Search

Endothelial cell (EC) interaction with antigraft antibodies (Abs) mediates EC injury and activation involved in vascular graft rejection. The aim of this study was to identify EC genes regulated in response to antigraft Ab binding that contribute to the endothelium alterations implicated in graft rejection or survival. By means of RNA differential display, 13 cDNA fragments corresponding to genes differentially expressed in ECs incubated with antigraft Abs were identified. Among these cDNAs were found the tissue inhibitor of metalloproteinase-1 (TIMP-1) and a desintegrin and metalloproteinase (ADAM-10). We demonstrated that TIMP-1 and ADAM-10 mRNA and protein expression was rapidly upregulated in ECs in response to antigraft Ab binding. Our data showed that TIMP-1 was upregulated in response to human IgG but not IgM and anti-galactosyl (Gal) alpha1-3Gal human xenogeneic Abs. In contrast, upregulation of ADAM-10 in ECs was shown to be mostly mediated by anti-Galalpha1-3Gal IgM Abs. Specific effects of human IgG and IgM xenogeneic Abs on endothelial transcripts indicate that different isotypes and specificities of Abs may mediate different EC changes. Our results suggest that interaction of ECs with antigraft Abs, according to their specificity, selectively induces synthesis and release of metalloproteinases and inhibitors, controlling proteolytic processes and immunological events that respectively contribute to graft rejection or survival.

Antiendothelial cell antibodies (AECA) have been detected in the sera of patients of autoimmune diseases showing vasculitis. Using IgM-ELISA, we found AECA in 42 (56%) of 75 sera samples from patients with Behçet's disease in a previous study. All of the 15 AECA-positive sera of Behçet's disease patients had an increased expression of the intercellular cell adhesion molecule-1 (ICAM-1), 93.3% of the sera induced the vascular cell adhesion molecule-1 (VCAM-1), and 100% of the serum induced the E-selectin molecule on human dermal microvascular endothelial cells (HDMEC). After stimulation of HDMEC with AECA-positive sera of Behçet's disease patients, the expression of ICAM-1 and VCAM-1 on HDMEC increased significantly at 4 hours, reaching a peak at 16 hours. Expression of E-selectin was induced at 1 hour after stimulation with a peak at 4 hours and it decreased thereafter. Adherence of T lymphocytes to HDMEC increased significantly after stimulation with AECA-positive sera from Behçet's disease patients. Also, the adherence of T lymphocytes to HDMEC increased at 4 hours and returned to its normal level at 48 hours. These results show that AECA-positive sera of Behçet's disease patients are capable of activating HDMEC to promote the adherence of T lymphocytes to increase the expression of ICAM-1, VCAM-1, and E-selectin on the cell surfaces. The whole process may play an important role in the pathogenesis of vasculitis in Behçet's disease.

Background: Natural killer (NK) cells are major component of the cellular response in xenotransplantation. NK cell activation and NK cell-mediated cytotoxicity can be a direct barrier to the potential use of xenogeneic organs in human transplantation., Methods: To determine if HLA-G would protect porcine xenogeneic cells from human NK cell lysis, human full-length HLA-G genomic DNA was transfected into porcine aortic endothelial cell (PAECs) by the lipofection method. Surface expression of HLA-G in transfected PAECs was confirmed by immunofluorescense staining with anti-HLA class I framework antibody, PA2.6. Fresh human peripheral blood lymphocytes were used as NK effector cells with HLA-G-transfected PAECs as targets in a 51Cr release assay. The inhibition of human polyclonal NK cells by HLA-G expression on PAECs was confirmed by antibody blocking using purified F(ab')2 portion of anti-human HLA class I antibody PA2.6., Results: Expression of HLA-G on PAECs conferred a significant protection against NK-mediated lysis (range: 52-100% inhibition) when peripheral blood lymphocytes from seven healthy donors, bearing either homozygous HLA-Cw3 or -Cw4 used as NK effector cells. Such protection was inhibited by purified F(ab')2 portion of anti-HLA class I antibody, indicating that the protection of PAECs was directly mediated by HLA-G expression., Conclusions: Expression of HLA-G on PAECs protected xenogeneic PAECs against human polyclonal NK cell-mediated lysis. These results indicate that the expression of HLA-G alone in the absence of other nonclassical HLA class I molecules is sufficient to inhibit human NK cell lysis. These findings suggest methods utilizing HLA-G to overcome NK cell-mediated cytotoxicity against porcine endothelial cells, considered the first cell type effected during xenograft cellular rejection.

