAKO) was used to detect the ribo- chloroform extraction followed by column isolation using the probes and served to catalyze the deposition of biotinylated Qiagen RNA Extraction Kit. cDNA was generated using the tyramide (Gen-Point, DAKO) according to manufacturer’s in- High Capacity cDNA Archive Kit. Real-time-PCR for normal structions. Additional amplification was accomplished through endothelial cell markers was performed with Sybr Green PCR additional rounds of strept-av-horseradish peroxidase (Gen- Master Mix using the primers listed in Table 1 on an ABI Prism Point, DAKO) and biotinylated tyramide. Final detection was 7700 Sequence Detection System (Applied Biosystems). Rela- accomplished through rabbit antibiotin conjugated to alkaline tive mRNA expression of normal endothelial cell markers was phosphatase (DAKO). Alkaline phosphatase was detected with determined by the ␦-delta Ct method where the cycle threshold Fast Red (DAKO) for 10–60 min at RT, then counterstained in (Ct) of the sample is subtracted by the Ct of 18S and compared hematoxylin. The nuclei were blued with ammonium hydroxide with the reference expression by HMVEC. Expression of the for 30 s, then mounted with crystal-mount (BioMeda, Foster tumor endothelial cell markers was determined using FAM- City, CA). For the negative control, sense riboprobes were used labeled probes and Taqman Universal PCR Master Mix, expres- to detect any nonspecific sequences. Additionally, a slide that sion was normalized to 18S and compared with the reference was exposed to RNase to destroy the mRNA was hybridized expression by 2H11 cells. with the antisense riboprobe to detect any nonspecific hybrid- Flow Cytometry Analysis. Cells were suspended by ex- ization. posure to versene and 0.005% trypsin and then washed with PBS containing 5% FBS. Cells from each of the seven murine endothelial lines (2 ϫ 105) were incubated with 1 ␮g of primary RESULTS antibody diluted in PBS containing 5% FBS in 100 ␮l total In situ hybridization was performed on tissue sections of volume for 1 h. The cells were washed twice in PBS containing three murine syngeneic tumors, B16 melanoma, Lewis lung 5% FBS and incubated with a 1:50 dilution of the appropriate carcinoma, and CT-26 colon carcinoma grown s.c. in the ap- fluorescently conjugated secondary antibody for 1 h. Samples propriate hosts to determine the mRNA expression of the mu- were analyzed using an FACSCaliber flow cytometer (Becton rine homologs of the known endothelial cell marker, VEGFR2, Dickinson, Franklin Lakes, NJ). The results were compared with and the murine homologs of several tumor endothelial cell appropriate isotype controls. Positive expression was deter- markers in the intratumoral vessels. The tumor endothelial cells mined by dividing the number of cells stained with the antibody were found to express VEGFR2 as well as murine (m)TEM1, by the total number of cells assayed multiplied by 100 (percent- mTEM3, mTEM5, and mTEM8 in all three in vivo mouse positive cells). Each determination was based on 10,000 events. models of tumor vasculature as was previously shown by Car- In Situ Hybridization. Briefly, cDNA fragments were son-Walter et al. (Ref. 20; Fig. 1). Like Carson-Walter et al., we generated by PCR amplification of fragments ranging from 200 found that the B16 melanoma was negative for mTEM7; how- to 650 bp, using primers with T7 promoters incorporated into ever, this marker was expressed by the vasculature of both the the antisense primers (19, 20). Digoxigenin riboprobes were Lewis lung carcinoma and the CT26 colon carcinoma. The generated by in vitro transcription in the presence of digoxige- confirmation that the murine homologs of several markers iden- nin, according to manufacturer’s instructions (Roche, Indianap- tified in endothelial cells isolated from human tumors are ex- olis, IN). Murine syngeneic tumors, B16 melanoma, Lewis lung pressed in these transplantable murine syngeneic tumor models carcinoma, and CT-26 colon carcinoma grown s.c. were har- validates their use in the study of tumor endothelial biology and

Fig. 1 In situ hybridization of the following s.c. grown syngeneic mouse tumors: B16 melanoma, Lewis lung carcinoma, and CT-26 colon carcinoma. Tissue sections (5 ␮m) of each tumor were stained with hematoxylin and exposed to riboprobes for mouse vascular endothelial growth factor receptor 2 (VEGFR2), murine tumor endothelial cell (mTEM) 1, mTEM 7 (A), and mTEM 8, mTEM 9, and mTEM 3 (B) and visualized by amplification with fast red chromagen. Images were taken at ϫ20 magnification. LLC, Lewis lung carcinoma.

