Effect of cell density on thrombin binding to a specific site on bovine vascular endothelial cells

Heterogeneity of endothelial cells. Specific markers

During embryonic development, endothelial cells differentiate from a common precursor called angioblast and acquire organ-specific properties. One of the important determinants of endothelial cell differentiation is the local environment, and especially the interaction with surrounding cells. This interaction may occur through the release of soluble cytokines, cell-to-cell adhesion and communication, and the synthesis of matrix proteins on which the endothelium adheres and grows. The acquisition and maintenance of specialized properties by endothelial cells is important in the functional homeostasis of the different organs. For instance, in the brain, alteration of the blood-brain barrier properties may have important consequences on brain functional integrity. One of the major limitations to the study of endothelial cell heterogeneity is the fact that these cells are still difficult to isolate and culture from the microcirculation of different organs, and once in culture, they tend to lose their specialized properties. This finding suggests that we have to develop new culture systems, which possibly include coculture with other cell types. An important issue is to develop tools that can help in recognizing endothelial cells and their differentiated phenotype both in vivo and in tissue culture. In this review we give a short overview of the differentiated properties of the endothelium, considering a few examples of highly specialized endothelial cells, such as the brain or bone marrow microcirculation or high endothelial venules. We made a particular effort to list the most common markers of endothelial cell phenotypes. These molecules and related antibodies may be valuable tools for endothelial cell isolation and characterization.

The influence of hypoglycaemic agents on the growth and metabolism of human endothelial cells

Using human umbilical vein endothelial cells, and human microvascular endothelial cells from omental and subcutaneous fat obtained at laparotomy, we studied the effects of sulphonylureas and the biguanide metformin on endothelial cell proliferation, prostacyclin production, ecto-5'-nucleotidase activity, and von Willebrand factor release. Each drug produced a concentration-dependent proliferation of umbilical vein but not of microvascular endothelial cells. The stimulation of umbilical vein endothelial cell proliferation by sulphonylureas, but not by metformin, was serum- and insulin-dependent. Sulphonylureas and metformin had no effect on the proliferation of human dermal fibroblasts, smooth muscle cells derived from the umbilical artery, or 3T3 cells, until concentrations greater than 100 fold those found in vivo were reached, when there was inhibition of proliferation. These agents had no effect on prostacyclin or von Willebrand factor production, or on ecto-5'-nucleotidase activity, until high concentrations were used, at levels which also inhibited proliferation. The results suggest that the sulphonylureas and metformin, may, at concentrations found in vivo, induce changes in the turnover of endothelial cells from large vessels, but not of microvascular endothelial cells.

The binding of radioiodinated thrombin to rat promegakaryoblasts

Rat promegakaryoblasts have recently been shown to possess a number of properties of mature platelets. We have studied the binding of 125I-thrombin to these cells grown in culture. The binding was found to be saturable, specific, reversible, and of high affinity. Such cell lines may be useful models for the study of platelet-agonist interactions.

Bradykinin and thrombin effects on polyphosphoinositide hydrolysis and prostacyclin production in endothelial cells

Prostacyclin (PGI2) production by thrombin- and bradykinin-stimulated bovine aortic endothelial cells (BAEC) and human umbilical vein endothelial cells (HUVEC) was related to the receptor-linked activation of inositide hydrolysis. Bradykinin caused a rapid and transient 3-fold increase in the formation of inositol polyphosphates in BAEC. The increase in InsP3 reflected changes mainly in the Ins(1,4,5)P3 isomer. Thrombin was less effective than bradykinin in increasing InsP3 levels and appeared to only minimally stimulate the production of PGI2 in BAEC. In HUVEC, thrombin caused a 5-fold elevation of Ins(1,4,5)P3, closely related to a rise in PGI2 production. However, bradykinin did not affect inositol phosphates and PGI2 production in HUVEC. Other inositol phosphates were also assessed to obtain information on putative metabolism of Ins(1,4,5)P3. The present study supports the notion that formation of Ins(1,4,5)P3 is linked to an increase in PGI2 production in endothelial cells and furthermore provides evidence for a large degree of heterogeneity in the responses of BAEC and HUVEC to thrombin and bradykinin.

