Differentiation of endothelial barrier function during normal angiogenesis requires homotypic VE-cadherin adhesion

Endothelial cells express two principal cadherins: VE-cadherin and N-cadherin. We established previously that only VE-cadherin expression was increased during differentiation of barrier function by angiogenic endothelium of the chick chorioallantoic membrane (CAM). Presently anti-VE-cadherin mAb, applied to the CAM at day 4.5 of gestation, served to inhibit the abrupt reduction of macromolecular extravasation that occurs normally at day 5.0. Neither anti-N-cadherin nor nonimmune IgG, on the other hand, prevented this temporal decrease of endothelial permeability. Despite the differential permeability responses, morphometric evaluations defined a reduction of mean paracellular cleft width after the application of either anti-VE-cadherin or anti-N-cadherin. Hence, alteration of molecular sieving characteristics within the junctional clefts, rather than modification of cleft dimensions; likely served as the principal modulator of macromolecular extravasation after inhibition of homotypic VE-cadherin adhesion. These results provide support to the concept that VE-cadherin contributes to the normal differentiation of endothelial barrier function during CAM angiogenesis in vivo.     

Protein kinase C activity contributes to endothelial hyperpermeability during early angiogenesis in the chick chorioallantoic membrane

During angiogenesis in the chick chorioallantoic membrane (CAM), microvascular proliferation continues through day 12 of the 18-day CAM lifespan. Up to day 4.5, the neovascularization is associated with endothelial hyperpermeability and differentiation of restrictive barrier function occurs abruptly at day 5.0. Although exogenous activation of cAMP/protein kinase A (PKA) signaling served to decrease macromolecular extravasation at day 4.5, endogenous signaling cascades responsible for the temporal hyperpermeability remain uncertain. Here, we evaluated protein kinase C (PKC) function in the CAM endothelium at day 4.5 and day 5.0. The specific, broad-based PKC inhibitor calphostin C reduced basal levels of FITC-dextran 40 extravasation at day 4.5. Bisindolymaleimide (BIM), which inhibits selective PKC isoforms, also reduced temporal FITC-dextran 40 efflux, but to a lesser extent than calphostin C. Activation of PKC activity by phorbol-12, 13-didecanoate (PDD) or phorbol-12, 13-dibutyrate (PDBu) at day 5.0 served to partially de-differentiate barrier properties of the angiogenic endothelium. The associated elevation of FITC-dextran 40 extravasation occurred without interendothelial gap formation along the junctional clefts. Together, these results are consistent with the interpretation that PKC activity contributes, in part, to CAM endothelial hyperpermeability at day 4.5. Furthermore, down-regulation of PKC signaling correlates temporally with the ontogeny of restrictive barrier function at day 5.0.     

Endocardial endothelium in the rat: junctional organization and permeability

Selective permeability of endocardial endothelium has been suggested as a mechanism underlying the modulation of the performance of subjacent myocardium. In this study, we characterized the organization and permeability of junctional complexes in ventricular endocardial endothelium in rat heart. The length of intercellular clefts viewed en face per unit endothelial cell surface area was lower, and intercellular clefts were deeper in endocardial endothelium than in myocardial vascular endothelium, whereas tight junctions had a similar structure in both endothelia. On this basis, endocardia endothelium. might be less permeable than capillary endothelium. However, confocal scanning laser microscopy showed that intravenously injected dextran 10000 coupled to Lucifer Yellow penetrated first the endocardial endothelium and later the myocardial capillary endothelium. Penetration of dextran 10000 in myocardium occurred earlier through subepicardial capillary endothelium than through subendocardial capillary endothelium. Penetration of tracer might thus be influenced by hydrostatic pressure. Dextran of MW 40000 did not diffuse through either endocardial endothelium or capilary endothelium. The ultrastructure of endocardial endothelium may constitute an adaptation to limit diffusion driven by high hydrostatic pressure in the heart. Differences in paracellular diffusion of dextran 10000 between endocardial endothelium and myocardial vessels, may result from differing permeability properties of the endocardium and underlying myocardium.     