Fig. 2 Generation times for each of the seven murine endothelial cell lines at three concentrations of fetal bovine serum (FBS; 2, 5, and 10%) determined over a 96 h period. Generation times (h) were calculated by applying an exponential curve fit to the growth curve data and extrapolating cell number doubling time in the presence of each serum concentration. The data are the means and SD of three independent experiments.

provides guidance for selection of murine endothelial cell lines The mRNA expression of the murine homologs of recog- that would be appropriate models for tumor angiogenesis in cell nized cell surface endothelial cell markers in each of the seven culture. murine endothelial cell lines was compared with the expression The generation times for each of the seven murine endo- of the same marker in primary HMVEC using primers that thelial cell lines 2H11, 2F2B, 3B11, IP2E4, SVEC4–10, recognized both murine and human mRNAs (Fig. 4). The SVEC4–10EE2, and SVEC4–10EHR1 were determined (Fig. mRNA expression of each marker in the murine endothelial cell 2). The effect of serum on the proliferation of each cell line was lines is represented as the expression relative to HMVECs. The assessed at concentrations of 2, 5, and 10% FBS. Cell number expression of VEGFR2, VEGFR1, and Tie1 by the murine was determined at 24, 48, 72, and 96 h using standard curves endothelial cells was similar to the expression in HMVEC. The derived from a metabolic luminescent end point. The effect of mRNA for endgolin/CD105 and endothelin receptor B was serum concentration on the cellular proliferation was small found at much lower levels in the murine cell lines than in except with the slower growing SVEC4–10EHR1 cells with HMVEC, and the mRNA for sialomucin/CD34 was found at generation time increased 1.8-fold when the cells were grown at much higher levels in the murine endothelial cells than in the lowest serum concentration. In general the primary deriva- HMVEC. The SVEC4–10 parent cell line expressed VEGFR1, tive cell lines, SVEC4–10EE2 and SVEC4–10EHR1, from the VEGFR2, endothelin receptor B, and Tie1 but not endoglin/ parental SVEC4–10 line were slower growing than the second- CD105 or sialomucin/CD34. The first derivative cell lines, ary derivative lines, 2F2B, 2H11, 3B11, and IP2E4. The cell line SVEC4–10EE2 and SVEC4–10EHR1, had similar mRNA ex- with the shortest doubling time was the 2H11 cells with a pression patterns to the parental cell line except that expression generation time of 18.7 h at a serum concentration of 10% FBS of VEGFR1 is markedly decreased, and expression of endoglin/ and a generation time of 23.8 h at a serum concentration of 2% CD105 is increased in the SVEC4–10EE2 cells. The 2F2B cell FBS. line that was derived from the SVEC4–10EE2 cells also has The assembly of cells into tubes/networks on a layer of decreased expression of VEGFR1 mRNA relative to the other extracellular matrix components is characteristic of endothelial murine endothelial cell lines. Only the 2H11 cell line and the cells in culture. The capability of each of the seven murine two lines derived from ascites, 3B11 and IP2E4, expressed all endothelial cell lines to assemble into tubes/networks was as- six of the cell surface endothelial cell markers. sessed in a tube formation assay on a layer of Matrigel overa5h Because the goal is to identify a murine endothelial cell period (Fig. 3). The parental cell line SVEC4–10 was capable of line that could be useful as a model for tumor endothelial cells, tube formation. Neither one of the initial derivative lines, the mRNA expression of cell surface markers predicted to be SVEC4–10EE2 or SVEC4–10EHR1, were able to form tubes selective for tumor endothelial cells was evaluated (Fig. 5). The on Matrigel. The second derivative line, 2F2B, that was derived expression of the mouse homologs of five tumor endothelial from the SVEC4–10EE2 was able to form tubes on Matrigel, markers identified in endothelial cells isolated from human and each of the lines derived from the SVEC4–10EHR1 cells colon carcinoma was determined in six murine endothelial cell were also active in the tube formation assay. lines, and values are relative to the expression in 2H11 cells.

Fig. 3 Inverted, phase microscope images of each of the seven murine endothelial cell lines (2 ϫ 104 cells/well) in a tube formation assay on a thick layer of Matrigel for5hat37°C.

