A General Strategy for Isolation of Endothelial Cells From Murine Tissues
Characterization of Two Endothelial Cell Lines From the Murine Lung and Subcutaneous Sponge Implants
Abstract A rapid, reproducible method for the isolation of murine endothelial cells (ECs) has been developed. Murine ECs were highly enriched by collagenase digestion of mechanically minced lung and subcutaneous sponge implants followed by specific selection with rat anti-mouse CD31 (ie, PECAM-1) monoclonal antibody–coated magnetic beads (Dynabeads). Pure EC populations were isolated from primary cultures by a second cycle of immunomagnetic selection. The cells from the lung were then cloned by a limiting-dilution method to exclude the possibility of nonendothelial cell contamination. Of the 300 cells plated, 29 clones (≈10%) were obtained. The clones were positive for CD31 as measured by flow cytometry, and one clone from the lungs (1G11) and the cells from sponge implants (designated as SIECs) were then subjected to subsequent culture in vitro for 40 and 30 passages (up to 5 months), respectively. Characterization was performed on cells between passage 3 and 10. Both cell types formed contact-inhibited monolayers on gelatin and capillary-like “tubes” on Matrigel. However, 1G11 cells exhibited a “cobblestone” morphology, whereas SIECs had a fibroblast-like appearance at confluence. By flow cytometry and enzyme-linked immunosorbent assay, these cells constitutively expressed CD31, VE-cadherin (cadherin-5), CD34, ICAM-1, VCAM-1, and P-selectin. After stimulation with 30 ng/mL of tumor necrosis factor-α, the cells became positive for E-selectin (at 4 hours poststimulation) and the expression of ICAM-1, VCAM-1, and P-selectin was upregulated (after 24 hours of stimulation). The presence of VE-cadherin in 1G11 cells and SIECs was confirmed by fluorescence microscopy and Northern blot analysis. The phenotype and morphology of both cell types were stable during 5 months of culture, and there was no evidence of overgrowth by contaminating cells. Taken together, the approach outlined herein may provide a general strategy for the isolation and culture of ECs from a variety of murine tissues. The general strategy outlined here is simple, effective, and flexible, allowing the inclusion of further positive or negative selection steps.
Endothelial cells lining the blood vessels and leukocytes interact closely during the activation and expression of immunity as well as in inflammatory reactions and hemostasis. This interplay is mainly mediated by two communication systems: the cytokines, released by leukocytes and ECs, and the regulated expression of cell adhesion molecules on the surface of both cell types.1 Much of our current knowledge about these interactions has been derived from studies with ECs of human or bovine origin from large vessels. A variety of mouse EC lines have been described,2 3 4 5 6 7 8 9 10 11 12 13 14 but either many of these have lost important EC molecules or functions (eg, CD31) or transforming agents may have affected cell functions. Information on mouse microvascular ECs is scant and fragmentary, and a general strategy for isolation and culture of ECs from diverse tissues is missing.
Murine ECs have proven difficult to obtain and maintain in culture. The most useful techniques have involved either the perfusion of target organs with digestive enzymes or the digestion and/or homogenization of an entire organ.6 15 These methods are technically demanding and time consuming and generally yield relatively impure EC populations. Consequently, overgrowth of contaminating nonendothelial cells may occur in culture. In addition, a large number of mice are required for these EC preparations. In recent years, methodological advances, such as the use of Dynabeads and flow cytometry in conjunction with mAbs or appropriate lectins, have provided powerful tools to improve the purity of EC cultures.16 17 However, these purification systems have been rarely used to isolate murine ECs directly from tissue homogenates.14 Long-term maintenance of normal ECs remains a difficult task. As an alternative, immortalization (eg, with the PmT) has been used to obtain microvascular endothelial cultures from different tissues.2 3 4 However, these immortalized cells are transformed and tumorigenic and can differ considerably from their normal counterparts, as illustrated, for instance, by recent results of in vitro versus in vivo expression of the long pentraxin PTX3.18
CD31 is a 130-kDa integral membrane glycoprotein that belongs to the immunoglobulin superfamily of cell adhesion molecules19 20 and is present on platelets, leukocytes, and ECs but at 10-fold higher levels on ECs.21 22 23 In humans, CD31 represents an excellent panendothelial marker,24 and mAbs specific for CD31 have been used to obtain ECs from mixed cell populations or as a reliable marker to identify ECs isolated by other techniques.25 26 Recently, we developed mAbs against mouse CD3127 that specifically recognize blood vessels on tissue sections and react with cultured murine ECs transformed by PmT. In the present study, we developed a methodology for purifying murine ECs by using these mAbs in conjunction with magnetic Dynabeads. This technique permitted the recovery and culture of ECs from the lungs or subcutaneous sponges, as an alternative source of vascularized tissue. The ECs thus obtained have been successively cultured for 40 and 30 passages, respectively, without detectable changes in morphology and phenotype. Therefore, the method described herein may represent a general strategy for the isolation and long-term maintenance of mouse ECs from different organ sites.