Coronary bypass vessels, saphenous vein (SV) and internal thoracic artery (ITA), differ in susceptibility to atherosclerosis and medium- to long-term patency. Whereas most ITA remain patent (90% at 10 years), 20% of SV grafts fail in the first year and approximately 45% fail within 10 years. Reasons for these differences are not fully understood. Loss of SV patency may reflect early metabolic events, particularly increased proteoglycan (PG) synthesis which contributes to intimal volume and promotes atherogenesis through retention of atherogenic lipoproteins. We determined, in vitro, the PG metabolic activity of SV, ITA, and human coronary arteries through autoradiographic detection of incorporated [3H]glucosamine. SV had significantly higher levels of PG synthesis than ITA, especially in the subendothelial zone and after time (7 days) in culture. Patterns of synthesis in coronary vessels were similar to SV with high levels of incorporation in the subendothelial zone of thickened intima (> 100 microm). Increased subendothelial labelling in SV was due to increased PG synthesis, not decreased degradation. ITA showed no propensity for upregulation of subendothelial PG synthesis. Immunohistochemistry showed TGF-beta1 and TGF-beta2 localised primarily to the subendothelial zone of SV and coronary arteries. With time in culture immunostaining increased in parallel with increased PG synthesis. Subendothelial TGF-beta1 and TGF-beta2 were absent in ITA. A panspecific TGF-beta neutralising antibody reduced subendothelial PG synthesis in SV and coronary arteries by 50 and 60%, respectively. These results support the idea that vessels susceptible to atherosclerosis show increased accumulation of subendothelial PG mediated by TGF-beta.

Granulocyte recruitment during the acute inflammatory response is initiated by their rolling along the endothelial lining of postcapillary venules. To determine whether expression of L-selectin alone is sufficient for rolling, the murine pre-B lymphocytic cell line 300.19, which does not bind E- or P-selectin, was transfected with human L-selectin cDNA, which led to stable L-selectin expression at a level similar to that of blood lymphocytes. Fluorescent-labeled cells were infused retrogradely into a side branch of the superior mesenteric artery of anesthetized rats. In venules of the mesenteric membrane, leukocyte rolling occurs without intentional stimulation. On average, 17% +/- 6% of L-selectin transfectants rolled in the observed venules, while mock-transfected cells did not roll. Rolling of L-selectin transfectants began approximately 20 minutes after surgery and continued for at least 120 minutes. In contrast, HL-60 promyelocytes, which express sialyl-Lewis(x) tetrasaccharide (sLe(x)) but not L-selectin and that bind to E- and P-selectin in vitro, did not roll between 20 and 120 minutes, but some HL-60 cells rolled at very early (< 20 minutes) and late (> 2 hours) time points. Pretreatment with either of two function-blocking monoclonal antibodies recognizing the lectin domain of L-selectin completely blocked rolling of L-selectin transfectants and sharply reduced (by 70%) rolling of isolated human granulocytes. Taken together, these results show that L-selectin can mediate leukocyte rolling by virtue of its lectin activity.

We investigated the role of cadherins in the solute barrier maintained by endothelial cells in vitro. Cell-column chromatographic measurement of endothelial barrier showed that reducing normal extracellular calcium from 1.2 to 0.12 mM increased endothelial permeability to 250% of baseline after 20 min. Restoring normal calcium restored the barrier within 15 min which remained stable for at least 60 min. We used sulfo-NHS-biotin and anti-cadherin antibodies to characterize endothelial proteins with possible roles in the maintenance of endothelial barrier. The non-specific probe sulfo-NHS-biotin identified at least ten endothelial cell surface proteins, with greatest labelling occurring at molecular weights of 125 and 145 kD. Six proteins, including the 125 and 145 kD proteins, associated with the cytoskeleton. Western blotting for the presence of classical cadherins containing the conserved cytoplasmic sequence CDPTAPPYDSLLVFDYEG detected two bands at 145 and 125 kD which associated with the cytoskeleton. Western blotting with an antibody, which recognizes FHLRAHAVDINGNQV, an extracellular homotypic binding region of N-cadherin, detects three bands. Of these three, one protein had a molecular weight of 125 kD and was associated with the cytoskeleton. Immunofluorescence with both N-cadherin and anti-peptide 1 antibodies found staining at endothelial cell borders. The utility of a newly developed cell-column calcium switch assay was tested by verifying the functional role of the previously described epithelial cadherin, uvomorulin, in epithelial barrier. We then applied this method to endothelial cell columns and found the N-cadherin antibody interfered with the reforming of interendothelial junctions. These results suggest that, as in epithelial cells, cadherins in bovine endothelial cells have a functional role in forming the calcium sensitive endothelial junction and may play an important role in the formation of normal barrier.