Overall, the 2H11 cell line was the highest expressor of these standard endothelial cell functional assays. In a proliferation markers. The two lines derived from ascites, 3B11 and IP2E4, assay, 2H11 cells were exposed to a control anti-2,4-dinitrophe- were also good expressors of all of the markers. The related nol antibody or to an anti-TEM antibody at a concentration of lines SVEC4–10EE2 and 2F2B had variable expression of the 100 ␮g/ml or were exposed to SU5416 (10 or 50 ␮M) for 48 h markers and were particularly low expressors of mTEM7 and (Fig. 6). Although neither antibody altered the proliferation of mTEM1. On the basis of this analysis, the murine 2H11 endo- the 2H11 cells, SU5416 inhibited proliferation in a concentra- thelial cell line appeared to be the most promising model for tion-dependent manner. The 2H11 cells form tubes in 4 to5hon tumor endothelial cells. a Matrigel surface. Exposure of the cells to SU5416 (50 ␮M) Antibodies specific for the murine homologs for several during that time was able to disrupt tube formation (Fig. 6). recognized endothelial cell surface proteins are available. The Exposure of the cells to anti-TEM antibody also disrupted tube cell surface expression of five endothelial cell markers was formation although exposure to the control anti-2,4-dinitrophe- assessed for the seven murine endothelial cell lines and, using nol antibody did not. human specific antibodies, for HMVEC, by flow cytometry (Table 2). As might be expected from the mRNA expression results, none of the seven murine endothelial cell lines expressed DISCUSSION endoglin/CD105, although HMVEC had strong expression of The identification of a murine endothelial cell line that has the endoglin protein. Three of the murine endothelial cell lines, tumor endothelial marker expression is important for the appro- 2H11, 3B11 and IP2E4, expressed sialomucin/CD34 protein as priate analysis of tumor angiogenesis and potential antiangio- was reflected by the mRNA expression in these same three cell genic anticancer strategies in vitro. These experiments have lines. All of the murine endothelial cells had some expression of identified the 2H11 immortalized murine endothelial cells to be gPIIIB/CD36. Only the 2H11 cell line expressed P1H12/ a relevant murine model for tumor endothelial cells because CD146. Although all of the murine endothelial cell lines ex- 2H11 cells express many of the murine homologs of standard pressed vascular cellular adhesion molecule 1 (VCAM1)/ endothelial cell markers, including sialomucin/CD34, GPIIIB/ CD106, the expression level for VCAM1 by the 2H11 and CD36, endogolin/CD105, P1H12/CD146, and VCAM1/CD106, SVEC4–10EHR1 cell lines was most similar to that of the as well as several murine homologs of tumor endothelial mark- HMVEC. ers. Like normal endothelial cells, the 2H11 cells will form a The 2H11 mouse endothelial cell line has been tested in cellular network when grown on extracellular matrices.

Fig. 4 Relative mRNA expression of the murine homologs of recognized endothelial cell surface markers determined by real-time-PCR. The results were normalized to 18S mRNA expression and compared with the expression of the same human endothelial cell surface markers by human microvascular endothelial cell(s) (HMVEC). The data are the means for two independent experiments. VEGFR, vascular endothelial growth factor receptor; ENDRB, endothelin-B receptor.

