Collagen-Elastin and Collagen-Glycosaminoglycan Scaffolds Promote Distinct Patterns of Matrix Maturation and Axial Vascularization in Arteriovenous Loop-Based Soft Tissue Flaps.

INTRODUCTION: Autologous free flaps are the criterion standard for reconstructions of complex soft tissue defects; however, they are limited by donor-site morbidities. The arteriovenous (AV) loop model enables the generation of soft tissue constructs based on acellular dermal matrices with a functional microvasculature and minimal donor site morbidity. The ideal scaffold for AV loop-based tissue engineering has not been determined. METHODS: AV loops were placed into subcutaneous isolation chambers filled with either a collagen-elastin scaffold or a collagen-glycosaminoglycan scaffold in the thighs of rats. Matrix elasticity, neoangiogenesis, cell migration, and proliferation were compared after 14 and 28 days. RESULTS: Mean vessel count and area had increased in both matrices at 28 compared with 14 days. Collagen-elastin matrices showed a higher mean vessel count and area compared with collagen-glycosaminoglycan matrices at 14 days. At 28 days, a more homogeneous vascular network and higher cell counts were observed in collagen-elastin matrices. Collagen-glycosaminoglycan matrices, however, exhibited less volume loss at day 28. CONCLUSIONS: Collagen-based scaffolds are suitable for soft tissue engineering in conjunction with the AV loop technique. These scaffolds exhibit distinct patterns of angiogenesis, cell migration, and proliferation and may in the future serve as the basis of tissue-engineered free flaps as an individualized treatment concept for critical wounds.

Recruitment of human cord blood-derived endothelial colony-forming cells to sites of tumor angiogenesis.

BACKGROUND AIMS: Endothelial progenitor cells (EPCs) specifically home to sites of malignant growth, rendering them attractive for anti-cancer therapies. Data are conflicting on the phenotype and quantitative contribution toward tumor angiogenesis based on differing culture assays to outgrow EPCs. To evaluate the origin and early phenotype of EPCs and to define a population with enhanced tumor-targeting capacity, we evaluated a hierarchy of cord blood-derived EPCs modeling the multi-step nature of tumor homing. METHODS: CD34(+) mononuclear cells were isolated from fresh cord blood and cultured to derive endothelial colony-forming cells (ECFCs). Human umbilical vein endothelial cells (HUVECs) served as control. Using intra-vital microscopy, the recruitment was analyzed in mice bearing C6 xenografts. Adhesion, migration, transmigration and differentiation were further addressed. RESULTS: Within the primary passage, ECFCs underwent a rapid maturation from a CD45(+) and CD31(+) phenotype to a CD45(-) and endothelial marker positive phenotype. Assessing in vivo tumor recruitment, ECFCs had the highest activity in all steps analyzed. In vitro, ECFCs demonstrated significantly higher adhesion under static and flow conditions. Similarly, ECFCs exhibited highest migratory and trans-migratory activity toward tumor-conditioned medium. On subcutaneous implantation, only ECFCs formed blood vessels covered with perivascular cells, similar to HUVECs. CONCLUSIONS: Our study indicates that ECFCs emerge from a CD45(+) and CD31(+) progenitor and rapidly mature in culture. ECFCs have a significantly higher potential for tumor targeting than non-cultured CD34(+) cells and HUVECs. They are ideal candidates for future cell-based anti-cancer therapies.

Cultivation in human serum reduces adipose tissue-derived mesenchymal stromal cell adhesion to laminin and endothelium and reduces capillary entrapment.

The increasing use of mesenchymal stromal cells (MSC) in clinical cellular therapy requires a safe and controlled production process compliant with Good Manufacturing Practice guidelines. Pooled blood group AB human serum (HS) has been used to replace fetal bovine serum (FBS), critically rated by the regulatory agencies, since it can support the expansion of adipose tissue-derived mesenchymal stromal cells (ASC). However, it remains unknown whether the choice of serum affects application-relevant characteristics of ASC. A microarray-based screen has revealed differentially expressed adhesion and extracellular matrix-associated molecules in HS- and FBS-ASC. Since cell therapy relies on the cells' efficacy to home and engraft, HS- and FBS-ASC were compared by analyzing adhesion, migration, and transmigration as well as short-term homing in vivo. HS-cultivated ASC demonstrated a higher adhesion to plastic, but reduced adhesion to extracellular matrix molecules, that is, laminin, and to endothelial cells both under static and flow conditions. Migration and transmigration assays confirmed the attraction of ASC by the tumor conditioned medium irrespective of the supplement. Coinjecting differently labeled HS- and FBS-ASC into nonobese diabetic, severe combined immunodeficiency mice revealed reduced numbers of HS-ASC in lungs and liver. This has been interpreted as reduced capillary entrapment. Our data indicate that varying the serum supplement may alter application-relevant characteristics of ASC, such as adhesion, as well as lung entrapment after infusion. Appropriate injury models and further molecular analyses are required to provide mechanistic insight into the differential effects of HS versus FBS on ASC cultures.