Transforming growth factor beta 1-induced changes in cell migration, proliferation, and angiogenesis in the chicken chorioallantoic membrane

Application of TGF beta 1 (10-100 ng) to the chicken chorioallantoic membrane (CAM) for 72 h resulted in a dose-dependent, gross angiogenic response. The vascular effects induced by TGF beta 1 were qualitatively different than those induced by maximal doses of basic FGF (bFGF) (500 ng). While TGF beta 1 induced the formation of large blood vessels by 72 h, bFGF induced primarily small blood vessels. Histologic analysis revealed that TGF beta 1 stimulated pleiotropic cellular responses in the CAM. Increases in fibroblast and epithelial cell density in the area of TGF beta 1 delivery were observed as early as 4 h after TGF beta 1 treatment. By 8 h, these cell types also demonstrated altered morphology and marked inhibition of proliferation as evidenced by 3H-thymidine labeling. Thus, the TGF beta 1-stimulated accumulation of these cell types was the result of cellular chemotaxis from peripheral areas into the area of TGF beta 1 delivery. Microscopic angiogenesis in the form of capillary sprouts and increased endothelial cell density first became evident at 16 h. By 24 h, capillary cords appeared within the mesenchyme of the CAM, extending towards the point of TGF beta 1 delivery. 3H-thymidine labeling revealed that the growth of these capillary cords was due to endothelial cell proliferation. Finally, perivascular mononuclear inflammation did not become evident until 48 h of treatment, and its presence correlated spatially and temporally with the gross and histological remodelling of newly formed capillary cords into larger blood vessels. In summary, these data suggest that, in the chicken CAM, TGF beta 1 initiates a sequence of cellular responses that results in growth inhibition, cellular accumulation through migration, and microvascular angiogenesis.     

The three-dimensional organization of tight junctions in a capillary endothelium revealed by serial-section electron microscopy

Estimates of capillary permeability for hydrophilic solutes are generally interpreted in terms of Pappenheimer's pore theory. The intercellular clefts of the capillary endothelium are considered a likely structural equivalent to the postulated system of small hydrophilic pores. However, correlation of permeabilities and cleft structure requires more knowledge of the detailed structure of the tight junctions which appear to obliterate the clefts. In this study the organization of tight junctions in endothelium of rat heart capillaries has been investigated by serial-section electron microscopy. Cross-sectioned intercellular clefts were photographed in a series of 190 consecutive sections (average thickness approximately equal to 40 nm) and in a series of 16 consecutive sections (average thickness approximately equal to 12.5 nm). Seventy-one junctional segments, each extending over 5-32 consecutive sections, were reconstructed. The endothelial junctions were organized as irregular networks of lines of contact between neighboring cells. Six pathways circumventing the lines of contact were followed through the entire junctional region of the clefts providing a tortuous pathway connecting the luminal and abluminal aspects of the clefts. Moreover, the individual lines of contact were provided with discrete discontinuities, apparently 4 nm wide. The observations support the notion that the paracellular pathway in capillary endothelium is permeable not only to small solutes but also to certain macromolecules.     

A microcarrier-based cocultivation system for the investigation of factors and cells involved in angiogenesis in three-dimensional fibrin matrices in vitro

Angiogenesis in situ includes coordinated interactions of various microvascular cell types, i.e., endothelial cells, pericytes and perivascular fibroblasts. To study the cellular interactions of microvascular cells in vitro, we have developed a microcarrier-based cocultivation system. The technical details of this method include seeding of endothelial cells on unstained cytodex-3 microcarriers and seeding of pericytes, fibroblasts or vascular smooth muscle cells on microcarriers which have been labeled by trypan blue staining. A mixture of both unstained and trypan blue-stained microcarriers was subsequently embedded in a three-dimensional fibrin clot. The growth characteristics of each cell type could be conveniently observed since the majority of cells left their supporting microcarriers in a horizontal direction to migrate into the transparent fibrin matrix. As differently stained microcarriers were randomly arranged in the fibrin matrix, the characteristic patterns of the microcarriers allowed location of particular points of interest at different developmental stages, facilitating the observation of cellular growth over the course of time. One further advantage of this microcarrier-based system is the possibility of reliably quantifying capillary growth by determination of average numbers of capillary-like formations per microcarrier. Thus, this model allows convenient evaluation of the effects of non-endothelial cells on angiogenesis in vitro. By using this coculture system, we demonstrate that endothelial capillary-like structures in vitro do not become stabilized by contacting vascular smooth muscle cells or pericytes during the initial stages of capillary formation.     