The variety of endothelial cell markers that were studied al. (31) found this receptor on the surface of vascular endothelial represent many aspects of endothelial cell function, but not all and smooth muscle cells and defined its role in mediating are specific for endothelial cells. Platelet endothelial cell adhe- vasoregulatory activity. P1H12/CD146 is involved in calcium- sion molecule-1/CD31 is also expressed by platelets, mono- independent homotypic microvascular endothelial cell adhesion cytes, neutrophils and selected T-cell subsets. The platelet en- and has become a widely used marker for microvascular endo- dothelial cell adhesion molecule-1/CD31 protein plays a major thelial cells (19, 32, 33). Tie1 and Tie2 are tyrosine kinase role in the cell-cell interactions of endothelial cells and is widely receptors for angiopoietin 1 and 2 believed to be specifically accepted as a pan-endothelial marker of all types of endothelial expressed on endothelial cells. The angiopoietin/Tie1 and Tie2 cells (21, 22). Sialomucin/CD34 also participates in cell-cell pathways are involved in embryonic and tumor angiogenesis interactions by playing a role in adherens junction formation and mediating endothelial cell motility and recruitment of peri- is primarily expressed by the tumor neovasculature (23). endothelial cells (34–37). The adhesion molecule, VCAM-1/ GPIIIB/CD36 expression has been identified on human dermal CD106, has been used as a marker for endothelial cells. In some microvascular endothelial cells as well as other nonendothelial systems, levels of VCAM-1 correlate with vascular injury and cell types, including platelets and monocytes (24). The GPIIIB tumor progression (33, 38–40). The VEGF pathway including glycoprotein binds to extracellular matrix proteins including VEGF and the receptors, VEGFR1 (flt-1) and VEGFR2 (kinase thrombospondin and collagen (25, 26) and is believed to play a insert domain-containing receptor/flk-1), is critical in embry- role in the vascular complications associated with malaria (27). onic, normal, and tumor angiogenesis (41). The VEGF receptors Endoglin/CD105 is related to the transforming growth factor-␤ are expressed on varied cells, including monocytes (VEGFR1), type III receptor and has been found to be expressed by endo- neuronal precursor cells (VEGFR1 and VEGFR2), and podo- thelial cells (28). It has been shown that endoglin/CD105 plays cytes (VEGFR2), but are most highly expressed on resting and a role in normal vascular architecture and has been found to be active endothelial cells (42–44). elevated in tumor endothelial cells in some systems (29, 30). St. Croix et al. (19) identified genes that were up-regulated Using antibodies selective for endothelin receptor B, Shetty et in endothelial cells isolated from a human colon carcinoma as

Fig. 5 Relative mRNA expression of the murine homologs of tumor endothelial cell surface markers determined by real-time-PCR. The results were normalized to 18S mRNA expression and are expressed relative to the expression level of each marker in 2H11 cells. The data are the means for two independent experiments. mTEM, murine tumor endothelial cell.

compared with endothelial cells isolated from normal colon gene expression profile more relevant to malignant disease is mucosa from the same patient. The genes expressed by the colon critical. Carson-Walter et al. (20) identified several human tu- tumor endothelial cells differed significantly from the genes mor endothelial markers predicted to be associated with the cell expressed by HMVEC and human umbilical vein endothelial surface and verified the differential expression pattern of the cell, the cells traditionally used when studying angiogenesis in homologs, mTEM1, mTEM5, and mTEM8 in murine tumor vitro. Therefore, the identification of cells or cell lines with a endothelial cells. The immortalized murine endothelial cell line,

Table 2 Flow cytometry detection of murine homologs recognized endothelial cell surface proteins. Murine endothelial cell lines were assessed for the expression of endothelial cell marker proteins using antibodies directed toward the mouse proteins. The same endothelial cell markers were assessed on HMVECa using antibodies directed toward the human protein. The data are presented as the percentage of the cell population expressing the protein. Endothelial cell surface marker Sialomucin GPIIIB Endoglin P1H12 VCAM1 Cell type CD34 CD36 CD105 CD146 CD106 2H11 31 31 5 18 20 2F2B 2 7 2 3 92 3B11 61 24 2 7 86 IP2E4 12 28 4 7 86 SVEC4–10 4 22 2 5 97 SVEC4-10EE2 2 19 2 5 99 SVEC4-10EHR1 2 24 2 9 36 HMVEC 0 3 97 96 28 a HMVEC, human microvascular endothelial cell; VCAM, vascular cellular adhesion molecule.

Fig. 6 A, number of 2H11 cells after 48-h exposure to anti-2,4-dinitrophenol (anti-DNP) control antibody (100 ␮g/ml), anti-tumor endothelial cell (anti-TEM) antibody (100 ␮g/ ml), or SU5416 (10 or 50 ␮M). B, inverted, phase microscope images of each of 2H11 murine endothelial cells (2 ϫ 104 cells/well) in a tube formation assay on a thick layer of Matrigel for4hat37°C alone (PBS) or along with exposure to anti-DNP control antibody (100 ␮g/ml), anti-TEM antibody (100 ␮g/ml), or SU5416 (50 ␮M).

