Endometrial Endothelial Cells Express Estrogen and Progesterone Receptors and Exhibit a Tissue Specific Response to Angiogenic Growth Factors

Microcirculation (1999) 6, 127–140 © 1999 Stockton Press All rights reserved 1073-9688/99 $12.00 http://www.stockton-press.co.uk

Endometrial Endothelial Cells Express Estrogen and Progesterone Receptors and Exhibit a Tissue Specific Response to Angiogenic Growth Factors M. LUISA IRUELA-ARISPE, JUAN CARLOS RODRIGUEZ-MANZANEQUE, AND GRAZIELLA ABU-JAWDEH Department of Pathology, Harvard Medical School, and Beth Israel Deaconess Medical Center, Boston, MA, USA ABSTRACT Objective: To develop a reliable method for the isolation and longterm culture of microvessel endothelial cells from human endometrium and to evaluate their response to angiogenic growth factors and steroid hormones in comparison to endothelial cells derived from other organs. Methods: Endometrial tissue from hysterectomy specimens were digested sequentially with collagenase and trypsin, cultured for 24 h, then selected by adhesion to anti-CD-34 coated magnetic beads. Alternatively, anti-CD-34coated beads could also be substituted by Ulex europaeus agglutinin-1, antiPECAM, or anti-E-selectin-coated beads. Characterization of the isolated cultures included expression of endothelial cell markers, regulation of E-selectin in response to TNF-␣, proliferative response to angiogenic growth factors, and expression of progesterone and estrogen receptors. We also analyzed the relative binding affinity of VEGF on endometrial endothelial cells in comparison to other endothelial cell types. Results: Selection on anti-CD-34-coated beads eliminated contaminating cells and resulted in a homogeneous population of human endometrial endothelial cells (HEEC), as assessed by expression of PECAM, von Willebrand’s factor, and uptake of acetylated-LDL. HEEC also upregulated E-selectin in response to TNF-␣ in a manner similar to that seen for other endothelial cell types. Expression of progesterone and estrogen receptor was revealed by immunocytochemistry and RT-PCR consistently until passage 5. Endometrial endothelial cells were more responsive to growth stimulation by VEGF than were dermal endothelial cells isolated under similar conditions. Further characterization indicated that VEGF bound more avidly to HEEC than to other endothelial cell types. Conclusions: Human endometrial endothelial cells were isolated to homogeneity by a two-part protocol and successfully passaged under culture conditions similar to those used for other endothelial cell types. The HEEC were very responsive to VEGF growth-stimulation likely due to elevated affinity, or increased levels of, KDR and FLT-1 on the cell surface. These results indicate that HEEC are capable of maintaining a mature phenotype in culture and might provide a model for understanding the response of these cells to the recurrent cycles of proliferation imposed on the endometrium during menstruation. KEY WORDS: angiogenesis, capillaries, microvessels, immunopurification, steroid hormones, VPF/VEGF Supported by the Pathology Foundation, Department of Pathology, Beth Israel Deaconess Medical Center, and from the NCI RO3#CA70559-01. For reprints of this article, contact Dr. Luisa Iruela-Arispe, Department of Molecular Cell and Developmental Biology, Molecular Biology Institute, 611 Circle Drive East, Los Angeles, CA 90095 USA. Received 19 June 1998; accepted 8 December 1998

INTRODUCTION

Our understanding of endothelial cell physiology and pathology depends, in large part, on studies performed in vitro with primary cultures of endothelial cells. Through studies of endothelial cells isolated from various sources, it has become increasingly apparent that considerable differences exist between

Endometrial endothelial cells and angiogenesis M.L. Iruela-Arispe et al. 128

macro- and microvascular endothelial cells. For instance, mediators that regulate a particular function in umbilical-vein-derived endothelial cells are not found to be equally effective on capillary endothelium (8,41). Others as well as we have shown that capillary endothelium express platelet-derived growth factor BB (PDGF BB) receptors and respond to this growth factor, whereas large-vessel endothelium do not (4,5,6,58). Heterogeneity can also be manifested by microenvironmental factors, because the response of endothelial cells is dependent on the extracellular context or stage of activation during angiogenesis (23,30). For example, transforming growth factor-␤1 inhibits proliferation of subconfluent endothelial cells, but stimulates mitosis when the same cells have organized capillary-like structures in vitro (36). In addition, a large number of studies has revealed organ-specific characteristics among microvascular endothelial cells. Clear differences exist with respect to cell-surface molecules (7,44), antigenic determinants (3,42), permeability, and metabolic properties (15,19,60). These findings indicate significant heterogeneity among vascular beds, and suggest that questions related to organ-specific physiology might be better addressed by studies of organ-specific endothelial cells. In this article, we describe a procedure for the isolation of human endometrial endothelial cells. The human endometrium is one of a limited number of organs that undergo recurrent cycles of physiological vascular regeneration. Our purpose is to utilize endometrial endothelial cells in studies of angiogenic progression and inhibition, as well as in the elucidation of other responses that are unique to endothelial cells from this tissue. Endothelial microvascular cells have been isolated from a variety of organs including skin (16,34,46), retina (10), synovium (37), brain (19), placenta (20,25), nasal mucosa (24), adrenal (22), mammary gland (32), heart (52), and lung (12,28,45). The endothelium of human endometrium, however has not been studied in vitro. This tissue presents dynamic physiological states on the vasculature, with recurrent cycles of growth and growth-inhibition that provide an interesting model in which to ask questions related to the molecular regulation of angiogenesis. We hypothesized that endometrialderived endothelial cells are more responsive to regulation by angiogenic factors and inhibitors than other endothelial cells. We also predict that these cells express estrogen and progesterone receptors and that its expression might provide modulation to growth factor response. Here, we test this hypothesis

and evaluate the effect of several growth factors that might be active in regulating the growth of these cells in vivo. The isolation of capillary-derived endothelial cells is a technically challenging proposition. This is due to the frequent contamination with neighboring cell types in the tissue, primarily pericytes, fibroblasts, and smooth muscle cells. A number of techniques have been used to increase the ratio of endothelial cells to other contaminating cell types in primary cultures. These include differential plating (10,13), mechanical weeding (46), media rich in D-valine (16), and filtration of microvessels through nylon membranes (10). More recently, flow cytometry (28) and binding to paramagnetic spheres (Dynabeads) coated with endothelial cell surface markers (31,32,37,55) have been used with great success. We have studied the feasibility of using paramagnetic spheres coated with anti-CD-34, and other endothelial-specific cell surface markers, to isolate homogeneous endothelial cultures from the human endometrium. Here, we describe this method and present data regarding the expression of endothelial cell markers in the purified cell populations. In addition, we demonstrate that human endometrial endothelial cells (HEEC) are more responsive to VEGF than are other endothelial cells from various tissues, and further demonstrate that the effect is likely due to an increased affinity, or elevated number of VEGF receptors, on the cell surface. These data provide evidence that cultured endometrial endothelial cells constitute a practical model system in which to explore questions related to the recurrent nature of the angiogenic response typical of the human endometrium and establish clear functional differences between HEEC and endothelial cells from other organs. MATERIALS AND METHODS Isolation of Endothelial Cells