8-prenylnaringenin, a novel phytoestrogen, inhibits angiogenesis in vitro and in vivo

8-Prenylnaringenin is a recently discovered phytoestrogen. Using an in vitro model of angiogenesis in which endothelial cells can be induced to invade a three-dimensional collagen gel within which they form capillary-like tubes, we demonstrate that 8-prenylnaringenin inhibits angiogenesis induced by basic fibroblast growth factor (bFGF), vascular endothelial growth factor (VEGF), or the synergistic effect of the two cytokines in combination, with an IC(50) of between 3 and 10 microM. This effect was seen with bovine microvascular endothelial cells derived from the adrenal cortex (BME cells) and with endothelial cells from the bovine thoracic aorta (BAE cells). The inhibitory effects of 8-prenylnaringenin were found to be roughly equipotent to those of genistein that has previously been shown to inhibit angiogenesis in vitro. Early chorioallantoic membrane (CAM) assay results showed reductions in both vessel lengths and vein diameters, with similar potency in the 8-prenylnaringenin and genistein groups. Similar effects on the CAM vessels were seen when the two substances were co-added. These findings suggest that 8-prenylnaringenin has potential therapeutic applications for diseases in which angiogenesis is an important component.     

Quantitative immunocytochemical studies of endogenous albumin in rat aortic endothelial and mesothelial cells

Endogenous albumin was revealed over cellular structures of rat ascendent aorta endothelia and mesothelium, with high resolution and specificity, by applying the protein A-gold immunocytochemical approach. This approach allows albumin distribution to be studied under steady-state conditions. The cellular layers evaluated were the aortic endothelium, the capillary endothelium (vasa vasorum), and the mesothelium externally lining the aorta at this level. Gold particles, revealing albumin antigenic sites, were preferentially located over plasmalemmal vesicles and intercellular clefts of endothelial and mesothelial cells, though with different labeling intensities. The interstitial space was also labeled. Morphometrical evaluation of plasmalemmal vesicles demonstrated a higher surface density for these structures in capillary endothelial cells (12%) compared with those in aortic endothelial (5%) and mesothelial cells (2%). Quantitation of gold labeling intensities over these structures revealed a higher labeling over plasmalemmal vesicles of capillary endothelium than over those of aortic endothelium and mesothelium. This result, together with the higher surface density of plasmalemmal vesicles found in capillary endothelium, suggest an important role of these structures in the transendothelial passage of endogenous albumin, particularly for capillary endothelium. On the other hand, labeling densities over mesothelial clefts were found to be higher than those of capillary and aortic endothelia. Results from this study concur with the proposal of a differential passage of albumin according to the cell lining considered, and suggest to a role for mesothelial intercellular clefts in contributing to the presence of albumin in interstitial spaces.     

Immunogold study of interendothelial junction-associated and glucose transporter proteins during postnatal maturation of the mouse blood-brain barrier