We report the results of studies performed in vitro and in vivo that were designed to explore individual, sequential, and concurrent Ag-antibody interactions at the surface of rabbit endothelial cells. Divalent heterologous antibodies to rabbit lung angiotensin-converting enzyme and to rabbit lung thrombomodulin were employed. In cultured monolayers, both antibodies redistributed the specific Ag and co-redistributed the immunologically unrelated Ag inducing partial or complete disappearance of the Ag from the cell surface (antigenic modulation) in 15 to 60 min. When injected into living rabbits, each antibody induced a rapid (1 to 3 min) redistribution and subsequent modulation of the specific and of the unrelated Ag at the surface of alveolar endothelial cells. Immune complexes, and the unrelated Ag, were shed in the circulation, attaining peak levels 3 to 4 min after the injection; were rapidly bound by platelets, E, and polymorphonuclear leukocytes; and were subsequently found in phagocytic cells in the spleen and in the liver. Thrombomodulin co-shed by angiotensin-converting enzyme antibody and, to a lesser degree, angiotensin-converting enzyme co-shed by thrombomodulin antibody, crossed the glomerular capillary walls and were reabsorbed by the epithelial cells of the proximal tubules within 2 to 3 min. The results show that immunologically unrelated Ag can be passively entrapped during formation of immune complexes at the cell surface, and provide new information on the kinetics of clearance of immune complexes containing endogenous, structural Ag.

Albumin, through its binding to the endothelial glycocalyx, functions as a major determinant of capillary permeability and as a carrier for various small molecules in its transcytosis across continuous endothelium via plasma-lemmal vesicles. Several albumin-binding proteins (ABP) have been identified: three membrane-associated ABP, which we call gp60, gp30, and gp18, and one secreted protein, acidic and rich in cysteine (SPARC). In this study, we used antiserum raised against bovine SPARC (BON) to investigate the possible interrelationships among ABP to better understand their role in binding and transcytosis. BON not only interacted with SPARC secreted by cultured endothelium but also recognized gp60 in lysates of cultured rat, human, and bovine endothelial cells. Purified SPARC inhibited BON interaction with gp60. BON immunoglobulin (Ig)G specifically inhibited albumin binding to both SPARC and gp60 extracts. This effect was eliminated by preabsorption of BON to immobilized SPARC. BON also significantly inhibited albumin binding to cultured microvascular endothelial cells via its interaction with gp60. Anti-SPARC peptide sera were also tested, and one serum raised against a peptide encompassing an NH2-terminal region of SPARC recognized both SPARC and gp60 but did not inhibit albumin binding; gp30 and gp18 were not recognized by any of these anti-SPARC antibodies. These results suggest that SPARC and gp60 are functionally and immunologically related ABP that may share a common albumin-binding domain. gp60 appears to be the major mediator of albumin binding to microvascular endothelium.

Anti-myeloperoxidase autoantibodies are found in association with idiopathic necrotizing glomerulonephritis and systemic vasculitis. It is not known if their presence is an epiphenomen or an integral part of the pathogenic process. To further delineate their hypothesized pathogenicity, we studied their ability to stimulate neutrophils to damage human umbilical vein endothelial cells in vitro. Anti-myeloperoxidase antibodies from human, rabbit and mouse sources were utilized. These antibodies stimulated neutrophils to damage endothelial cells as determined by 51Cr release. The effect was dependent on priming the neutrophils with tumor necrosis factor-alpha, and further enhanced with the addition of endotoxin. The amount of endothelial cell damage was dependent on the dose of anti-myeloperoxidase, the source of the neutrophils, the concentration of TNF, and the presence of endotoxin. Under identical conditions, control antibodies did not stimulate neutrophils to damage endothelial cells. The effect was confirmed by labeling the endothelial cells with 3H-adenine which yielded the same results. These results provide further in vitro evidence that anti-myeloperoxidase autoantibodies may play a significant role in the pathogenesis of idiopathic pauci-immune glomerulonephritis and vasculitis.