Endometria from pre-menopausal hysterectomy specimens, removed for reasons other than endometrial abnormalities, were used for the isolation of endothelial cells. The uteri were from patients 30 to 50 y old and were processed under sterile conditions 30 min to 2 h after surgery. The use of human specimens for these experiments was reviewed and approved by the Committee on Clinical Investigations, Beth Israel Deaconess Medical Center, Boston, Massachusetts. Briefly, incisions were made along the lateral aspects

Endometrial endothelial cells and angiogenesis M.L. Iruela-Arispe et al. 129

of the uterus and strips of endometrium were removed from the upper portion of the posterior wall. The strips of endometrium were divided into cubes of 0.5–0.7 cm in diameter. Tissue cubes were then either processed for RNA extraction, fixed for histology/immunocytochemistry, or immersed in Hanks solution for isolation of cells.

cultures reached 80% confluency, a second magnetic purification was performed. After two immunemagnetic purifications, we consistently found that 100% of the cells were able to endocytose Ac-LDL. For subcultures, confluent HEEC were split at a 1:4 ratio and maintained for four additional passages (5–6 total).

For cell isolation, tissue cubes were rinsed in phenol red-free DMEM containing antibiotics (penicillin and streptomycin at 100 ␮g/mL) and amphotericin-B (Gibco/BRL, Grand Island, NY). Collagenase (Sigma, St. Louis, MO) at a final concentration of 0.1 mg/mL was added to the suspension and kept under constant agitation at room temperature for 1–2 h. Released cells were spun and pellets were rinsed 3 times by resuspension and centrifugation in fresh DMEM. A second mild digestion with trypsinEDTA was frequently necessary to further break up the microvessels. Filtration through nylon mesh (70 ␮M) (Fisher, Santa Clara, CA) increased the efficacy of attachment of endothelial cells to the coated fibronectin dishes, by removal of partially digested matrix fragments and epithelial components. After digestion, cells were washed in EBM-2 (Clonetics, Walkersville, MI) containing 1% FCS, by three sequential resuspension and centrifugation cycles. The final pellet was reconstituted in 3 mL of EBM-2 and plated on petri dishes previously coated with fibronectin (Sigma) (50 ␮g/mL) on the same media supplemented with 10% FCS, 100 U/mL penicillin, and 100 ␮g/mL streptomycin. After 2 h, nonadherent cells were removed. Adherent cells were washed and grown until confluency. These cultures were serially subcultured, and the purity of the cultured cells was examined by acetylated-LDL endocytosis (BTI, Stoughton, MA).

Human dermal endothelial cells (HDEC) were a gift from Michael Detmar (Department of Pathology, Beth Israel Deaconess Medical Center, Boston, MA) (55). Human umbilical vein endothelial cells (HUVEC) were a gift from Don Senger (Department of Pathology, Beth Israel Deaconess Medical Center, Boston, MA). Human coronary endothelial cells (HCEC) were obtained from Clonetics. Experiments were performed with cultures between passages 3 and 5. Subculturing ratio was always 1:4. Several independent isolates were used for each experiment to ensure consistency in the results.

Further purification was accomplished by an affinity method using antibodies directed against the cellsurface protein CD-34 (Affinity Bioreagents, Neshanic Station, NJ). Cultured cells were incubated with anti-CD-34-coated magnetic beads for 10–15 min (for preparation of the beads, see below). After several thorough washes with serum-free EBM, cultures were released from the substrate by brief treatment with trypsin, and cells with attached beads were isolated using a magnetic particle concentrator. Isolated cells were then plated and cultured in EBM-2 supplemented with 15% FCS, 1 ␮g/mL hydrocortisone acetate, 5 × 10−5 M N6, 2⬘-0-dibutyrul-adenosine 3⬘:5⬘-cyclic monophosphate (Sigma), 1000 U/mL penicillin, and 100 ␮g/mL streptomycin (Gibco/BRL). After 1–2 days or when

Immunomagnetic isolation of HEEC was also performed, with equal success, using anti-CD-31 or PECAM antibodies, anti-E-selectin-coated beads, or Ulex europaeus agglutinin-1, which binds specifically to terminal L-fucosyl residues on the surface of human endothelial cells (37). For E-selectin, pretreatment of cultures with TNF-␣ (50 ng/ml) (TNF-␣ was a kind gift from Dr. Michael Detmar, Beth Israel Deaconess Medical Center, Boston, MA) for 6 h was required for upregulation of this molecule at the cell surface. Preparation of Coated Magnetic Beads

Magnetic beads (15 mg), coated with sheep antimouse IgG or tosylactivated (for conjugation with lectin) (Dynal, Lake Success, NY), were incubated end-over-end with one of the following antibodies to endothelial cell surface markers: (a) mouse antihuman E-selectin monoclonal antibody (Genzyme, Cambridge, MA); (b) mouse anti-human CD34 monoclonal antibody (Affinity Bioreagents); (c) mouse anti-human PECAM monoclonal antibody; or (d) Ulex europaeus lectin (Vector Labs, Burlingame, CA), and were stored at 4 °C for up to 1 mo. Incubation took place overnight at 4 °C, after which magnetic beads were washed 4 times with serumfree media containing 0.2 mg/mL BSA under sterile conditions using a magnetic particle concentrator. Characterization of Endothelial Cells

Cells were plated on fibronectin-coated coverslips, washed with Hanks solution, and fixed in cold 4% paraformaldehyde for 30 min at 4 °C. Fixed cells

Endometrial endothelial cells and angiogenesis M.L. Iruela-Arispe et al. 130

were rinsed with PBS, blocked with 10% normal goat serum (Sigma) for 2 h, and incubated for 1 h with primary antibodies. Control coverslips were incubated with buffer alone. The preparations were then washed 5 times with PBS and incubated for 1 h with the appropriate biotinylated secondary antibody (Vector) at 1:200 dilution. Cells were again washed 5 times with PBS and incubated with avidin-fluorescein isothiocyanate (Vector) at 1:200 dilution. Cells were washed as above, mounted in 50% glycerol in PBS, and photographed with a Zeiss photomicroscope with an Ecktakrome film 1600 ASA. Purified cultures were also examined for the presence of scavenger receptors for acetylated low density lipoprotein using 1,1⬘dioctadecyl-3,3,3⬘,3⬘tetramethyl-indocarbocyanine perchlorate acetylated low density lipoprotein (Dil-Ac-LDL) (Biochemical Technologies, Inc., Stoughton, MA). Cells were incubated in DMEM containing Dil-AcLDL (10 ␮g/mL) at 37 °C for 4 h. Thereafter, cells were fixed for 30 min in 3% paraformaldehyde solution, mounted, and observed by fluorescent microscopy. For immunocytochemical studies, human dermal endothelial cells and stromal endometrial cells were used as positive and negative controls, respectively. Isolation of Total RNA and Northern Blot Analysis