Association of a plasminogen activator inhibitor (PAI-1) with the growth substratum and membrane of human endothelial cells

We have studied the distribution of the plasminogen activator inhibitor type 1 (PAI-1) in cultures of confluent human umbilical vein endothelial cells. Plasminogen activator inhibitor activity measured by the 125I-fibrin plate assay was detected in the cytosol (2.85 +/- 0.16 U), 100,000 g particulate fraction (1.26 +/- 0.30 U), and in the growth substratum (9.82 +/- 1.80 U). Characterization of the protein responsible for this activity by reverse fibrin autography, immunoprecipitation, and immunoblotting demonstrated that it had an Mr of 46,000 and was antigenically related to PAI-1. Only the active form of the inhibitor was found in all three fractions. Inhibitor in the cytosol and particulate fraction converted to the latent form during 37 degrees C incubation while the substratum inhibitor remained fully active. Extracellular PAI-1 was detected in the growth substratum before its appearance in conditioned medium and represented the major protein deposited beneath the cells. The inhibitor was only transiently localized in the substratum, disappearing within 6 h and concomitantly appearing in the culture medium. Incubation of isolated metabolically labeled substratum with tissue plasminogen activator (tPA) resulted in the appearance and release of an immunologically related inactive 44,000 Mr form as well as the tPA-PAI-1 complex (110,000 Mr). PAI-1 was also converted into its 44,000-Mr form and released by treatment of the substratum with human leukocyte elastase. The rapid deposition and predominance of PAI-1 in the underlying compartment of endothelial cells may explain how the basement membrane is protected from proteolytic degradation by plasmin-generating enzymes.

Functional heterogeneity of vascular endothelial cells

This review has highlighted some of the well-described differences in endothelial cells derived from different sites of the vascular tree. In presenting a select group of endothelial properties, there was no intention to imply that these are the only properties of endothelial cells that exhibit heterogeneity. Nonetheless, having described endothelial heterogeneity in regard to a number of defined properties, we are left with persistent questions including: Are these divergent properties of endothelial cells fixed? Alternatively, can we alter the properties of endothelial cells by affecting the signals from the environment? A number of reports strongly suggest that endothelial cells are capable of acquiring new properties. Stewart and Wiley investigated the role of the neural tissue environment on the differentiation of brain capillary endothelial cells. These authors transplanted ectopic sites, i.e. vascular segments of brain from very young quail embryos to chick coeliac cavity, and a quail somites to chick brain ventricles. The distinctive morphology of quail cells provided a cell marker to differentiate host from graft. The results of this study demonstrated that mesenteric or somatic vessels growing into grafted brain developed functional, structural and histochemical features specific for neural capillaries. Conversely, capillaries in mesodermal tissue that had been grafted to the brain were devoid of the neural capillary characteristics, indicating that brain vessels do not form a barrier when they are made to vascularize non-neural tissue. Milici and Carley reported that bovine adrenal capillary cells cultured on plastic exhibited occasional diaphragmed fenestrations and no transendothelial channels. However, if these same cells were cultured on a basement membrane (matrix) laid down by MDCK cells (a canine nephron epithelial cell line), the cells responded by increasing the number of diaphragmed fenestrations and transendothelial channels. This cell culture study supported an earlier whole animal study in which the importance of the epithelium and/or epithelial basal lamina in the maintenance of endothelial ultrastructure was demonstrated in a developmental study of rat intestinal capillaries. In this earlier study, it was noted that epithelial development coincided with the formation of fenestrations by the endothelium. Enzymatic activities of endothelial cells can also be altered by environmental signals. For example, primary cerebral microvascular endothelial cells exhibit barrier features and are enriched in gamma-glutamyl transpeptidase activity, yet rapidly lose the activity when subcultured.(ABSTRACT TRUNCATED AT 400 WORDS)