Systemic peripheral artery relaxation by KCNQ channel openers and hydrogen sulfide.

BACKGROUND: Perivascular adipose tissue secretes an adipocyte-derived relaxing factor (ADRF) that opens voltage-dependent K (Kv) channels in peripheral arteries. We studied the role of KCNQ-type Kv channels and tested the hypothesis that hydrogen sulfide (H2S) could be an ADRF. METHODS: We performed isometric contraction studies on systemic arteries of rats and mice. RESULTS: In mesenteric arteries and aortas without perivascular adipose tissue, the KCNQ channel openers retigabine, VRX0530727, VRX0621238, and VRX0621688 produced concentration-dependent vasorelaxation; VRX0621688 was the most potent vasodilator. The KCNQ inhibitor XE991 (30 micromol/l) blocked the effects of both the drugs and ADRF. Inhibitors of cystathionine gamma lyase (CSE) beta-cyano-L-alanine (BCA, 5 mmol/l) and 4-propargyl glycine (PPG, 10 mmol/l) also blocked the relaxations. CSE is expressed in perivascular adipose tissue and endogenously generates H2S. The H2S donor NaHS produced concentration-dependent vasorelaxation, which was also blocked by XE991. The vasodilatory capacities of retigabine, VRX0530727, VRX0621238, and VRX0621688 were preserved following inhibition of H2S generation in perivascular fat. CONCLUSION: We suggest that KCNQ channel opening is a powerful mechanism to produce vasorelaxation of systemic arteries in rats and mice. Furthermore, KCNQ channels play a major role in the paracrine control of vascular tone by perivascular adipose tissue, which is at least in part mediated or modulated by H2S. In conditions of reduced H2S release from perivascular adipose tissue, these paracrine effects can be mimicked by synthetic KCNQ channel openers.

Semiautomatic quantification of angiogenesis.

BACKGROUND: Angiogenesis is of major interest in developmental biology and cancer research. Different experimental approaches are available to study angiogenesis that have in common the need for microscopy, image acquisition, and analysis. Problems that are encountered hereby are the size of the structures, which requires generation of composite images and difficulties in quantifying angiogenic activity reliably and rapidly. Most graphic software packages lack some of the required functions for easy, semiautomatic quantification of angiogenesis and, consequently, multiple software packages or expensive programs have to be used to cover all necessary functions. METHODS AND RESULTS: A software package (AQuaL) to analyze angiogenic activity was developed using Java, which can be used platform-independently. It includes image acquisition relying on the Java Media Framework and an easy to use image alignment tool. Multiple overlapping images can be aligned and saved without limitations and loss of resolution into a composite image, which requires only the selection of a single point representing a characteristic structure in adjacent images. Angiogenic activity can be quantified in composite images semiautomatically by the assessment of the area overgrown by cells after filtering and image binarization. In addition, tagging of capillary-like structures allows quantification of their length and branching pattern. Both developed methods deliver reliable and correlating data as exemplified in the aortic ring angiogenesis assay. CONCLUSION: The developed software provides modular functions specifically targeted to quantify angiogenesis. Whereas the area measurement is time saving, length measurement provides additional information about the branching patterns, which is required for a qualitative differentiation of capillary growth.

Inhibition of the tyrosine phosphatase SHP-2 suppresses angiogenesis in vitro and in vivo.