Total RNA was isolated by the single-step procedure of Chomcznski and Sacchi (14) and the integrity of the samples was assessed on agarose minigels. Total RNA was transferred to nylon membranes (Nytran) (Amersham, Arlington Heights, IL), linked to the membranes by ultraviolet light (Stratalinker), and prehybridized at 42 °C for 2–5 h in a solution containing 50% formamide, 6× SSPE, 1× Denhart’s solution, 0.1% SDS, and 100 ␮g/mL of heatdenatured salmon sperm DNA. Hybridization with [32P]-labeled cDNA proceeded in the same solution at 42 °C for 12–18 h (32). Stringency of washes varied according to the probe. Membranes were then exposed to Kodak X-Omat AR film for 1–3 days. Probes were prepared from cDNA restriction fragments labeled with [32P]-dCTP by random priming using a Multiprime kit (Amersham) and purified on Sephadex G-50 (Promega, Madison, WI). Loading and transfer efficiency were evaluated with a probe to 28S rRNA. Densitometry or Phosphor imager scans of these signals were used to normalize the values obtained from other probes. The expression of E-selectin mRNA was investigated

on cultures treated with TNF-␣ (50 ng/mL) for 30 min, 2 h, 4 h, and 6 h. Cell Proliferation Assays

Confluent cultures were incubated for 48 h in the absence of serum or growth factors. Under these conditions, we have found that endothelial cells enter a state of mitotic quiescence. Quiescent cells were then reseeded on plates previously coated with Vitrogen (50 ␮g/mL) and treated with 1% FCS alone or in the presence of VEGF (Peprotech, Rocky Hill, NJ), FGF-2 (a kind gift from Dr. Gera Neufelt, Israel), or PDGF-BB (Peprotech). Incubation times varied as indicated in the figure legends. For assays longer than 48 h, media was replaced every 2 days. At the end of treatment, cells were either counted with hemocytometer or coulter counter, or pulsed with 1 ␮Ci/mL of [3H]-thymidine (NEN, Boston, MA). Thereafter, cultures were washed twice with serum-free DMEM, fixed with cold 10% trichloroacetic acid for 10 min, washed with cold ethanol, and air dried. Incorporation of [3H]-thymidine into acid-insoluble material was determined by scintillation counting. Binding Assays

VEGF (Peprotech) was iodinated according to the procedure of Hunter and Greenwood (33). Briefly, 1␮g of VEGF was mixed with 1␮Ci of Na[125I] and 10␮L of chloramine T (Sigma) (0.2 mg/mL) up to a final volume of 150␮L in 0.5M phosphate buffer pH 7.6. Approximately 1 min later, the reaction was stopped by the addition of 10␮L of saturated acetyltyrosine and 10 ␮L of 10 mMKI. Free 125I was removed on a desalting PD10 column (Biorad) preequilibrated with PBS containing 1 mg/mL of gelatin. Fractions were counted before and after precipitation with 10% trichloroacetic acid. Endothelial cells (HEEC and HDEC) were seeded on dishes and allowed to grow to confluency (7 × 104 cells). Cells were then transferred to 4 °C and washed with cold binding buffer (EBM-2 containing 1 mg/mL gelatin). Binding was performed at 4 °C for 1 h. Experiments were done in triplicate. Competition was performed with increased levels of unlabeled VEGF. Non-specific binding was determined in the presence of an excess of unlabeled VEGF (500 ng/mL). At the end of the incubation period, cultures were washed 4 times with cold PBS and one last time with cold PBS containing 2M NaCl. At this point, cells were lysed with 500␮L of 1% sodium dodecyl sulfate and 1mM EDTA. At saturating concentrations, the nonspecific binding was less

Endometrial endothelial cells and angiogenesis M.L. Iruela-Arispe et al. 131

with Dynabeads, cells coated with numerous magnetic beads could be seen under contrast microscopy. Cells spread to the fibronectin-coated dishes within 2 h. Presence of the beads did not inhibit adhesion or spreading, and the beads were rapidly diluted out and lost during the first two passages. Using this technique, HEEC grew to confluency, showed contact inhibition, and presented a typical cobblestone morphology. The use of EBM-2 media supplemented with 15% fetal calf serum, hydrocortisone, and cAMP was essential to prevent senescence and keep a doubling time of 30 h to 42 h for four-to-five passages. Figure 1A shows a phase-contrast micrograph of the

Figure 1. Isolation of endothelial cells from human endometrial specimens. (A) Phase- contrast micrograph of cells isolated from the endometrium 24 h after initial plating. Two morphologies are apparent: cobblestone (closed arrows), likely to represent endothelial cells and elongated (open arrows), likely from stromal cells. (B) Expression of CD34 in human endometrium. Endothelial cells from capillaries are positive as indicated by the brown peroxidase reaction (arrows). (C) Confluent culture of HEEC after selection with anti-CD-34-coupled magnetic beads. Note the cobblestone appearance of the monolayer that is free of beads after two passages. (D) Acetylated-LDL binding of endothelial cells after immune-selection with CD34. All cells in the culture are positive.

than 20%. Binding was analyzed according to Scatchard’s procedure with the use of the ligandfitting program. RESULTS

We have developed a two-step procedure for the isolation of HEEC, which consists of differential plating and magnetic isolation with endothelial-specific cellsurface markers. The use of coated Dynabeads to enrich endothelial cell preparations dramatically improved the purity of the cultures. Rarely, a second magnetic isolation was necessary. After selection

Figure 2. Expression and distribution of endothelial cell markers on endometrial tissue. Endometrial tissue was fixed in paraformaldehyde and embedded in paraffin. Sections (5␮m) were subjected to immunocytochemical analysis using the following primary antibodies: (A) antivon Willebrand factor (vWF); (B) Biotinylated Ulex europeus lectin; (C) anti-CD-31; and (D) anti-CD-34. Detection was accomplished by use of a biotinylated secondary antibody followed by avidin-FITC. Staining with Hoechst revealed nuclei in blue. Arrows indicate positive vessels.

Endometrial endothelial cells and angiogenesis M.L. Iruela-Arispe et al. 132

cultured cells after differential plating, but prior to magnetic bead selection. Note two clearly distinct morphologies: (a) cobblestone (closed arrow), presumably endothelial cells; and (b) elongated (open arrow), presumably fibroblasts. By endocytosis of Ac-LDL, we verified that approximately 70% of the cells in most preparations were of endothelial origin. Further purification was accomplished by an affinity method using antibodies directed against the cellsurface protein CD-34, CD-31, or Ulex lectin. As depicted by Fig. 1B, CD-34 is specific to the endothelium. CD-34 is a cell-surface protein present in both endothelial and hematopoietic cells, the lattercell type was not present in our original cultures. Cultured cells were incubated with anti-CD-34 magnetic beads and released from the substrate by brief treatment with trypsin. Cells bound to beads were subsequently plated and subcultured. Microscopic examination post-confluency showed a clearly distinct cobblestone monolayer (Fig. 1C). Cultures were also incubated with acetylated-LDL. After immune-magnetic purification, we consistently found that 100% of the cells were able to uptake Ac-LDL, as shown by the punctuate fluorescence indicating the endocytosis of Ac-LDL into secondary lysosomes (Fig. 1D). To confirm the purity of the endothelial isolations and to characterize the cultures further, we performed a series of immunocytochemical analysis and functional assays. For the immunocytochemistry, we initially tested the expression of the endothelial markers in the endometrial tissue. Figure 2 shows immunolocalization of von Willebrand factor (vWF) (Fig. 2A), Ulex europeus lectin (Fig. 2B), CD-31 (Fig. 2C), and CD-34 (Fig. 2D). All antibodies provided vascular patterns although interesting differences were observed. Traditionally, endothelial cells have been characterized immunocytochemically with antibodies against vWF. Although antibodies against this protein generally provide an excellent label of endothelial cells in large blood vessels, binding to capillary endothelium has been inconsistent. This phenomenon was clearly evident in the endometrium. Larger and medium-size vessels were well-stained, however the number of positive vessels was clearly lower than in adjacent sections stained with Ulex, CD-34, or CD31. One can associate the inconsistency in staining to possible regulation of vWF during the endometrial cycle, as has been previously suggested (2,38), or to the fact that vWF is differentially regulated in vascular beds of different size and organs, as indicated by recent data on promoter expression (1). We tend to favor the second hypothesis, because analysis of