Reagents and Antibodies
DMEM, RPMI 1640, l-glutamine, FCS, sodium pyruvate, and nonessential amino acids were from Seromed Biochrom; plastics for cell culture were from Falcon; anti-mouse CD31 (MEC13.3 and MEC7.46)27 and CD34 (MEC14.7)27A mAbs raised in rats were developed in this department (C.G. et al, unpublished data, 1996); biotinylated MEC7.46 was from HyCult biotechnology bv; anti–ICAM-1 (CRL1878) and anti–VCAM-1 (M/K2.7) mAbs raised in rats were from the American Type Culture Collection (Rockville, Md.); rat mAb anti–E-selectin (21KC10), rabbit pAbs against mouse P-selectin and the cytoplasmic domain of VE-cadherin were generous gifts of Dr D. Vestweber (Münster, Germany); horseradish peroxidase–conjugated sheep anti-rat immunoglobulin and donkey anti-rabbit immunoglobulin were from Amersham (Buckinghamshire, England); and FITC-conjugated mAb anti-rat immunoglobulin and sheep anti-rabbit immunoglobulin were from Sera-Lab (Sussex, England).
The mouse EC line H5V, established in this laboratory,3 and the B16BL6 melanoma line were routinely maintained in DMEM with 10% FCS. The murine fibroblast cell line L929 was maintained RPMI 1640 with 10% FCS.
C57BL/6NCrLBR female mice (18 to 20 g body weight) were purchased from Charles River, Calco, Italy. Procedures involving animals and their care were conducted in conformity with institutional guidelines that are in compliance with national and international laws and policies (EEC Council Directive 86/609, OJ L 538,1, December 12, 1987; National Institutes of Health [NIH] Guide for the Care and Use of Laboratory Animals, NIH publication No. 85-23, 1985).
Implantation of Sponges
Vascularized sponges from subcutaneous sites were obtained using the methods described by MacPhee et al28 with some modifications. A fragment of sponge (Spongostan Anal), hydrated overnight at 4°C in PBS (GIBCO-BRL) and containing 1 mg/mL of EC growth supplements (prepared from bovine brain as described29 ) and heparin (Sigma Chemical Co), was inserted subcutaneously in the right flank of an anesthetized mouse. After 7 days the sponges were removed for EC isolation.
EC Isolation and Culture
Tissues (lung and sponges) were removed aseptically, rinsed in Hanks’ balanced salt solution (GIBCO), minced into ≈1×2-mm squares, and digested in 20 mL of collagenase A (1 mg/mL, Boehringer Mannheim) at 37°C for 45 minutes with occasional agitation. The cellular digest was filtered through sterile 31-μm nylon mesh, centrifuged at 400g for 10 minutes, and washed twice in 10% FCS–DMEM; the cell pellet was resuspended in 4 mL of 10% FCS–DMEM.
Dynabeads (Dynal AS) coated with sheep anti-rat IgG (30-μL aliquot per 5-mL tube) were incubated in 1 mL of MEC13.3 supernatant at 4°C overnight and then washed three times with 10% FCS–DMEM; 1 mL of cell suspension was put into the tube containing the washed beads. After 30 minutes at 4°C with occasional agitation, the bead-bound cells were recovered, washed five times with 10% FCS–DMEM and once with FCS-free DMEM, and then digested for 5 to 10 minutes at 37°C in 1 mL of trypsin/EDTA (GIBCO) to release the beads. The bead-free cells were centrifuged in 10% FCS–DMEM and then resuspended in 7 mL of growth medium (see below) for culture.