2H11, expresses relatively high levels of mTEM1, mTEM5, SCID mice that maintain the hyperproliferative characteristics mTEM7, and mTEM8, suggesting that these cells may be a of the psoriasis as well as a functional human vasculature. These useful model of tumor endothelial cells for cell-based assays. methods allow human angiogenesis in a murine host but are not The mouse remains the model species of choice in cancer models of angiogenesis associated with malignant disease. experimental therapeutics. It is critical to the selection of po- Animal models including genetically engineered mice and tential therapeutics to be aware of the similarities and differ- xeno-transplanted mice are being developed to facilitate the ences between the human and murine molecular targets. Be- assessment of therapies directed toward human angiogenesis cause of the inter-species differences in specific molecular targets in the mouse. However, the mouse remains important in targets, it is often necessary to development potential therapeu- early preclinical development of potential antiangiogenic ther- tic agents directed toward the murine protein. Whether synge- apies. Cell-based models, including endothelial cell prolifera- neic murine tumors or human tumor xenograft models are used, tion, migration, and tube formation, are well-established as the the stromal compartment of the tumors is murine. Several strat- primary screen for potential antiangiogenic activity. Performing egies have been used to transplant functional human endothelial these assays in both human and murine endothelial cells that cells into immunodificient SCID mice such as s.c. implantation have characteristics of human tumor endothelial cells from of a Matrigel matrix, collagen/fibronectin matrices, or polylactic patients is the ideal. The immortalized murine 2H11 endothelial acid sponges containing genetically altered human umbilical cell line appears to be an appropriate murine endothelial cell vein endothelial cell or HMVEC (45–49). Alternatively, the model of tumor angiogenesis. Continuing use of these murine SCID mouse has been successfully engrafted with human stem cells in cell-based angiogenesis assays will be needed to estab- cells (50). Raychaudhuri et al. (51) developed a model that lish their usefulness in understanding endothelial cell biology involves the transplantation of human psoriasis plaques into and drug discovery.