Thrombin-cellular interactions

It is clear that there are a number of different types of reactions between thrombin and the cell surface (TABLE 6). In one type, thrombin binds to cell-surface receptors resulting in cellular activation. In other types of reactions, there are at least two components to the thrombin-specific pathway of cellular activation: a classical receptor to which thrombin binds, and a protein that is cleaved. In both types of reactions, thrombin binding and/or proteolysis is linked to changes in GTP-binding proteins, protein kinase C, or other pathways. In most cases, the receptor and membrane substrates involved in cellular activation are not well characterized. In another type of reaction, the interaction between thrombin and proteins in the extracellular fluid is regulated by cell-surface receptors. Binding of thrombin to these receptors can result in acceleration or inhibition of the reactions with the soluble proteins. In the fourth type of reaction, thrombin cleaves a cell-membrane protein that is involved in reactions with plasma proteins. Recognition of the different types of interactions between thrombin and the cell surface is necessary for the correct interpretation of experimental observations. Although the term receptor has classically referred to a cell-surface component to which an agonist binds, it is now clear that there are additional membrane components that specifically bind potential agonists not leading to cellular activation.

Thrombin stimulates neutrophil adherence by an endothelial cell-dependent mechanism: characterization of the response and relationship to platelet-activating factor synthesis

Thrombin, a serine coagulation protease that is generated at sites of tissue injury and inflammation, stimulates the adherence of PMNs and neutrophils to EC. We found that thrombin enhanced the adhesion of neutrophils to primary monolayers of human umbilical vein EC when assayed by the binding of 111Indium-labeled PMNs to the EC, the recovery of unlabeled PMNs after incubation with thrombin-treated EC, and by phase contrast and scanning electron microscopy (SEM). SEM demonstrated that thrombin caused PMNs to intimately adhere to the EC plasma membrane and under some conditions to become polarized. The thrombin-stimulated adherence was a rapid, time-dependent response with an onset within 1 minute of addition of thrombin, a peak at 5-10 minutes, and a decline thereafter. The response was concentration-dependent over the range 0.01-2 U/ml thrombin, and required active thrombin. Prothrombin, factor Xa, and fibrinogen were not effective. Thrombin-stimulated PMN adherence was dependent on the EC, because thrombin did not significantly stimulate neutrophils to adhere to albumin-coated petri dishes, subendothelial matrices, or primary cultures of smooth muscle cells. Human EC, when treated with thrombin, also produce platelet-activating factor (PAF; 1-alkyl-2-acetyl-sn-glycero-3-phosphocholine). The concentration-response relationships and time courses for thrombin-stimulated PMN adherence and PAF production were tightly correlated. Furthermore, PAF itself stimulated the adherence of PMNs to EC, and pretreatment of PMNs with PAF selectively inhibited their adherence response to thrombin. These findings demonstrate two novel biologic activities of thrombin, the stimulation of EC-dependent adherence of PMNs and the production of PAF by EC, and suggest that they are functionally related. In addition, they suggest that thrombin may act as a plasma-derived humoral mediator of inflammation under some conditions.

Differences in reactivity of confluent and nonconfluent cultures of human endothelial cells toward thrombin-stimulated platelets or heparinized salt solution

This study examined whether nonconfluent endothelial cell cultures reacted differently than confluent ones toward thrombin-stimulated platelets or a heparinized salt solution. The adherence to the endothelial cell cultures of 51Cr-labeled human platelets stimulated at different thrombin concentrations was studied. There was significantly higher adherence of stimulated platelets to nonconfluent cultures compared with confluent ones. This was confirmed by scanning electron microscopy, which also revealed a tendency for the platelets to adhere at the cell periphery. Electron microscopy also showed that thrombin-stimulated platelets induced endothelial cell contraction. Part of the peripheral endothelial cell surface toward the bottom of the culture dish was inverted, facing the lumen of the dish. This phenomenon was particularly seen in nonconfluent cultures. When 51Cr-labeled endothelial cultures were incubated with a mildly injurious fluid as heparinized sodium acetate and 20% serum, at 20 degrees C for 30 min, the nonconfluent cultures showed significantly more cell detachment and release of 51Cr than the confluent ones. We conclude that under the conditions of the present experiments there are differences in the reactivity of confluent and nonconfluent endothelial cell cultures. These differences probably reflect biological dissimilarities. In experiments where properties of cultured endothelium are studied, care should be taken that the degree of confluency is standardized.