Cells in culture were collected by trypsin/EDTA digestion, centrifuged in 10% FCS–DMEM, and mixed with washed beads at a ratio of 1 to 3 beads per cell. Bead-binding cells were separated as described above. The isolates were cultured in 25-cm2 flasks precoated overnight with 1% gelatin (type B from bovine skin, Sigma) in PBS. The growth medium was 20% FCS–DMEM, 2 mmol/L l-glutamine, 2 mmol/L sodium pyruvate, 20 mmol/L HEPES, 1% nonessential amino acids, 100 μg/mL streptomycin, 100 UI/mL penicillin, freshly added heparin, and EC growth supplement at final concentration of 100 μg/mL. Confluent cells were passed routinely at a split ratio of 1 to 3 after trypsin/EDTA digestion and cultured under the same conditions. Cell cloning was performed as described.30
In Vitro Angiogenesis Assay on Matrigel
ECs (5×105) in 0.5 mL of 20% FCS–DMEM were put onto 0.3 mL of polymerized Matrigel (10 mg/mL, Collaborative Research, Inc) in a 24-well plate and incubated at 37°C overnight.
ECs (2×104) in 0.2 mL of growth medium were incubated for 2 days in gelatin-coated 96-well plates. For cytokine stimulation, TNF-α (BASF/Knoll) was used at a concentration of 30 ng/mL. The ELISA was performed as described.5 For VE-cadherin and P-selectin, the cell monolayers were fixed and permeabilized as described31 before specific antibodies were added. Samples were tested in triplicate. Four to six experiments were performed for each marker except for P-selectin (two experiments).
Immunofluorescent staining for VE-cadherin was done as described.32 Confluent monolayers grown on glass coverslips were used. For flow cytometry, confluent monolayers were detached by a 1-minute exposure to 37°C prewarmed trypsin/EDTA. Cells were washed, incubated with the different antibodies, and prepared for FACS analysis as described.27 For VE-cadherin, cells were fixed and permeabilized as described above. Samples were read by the FACStar flow cytometer (Becton-Dickinson), and the data were analyzed by lysys II software.
Northern Blot Analysis
Northern blot analysis for VE-cadherin was performed as described.31 In brief, total RNA was extracted and purified by the guanidinium isothiocyanate/CsCl2 method. Conditions for electrophoresis, blotting, and hybridization of purified RNA have been described.31 The EC lines described herein are available to interested scientists on request.
Specificity and Cell Recovery Efficiency of the Immunomagnetic Isolation Method
The immunomagnetic isolation technique was developed using a rat anti-murine CD31 mAb precoated with commercially available Dynabeads, which were conjugated with sheep anti-rat immunoglobulin antibodies. To evaluate the specificity and recovery efficiency of this method, the CD31+ murine EC line H5V (5×106 cells, Fig 1A⇓) was mixed with the VLA-4+ (α4β1-integrin) B16BL6 (Fig 1B⇓) melanoma cell line in a ratio of 1 to 1. After selection with CD31 mAb-coated Dynabeads, the percentage of bead-binding cells in the isolates was >98% as measured microscopically, and the recovery rate was 65% of the original H5V cell number. These cells were then digested with trypsin to release the beads and cultured for another 7 days. Flow cytometric analysis showed that there were virtually no VLA-4+ cells in the resultant cell population and that they did express CD31 (Fig 1C⇓). The intensity of CD31 staining in the isolates was lower than that in the original H5V population (Fig 1A⇓), this was not the result of melanoma cell contamination because treatment of H5V cells alone with the identical procedure resulted in the same CD31 profile (data not shown). The decrease in CD31 intensity probably reflects steric hinderance on the part of the MEC 13.3 antibody, which was still bound to the cells after trypsin treatment. If MEC 13.3–selected cells were stained with biotinylated MEC 7.46, an anti-CD31 mAb that does not compete with MEC 13.3, no decrease in CD31 expression was observed.
Isolations were also performed using a mixture of H5V and L929 fibroblasts at a ratio of 1 to 20 (H5V to L929) with comparable results. These preliminary experiments suggest that this method can be used to isolate CD31+ cells from a mixed cell population.