Endothelial cell survival is indispensable to maintain endothelial integrity and initiate new vessel formation. We investigated the role of SHP-2 in endothelial cell survival and angiogenesis in vitro as well as in vivo. SHP-2 function in cultured human umbilical vein and human dermal microvascular endothelial cells was inhibited by either silencing the protein expression with antisense-oligodesoxynucleotides or treatment with a pharmacological inhibitor (PtpI IV). SHP-2 inhibition impaired capillary-like structure formation (p < 0.01; n = 8) in vitro as well as new vessel growth ex vivo(p < 0.05; n = 10) and in vivo in the chicken chorioallantoic membrane (p < 0.01, n = 4). Additionally, SHP-2 knock-down abrogated fibroblast growth factor 2 (FGF-2)-dependent endothelial proliferation measured by MTT reduction (p < 0.01; n = 12). The inhibitory effect of SHP-2 knock-down on vessel growth was mediated by increased endothelial apoptosis (annexin V staining, p < 0.05, n = 9), which was associated with reduced FGF-2-induced phosphorylation of phosphatidylinositol 3-kinase (PI3-K), Akt and extracellular regulated kinase 1/2 (ERK1/2) and involved diminished ERK1/2 phosphorylation after PI3-K inhibition (n = 3). These results suggest that SHP-2 regulates endothelial cell survival through PI3-K-Akt and mitogen-activated protein kinase pathways thereby strongly affecting new vessel formation. Thus, SHP-2 exhibits a pivotal role in angiogenesis and may represent an interesting target for therapeutic approaches controlling vessel growth.

Chemokine fractalkine mediates leukocyte recruitment to inflammatory endothelial cells in flowing whole blood: a critical role for P-selectin expressed on activated platelets.

BACKGROUND: The membrane-bound chemokine fractalkine (CX3CL1) is expressed on various cell types such as activated endothelial cells and has been implicated in the inflammatory process of atherosclerosis. The aim of the present study was to dissect the role of fractalkine in leukocyte recruitment to inflamed endothelium under arterial shear forces. METHODS AND RESULTS: With the use of immunofluorescence and laminar flow assays, the present study shows that human umbilical vein endothelial cells stimulated with tumor necrosis factor-alpha and interferon-gamma abundantly express CX3CL1 and promote substantial leukocyte accumulation under arterial flow conditions. In the presence of high shear, firm adhesion of leukocytes to inflamed endothelial cells is reduced by approximately 40% by a function-blocking anti-fractalkine antibody or by an antibody directed against the fractalkine receptor (CX3CR1). With the use of intravital video-fluorescence microscopy we demonstrate that inhibition of fractalkine signaling attenuates leukocyte adhesion to the atherosclerotic carotid artery of apolipoprotein E-deficient mice, which suggests that the CX3CL1-CX3CR1 axis is critically involved in leukocyte adhesion to inflamed endothelial cells under high shear forces both in vitro and in vivo. Surprisingly, platelets were strictly required for fractalkine-induced leukocyte adhesion at high shear rates. Correspondingly, specific inhibition of platelet adhesion to inflamed endothelial cells also significantly reduced leukocyte accumulation. We show that both soluble and membrane-bound fractalkine induces platelet degranulation and subsequent surface expression of P-selectin, which thereby promotes direct platelet-leukocyte interaction. CONCLUSIONS: Fractalkine expressed by inflamed endothelial cells triggers P-selectin exposure on adherent platelets, which thereby initiates the local accumulation of leukocytes under arterial shear, an essential step in the development of atherosclerotic lesions.

The tyrosine phosphatase, SHP-1, is a negative regulator of endothelial superoxide formation.

OBJECTIVES: We investigated the role of SH2-domain containing phosphatase-1 (SHP-1) in endothelial reduced nicotinamide adenine dinucleotide (phosphate) (NAD[P]H)-oxidase-dependent oxidant production. BACKGROUND: Superoxide (O2*-) generation by endothelial NAD(P)H-oxidase promotes endothelial dysfunction and atherosclerosis. Signaling pathways that regulate NAD(P)H-oxidase activity are, however, poorly understood. METHODS: SH2-domain containing phosphatase-1 was inhibited using site-directed magnetofection of antisense oligodesoxynucleotides (AS-ODN) or short interfering ribonucleic acid (siRNA) in vitro in human umbilical vein endothelial cells (HUVEC) and in isolated hamster arteries; O2*- was measured by cytochrome c reduction in vitro. Activities of NAD(P)H-oxidase activity, phosphatidyl-inositol-3-kinase (PI3K), and SHP-1 were assessed by specific assays; Rac1 activation was assessed by a pull-down assay. RESULTS: Basal endothelial O2*- release was enhanced after inhibition of endothelial SHP-1 (p < 0.01), which could be prevented by specific inhibition of NAD(P)H-oxidase (p < 0.01); SHP-1 activity was high under basal conditions, further increased by vascular endothelial growth factor (10 ng/ml, p < 0.05), and abolished by SHP-1 AS-ODN treatment (p < 0.01), which also increased NAD(P)H-oxidase activity 3.3-fold (p < 0.01). Vascular endothelial growth factor also induced O2*- release (p < 0.01), which was even more enhanced when SHP-1 was knocked down (p < 0.05). The effect of SHP-1 was mediated by inhibition of PI3K/Rac1-dependent NAD(P)H-oxidase activation (p < 0.01); SHP-1 AS-ODN augmented tyrosine phosphorylation of the p85 regulatory subunit of PI3K (p < 0.05) and Rac1 activation. The latter was prevented by wortmannin, a blocker of PI3K. CONCLUSIONS: In HUVEC, SHP-1 counteracts basal and stimulated NAD(P)H-oxidase activity by negative regulation of PI3K-dependent Rac1 activation; SHP-1 thus seems to be an important part of endothelial antioxidative defense controlling the activity of the O2(*-)-producing NAD(P)H-oxidase.