Urokinase binding to bovine corneal endothelial cells

Previously, we showed that the clotting factor thrombin binds to bovine corneal endothelial cells forming a covalent complex with a protein released by these cells into the culture media. We report that the plasminogen activator, urokinase, an enzyme involved in dissolution of blood clots, binds specifically to bovine corneal endothelial cells. [125I]-urokinase is rapidly bound by corneal endothelial cells. The serine active site of urokinase is required for binding since [125I]-urokinase inactivated with diisoprophylfluorophosphate does not bind to the cells. Pre-incubation of corneal cells with unlabeled urokinase for 20 hr results in increased release of binding sites for [125I]-urokinase into the culture media and at least a 3.5-fold increase in binding by the cells. [125I]-urokinase forms a 73 300 dalton sodium dodecyl sulfate complex with a corneal endothelial cell protein. Although thrombin competes with urokinase for binding to corneal endothelial cells, urokinase has at least a 10-fold higher affinity for binding to the cells. Furthermore, these cells make a plasminogen activator identical in molecular weight to urokinase. Thus, under physiological conditions, urokinase rather than thrombin binds to corneal endothelial cells. Binding may serve to regulate endogenous urokinase activity in the anterior chamber of the eye by limiting its extracellular proteolytic activity.

Evidence that cell surface heparan sulfate is involved in the high affinity thrombin binding to cultured porcine aortic endothelial cells

It has been postulated that thrombin binds to endothelial cells through, at least in part, cell surface glycosaminoglycans such as heparan sulfate, which could serve as antithrombin cofactor on the endothelium. In the present study, we have directly evaluated the binding of 125I-labeled bovine thrombin to cultured porcine aortic endothelial cells. The thrombin binding to the cell surface was rapid, reversible, and displaced by enzymatically inactive diisopropylphosphoryl-thrombin. The concentration of thrombin at half-maximal binding was approximately 20 nM. Both specific and nonspecific binding of 125I-thrombin to the endothelial cell surface was partially inhibited in the presence of protamine sulfate, after the removal of cell surface heparan sulfate by the treatment of cells with crude Flavobacterium heparinum enzyme or purified heparitinase. The binding as a function of the concentration of thrombin revealed that the maximal amount of specific binding was reduced by approximately 50% with little alteration in binding affinity by these enzymatic treatments. The reversibility and active-site independence as well as the rate of the binding did not change after heparitinase treatment. Whereas removal of chondroitin sulfates by chondroitin ABC lyase treatment of cells did not affect the binding, identical enzymatic treatments of [35S]sulfate-labeled cells showed that either heparan sulfate or chondroitin sulfate was selectively and completely removed from the cell surface by heparitinase or chondroitin ABC lyase treatment, respectively. Furthermore, proteolysis of cell surface proteins by the purified glycosaminoglycan lyases was excluded by the identical enzymatic treatments of [3H]leucine-labeled or cell surface radioiodinated cells. Our results provide the first direct evidence that heparan sulfate on the cell surface is involved in the high-affinity, active site-independent thrombin binding by endothelial cells, and also suggest the presence of thrombin-binding sites that are not directly related to heparan sulfate.

Correlation between trypsin binding to a specific receptor and prostacyclin production in cultured vascular endothelial cells

The correlation between the binding and processing of trypsin and its effect on prostacyclin (PGI2) production in cultured adult bovine aortic endothelial (ABAE) cells was studied. ABAE cells demonstrated an ability to produce PGI2 in a dose-response manner to trypsin at the range of 0.1-2.0 micrograms/ml with a saturation at a concentration of 1 microgram/ml. Likewise, 125I-trypsin binding to the cultured cells increased in a dose-response way and reached saturation at a concentration of about 1 microgram/ml; 125I-trypsin was bound to a specific high-affinity cell-surface receptor with a dissociation constant (Kd) of 1.5 X 10(-8) M and each of the confluent ABAE cells has about 1.2 X 10(5) such receptors sites. The cell-surface receptor for trypsin displayed specific characteristics and an excess amount of unlabeled trypsin successfully abolished 125I-trypsin binding while thrombin in excess failed to compete for 125I-trypsin binding. Only a small fraction of the cell-surface-bound 125I-trypsin was internalized and subsequently degraded by ABAE cells as compared to the process of 125I-trypsin internalization by human skin fibroblasts (HSF). This study demonstrated that the stimulatory effect of trypsin on prostacyclin production and release by ABAE cells might be mediated by a specific cell-surface receptor for trypsin on these cells distinct from the thrombin receptor.