Isolation of CD31+ Cells From Murine Tissues
Immunomagnetic selection for CD31+ cells was then performed after collagenase digestion of minced murine lungs and subcutaneous sponge implants, as detailed in “Methods.” After trypsin digestion to release the beads, the isolates were cultured for 10 to 14 days until they reached confluence. The primary cultures were again selected by the same method, and cultures at the second passage were analyzed by flow cytometry. Both cell cultures obtained from the lung (Fig 2A⇓) or sponge implants (not shown) were positive for CD31. Cells isolated from the lung were also cloned by using a limiting-dilution method. Of the 300 cells plated, 29 clones were obtained within 2 weeks. Screening by flow cytometry revealed that these cell clones were all positive for CD31. These data suggest that after two cycles of selection, the resultant cell populations consisted of ECs. This finding was also confirmed by the fact that they could be maintained in vitro for a long time without any detectable changes in phenotype and morphology (see below). EC cultures were successfully established, even when a very mixed cell population, such as that obtained from lung digestion, was used; in this case the CD31+ cells recovered after the first cycle of immunomagnetic selection were 0.58% of the input population (5×108 cells). Recovery of CD31+ cells after the second cycle of selection was 45%.
Characterization of Murine ECs
One cell clone (1G11) and the cells isolated from sponge implants (designated as SIECs, or sponge-induced ECs) at passages 3 to 10 were used for characterization. At confluence, SIECs displayed an elongate shape (Fig 3A⇓), while 1G11 cells (Fig 3B⇓) and the other clones from lung ECs showed a “cobblestone” morphology. In the in vitro angiogenesis assay on the extracellular matrix Matrigel, 1G11 and SIEC cells formed capillary-like structures within 18 hours (data not shown).
The expression of several cell molecules was evaluated by FACS analysis and ELISA. Fig 2⇑ (B through D) shows the profiles of the constitutive expression of CD31, CD34, VCAM-1, ICAM-1, and VE-cadherin on 1G11 cells. Three other clones were investigated for these markers with similar results. Constitutive expression of these molecules and of P-selectin was also observed on ELISA for 1G11 and SIECs. After stimulation with TNF-α, cells became positive for E-selectin and expressed a significantly increased level of P-selectin (data not shown). Absorbance values for E- and P-selectins were always significantly higher in TNF-α–stimulated ECs than in medium-exposed ECs (P<.01 by Student’s t test).
VE-cadherin is a molecule restricted to ECs. Antibodies directed to mouse VE-cadherin33 were used to investigate the distribution of this molecule on 1G11 cell monolayers by immunofluorescence. As expected, this molecule was visualized as a strong, membrane fluorescence at points of cell-to-cell contact (Fig 3C⇑), but simultaneous staining with irrelevant antibodies was negative (Fig 3D⇑). This expression pattern of VE-cadherin on murine ECs is consistent with that on their human counterparts.31 More recently, the gene encoding mouse VE-cadherin has been cloned.33 Total cytoplasmic RNA from H5V, 1G11, and SIE cells was subjected to Northern blot analysis. The results illustrated in Fig 4⇓ indicate that these cells express VE-cadherin–specific transcripts. 1G11 and SIE cells have been successively cultured in vitro for 40 and 30 passages, respectively. During this time, expression of CD31, ICAM-1, VCAM-1, and E-selectin, as examined by ELISA or flow cytometry, has been stable (data not shown).
The present study describes a strategy for the isolation of ECs from diverse mouse tissues, including lung and subcutaneous sponge implants. The strategy involves enzyme digestion and two rounds of positive selection with anti-CD31 mAbs and immunomagnetic beads. From preliminary experiments with artificial mixtures of cells with known different markers, it was possible to recover ≈65% of the CD31+ cells from the mixtures. This result shows that with single-cell suspensions, as are usually obtained after in vitro exposure of cells to enzymes, the isolation system with Dynabeads and mAb was relatively efficient. However, when enzyme digestion is performed on minced tissues, the resulting suspension often contains small, heterotypic aggregates, which can cause mixed cultures. With a subsequent round of cell selection in culture, we obtained a pure EC population expressing CD31.
The purity and functional properties of the cell lines and clones obtained were extensively characterized. The cells showed constitutive expression of VE-cadherin, CD31, CD34, and P-selectin and cytokine-induced expression of E-selectin. These molecules have been detected by different methods (ie, ELISA, cytochemistry, and FACS analysis). The presence of smooth muscle cells was reasonably excluded, since these cells are negative for CD31 and VE-cadherin.27 31
VE-cadherin and E-selectin are among the most specific markers for ECs, since no other cell type has been found to express them.24 33 34 Using antibodies that recognized mouse VE-cadherin, we showed that this molecule is concentrated at appositional surfaces of cultured mouse ECs and is expressed at intercellular boundaries of confluent monolayer, as reported for human ECs.29 34 The presence of VE-cadherin in cultured mouse ECs was further confirmed by Northern blot analysis using a specific cDNA probe. These findings provide strong evidence that murine ECs express VE-cadherin on their surfaces. Moreover, these cells also express CD31, CD34, and P-selectin. These molecules are shared by ECs and hematopoietic elements.