12 independent endometrial specimens from different stages of the endometrial cycle showed the same pattern (data not shown). In addition, we found that anti-CD-31 did not stain subepithelial capillaries, while both Ulex and anti-CD-34 stained these vessels. PECAM-1/CD-31 is a member of the celladhesion molecule subfamily of Ig-like proteins and is constitutively expressed at the cell surface of endothelial cells and platelets (51). Nevertheless, the subendothelial capillaries are considered to be sinusoidal in nature, therefore, it is conceivable that the nature of their cell-juctional interactions is different from that of continuous endometrial capillaries, which showed expression of PECAM-1. Both Ulex and CD-34 showed ubiquitous expression among different capillaries in the endometrium. In addition, dermal endothelial cells and endometrial stromal cells were used as positive and negative controls, respectively. Immunolocalization of vWF showed a punctuate intracellular pattern in most, but not all, endometrial and dermal-derived endothelial cells (Figs. 3A and B). This pattern was consistent with the spotty distribution observed in the endometrial tissue sections (Fig. 2). PECAM staining was detected at the cell membrane and more intensely at the cell–cell junctions (Figs. 3C and D). Expression of CD-34 was consistent in all the cultured cells (Fig. 3E). We also stained stromal fibroblasts, as controls; no signal was detected with antibodies toward CD34, PECAM, or Ulex in these cells (Fig. 3F). Human endometrial endothelial cells were also stained with desmin and smooth muscle ␣-actin to preclude contamination by pericytes or smooth muscle cells (data not shown). Expression of E-selectin can provide further support to the endothelial nature of our isolates. E-selectin is expressed on endothelial cells and in platelets. In most endothelial cultures, E-selectin levels are low or undetectable (9), nevertheless, transcripts are significantly increased upon treatment with TNF-␣ (55). Human endometrial endothelial cell cultures showed low-basal level of E-selectin mRNA, which showed a significant increase as early as 30 min after treatment with TNF-␣ (Fig. 4). Upregulation peaked at 2 h and was still detected by 6 h. Cultures were refractive to a second TNF-␣ treatment administered within 24 h (data not shown). E-selectin protein was also increased by TNF-␣, but with slightly delayed kinetics. Upregulation of E-selectin by TNF-␣ was also used successfully to purify HEEC using magnetic beads coated with E-selectin antibodies. The advantage of this procedure is that the beads are quickly lost due to shedding of E-selectin from the endothelial surface. The potential criticism

Endometrial endothelial cells and angiogenesis M.L. Iruela-Arispe et al. 133

of this technique is that it would preferentially select only responsive endothelial populations. Endometrial Endothelial Cells Express Both Progesterone and Estrogen Receptors

Investigation of steroid receptors on endometrialderived endothelial cells is of particular interest in this tissue. The cyclic nature of endometrial repair and descamation is temporarily related to the serum levels of 17-␤-estradiol and progesterone. It has been repeatedly postulated that the angiogenic response of the endometrium is likely under hormonal control, nevertheless, there are no studies to support this claim directly. The presence of estrogen and progesterone receptors has been clearly and repeatedly demonstrated on smooth muscle cells (43), in fact, menstruation is believed to result from the constriction of the smooth muscle cells from the spiral arteries in response to low levels of progesterone and estradiol. Identification of steroid receptors on endothelial cells, however, has followed a more controversial path, with studies supporting and contradicting their expression on the endothelium. To some extent, these controversies could be founded on intrinsic organ-specific differences among vascular beds. We performed immunohistochemical analysis of progesterone and estrogen receptor expression in sections of endometrium and in isolated endothelial cells. Figure 5 shows immunolocalization of both progesterone (Fig. 5A) and estrogen (Fig. 5B) receptor on endothelial endometrial tissue sections. The expression is retained on the purified isolated cultures (Figs. 5C and D). Presence of these receptors was also confirmed at the transcriptional level. We conducted reverse transcriptase followed by PCR analysis on a series of total RNA samples from HEEC at different passage numbers. We were able to detect a positive band for both estrogen and progesterone receptors on HEEC cells (Fig. 5) consistently up to passage 5. The data confirms findings from the immunocytochemical analysis, and also demonstrates that expression of these receptors is lost upon serial passage in culture. Human Endometrial Endothelial Cells Express Receptors for VEGF and Respond Rapidly to VEGF Stimulation

Expression of VEGF receptors KDR and FLT-1 is largely confined to the endothelium, although a few exceptions to this rule have been noted in the literature (11). To further support the endothelial nature of purified HEEC cultures, we tested for VEGF receptor on endothelial and stromal cells isolated from endometrium. Figure 6 shows Northern blot analysis

for VEGF and for its receptors, KDR and FLT-1. Surprisingly, levels of both KDR and FLT-1 mRNA were elevated on HEEC when compared to dermal endothelial cells. In contrast, VEGF mRNA was detected at equivalent levels in both endometrial and dermal-derived fibroblasts. We next evaluated the proliferative response of HEEC to several angiogenic growth factors, VEGF, FGF-2, and PDGF-BB, and compared this response to that of endothelial cells derived from other organs (HDEC, HCEC, and HUVEC). All cells were used at the same passage number. Interesting differences were noted among the cell types (Fig. 7A and Table 1). Clearly, HEEC were the most responsive to VEGF, followed by HDEC. The response of endometrial-derived cells to FGF-2 was not as impressive as that of HDEC. The data suggest that clear differences exist in the response of organ-specific endothelial cells to particular growth factors. Interestingly, HEEC did not respond to PDGF-BB. PDGF-␤ receptors have been identified in endothelial cells from microvessels, being excluded from large vessels (4,5,6,58); therefore, it is not surprising that both HUVEC and HCEC respond poorly to this growth factor. Nevertheless, cells isolated from dermis and endometrium both contain a large component of microvascular-derived endothelium. We examined the expression of PDGF-␤ receptors in endometrial sections by immunocytochemistry and found that the microvasculature was devoid of signal (data not shown) providing an explanation to our findings. Therefore, the response of microvessel endothelial cells is potentially heterogeneous to this ligand, or at least excluded from endometrial tissue. Response of HEEC to VEGF was further evaluated by cell-cycle analysis and binding assays. Human endometrial endothelial cells and HDEC were made quiescent, and response to VEGF was evaluated by thymidine incorporation for 80 h (Fig. 7B). Human endometrial endothelial cells entered S phase by 20 h after treatment, in contrast to 30–35 h for HDEC. By 80 h, HEEC had completed two cell cycles, while HDEC had completed only one. Binding assays to VEGF were also performed on HEEC, HCEC, HDEC, and HUVEC (Fig. 8). An impressive difference was detected between HEEC and the other cell types. While complete competition was seen at 100 ng/mL of cold VEGF for HCEC, HUVEC, and HDEC; twice the amount of unlabeled competitor was required to eliminate binding of labeled VEGF to HEEC. The data suggest an increased number of receptors for VEGF on HEEC (as also indicated by Fig. 6) or a significantly higher