The regulation of tissue plasminogen activator activity by human fibroblasts

We have found that live and ethanol-fixed fibroblasts, when covered with conditioned medium containing tissue plasminogen activator, associate with the enzyme and remove it from the medium. Binding of tissue plasminogen activator to fixed cells showed equilibrium kinetics with maximal uptake corresponding to 2.4 units of enzyme per 10(6) fixed cells. Enzyme bound to fixed cells could activate plasminogen and produce plaques of caseinolysis in casein-plasminogen-agar overlays. Electrophoretic analysis showed it covalently attached to a fibroblast component with a molecular weight of 40,000-50,000. Sequestration of tissue plasminogen activator by live fibroblasts showed nonsaturable first order kinetics with a rate constant of 0.465/hr. We conclude that active enzyme is bound to a surface receptor, then internalized and degraded. Fibroblasts did not release the binding molecule into the medium; binding of tissue plasminogen activator from the medium was unaffected by heparin or thrombin. This phenomenon differs from that described by Baker et al. and ascribed to "proteasenexin."

Variables controlling somatomedin production by cultured human fibroblasts

Somatomedin is secreted by multiple types of cultured cells including human fibroblasts. Since the control of somatomedin (Sm) production by fibroblasts may be an important regulatory step in cell division, we undertook studies to define the variables in tissue culture experimental design that may have significant effects on the secretion of Sm. Cell density was an important parameter in determining basal Sm production rates. Cultures plated at 2.5 X 10(4) cells/well produced 0.38 +/- 0.06 U/ml/10(5) cells whereas an increase in culture density to 6.2 X 10(4) cells/well was associated with a decrease in Sm production per cell to 0.23 +/- 0.04 U/ml/10(5) cells (P less than .01). Cultures stimulated by platelet-derived growth factor (PDGF) showed a similar reduction in Sm production with increasing cell density. Epidermal growth factor was stimulatory (5.4-fold increase) in low-density cultures but had no effect when high-density cultures were used. If the experiments were initiated between 2 and 4 days after the last media change there were no significant differences in basal or PDGF-stimulated Sm concentrations. Between days 5 and 9 however, there was a progressive increase in the basal Sm production rate. The duration of incubation was an important variable since Sm production increased during the first 12 h in noncycling cells and showed an accelerated increase between 4 and 8 h in cycling cells. Cells between the eighth and 12th passage had similar basal Sm production (0.22 +/- 0.04 U/ml/10(5) cells) rates; cells between the 19th and 20th passages had significantly higher basal Sm production rates (0.41 +/- 0.05 U/ml/10(5) cells) (P less than .01). These results suggest that several variables, particularly cell density and passage number, are critical variables when quantitating the effect of hormones and growth factors on Sm production by cultured fibroblasts.

Binding of coagulation factors IX and X to the endothelial cell surface

Bovine coagulation factors IX and X bind to independent sites on bovine aortic endothelial cells. Binding studies with cells maintained serum-free showed that there are at least two classes of binding sites for factor IX and factor X with a dissociation constant of 4.9 x 10(-9) M and 2.1 x 10(-8) M for the respective high affinity sites. Ca+2 was required for specific binding and was reversed by addition of EDTA or EGTA. Competition experiments showed that factor IX and factor IXa bind to the same sites, which are different from the factor X binding sites. Neither binding of factor IX or factor X is inhibited by addition of prothrombin or protein C. Indirect immunofluorescence of factor IX indicated that binding was diffuse on the cell surface.