Membrane-potential-dependent inhibition of platelet adhesion to endothelial cells by epoxyeicosatrienoic acids.

OBJECTIVE: Epoxyeicosatrienoic acids (EETs) are potent vasodilators produced by endothelial cells. In many vessels, they are an endothelium-derived hyperpolarizing factor (EDHF). However, it is unknown whether they act as an EDHF on platelets and whether this has functional consequences. METHODS AND RESULTS: Flow cytometric measurement of platelet membrane potential using the fluorescent dye DiBac4 showed a resting potential of -58+/-9 mV. Different EET regioisomers hyperpolarized platelets down to -69+/-2 mV, which was prevented by the non-specific potassium channel inhibitor charybdotoxin and by use of a blocker of calcium-activated potassium channels of large conductance (BK(Ca) channels), iberiotoxin. EETs inhibited platelet adhesion to endothelial cells under static and flow conditions. Exposure to EETs inhibited platelet P-selectin expression in response to ADP. Stable overexpression of cytochrome P450 2C9 in EA.hy926 cells (EA.hy2C9 cells) resulted in release of EETs and a factor that hyperpolarized platelets and inhibited their adhesion to endothelial cells. These effects were again inhibited by charybdotoxin and iberiotoxin. CONCLUSIONS: EETs hyperpolarize platelets and inactivate them by inhibiting adhesion molecule expression and platelet adhesion to cultured endothelial cells in a membrane potential-dependent manner. They act as an EDHF on platelets and might be important mediators of the anti-adhesive properties of vascular endothelium.

Nitric oxide enhances de novo formation of endothelial gap junctions.

OBJECTIVE: Gap junctions (formed by connexins, Cx) are important for functional coordination of cells in the vascular wall. However, little is known about their physiological regulation in this tissue. We examined the effects of nitric oxide (NO), an important mediator of vasomotion, wound healing and angiogenesis, on the formation of gap junctions in endothelial cells (human umbilical vein endothelial cells, HUVEC). METHODS: Flow cytometry was used to determine dye transfer through newly formed gap junctions between acutely coincubated HUVECs. Parallel experiments in wild-type HeLa cells (no connexins) and transfected HeLa cells exclusively expressing Cx43, Cx40 or Cx37 were performed to determine the specific role of Cx subtypes. The intracellular distribution of Cx40 was examined after fractionation with triton by Western blotting. Intracellular levels of cGMP and cAMP were measured by radioimmunoassay. RESULTS: The NO donor SNAP (1 microM) enhanced gap-junctional coupling in HUVECs by about 40%. This was associated with an enhanced incorporation of Cx40 into the membrane. Both effects were restricted to Cx40 as analyzed in experiments with Cx-selective HeLa cells. The NO-induced increase in cell coupling was elicited by a corresponding rise of cGMP, which secondarily increased intracellular cAMP levels. The latter was an integral part of the signal cascade, since the protein kinase A (PKA) inhibitor H89 blocked the SNAP-induced incorporation of Cx40 into the plasma membrane. CONCLUSIONS: We conclude that NO is a potent modulator of gap-junctional coupling in endothelial cells. It enhances de novo formation of endothelial gap junctions by increasing incorporation of Cx40 into the plasma membrane due to PKA activation.

Differential regulation of xanthine and NAD(P)H oxidase by hypoxia in human umbilical vein endothelial cells. Role of nitric oxide and adenosine.