REFERENCES 22. Wong D, Dorovini-Zis K. Platelet/endothelial cell adhesion molecule-1 (PECAM-1) expression by human brain microvessel endothelial 1. Teicher BA, editor. Tumor Models in Cancer Res. New Jersey: The cells in primary culture. Brain Res 1996;731:217–20. Humana Press, Inc., 2001. 23. Krause DS, Fackler MJ, Civin CI, May WS. CD34: structure, 2. Modzelewski RA, Davies P, Watkins SC, Auerbach R, Chang M-J, biology, and clinical utility. Blood 1996;87:1–12. Johnson CS. Isolation and identification of fresh tumor-derived endothelial cells from a murine RIF-1 fibrosarcoma. Cancer Res 1994;54: 24. Swerlick RA, Lee KH, Wick TM, Lawley TJ. Human dermal 336–9. microvascular endothelial but not human umbilical vein endothelial cells express CD36 in vivo and in vitro. J Immunol 1992;148:78–83. 3. Morikawa S, Baluk P, Kaidoh T, Haskell A, Jain RK, McDonald 25. Asch AS, Barnwell JW, Silverstein RL, Nachman RL. Isolation of DM. Abnormalities in pericytes on blood vessels and endothelial sprouts thrombospondin membrane receptor. J Clin Investig 1987;79:1054–61. in tumors. Am J Pathol 2002;160:985–1000. 26. Tandon NN, Kralisz U, Jamieson GA. Identification of glycoprotein 4. Thurston G, Suri C, Smith K, et al. Leakage-resistant blood vessels IV (CD36) as a primary receptor for platelet-collagen adhesion. J Biol in mice transgenically overexpressing angiopoietin-1. Science (Wash D Chem 1989;264:7576–83. C) 1999;286:2511–4. 27. Oquendo P, Hundt E, Lawler J, Seed B. CD36 directly mediates 5. Brown EB, Campbell RB, Tsuzuki Y, et al. In vivo measurement of cytoadherence of plasmodium falciparum parasitized erythrocytes. Cell gene expression, angiogenesis and physiological function in tumors 1989;58:95–101. using multiphoton laser scanning microscopy. Nat Med 2001;7:864–8. 28. Ge AZ, Butcher EC. Cloning and expression of a cDNA encoding 6. Lu L, Wang S, Auerbach R. In vitro and in vivo differentiation into mouse endoglin, an endothelial cell TGF-␤ ligand. Gene 1994;138: B Cells, T cells and myeloid cells of primitive yolk sac hematopoietic 201–6. precursor cells expanded Ͼ100-fold by coculture with a clonal yolk sac 29. Arthur HM, Ure J, Smith AJ, et al. Endoglin, an ancillary TGF␤ endothelial cell line. Proc Natl Acad Sci USA 1996;93:14782–7. receptor, is required for extraembryonic angiogenesis and plays a key 7. Capillo M, Mancuso P, Gobbi A, et al. Continuous infusion of role in heart development. Dev Biol 2000;217:42–53. endostatin inhibits differentiation, mobilization, and clonogenic poten- 30. Miller DW, Graulich W, Karges B, et al. Elevated expression of tial of endothelial cell progenitors. Clin Cancer Res 2003;9:377–82. endoglin, a component of TGF-␤-receptor complex, correlates with 8. Shirakawa K, Furuhata S, Watanabe I, et al. Induction of vasculo- proliferation of tumor endothelial cells. Int J Cancer 1999;81:568–72. genesis in breast cancer models. Br J Cancer 2002;87:1454–61. 31. Shetty SS, Okada T, Webb RL, DelGrande D, Lappe RW. Func- 9. Luttun A, Carmeliet G, Carmeliet P. Vascular progenitors: from tionally distinct endothelin B receptors in vascular endothelium and biology to treatment. Trends Cardiovasc. Med 2002;12:88–96. smooth muscle. Biochem Biophys Res Comm 1993;191:459–74. 10. Shirakawa K, Shibuya M, Heike Y, et al. Tumor-infiltrating endo- 32. Solovey AN, Gui L, Chang L, Enenstein J, Browne PV, Hebbel RP. thelial cells and endothelial precursor cells in inflammatory breast Identification and functional assessment of endothelial P1H12. J Lab cancer. Int J Cancer 2002;99:344–51. Clin Med 2001;138:322–31. 11. Lyden D, Hattori K, Dias S, et al. Impaired recruitment of bone- 33. Mutin M, Dignat-George F, Sampol J. Immunologic phenotype of marrow-derived endothelial and hematopoietic precursor cells blocks cultured endothelial cells: quantitative analysis of cell surface mole- tumor angiogenesis and growth. Nat Med 2001;7:1194–201. cules. Tissue Antigens 1997;50:449–58. 12. Heissig B, Hattori K, Dias S, et al. Recruitment of stem and 34. Davis S, Aldrich TH, Jones PF, et al. Isolation of angiopoietin-1, a progenitor cells from the bone marrow niche requires MMP-9 mediated ligand for the TIE2 receptor, by secretion-trap expression cloning. Cell release of kit-ligand. Cell 2002;109:625–37. 1996;87:1161–9. 13. Asahara T, Masuda H, Takahashi T, et al. Bone marrow origin of 35. Suri C, Jones PF, Patan S, et al. Requisite role of angiopoietin-1, a endothelial progenitor cells responsible for postnatal vasculogenesis in ligand for the TIE2 receptor, during embryonic angiogenesis. Cell physiological and pathological neovascularization. Circ Res 1999a;85: 1996;87:1171–80. 221–8. 36. Maisonpierre PC, Suri C, Jones PF, et al. Angiopoietin-2, a natural 14. Asahara T, Takahashi T, Masuda H, et al. VEGF contributes to antagonist for Tie-2 that disrupts in vivio angiogenesis. Science (Lond) postnatal neovascularization by mobilizing bone marrow-derived endo- 1997;277:55–60. thelial progenitor cells. EMBO J 1999b;18:3964–72. 37. Koblizek TI, Weiss C, Yancopolous GD, Deutsch U, Risau W. 15. O’Connell KA, Edidin M. A mouse lymphoid endothelial cell line Angiopoietin-1 induces sprouting angiogenesis in vitro. Curr Biol 1998; 8:529–32. immortalized by simian virus 40 binds lymphocytes and retains functional characteristics of normal endothelial cells. J Immunol 1990;144: 38. Grooby WL, Krishnan R, Russ GR. Characterization of ovine 521–5. umbilical vein endothelial cells and their expression of cell adhesion molecules: comparative study with human endothelial cells. Immunol 16. O’Connell K, Landman G, Farmer E, Edidin M. Endothelial cells Cell Biol 1997;75:21–8. transformed by SV40 T antigen cause Kaposi’s sarcomalike tumors in nude mice. Am J Pathol 1991;139:743–9. 39. Iiyama K, Hajra L, Iiyama M, et al. Patterns of vascular cell adhesion molecule-1 and intercellular adhesion molecule-1 expression 17. O’Connell KA, Rudmann AA. Cloned spindle and epitheliodid cells in rabbit and mouse atherosclerotic lesions and at sites predisposed to from murine Kaposi’s sarcoma-like tumors are of endothelial origin. lesion formation. Circ Res 1999;86:199–207. J Investig Dermatol 1993;100:742–5. 40. Charpin C, Garcia S, Andrac L, Horschowski N, Choux R, Lavaut 18. Bastaki M, Nelli EE, Dell’Era P, et al. Basic fibroblast growth MN. VCAM (IGSF) adhesion molecule expression in breast carcinomas factor-induced angiogenic phenotype in mouse endothelium: a study of detected by automated and quantitative immunocytochemical assays. aortic and microvascular endothelial cell lines. Arterioscler Thromb Hum Pathol 1998;29:896–903. Vasc Biol 1997;17:454–64. 41. Millauer B, Wizigmann-Voos S, Schnurch H, et al. High affinity 19. St. Croix B, Rago C, Velculescu V, et al. Genes expressed in human VEGF binding and developmental expression suggest FLK-1 as a major tumor endothelium. Science (Lond) 2000;289:1197–202. regulator of vasculogenesis and angiogenesis. Cell 1993;72:835–46. 20. Carson-Walter EB, Watkins DN, Nanda A, Vogelstein B, Kinzler 42. Sawano A, Iwai S, Sakurai Y, et al. Flt-1, vascular endothelial KW, St. Croix B. Cell surface tumor endothelial markers are conserved growth factor receptor 1, is a novel cell surface marker for the lineage in mice and humans Cancer Res 2001;61:6649–55. of monocyte-macrophages in humans. Blood 2001;97:785–91. 21. Scholz D, Schaper J. Platelet/endothelial cell adhesion molecule-1 43. Jin K, Zhu Y, Sun Y, Mao XO, Xie L, Greenberg DA. Vascular (PECAM-1) is localized over the entire plasma membrane of endothelial endothelial growth factor (VEGF) stimulates neurogenesis in vitro and cells. Cell Tissue Res 1997;290:623–31. in vivo. Proc Natl Acad Sci USA 2002;99:11946–50.