Finally, both cells demonstrated the ability to form capillary-like structure on Matrigel. This test of in vitro angiogenesis,35 though not absolutely specific for ECs, can help distinguish ECs from some common contaminating cell types, especially mesothelial cells,24 which can also display a cobblestone-like morphology. Thus, we conclude that the strategy described herein allows purification and culture of bona fide mouse tissue ECs that retain key functional properties of this cell lineage. The strategy outlined here to isolate and culture ECs is simple and effective. Moreover, it allows a flexible approach by the inclusion of additional positive and negative selection steps to further enrich ECs and eliminate contaminants. In preliminary studies, it allowed successful enrichment of ECs from mouse tumors.
Selected Abbreviations and Acronyms
|DMEM||=||Dulbecco’s modified Eagle’s medium|
|ELISA||=||enzyme-linked immunosorbent assay|
|FACS||=||fluorescence-activated cell sorting|
|FCS||=||fetal calf serum|
|PmT||=||Polyoma middle T|
|TNF||=||tumor necrosis factor|
This work was supported by the European Community Project EC-ALA/MED Countries International Scientific Cooperation, the “National AIDS Project,” and the “Italy-US Program on Therapy of Tumor” from Istituto Superiore di Sanità (to A.M.). The generous contribution of the Associazione Italiana Ricerca sul Cancro (AIRC) is gratefully acknowledged.
Mantovani A, Bussolino F, Dejana E. Cytokine regulation of endothelial cell function. FASEB J.. 1992;6:2591-2599.
Garlanda C, Parravicini C, Sironi M, De Rossi M, Wainstok de Calmanovici R, Carozzi F, Bussolino F, Colotta F, Mantovani A, Vecchi A. Progressive growth in immunodeficient mice and host cell recruitment by mouse endothelial cells transformed by polyoma middle-sized T antigen: implications for the pathogenesis of opportunistic vascular tumors. Proc Natl Acad Sci U S A.. 1994;91:7291-7295.
Hahne M, Jager U, Isenmann S, Hallmann R, Vestweber D. Five tumor necrosis factor-inducible cell adhesion mechanisms on the surface of mouse endothelioma cells mediate the binding of leukocytes. J Cell Biol.. 1993;121:655-664.
O’Connell KA, Edidin M. A mouse lymphoid endothelial cell line immortalized by simian virus 40 binds lymphocytes and retains functional characteristics of normal endothelial cells. J Immunol.. 1990;144:521-525.
Sato N, Sato T, Takahashi S, Kikuchi K. Establishment of murine endothelial cell lines that develop angiosarcomas in vivo: brief demonstration of a proposed animal model for Kaposi’s sarcoma. Cancer Res.. 1986;46:362-366.
Toda K-I, Tsujioka K, Maruguchi Y, Ishii K, Miyachi Y, Kuribayashi K, Imamura S. Establishment and characterization of a tumorigenic murine vascular endothelial cell line (F-2). Cancer Res.. 1990;50:5526-5530.
Auerbach R, Lu WC, Pardon E, Gumkowski F, Kaminska G, Kaminski M. Specificity of adhesion between murine tumor cells and capillary endothelium: an in vitro correlate of preferential metastasis in vivo. Cancer Res.. 1987;47:1492-1496.
Plendl J, Hartwell L, Auerbach R. Organ-specific change in Dolichos biflorus lectin binding by myocardial endothelial cells during in vitro cultivation. In Vitro Cell Dev Biol. 1993;29A:25-31.
Gumkowski F, Kaminska G, Kaminski M, Morrissey LW, Auerbach R. Heterogeneity of mouse vascular endothelium: in vitro studies of lymphatic, large blood vessel and microvascular endothelial cells. Blood Vessels.. 1996;24:11-23.
Modzelewski RA, Davies P, Watkins SC, Auerbach R, Chang MJ, Johnson CS. Isolation and identification of fresh tumor-derived endothelial cells from a murine RIF-1 fibrosarcoma. Cancer Res.. 1994;54:336-339.