The activation of endothelial cells is thought to contribute to the host response to infection and to the pathogenesis of autoimmune disease. It was recently shown that antibody and complement can activate endothelial cells leading to cleavage and release of heparan sulfate from the cells. We show here that release of heparan sulfate from endothelial cells is mediated by antibody and the complement fragment C5a and that assembly of the membrane attack complex and lysis of endothelial cells is not necessarily involved. These data suggest that the generation of C5a in conditions such as autoimmunity and infection in which anti-endothelial cell antibodies may also be present, might amplify tissue injury by a novel mechanism involving endothelial cell activation and loss of heparan sulfate mediated by antibody and C5a.

The effects of direct cell contact between endothelial (ECs) and mesangial cells (MCs) on MCs proliferation were examined in a coculture system in vitro. Mitomycin C treated ECs (M-ECs) were plated on culture dishes and MCs were cocultured with these M-ECs. Cell number was measured at the end of days 1, 3, and 5. In the coculture system with direct contact, the growth of cocultured MCs was modulated as follows: 1) the growth of MCs was inhibited up to day 3, and 2) a high level of proliferation was observed between days 3 and 5. This biphasic pattern of growth could not be detected in coculture of fibroblasts with MCs. In coculture without direct contact, using intercup chambers, the kinetics in cell proliferation between cocultured MCs and MCs alone were essentially the same. Conditioned media derived from cocultures up to day 3 in a contact-dependent manner inhibited the 3H-thymidine uptake of MCs. From these results, it would thus appear that MCs proliferation is regulated by intercellular contact with ECs.

Rocky Mountain spotted fever (RMSF) takes its name from the characteristic rash that occurs as a consequence of vasculitis associated with rickettsial invasion of the endothelium. The authors examined sera from 14 patients with serologically confirmed RMSF for the presence of antibodies (IgG and IgM) reactive with human umbilical vein-derived endothelial cells and with the phospholipids cardiolipin (CL) and phosphatidyl serine (PS). Sera from 7 patients (50%) exhibited antiendothelial antibodies, and 10 (71%) patient sera reacted with CL and/or PS. Because such antibodies may interfere with or augment endothelial thrombosis-related activities, acute and convalescent sera were tested for their effects on endothelial PGI2 secretion and protein C activation. Acute sera from two patients and convalescent sera from four patients stimulated protein C activation. Additionally, sera from five acute and nine convalescent cases inhibited basal endothelial PGI2 secretion, but sera from two acute and three convalescent cases stimulated thrombin-dependent PGI2 secretion. These results demonstrated that, in a significant proportion of patients, RMSF was accompanied by the appearance of antibodies that bound to endothelial cells and to phospholipids; some of these antibodies may have altered anticoagulant endothelial functions.

The aqueous and vitreous humors are normally avascular and may be sites of altered inflammatory responses. Inhibitors of endothelial cell and lymphocyte growth as well as of angiogenesis have been isolated from vitreous and an inhibitor of lymphocyte mitogenesis has been found in aqueous. Data are presented here indicating that aqueous humor also inhibits endothelial cell growth. In addition, neutralizing antibodies against TGF beta block the effects of materials isolated from these tissues on endothelial cell and lymphocyte growth. Similar data were obtained from aorta, another normally avascular site. These observations suggest that the avascularity and altered inflammatory responses in these sites may be regulated in part by TGF beta.