44. Fan L, Wakayama T, Yokoyama S, Amano O, Iseki S. Downregu- 48. Nor JE, Christensen J, Mooney DJ, Polverini PJ. Vascular endo- lation of vascular endothelial growth factor and its receptors in the thelial growth factor (VEGF)-mediated angiogenesis is associated with kidney in rats with puromycin aminonucleoside nephrosis. Nephron enhanced endothelial cell survival and induction of bcl-e expression. 2002;90:95–102. Am J Pathol 1999;154:375–84. 45. Bosma MJ, Carroll AM. The SCID mouse mutant: definition, 49. Yang J, Nagavarapu U, Relloma K, et al. Telomerized human characterization, and potential uses. Ann Rev Immunol 1991;9:323–50. microvasculature is functional in vivo. Nat Biotechnol 2001;19:219–24. 46. Brooks PC, Montgomery AM, Rosenfeld M, et al. Integrin ␣ v ␤ 3 50. Greiner DL, Hesselton RA, Shultz LD. SCID mouse models of antagonists promote tumor regression by inducing apoptosis of angio- human stem cell engraftment. Stem Cells 1998;16:166–77. genic blood vessels. Cell 1994;79:1157–64. 51. Raychaudhuri SP, Sanyal M, Raychaudhuri SK, Dutt S, Farber EM. 47. Schechner JS, Nath AK, Zheng L, et al. In vivo formation of Sever combined immunodeficiency mouse-human skin chimeras: a complex microvessels lined by human endothelial cells in an immuno- unique animal model for the study of psoriasis and cutaneous inflam- deficient mouse. Proc Natl Acad Sci USA 2000;97:9191–6. mation. Br J Dermatol 2001;144:931–9.

Jennifer Walter-Yohrling, Sharon Morgenbesser, Cecile Rouleau, et al.

Clin Cancer Res 2004;10:2179-2189.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/10/6/2179

Cited articles This article cites 50 articles, 15 of which you can access for free at: http://clincancerres.aacrjournals.org/content/10/6/2179.full#ref-list-1

Citing articles This article has been cited by 12 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/10/6/2179.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/10/6/2179. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.