Endometrial endothelial cells and angiogenesis M.L. Iruela-Arispe et al. 134

Figure 3. Expression of endothelial cell markers by cultured endometrial endothelial cells. Human endometrialderived (A, C, and E), dermal-derived (B, and D) endothelial cells or endometrial fibroblasts (F) were cultured on glass cover-slips, fixed in paraformaldehyde, and treated briefly with triton X-100. Cultures were then incubated with the following antibodies: (A and B) vWF; (C and D) PECAM-1; and (E and F) CD-34. Immunocomplexes were detected with biotinylated secondary antibodies followed by avidin-FITC. Nuclei were stained with Hoechst. Arrows indicate positive reaction.

affinity for this growth factor by a yet unknown mechanism. Estrogen and Progesterone Modulate the Response of HEEC to Growth Factors

To evaluate the functional significance of estrogen and progesterone receptors on HEEC, we treated cultures with increasing concentrations of 17-␤estradiol and progesterone in the presence and absence of VEGF and FGF-2. Figure 9 indicates the results of six independent proliferation experiments performed on HEEC cultures from passages 3 to 5. Also, 17-␤-estradiol potentiated, while progesterone inhibited the proliferative response mediated by both angiogenic growth factors. The effect mediated by estradiol was significant at physiological levels (0.1 to 1 nM) and was as high as 75% over baseline. These results provide functional relevance to estradiol on the vascular repair of the endometrium during the proliferative phase of the menstrual cycle. In

contrast, progesterone was inhibitory on endothelial cell proliferation. This finding is consistent with the suppressive signals predominant during the secretory phase, when circulating levels of this hormone are highest and when the vascular repair has ceased. The positive effect of estradiol on endothelial cell proliferation has been reported for HUVEC and coronary endothelial cells (40,50), however, the effects were more modest than those observed on HEEC. We have also performed similar experiments on HDEC, HLEC, and HUVECs. In our hands, the stimulatory effect mediated by estradiol in those cultures was not always consistent, and when seen was modest (12–15%) (data not shown). The effect of progesterone, has not been previously reported on the endothelium. DISCUSSION

Given the growing evidence that endothelial cells isolated from different organs express unique prop-

Endometrial endothelial cells and angiogenesis M.L. Iruela-Arispe et al. 135

Figure 4. Endometrial endothelial cells upregulate ESelectin mRNA upon treatment with TNF-␣. Human endometrial endothelial cells were cultured in the absence of serum for 16 h. TNF-␣ at 50 ng/mL (+) or vehicle (−) were added to the cultures. Cells were harvested at 30 min, 2 h, and 6 h. Total RNA was isolated, subjected to electrophoretic separation, and transferred to Nytran membranes. Hybridization with an E-selectin cDNA fragments revealed marked increase in expression upon treatment with TNF␣. A cDNA fragment from 36B4 ribosomal protein was used as control for loading and transfer efficiency.

erties, it is reasonable to predict that different capillary beds might respond to tissue-specific cues that regulate local patterns of growth and inhibition. Therefore, an in-depth study of organ-specific vascular beds can be expected to provide information about particular pathways that might have broad applicability to the pharmacological regulation of angiogenesis in that organ, without necessarily affecting microvessels in other tissues. In the present study, we derived a simple protocol for the isolation of endothelial cells from the endometrium, demonstrated that these cells exhibit a significantly higher binding for VEGF, and provide evidence for a functional modulation of steroid hormones on growth factors’ function. The endothelial nature of the isolated cells has been characterized by several conventional markers and by functional assays. In general, HEEC, display features similar to endothelial cell isolated from other organs. However, they appear to have an enhanced responsiveness to the angiogenic growth factor, VEGF, and do not respond to PDGFBB. As other reports have indicated (2,38), we found that expression of von Willebrand factor is heterogenous and identifies endothelial cells from large- and medium-size vessels in the endometrium preferen-

Figure 5. Identification of progesterone and estrogen receptors on human endometrial endothelial cells. Endometrial tissue sections were incubated with an anti-progesterone receptor (A) or anti-estrogen receptor (B) antibody followed by biotinylated secondary antibodies. Immune complexes were revealed after treatment with diaminobenzidine as a brown color. Counterstaining was performed with a nuclear fast-red dye. Arrows in both panels indicate positive endothelial cells. Isolated HEEC cultures were incubated with anti-progesterone receptor (C) or estrogen receptor (D) antibodies, followed by secondary conjugated to FITC. To validate the immunocytochemical results, total RNA isolated from HEEC cultures was also subjected to reverse transcriptase for the generation of first-strand cDNAs followed by PCR with progesterone receptor (C, lanes 1–4) or estrogen receptor (D, lanes 5–8) specific primers. Lanes are PCR reactions from HEEC cultures at progressive passage number: 1 and 5: passage 2; 2 and 6: passage 5; 3 and 7: passage 7; 4 and 8: passage 10.

Endometrial endothelial cells and angiogenesis M.L. Iruela-Arispe et al. 136

Figure 6. Endometrial-derived endothelial cells express KDR and FLT-1. Total RNA was isolated from endometrial endothelial cells (HEEC), dermal endothelial cells (HDEC), dermal fibroblasts (HDF), and endometrial stromal (HES). All strains were used between passages 3 and 5. Resulting Northern blots were hybridized with cDNA probes to KDR; FLT-1; VPF/VEGF; and 28S ribosomal subunit for evaluation of loading and transfer efficiency. Lanes: 1 ⳱ HEEC; 2 ⳱ HDEC; 3 ⳱ HDF; 4 ⳱ HES.

tially. In contrast, Ulex lectin and CD-34 showed a more homogeneous and consistent distribution in endometrial capillaries. The endometrium is a unique organ that undergoes physiologic cyclic remodeling. During its growth and maturation, intense vascular growth takes place, supporting stromal and glandular expansion. The endometrial vascular bed, and endometrial endothelial cells in particular, have not been wellcharacterized. Because these cells present unique features of recurrent growth and inhibition, we rationalized that inherent differences must exist between endometrial endothelial cells and endothelium from other vascular beds. The nature of the angiogenic stimulus in the endometrium is also poorly defined. Although a number of growth factors known to be angiogenic in other organs have been identified in the human endometrium, the specific effect of these cytokines on endometrial endothelial cells has never been investigated. Among the factors identified are FGF-2, VEGF, insulin-like growth factor, epidermal growth factor, and transforming growth factor-␤ (29). Interestingly, the levels of such growth factors is not increased, as would be expected, early in the prolif-