The distribution of glucose transporter (GLUT-1) and of interendothelial junction—associated proteins—zonula occludens protein (ZO-1), occludin, and β-catenin—was studied using quantitative immunogold procedure. Lowicryl K4M-embedded samples of the cerebral cortex of 1-, 7-, and 14-day-, and 6-week-old (young-adult) mice were used. Ultrathin sections were exposed to specific rabbit polyclonal antibodies followed by colloidal gold-labelled secondary antibodies. We found that the density of immunosignals for GLUT-1 in both luminal and abluminal plasma membranes of the endothelial cells, and those closely related to the interendothelial junctions was low in blood microvessels from newborn mice, dropped slightly at the 7th day, and increased through the 14th day to the level of mature blood-brain barrier (BBB) observed in 6-week-old mice. The expression of ZO-1 was high in newborn mice and increased at the 7th day to the level similar to that found in 14-day- and 6-week-old mice. The expression of occludin was less intense than that of ZO-1 and increased from birth, reaching at the 14th day the level typical for mature BBB found in young-adult animals. The immunosignals for occludin were sparsely distributed inside the junctional clefts. Such a distribution indicates that the tight junctional characteristics are limited to a few short segments of the entire interendothelial cleft. The density of immunosignals for β-catenin was lowest, and it had the tendency to a gradual, although inconsiderable, drop in the time course of BBB maturation. These findings suggest that the relatively high concentration of GLUT-1 in the interendothelial junctions results from the participation of abluminal plasma membranes of adjacent endothelial cells in the formation of the junctional complexes. The interendothelial junctions of newborn mice are equipped already with the main components of the tight junctions, and the concentration of these components (ZO-1, occludin) reaches the level of the mature BBB at the 14th day of postnatal life.     

Cell proliferation along vascular islands during microvascular network growth

ABSTRACT: BACKGROUND: Observations in our laboratory provide evidence of vascular islands, defined as disconnected endothelial cell segments, in the adult microcirculation. The objective of this study was to determine if vascular islands are involved in angiogenesis during microvascular network growth. RESULTS: Mesenteric tissues, which allow visualization of entire microvascular networks at a single cell level, were harvested from unstimulated adult male Wistar rats and Wistar rats 3 and 10 days post angiogenesis stimulation by mast cell degranulation with compound 48/80. Tissues were immunolabeled for PECAM and BRDU. Identification of vessel lumens via injection of FITC-dextran confirmed that endothelial cell segments were disconnected from nearby patent networks. Stimulated networks displayed increases in vascular area, length density, and capillary sprouting. On day 3, the percentage of islands with at least one BRDU-positive cell increased compared to the unstimulated level and was equal to the percentage of capillary sprouts with at least one BRDU-positive cell. At day 10, the number of vascular islands per vascular area dramatically decreased compared to unstimulated and day 3 levels. CONCLUSIONS: These results show that vascular islands have the ability to proliferate and suggest that they are able to incorporate into the microcirculation during the initial stages of microvascular network growth.     

The functional and structural border of the neurohemal region of the median eminence

Summary In stressed rats the tanycytes of the ventrolateral wall of the third ventricle exhibit by light microscopic immunohistochemistry a positive staining for neurohormones which is distinctly limited to the distal perivascular end of the tanycyte process. Since by electron microscopic immuncytochemistry the tanycyte cytoplasm does not show any reaction product, the light microscopic reaction most likely results from a labeling of the intercellular space in the direct vicinity of the subendothelial cleft. Whether this subendothelial space is permeable to neurohormones was tested by injection of HRP¹. In the region of the arcuate nucleus 30 min after intravenous application, the marker is affixed to the membranes of the perivascular tanycyte processes in the subendothelial cleft of capillaries possessing non-fenestrated endothelia. Occasionally, HRP penetrates for a short distance between the tanycytes. Then the labeling of the intercellular cleft ends abruptly. Here, several parallel ridges of tight junctions between the perivascular distal tanycyte processes are found by the freezeetching technique. Since HRP cannot reach the subendothelial clefts of this region by passing through capillary walls due to the presence of a blood-brain barrier, it is suggested that the marker penetrates from the median eminence this far via the subendothelial extracellular space. It is prevented from spreading further by the tight junctions of the perivascular tanycyte endings. The same way may be taken by the neurohormones. Hence, a border area exists adjacent to the dorsolateral aspect of the neurohemal region of the median eminence where the tanycytes isolate the neuropil from the cerebrospinal fluid not only by their apical tight junctions, but also by basal tight junctions from the subendothelial cleft. This communicates with the perivascular space of the portal vessels.Supported by the Deutsche Forschungsgemeinschaft (Grant Nr. Kr. 569/2) and Stiftung VolkswagenwerkDedicated to Professor Dr. R. Ortmann on the occasion of his 65th birthday.The skilful technical assistance of Miss K. Bielenberg, Mrs. A. Hinz and Mrs. H. Prien is thankfully acknowledged     