Jackson CJ, Garbett PK, Nissen B, Schrieber L. Binding of human endothelium to Ulex europaeus I-coated Dynabeads: application to the isolation of microvascular endothelium. J Cell Sci.. 1990;96:257-262.
Introna M, Vidal Alles V, Castellano M, Picardi G, De Gioia L, Bottazzi B, Peri G, Breviario F, Salmona M, De Gregorio L, Dragani TA, Srinivasan N, Blundell TL, Hamilton TA, Mantovani A. Cloning of mouse PTX3, a new member of the pentraxin gene family expressed at extrahepatic sites. Blood.. 1996;87:1862-1872.
Newman PJ, Berndt MC, Gorski J, White GC, Lyman S, Paddock C, Muller WA. PECAM-1 (CD31) cloning and relation to adhesion molecules of the immunoglobulin gene superfamily. Science.. 1990;247:1219-1222.
Xie Y, Muller WA. Molecular cloning and adhesive properties of murine platelet/endothelial cell adhesion molecule 1. Proc Natl Acad Sci U S A.. 1993;90:5569-5573.
Stockinger H, Gadd SJ, Eher R, Majdic O, Schreiber W, Kasinrerk W, Strass B, Schnabl E, Knapp W. Molecular characterization and functional analysis of the leukocyte surface protein CD31. J Immunol.. 1990;145:3889-3897.
Tanaka Y, Albelda SM, Horgan KJ, van Seventer GA, Shimizu Y, Newman W, Hallam J, Newman PJ, Buck CA, Shaw S. CD31 expressed on distinctive T cell subsets is a preferential amplifier of beta 1 integrin-mediated adhesion. J Exp Med.. 1992;176:245-253.
Newman PJ, Albelda SM. Cellular and molecular aspects of PECAM-1. Nouv Rev Fr Hematol. 1992;34(suppl):S9-S13.
Hewett PW, Murray JC. Human microvessel endothelial cells: isolation, culture and characterization. In Vitro Cell Dev Biol Anim. 1993;29A:823-830.
Vecchi A, Garlanda C, Lampugnani MG, Resnati M, Matteucci C, Stoppacciaro A, Schnurch H, Risau W, Ruco L, Mantovani A, Dejana E. Monoclonal antibodies specific for endothelial cells of mouse blood vessels: their application in the identification of adult and embryonic endothelium. Eur J Cell Biol.. 1994;63:247-254.
Garlanda C, Berthier R, Garin J, Stoppacciaro A, Ruco L, Vittet D, Gulino D, Matteucci C, Mantovani A, Vecchi A, Dejana E. Characterization of MEC14.7 a new monoclonal antibody recognizing mouse CD34: a useful reagent for identifying and characterizing blood vessels and hematopoietic precursors. Eur J Cell Biol.. 1997;73:368-377.
MacPhee MJ, Wiltrout RH, McCormick KL, Sayers TJ, Pilaro AM. A method for obtaining and culturing large numbers of purified organ-derived murine endothelial cells. J Leukoc Biol.. 1994;55:467-475.
Maciag T, Cerundolo J, Ilsley S, Kelley PR, Forand R. An endothelial cell growth factor from bovine hypothalamus: identification and partial characterization. Proc Natl Acad Sci U S A.. 1979;76:5674-5678.
Breviario F, Caveda L, Corada M, Martin Padura I, Navarro P, Golay J, Introna M, Gulino D, Lampugnani MG, Dejana E. Functional properties of human vascular endothelial cadherin (7B4/cadherin-5), an endothelium-specific cadherin. Arterioscler Thromb Vasc Biol.. 1995;15:1229-1239.
Ayalon O, Sabani H, Lampugnani MG, Dejana E, Geiger B. Spatial and temporal relationships between cadherins and PECAM-1 in cell junctions of human endothelial cells. J Cell Biol.. 1994;126:247-258.
Breier G, Breviario F, Caveda L, Berthier R, Schnurch H, Gotsch U, Vestweber D, Risau W, Dejana E. Molecular cloning and expression of murine vascular endothelial-cadherin in early stage development of cardiovascular system. Blood.. 1996;87:630-641.
Dejana E, Corada M, Lampugnani MG. Endothelial cell-to-cell junctions. FASEB J.. 1995;9:910-918.