Figure 7. Endometrial endothelial cells proliferate in response to VEGF and FGF-2. (A) Endothelial cell cultures (HEEC, HCEC, HUVEC, and HDEC) at identical passage number were made quiescent by serum deprivation for 48 h. Thereafter, 2 × 105 cells from each type were plated onto new dishes and treated with 1) 1% fetal calf serum (FCS); 2) 1% FCS + 20 ng/mL VPF/VEGF; 3) 1% FCS + 20 ng/mL FGF-2; 4) 1% FCS + PDGF-BB (20 ng/mL). After 4 days of incubation, cells were counted by coulter counter. Values were normalized to control (DMEM) ± SEM. n ⳱ 4 for all conditions. (B) Cell-cycle progression analysis of HEEC and HDEC. Cultures were made quiescent by 48 h, serum deprivation after confluency. Cells were passed into 12-well plates and harvested every 8 h for a total of 120 h. A pulse of [3H]-thymidine was 4 h prior to harvest time. Values indicate incorporation of TCA-precipitable counts ± SEM. n ⳱ 5 for all conditions.

erative phase when vascular repair and angiogenesis follow menstruation. Heparin-like activity has also been detected in endometrial fluids with increasing concentrations toward the end of the cycle in women, which could enhance the action of specific heparin-dependent angiogenic agents (29). Most of these cytokines also influence fibroblast proliferation, and in some cases regulate epithelial growth and differentiation (as transforming growth factor␤). Therefore, it is possible that the presence of some growth-stimulating factors might not be directly related to the endometrial angiogenic wave. The iso-

Endometrial endothelial cells and angiogenesis M.L. Iruela-Arispe et al. 137

Table 1. Relative proliferative increases of several endothelial cell types by angiogenic growth factors

VEGF FGF-2 PDGF BB

HEEC

HCEC

HUVEC

HDEC

72%* 30.2% 4%

28.6% 31.4% 5.7%

21.4% 39.3% 3.6%

51.2% 63.4% 44%

*Percentages were calculated from data presented in Fig. 7A.

lation of endometrial endothelial cells was a necessary step to evaluate the relative contribution of these growth factors to the cyclic vascular repair in this tissue. In addition, the endometrium is, unlike other organs, under the regulatory control of steroid hormones. The relative contribution of these hormones to the angiogenic response has been hypothesized, yet not investigated. Our results provide evidence that steroid hormones can modulate the action of growth factors on HEEC. Therefore, although levels of angiogenic growth factors are not drastically changed through the proliferative and secretory phases of the menstrual cycle, the contribution of estrogen and progesterone helps to undestand how, in combination, hormones and growth factors regulate vascular repair in the endometrium. Our study also showed that HEEC respond to both FGF-2 and VEGF, but not to PDGF-BB. A significantly higher growth response was seen in HEEC in response to VEGF. The finding led us to compare

Figure 9. Estrogen and progesterone modulate the proliferative responses of VEGF and FGF-2 in HEEC. Quiescent HEEC cells from passage 3 to 5 were seeded on 48-well plates and treated with 2 ng/mL FGF-2 and 25 ng/mL VEGF in the presence of steroids (A: estradiol and B: progesterone) or vehicle, as indicated. Incubation with growth factors and hormones was performed for 48 h and labeling with [3H]-thymidine (1␮Ci/mL) was performed during the last 12 h of incubation. Values indicate incorporation of TCA-precipitable counts ± SEM. n ⳱ 6 for all conditions.

affinity binding of this growth factor for HEEC and other endothelial cell types. As reported, impressive differences were found. Clearly, HEEC had a higher binding capacity for VEGF than any of the other endothelial cell types examined, providing an explanation for the marked proliferative response to this growth factor. Figure 8. VEGF-binding profile to several endothelial cell types. Endothelial cell (HEEC, HCEC, HUVEC, and HDEC) and stromal fibroblasts (HSF) were plated on 48well plates at equal number. Highest binding of 1 ng/mL [125I]-VEGF 165 to endothelial cultures is referred to as 100%. VEGF binding was competed by increasing amounts of unlabeled VEGF.

VEGF binds to two receptor tyrosine kinases, FLT-1 and KDR (17,49,56). Upon binding and formation of receptor dimers, activation is characterized by receptor phosphorylation followed by initiation of the cell-cycle cascade. Our observation of higher binding levels for VEGF in HEEC suggests that these cells

Endometrial endothelial cells and angiogenesis M.L. Iruela-Arispe et al. 138

have higher levels of receptors and are more responsive to growth stimulation than other cell types. Our data showed that in response to VEGF, HEEC enter S phase more quickly than HDEC, HUVEC, and HCEC suggesting a dramatically enhanced response to this growth factor. Most of our current knowledge of the endometrial vasculature is based on morphological studies (21,29,54). The structure of the primate endometrial vasculature is unique when compared to other mammals. At the myoendometrial junction, radial arteries branch into 1) basal arteries, which supply the basal endometrium, and 2) spiral arteries, which supply the functional endometrium (39). The basal arteries maintain the integrity of the basal layer through all phases of the menstrual cycle and appear to be unaffected by hormonal stimuli. In contrast, the spiral arteries appear responsive to steroids and undergo alterations in morphology and length that are well-correlated with fluctuations in circulating hormones (39). Falling levels of progesterone and estrogen promote contraction of the spiral arteries and result in episodes of intermittent hypoxia in the functional endometrium. The end result is necrosis and exfoliation of the functional endometrium with loss of blood. Following the menstrual phase, the basal microvessels give rise to a new capillary plexus. The cyclic nature of endothelial growth follows endometrial physiology and is thought to be regulated, at least indirectly by steroid hormones (21,54,57). Here, we demonstrate the presence of both estrogen and progesterone receptors in HEEC, indicating that these hormones play a direct role in the regulation of endothelial cell function in the endometrium. Several reports in the literature have indicated that 17␤-estradiol stimulates both migration and proliferation of human umbilical vein and human coronary endothelial cells (40,59). A supporting role for estradiol in angiogenesis was also demonstrated using in vitro assays (50). It is likely that this hormone displays similar functions on HEEC. The prompt vascular-growth inhibition associated with the secretory phase of the endometrial cycle is also unique to this organ. During the vascular growth, the rate of proliferation of endothelial cells has been compared to that of tumor cells (54); however, in contrast to tumor cells, vessel growth ceases after the reconstitution of the functional layer. Unlike any other organ examined, we found that expression of thrombospondin-1 (TSP-1) is temporally regulated in the endometrium, predominantly expressed during the secretory phase, a time when capillary growth is suppressed (35). Equally, a novel

gene named METH-2, with significant homology to TSP-1 and angio-inhibitory properties (Vazquez and Iruela-Arispe, submitted), is also expressed in a narrow window during the early secretory phase (days 17–21) (Iruela-Arispe, unpublished observations). Because of the potential importance of vascular regulation to control of several endometrial disorders, the identification of specific regulators of vascular growth and inhibition is of particular interest in this tissue. An increasing body of literature provides examples of biochemical and functional heterogeneity in endothelial surfaces. Data are now available that characterize differentiated microdomains in tissue vasculature (18,26,27,47,48,53,61). Given the properties of human endometrial endothelial cells, the expression of unique genes in this microvasculature might seem intrinsically obvious, and indicates a high probability that a search for such genes will provide significant insights into the function of this tissue. At present, the literature offers no data on this issue. The identification of endometrial-specific genes, such as METH-2, that differentiate endometrial endothelial cells from endothelial cells of other vascular beds can provide important basic information of the understanding of the pathologies uniquely observed in endometrial capillaries. In addition, research in this area can provide important tools for pharmacological targeting and for potential treatment of endometrial pathologies. The availability of human endometrial endothelial cells will make study of these, and other issues, more experimentally accessible. ACKNOWLEDGMENTS