The ultrastructural basis of the permeability of arterial endothelium to horseradish peroxidase

The thoracic aorta and basilar artery, in which the incidence of atherosclerosis is known to be different, were examined to elucidate the correlation between the structure of the intercellular cleft junction between adjacent endothelial cells and its permeability to HRP. Tannic acid or HRP in the vessel lumen passed through the intercellular clefts of the thoracic aorta into the subendothelial space, whereas in the basilar artery they were unable to penetrate beyond the tight junction of the intercellular clefts. Freeze-fracture replicas revealed that the tight junctions of the thoracic aorta consisted of one to two junctional strands in most areas of the cleaved planes, with discontinuities in some places, whereas those of the basilar artery consisted of a continuous belt-like meshwork of six anastomosing junctional strands on average. These observations confirm that the structure of endothelial junctions in arteries has a close correlation with the permeability of the intercellular clefts to HRP.     

The three-dimensional organization of smooth endoplasmic reticulum in capillary endothelia: its possible role in regulation of free cytosolic calcium

A rise in cytosolic free Ca in capillary endothelia leads to increased permeability. It has been proposed that this Ca(2+)-regulated modulation of junctional permeability of vascular endothelia involves structural elements comparable to those involved in stimulus-contraction coupling in smooth muscle. To explore this analogy the three-dimensional organization of smooth-surfaced cisternae, vesicular membrane profiles, and tight junctions was examined in endothelia of diaphragm and heart capillaries of the rat. Three-dimensional reconstructions, based on consecutive sections of the capillaries, have demonstrated a population of small, irregular membrane profiles, occurring in individual thin sections of the endothelial cytoplasm. These profiles represent an elaborate system of smooth-surfaced cisternae, structurally similar to the sarcoplasmic reticulum (SR) of smooth muscle cells. Slender processes from the cisternae are often situated in parallel to the tight junctions at a distance of about 100 nm. The great majority of the characteristic circular membrane profiles represents caveolae and racemose invaginations of the endothelial plasma membrane, often in close relation to the cisternae. It is hypothesized that the endothelial cisternae and invaginations of the cell membrane are involved in regulation of free cytosolic calcium in the same way as the SR and caveolae in smooth muscle cells. The junction-related cisternal processes may play a role in the Ca(2+)-regulated modulation of junctional permeability.     

An Ex Vivo Model for Anti-Angiogenic Drug Testing on Intact Microvascular Networks

New models of angiogenesis that mimic the complexity of real microvascular networks are needed. Recently, our laboratory demonstrated that cultured rat mesentery tissues contain viable microvascular networks and could be used to probe pericyte-endothelial cell interactions. The objective of this study was to demonstrate the efficacy of the rat mesentery culture model for anti-angiogenic drug testing by time-lapse quantification of network growth. Mesenteric windows were harvested from adult rats, secured in place with an insert, and cultured for 3 days according to 3 experimental groups: 1) 10% serum (angiogenesis control), 2) 10% serum + sunitinib (SU11248), and 3) 10% serum + bevacizumab. Labeling with FITC conjugated BSI-lectin on Day 0 and 3 identified endothelial cells along blood and lymphatic microvascular networks. Comparison between day 0 (before) and 3 (after) in networks stimulated by 10% serum demonstrated a dramatic increase in vascular density and capillary sprouting. Growing networks contained proliferating endothelial cells and NG2+ vascular pericytes. Media supplementation with sunitinib (SU11248) or bevacizumab both inhibited the network angiogenic responses. The comparison of the same networks before and after treatment enabled the identification of tissue specific responses. Our results establish, for the first time, the ability to evaluate an anti-angiogenic drug based on time-lapse imaging on an intact microvascular network in an ex vivo scenario.     