Endothelial cells (EC) synthesize and secrete von Willebrand factor (vWF), a multimeric glycoprotein required for normal hemostasis. Within human endothelial cells, vWF multimers of extremely high molecular weight are stored in rod-shaped organelles known as Weibel-Palade bodies. Inflammatory mediators, such as interleukin-1, induce in vitro a variety of procoagulant responses by EC, including the secretion of stored vWF. We postulated that other inflammatory mediators might act to balance this procoagulant reaction, thereby assisting in the maintenance of blood fluidity during immune activation. Both gamma-interferon (gamma-IFN) and tumor necrosis factor (TNF) were found to act independently and cooperatively to depress the stimulated release of vWF from EC. Analysis of stored vWF in either gamma-IFN and/or TNF-treated EC demonstrated a loss of high molecular weight multimers while immunofluorescent studies documented a loss of visible Weibel-Palade bodies. This suggests that gamma-IFN and TNF interfere with normal vWF storage. gamma-IFN acted in a dose-, time-, and RNA-dependent fashion, and its inhibition of vWF release was reversible with time. No effect of gamma-IFN on EC was noted when anti-serum to gamma-IFN was added. Unlike gamma-IFN, alpha-interferon did not effect EC vWF. Therefore, gamma-IFN and TNF may be important in decreasing vWF release during inflammatory or immunologic episodes.

A cellular enzyme-linked immunosorbent assay (ELISA) was used to detect circulating anti-endothelial cell antibodies (AECA) in sera of patients suffering from rheumatoid arthritis (RA). Felty's syndrome (FS), systemic lupus erythematosus (SLE), progressive systemic sclerosis (PSS) and those with a "lupus anticoagulant" (LA). IgG AECA were detected in RA, FS, SLE and LA sera, while IgM AECA were only detected in RA and FS. AECA were not specific for endothelial cells. However, IgG binding to endothelial cells and dermal fibroblasts was F(ab) mediated while in all other cell types tested, nonspecific binding most likely via the Fc region occurred. In RA and FS rheumatoid factor was shown to augment immunoglobulin binding to endothelial cells. Additional studies revealed that some of these pathological sera also contained cytotoxic IgG autoantibodies which fixed complement and damaged human umbilical vein endothelial cells in vitro. These studies suggest that a group of autoantibodies, present in a variety of collagen vascular disorders, react with endothelial cells and thus may be important aetiopathogenic factors in the vasculopathies associated with these disorders.

In blood coagulation on endothelium, an unperturbed vascular endothelial cell surface apparently provides activity equivalent to the phospholipid needed for generation of factor Xa and thrombin in soluble systems. Phospholipid in soluble systems also markedly enhances the ability of factor Xa to activate factor VII; therefore we investigated the influence of an unperturbed monolayer of human umbilical vein endothelial cells (HUVEC) upon factor VII activation. HUVEC were found to augment factor Xa-catalyzed activation of factor VII. This appeared to result from the binding of trace amounts of factor Xa to the cells. Adding active site-inhibited factor Xa to reaction mixtures, but not factor X, abolished the enhanced activation. Adding either anti-factor V antibodies or exogenous factor Va had no effect upon reaction rates. Thus factor Va does not function as a cofactor for the reaction. In further experiments the effect upon activation of factor VII and prothrombin was studied by varying the order of addition of factor Xa and factor Va to supernatants of HUVEC monolayers. Evidence was obtained that HUVEC, unlike platelets, possess a functional factor Xa binding site that is independent of factor Va. Since phospholipid is the only known cofactor for factor Xa/Ca2+-induced activation of factor VII, the demonstration of enhanced activation of factor VII in the presence of unperturbed cultured HUVEC supports a hypothesis that the functional equivalent of procoagulant phospholipid is available on the luminal surface of vascular endothelium in vivo.

Traditionally, the human lymphocyte antigens have been considered to be the major barrier to successful transplantation, and lymphocytes have been used as the target cell in evaluating histocompatibility. The presence in the serum of recipients of preformed antibodies, cytotoxic to donors lymphocytes, is associated with a high probability of hyperacute rejection. We identified 11 patients in whom, despite a compatible direct lymphocytotoxic cross-match, acute failure of the cardiac homograft was associated with histologic and immunologic findings consistent with hyperacute rejection. Direct immunofluorescence and immunohistochemical staining showed the presence of antibodies on the surface of vascular endothelial cells in each of these 11 patients. The serum of these recipients was found to contain antibodies against a panel of endothelial cells. In contrast, cytotoxic antibodies to vascular endothelial cells were not present in a control group of 18 heart transplant recipients who did not experience hyperacute rejection. Thus the presence of antibodies against vascular endothelial cells seems to be related to hyperacute rejection of the cardiac allograft.