The authors thank Dr. Michael Detmar, for human dermal endothelial cells; and Dr. Don Senger, for the human umbilical vein endothelial cells; Ms. Sarah Oikemus, for her excellent technical assistance; and Ms. Carol Foss, for her secretarial assistance. REFERENCES 1. Aird WC, Edelberg JM, Weiler-Guettler H, Simmons WW, Smith TW, Rosenberg RD. (1997). Vascular bed-specific expression of an endothelial cell gene is programmed by the tissue microenvironment. J Cell Biol 138:1117–1124. 2. Au CL, Rogers PA. Immunohistochemical staining of von Willebrand factor in human endometrium during normal menstrual cycle. Hum Reprod 8:17–23. 3. Auerbach R, Alby L, Morrissey LW, Tu M, Joseph J. (1985). Expression of organ-specific antigens on capillary endothelial cells. Microvasc Res 29:401–411. 4. Bar RS, Boes M, Booth BA, Dake BL, Henley S, Hart

Endometrial endothelial cells and angiogenesis M.L. Iruela-Arispe et al. 139

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

MN. (1989). The effects of platelet-derived growth factor in cultured microvessel endothelial cells. Endocrinology 124:1841–1848. Battegay EJ, Rupp J, Iruela-Arispe ML, Sage EH, Pech M. (1994). PDGF-BB modulates endothelial proliferation and angiogenesis in vitro via PDGF-␤receptors. J Cell Biol 125:917–928. Beitz JG, Kim IS, Calabresi P, Frackelton AR. (1991). Human microvascular endothelial cells express receptors for platelet-derived growth factor. Proc Natl Acad Sci USA 88:2021–2025. Belloni PN, Nicolson GL. (1988). Differential expression of cell surface glycoproteins on various organderived microvascular endothelia and endothelial cell cultures. J Cell Physiol 136:398–410. Belloni PN, Carney DH, Nicolson GL. (1992). Organderived microvessel endothelial cells exhibit differential responsiveness to thrombin and other growth factors. Microvasc Res 43:20–45. Bischoff J, Brasel C, Kraling B, Vranovska V. (1997). E-selectin is upregulated in proliferating endothelial cells in vitro. Microcirculation 4:279–287. Bowman PD, Betz AL, Goldstein GW. (1982). Primary culture of microvascular endothelial cells from bovine retina: selective growth using fibronectincoated substrate and plasma-derived serum. In vitro 18:626–632. Brown LF, Detmar M, Tognazzi K, Abu-Jawdeh G, Iruela-Arispe ML. (1997). Uterine smooth muscle cells express functional receptors (flt-1 and KDR) for vascular permeability factor/vascular endothelial growth factor. Lab Invest 76:245–255. Carley WW, Niedbala MJ, Gerritsen ME. (1992). Isolation, cultivation, and partial characterization of microvascular endothelium derived from human lung. Am J Respir Cell Mol Biol 7:620–630. Carson MP, Haudenschild CC. (1986). Microvascular endothelium and pericytes: high yield, low passage cultures. In Vitro Cell Dev Biol 22:344–354. Chomcznski P, Sacchi N. (1987). Single-step method of RNA isolation by acid guanidinium thiocyanatephenol-choloroform extraction. Anal Biochem 162: 156–159. Chung-Welch N, Patton WF, Yen-Patton GPA, Hechtman HB, Shepro D. (1989). Phenotypic comparison between mesothelial and microvascular endothelial cell lineages using conventional endothelial cell markers, cytoskeletal protein markers and in vitro assays of angiogenic potential. Differentiation 42:44–53. Davison PM, Bensch K, Karasek MA. (1983). Isolation and long-term serial cultivation of endothelial cells from the microvessels of the adult human dermis. In Vitro Cell Dev Biol 19:937–945. de Vries C, Escobedo JA, Ueno H, Houck K, Ferrara N, Williams LT. (1992). The fms-like tyrosine kinase, a receptor for vascular endothelial growth factor. Science 255:989–991. Defilippi P, van Hinsbergh V, Bertolotto A, Rossino P,

19.

20.

21. 22.

23. 24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

Silengo L, Tarone G. (1991). Differential distribution and modulation of expression of alpha1/beta1 integrin on human endothelial cells. J Cell Biol 114:855– 863. Dorovini-Zis K, Prameya R, Bowman PD. (1991). Culture and characterization of microvessel endothelial cells derived from human brain. Lab Invest 64: 425–436. Drake BL, Loke YW. (1991). Isolation of endothelial cells from human first trimester decidua using immunomagnetic beads. Hum Reprod (Oxf) 6:1156–1159. Findlay JK. (1986). Angiogenesis in reproductive tissues. J Endocrinol 111:357–366. Folkman J, Haudenschild CC, Zetter BR. (1979). Long-term culture of capillary endothelial cells. Proc Natl Acad Sci USA 76:5217–5221. Folkman J, Klagsbrun M. (1987). Angiogenic factors. Science 235:4442–447. Fukuda K, Imamura Y, Y. K, Ooyama T, Hanamure Y, Ohyama M. (1989). Establishment of human mucosal microvascular endothelial cells from inferior turbinate in culture. Am J Otolaryngol 10:85–91. Gallery EDM, Rowe J, Schrieber L, Jackson CJ. (1991). Isolation and purification of microvascular endothelium from human decidual tissue in the late phase of pregnancy. Am J Obstet Gynecol 165:191– 196. Gerritsen ME. (1987). Functional heterogeneity of vascular endothelial cells. Biochem Pharmacol 36: 2701–2711. Gerritsen ME, Niedbala MJ, Szczepanski A, Carley WW. (1993). Cytokine activation of human macroand microvessel-derived endothelial cells. Blood Cells 19:325–342. Gerritsen ME, Shen CP, McHugh MC, Atkinson WJ, Kiely JM, Milstone DS, Luscinskas FW, Gimbrone MA. (1995). Activation-dependent isolation and culture of murine pulmonary microvascular endothelium. Microcirc 2:151–163. Goodger-MacPherson AM, Rogers PAW. (1995). Blood vessel growth in the endometrium. Microcircularion 4:329–343. Hagemeier HH, Vollmer E, Goerdt S, SchulzeOsthoff K, Sorg C. (1986) A monoclonal antibody reacting with endothelial cells budding vessels in tumors and inflammatory tissues, and non-reactive with normal adult tissues. Int J Cancer 38:481–488. Hewett PW, Murray JC. (1996). Isolation of microvascular endothelial cells using magnetic beads coated with anti-PECAM-1 antibodies. In Vitro Cell Dev Biol Anim 31:473–481. Hewett PW, Murray JC, Price EA, Watts ME, Woodcock M. (1993). Isolation and characterization of microvessel endothelial cells for human mammary adipose tissue. In Vitro Cell Dev Biol 29A:325–331. Hunter WMK, Greenwood FC. (1964). A radioimmunoelectrophoretic assay for human growth hormone. Biochem J 91:43–56.