OBJECTIVES: Although in tissue injury following hypoxia/reoxygenation (H/R) an increased endothelial formation of superoxide anions (O(2)(-)) plays an important role, it is still not fully understood which of the potential enzymatic sources of endothelial O(2)(-) are crucially involved. In this study, we particularly examined the activities of NAD(P)H oxidase and xanthine oxidase (XO) after 8 h of exposure to mild hypoxia. We further studied whether enzyme activities can be modified by NO and adenosine during hypoxic treatment. METHODS AND RESULTS: In human umbilical vein endothelial cells O(2)(-) production was measured immediately after exposure to hypoxia ('early reoxygenation') or after 2 h of reoxygenation at normoxic conditions ('late reoxygenation'). In the early reoxygenation phase the O(2)(-) production was attenuated by 28.5% while it was enhanced by 58.2% after late reoxygenation. Using specific inhibitors of NAD(P)H oxidase and XO, gp91ds-tat and oxypurinol, respectively, we show that the constitutively active NAD(P)H oxidase was blocked following hypoxia while XO was activated. The presence of NO during hypoxia had no effect on NAD(P)H oxidase activity but it significantly inhibited the activation of XO. Inhibition of XO activation was, at least in part, caused by the release of adenosine from endothelial cells which induces an increased formation of NO by its A1 and A2 receptors. CONCLUSION: Our results indicate that during exposure to mild hypoxia for 8 h, a change in the enzymatic source of endothelial O(2)(-) occurs: a prolonged inhibition of NAD(P)H oxidase was found while an enhanced activity of XO occurs in the reoxygenation phase. These results suggest that different strategies of antioxidant therapy should be taken into consideration in oxidative stress related to chronic hypoxia when compared to normoxic atherosclerotic tissues with an activated vascular NAD(P)H oxidase as the main source of O(2)(-).

An angiogenic role for the human peptide antibiotic LL-37/hCAP-18.

Antimicrobial peptides are effector molecules of the innate immune system and contribute to host defense and regulation of inflammation. The human cathelicidin antimicrobial peptide LL-37/hCAP-18 is expressed in leukocytes and epithelial cells and secreted into wound and airway surface fluid. Here we show that LL-37 induces angiogenesis mediated by formyl peptide receptor-like 1 expressed on endothelial cells. Application of LL-37 resulted in neovascularization in the chorioallantoic membrane assay and in a rabbit model of hind-limb ischemia. The peptide directly activates endothelial cells, resulting in increased proliferation and formation of vessel-like structures in cultivated endothelial cells. Decreased vascularization during wound repair in mice deficient for CRAMP, the murine homologue of LL-37/hCAP-18, shows that cathelicidin-mediated angiogenesis is important for cutaneous wound neovascularization in vivo. Taken together, these findings demonstrate that LL-37/hCAP-18 is a multifunctional antimicrobial peptide with a central role in innate immunity by linking host defense and inflammation with angiogenesis and arteriogenesis.

Gap-junctional coupling between neutrophils and endothelial cells: a novel modulator of transendothelial migration.

Communication between leukocytes and endothelial cells is crucial for inflammatory reactions. Paracrine cross-talk and outside-in signaling (via adhesion molecules) have been characterized as communication pathways to date. As leukocytes and endothelial cells express connexins, we considered intercellular communication via gap junctions an intriguing additional concept. We found that gap-junctional coupling between neutrophils and endothelium occurred in a time-dependent, bidirectional manner and was facilitated by adhesion. After blockade of connexins, transmigration of neutrophils through the endothelial layer was enhanced, and the barrier function of cell monolayers was reduced during transmigration. Tumor necrosis factor alpha decreased coupling. In the presence of connexins, transmigration of neutrophils did not alter permeability. Thus, neutrophils couple to endothelium via gap junctions, functionally modulating transmigration and leakiness. Gap-junctional coupling may be a novel way of leukocyte-endothelial communication.

Depolarization of endothelial cells enhances platelet aggregation through oxidative inactivation of endothelial NTPDase.