Characterization and differential expression of an endothelial cell-specific surface antigen in continuous and sinusoidal endothelial, in skin vascular lesions and in vitro

Continuous and sinusoidal endothelial cells display marked morphological and functional heterogeneity as to their plasmalemmal vesicle content, to the kind of intercellular junctional complexes, to the existence and kind of fenestrae and gaps, to the existence and character of their basement membrane, to their ability for phagocytosis and to other functional parameters. Monoclonal antibody 1F10, raised against human umbilical vein endothelial cells (HUVE cells), reflects these differences in recognizing--without any nonendothelial side reactions--an endothelial cell surface antigen, abundantly expressed in continuous endothelia, low and inconsistently expressed in liver sinusoidal and dermal lymphatic endothelia and absent from splenic sinusoidal endothelial cells. In differentiated skin vascular tumors, 1F10 antigen is expressed in normal amounts while it is only low and inconsistently expressed in the dedifferentiated endothelial cells of Kaposi's sarcoma and hemangiosarcoma. HUVE cells in culture, in contrast to their in situ ancestors, express variable amounts of 1F10 antigen. When endothelial cell-conditioned medium (ECC medium) is supplied to HUVE cells in culture, no 1F10 antigen is expressed, while supplementation with fresh serum-containing medium (FSC medium) or cytokines, such as bFGF, suffices to maintain 1F10 expression in 10-70% of the cells. From this we conclude that developmental regulation, environmental influences and cytokine supply contribute to the differentiation and maintenance of the 1F10+ and 1F10-endothelial cell phenotypes, both in vivo and in vitro.     

Angiogenic factors and alveolar vasculature: development and alterations by injury in very premature baboons

Proper formation of the pulmonary microvasculature is essential for normal lung development and gas exchange. Lung microvascular development may be disrupted by chronic injury of developing lungs in clinical diseases such as bronchopulmonary dysplasia. We examined microvascular development, angiogenic growth factors, and endothelial cell receptors in a fetal baboon model of chronic lung disease (CLD). In the last third of gestation, the endothelial cell marker platelet endothelial cell adhesion molecule (PECAM)-1 increased 7.5-fold, and capillaries immunostained for PECAM-1 changed from a central location in airspace septa to a subepithelial location. In premature animals delivered at 67% of term and supported with oxygen and ventilation for 14 days, PECAM-1 protein and capillary density did not increase, suggesting failure to expand the capillary network. The capillaries of the CLD animals were dysmorphic and not subepithelial. The angiogenic growth factor vascular endothelial growth factor (VEGF) and its receptor fms-like tyrosine kinase receptor (Flt-1) were significantly decreased in CLD. Angiopoietin-1, another angiogenic growth factor, and its receptor tyrosine kinase with immunoglobulin and epidermal growth factor homology domains were not significantly changed. These data suggest that CLD impairs lung microvascular development and that a possible mechanism is disruption of VEGF and Flt-1 expression.     

Rare facial clefts: treatment during charity missions in developing countries

During 10 charity missions in developing countries, 14 patients of a total of 374 children with cleft lip and palate deformities were treated for rare facial clefts. There were three midline clefts (Tessier no. 0 cleft, n = 1; Tessier no. 14 cleft, n = 2), four oblique facial clefts (Tessier no. 3 cleft, n = 2; Tessier no. 5 cleft, n = 2), and seven lateral facial clefts (Tessier no. 7 cleft). Surgical treatment focused on cleft repair by soft-tissue reconstruction apart from two Tessier no. 14 clefts, in which the bony gap was also closed using bone grafts from the iliac crest. The postoperative course was uneventful except for one local wound infection that was treated successfully using oral antibiotics. This article summarizes the authors' experience with the surgical management of these malformations and considers the limitations under conditions of charity missions in developing countries. Furthermore, some rare forms of cleft formation are added to the existing literature.     