Endometrial endothelial cells and angiogenesis M.L. Iruela-Arispe et al. 140

34. Imicke E, Ruszczak ZB, Mayer-da Silva A, Detmar M, Orfanos CE. (1991). Cultivation of human dermal microvascular endothelial cells in vitro: Immunocytochemical and ultrastructural characterization and effect of treatment with synthetic retinoids. Arch Dermatol Res 283:149–157. 35. Iruela-Arispe ML, Porter P, Bornstein P, Sage H. (1996). Thrombospondin-1, an inhibitor of angiogenesis, is regulated by progesterone in the human endometrium. J Clin Invest 97:403–412. 36. Iruela-Arispe ML, Sage EH. (1993). Endothelial cells exhibiting angiogenesis in vitro proliferate in response to TGF-␤1. J Cell Biochem 52:414–430. 37. Jackson CJ, Garbett PK, Nissen B, Schrieber L. (1990). Binding of human endothelium to Ulex europaeus-1 coated dyna-beads: Application to the isolation of microvascular endothelium. J Clin Sci 96: 257–262. 38. Johannisson E. (1986). Effects of oestradiol and progesterone on the synthesis of DNA and antihaemophilic Factor VIII antigen in human endometrial endothelial cells. Hum Reprod 1:207–212. 39. Kaiserman-AbramofI, Padykula HA. (1989). Angiogenesis in the postovulatory primate endometrium: the coiled arteriolar system. Anat Record 224:479– 489. 40. Kim-Schulze S, McGowan KA, Hubchak SC, Cid MC, Martin MB, Kleinman HK, Greene GL, Schnaper HW. (1996). Expression of an estrogen receptor by human coronary artery and umbilical vein endothelial cells. Circulation 94:1402–1407. 41. Kumar S, West DC, Ager A. (1987). Heterogeneity in endothelial cells from large vessels and microvessels. Differentiation 36:57–70. 42. Kuzu I, Bicknell R, Harris AL, Johnes M, Gatter KC, Mason DY. (1992). Heterogeneity of vascular endothelial cells with relevance to diagnosis of vascular tumors. J Clin Pathol 45:143–148. 43. Lee WS, Harder JA, Yoshizumi M, Lee ME, Harber E. (1997). Progesterone inhibits arterial smooth muscle cell proliferation. Nature Med 3:10055–1008. 44. Lodge PA, Haisch CE, Huber SA, Huber SA, Martin B, Craighead JC. (1991). Biological differences in endothelial cells depending upon organ differentiation. Trans Proceedings 23:216–218. 45. Magee JC, Stone AE, Oldham KT, Guice KS. (1994). Isolation, culture, and characterization of rat lung microvascular endothelial cells. Am J Physiol 267:L433– L441. 46. Marks RM, Czerniecki M, Penny R. (1985). Human dermal microvascular endothelial cells: An improved method for tissue culture and a description of some singular properties in culture. In Vitro Cell Dev Biol 21:627–635. 47. Mason JC, Yarwood H, Sugars K, Haskard DO. (1997). Human umbilical vein and dermal microvascular endothelial cells show heterogeneity in response to PKC activation. Am J Physiol 273:C1233–C1240.

48. McCarthy SA, Kuzu I, Gatter KC, Bicknell R. (1991). Heterogeneity of the endothelial cell and its role in organ preference of tumour metastasis. TiPS 121: 462–467. 49. Millauer B, Wizigmann-Voos S, Schnu¨rch H, Martinez R, Moller NPH, Risau W, Ullrich A. (1993). High affinity VEGF binding and developmental expression suggest flk-1 as a major regulator of vasculogenesis and angiogenesis. Cell 72:835–846. 50. Morales DE, McGowan KA, Grant DS, Maheshwari S, Bhartiya D, Cid MC, Kleinman HK, Schnaper HW. (1995). Estrogen promotes angiogenic activity in human umbilical vein endothelial cells in vitro and in a murine model. Circulation 91:755–763. 51. Newman PJ. (1997). The biology of PECAM-1. J Clin Invest 100:S25–S29. 52. Nishida M, Carley WW, Gerritsen ME, Ellingsen O, Kelly, Smith TW. (1993). Isolation and characterization of human and rat cardiac microvascular endothelial cells. Am J Physiol 264:H639–H652. 53. Petzelbauer P, Bender JR, Wilson J, Pober JS. (1993). Heterogeneity of dermal microvascular endothelial cell antigen expression and cytokine responsiveness in situ and in cell culture. J Immunol 151:5062–5072. 54. Reynolds LP, Killilea SD, Redmar DA. (1992). Angiogenesis in the female reproductive system. FASEB J 6:886–892. 55. Richard L, Velasco P, Detmar M. (1998). A simple immunomagnetic protocol for the selective isolation and long-term culture of human dermal microvascular endothelial cells. Exp Cell Res 240:1–6. 56. Shibuya M, Yamaguchi S, Yamane A, Ikeda T, Tojo A, Matsushime H, Sato M. (1990). Nucleotide sequence and expression of a novel human receptortype tyrosine kinase gene (flt) closely related to the fms family. Oncogene 5:519–524. 57. Shweike D, Itin A, Neufelt G, Gitary-Goren H, Keshet E. (1993). Patterns of expression of vascular endothelial growth factor (VEGF) and VEGF receptors in mice suggest a role in hormonally regulated angiogenesis. J Clin Invest 91:2235–2243. 58. Smits A, Hermansson M, Nister M, et al. (1989). Rat brain capillary endothelial cells express functional PDGF-␤-type receptors. Growth Factors 2: 1–8. 59. Spyridopoulos I, Sullivan AB, Kearney M, Isner JM, Losordo DW. Estrogen-receptor-mediated inhibition of human endothelial cell apoptosis. Estradiol as a survival factor. Circulation 95:1505–1514. 60. Stoltz D, Jacobson B. (1991). Macro- and microvascular endothelial cells in vitro: maintenance of biochemical heterogeneity despite loss of ultrastructural characteristics. In Vitro Cell Dev Biol 27A:169–182. 61. Zetter BR. (1988). Endothelial heterogeneity: influence of vessel size, organ location, and species specificity. In: Biology of vascular enthothelial (U Ryan, Ed.) CRC Press. Boca Raton, FL: 63–80.