OBJECTIVE: The objective of this study was to investigate whether depolarization of cultured endothelial cells (human umbilical vein endothelial cells, HUVECs) affects their ectonucleotidase activity through superoxide (O2-) production. METHODS AND RESULTS: Depolarization by the cation channel gramicidin (100 nmol/L) or tetrabutylammonium chloride (1 mmol/L) induced O2- release from HUVECs (n=4), which was decreased by superoxide dismutase (SOD, 500 U/mL). The activity of endothelial ectonucleotidases was assessed by the production of inorganic phosphate from ADP, which was decreased by chronic depolarization by 25% (n=6, P<0.05) and the amount of AMP derived from ADP in the presence of the 5'-nucleotidase inhibitor alpha,beta-methylene-5'-diphosphate (100 micromol/L). AMP was decreased by chronic depolarization from 0.54+/-0.16 to 0.39+/-0.11 micromol/min/mg protein (n=6, P<0.05). This was abolished in the continuous presence of SOD (n=6). NTPDase protein was predominantly expressed in HUVECs (n=4). Protein abundance, viability of cells, and apoptosis rates were not altered by depolarization (n=10). Only in the presence of depolarized HUVECs, but not with control cells or with HUVECs depolarized in the presence of SOD, did 5 micromol/L of ADP cause irreversible platelet aggregation. Increases in transmural pressure induced endothelial depolarization in intact hamster small arterioles. CONCLUSIONS: Depolarization causes the endothelial production of O2-, which inhibits the activity of endothelial ectonucleotidases. Increases in transmural pressure induce endothelial depolarization. In chronically hypertensive diseases, depolarization might favor platelet aggregation.

NAD(P)H oxidase-dependent platelet superoxide anion release increases platelet recruitment.

Platelets, although not phagocytotic, have been suggested to release O. Since O-producing reduced nicotinamide adenine dinucleotide (phosphate) (NAD(P)H) oxidases can be specifically activated by certain agonists and are found in several nonphagocytotic tissues, we investigated whether such an enzyme is the source of platelet-derived O. We further studied which agonists cause platelet O release and whether platelet-derived O influences thrombus formation in vitro. Collagen, but not adenosine 5'-diphosphate (ADP) or thrombin, increased O formation in washed human platelets. This was a reduced nicotinamide adenine dinucleotide (NADH)-dependent process, as shown in platelet lysates. Consistent with a role of a platelet, NAD(P)H oxidase expression of its subunits p47(phox) and p67(phox) and inhibition of platelet O formation by diphenylene-iodoniumchloride (DPI) and by the specific peptide-antagonist gp91ds-tat were observed. Whereas platelet-derived O did not influence initial aggregation, platelet recruitment to a preformed thrombus following collagen stimulation was significantly attenuated by superoxide dismutase (SOD) or DPI. It was also inhibited when ADP released during aggregation was cleaved by the ectonucleotidase apyrase. ADP in supernatants of collagen-activated platelets was decreased in the presence of SOD, resulting in lower ADP concentrations available for recruitment of further platelets. Exogenous O increased ADP- concentrations in supernatants of collagen-stimulated platelets and induced irreversible aggregation when platelets were stimulated with otherwise subthreshold concentrations of ADP. These results strongly suggest that collagen activation induces NAD(P)H oxidase-dependent O release in platelets, which in turn enhances availability of released ADP, resulting in increased platelet recruitment.

Shear stress-induced release of basic fibroblast growth factor from endothelial cells is mediated by matrix interaction via integrin alpha(v)beta3.

Considering that chronic elevation of shear stress results in remodeling of the vasculature, we analyzed whether mechanical load could mediate basic fibroblast growth factor (bFGF) release and whether bFGF would act as mediator of shear stress-induced endothelial proliferation and differentiation. Supernatant media of shear stress-exposed endothelial cells (EC) contained significantly higher amounts of bFGF than medium from static cells. Released bFGF was fully intact with regard to its function as an inductor of proliferation and differentiation. Shear stress-conditioned media induced capillary-like structure formation, whereas static control medium did not. Likewise, only shear stress-conditioned medium induced proliferation of serum starved EC. Both capillary-like structure formation and proliferation could be inhibited by neutralization of bFGF or its receptor. The release of bFGF was subject to specific, integrin-mediated control, since inhibition of alpha(v)beta(3) integrin prevented it, whereas inhibition of alpha(5)beta(1) integrin had no effect. We conclude that shear stress induces the release of bFGF from EC in a tightly controlled manner. The release is dependent on specific cell-matrix interactions via alpha(v)beta(3) integrins. The effects on cell proliferation and differentiation suggest that release of bFGF is functionally significant and may represent a necessary initial step in adaptive remodeling processes induced by shear stress.