Antiangiogenic potential of 10-hydroxycamptothecin

To investigate the antiangiogenic potential of 10-hydroxycamptothecin (HCPT), the proliferation of human microvascular endothelial cells (HMEC) and seven human tumor cell lines were detected by SRB assay, and the endothelial cell migration and tube formation were assessed using two in vitro model systems. Also, inhibition of angiogenesis was determined with a modification of the chick embryo chorioallantoic membrane (CAM) assay in vivo. Morphological assessment of apoptosis was performed by fluorescence microscope. HCPT 0.313-5 micromol x L(-1) treatment resulted in a dose-dependent inhibition of proliferation, migration and tube formation in HMEC cells, and HCPT 6.25-25 nmol x egg(-1) inhibited angiogenesis in CAM assay. HCPT 1.25-5 micromol x L(-1) elicited typical morphological changes of apoptosis including condensed chromatin, nuclear fragmentation, and reduction in volume in HMEC cells. HCPT significantly inhibited angiogenesis both in vitro and in vivo at relatively low concentrations, and this effect was related with induction of apoptosis in HMEC cells. These results taken collectively suggest that HCPT may be a potent antiangiogenetic and cytotoxic drug and further investigation is warranted.     

Contact-dependent inhibition of angiogenesis by cardiac fibroblasts in three-dimensional fibrin gels in vitro: implications for microvascular network remodeling and coronary collateral formation

Angiogenesis and coronary artery collateral formation can improve blood flow and thereby prevent myocardial ischemia. The role of perivascular fibroblasts in neovascularization remains incompletely understood. Here we investigated the effects of epicardial and myocardial fibroblasts on angiogenesis in vitro by using a serum-free microcarrier-based fibrin gel angiogenesis system. To clearly distinguish between different cell types, we either stained endothelial cells or fibroblasts in the living with 1,1′-dioctadecyl-3,3,3′,3′-tetramethyl-indocarbocyanine-perchlorate (DiI). In cocultures, low numbers of heart fibroblasts stimulated endothelial sprouting, and capillary growth was also induced by fibroblast-conditioned media, indicating a paracrine mechanism. Capillary formation was decreased by increasing the density of fibroblasts in the cocultures, indicating contact-dependent inhibition. Using time-lapse studies, it turned out that close contacts between fibroblasts and endothelial cells resulted in rapid retraction of endothelial cells or, rarely, in cell death. Depending on the local ratio of fibroblasts to endothelial cell numbers, fibroblasts determined the location of capillary growth and the size of developing capillaries and thereby contributed to capillary network remodeling. In contrast to primary heart fibroblasts, NIH 3T3 fibroblasts did not display contact-dependent inhibition of endothelial sprouts. NIH fibroblasts were frequently seen in close association with endothelial capillaries, resembling pericytes. Contact-dependent inhibition of angiogenesis by epicardial fibroblasts could not be reversed by addition of neutralizing anti-TGF-β1 antibodies, by addition of serum, of medium conditioned by hypoxic tumor cells or myocardium, by various cytokines or by growing cocultures under hypoxic conditions. Our results implicate a pivotal role of periendothelial mesenchymal cells for the regulation of microvascular network remodeling and collateral formation. Received: 15 September 1997 / Accepted: 6 April 1998     

Chlamydia pneumoniae infection alters the junctional complex proteins of human brain microvascular endothelial cells

Chlamydia pneumoniae has been identified and associated with multiple sclerosis (MS) and Alzheimer's disease (AD) pathogenesis, although the relationship of this organism in these diseases remains controversial. We have hypothesized that one potential avenue of infection is through the junctional complexes between the blood-brain barrier (BBB) endothelia. C. pneumoniae is characteristically a respiratory pathogen, but has been implicated in atherosclerosis, coronary artery disease, and neuroinflammatory conditions. C. pneumoniae infection may lead to endothelial damage, junctional alterations, and BBB breakdown. Therefore, in this study, C. pneumoniae infection of human brain microvascular endothelial cells (HBMECs) resulted in increased expression of the zonula adherens proteins beta-catenin, N-cadherin, and VE-cadherin, and decreased expression of the tight junctional protein occludin, as determined by immunocytochemistry and Western blot analyses. These events may underlie a mechanism for the regulation of paracellular permeability while maintaining barrier integrity during C. pneumoniae infection associated with neuropathologies such as MS and AD.