Rejection of pig liver xenografts in patients with liver failure: implications for xenotransplantation.

The pathophysiological state of rejection in liver xenotransplantation is poorly understood. Data from clinical pig liver perfusion suggest that pig livers might be rejected less vigorously than pig hearts or kidneys. Pig livers used in clinical xenoperfusions were exposed to blood from patients with liver failure. We have shown in an animal model that transplant recipients with liver failure are less capable of initiating hyperacute rejection of a xenografted liver than a healthy transplant recipient. The goal of this report is to examine the pathological characteristics of pig livers used in 2 clinical pig liver perfusions and combine this information with in vitro studies of pig-to-human liver xenotransplantation to determine whether the findings in the perfused pig livers could be explained in part by the diminished capacity of the patient with liver failure to respond to xenogeneic tissue. Pathological analysis of the perfused pig livers showed immunoglobulin M deposition in the sinusoids with little evidence of complement activation. Our in vitro studies showed that serum from patients with liver failure caused less injury to pig liver endothelium than serum from healthy subjects. Serum from patients with liver failure had similar levels of xenoreactive antibodies as serum from healthy humans. Incubation of serum from patients with liver failure with pig hepatic endothelial cells generated less iC3b, Bb fragment, and C5b-9 than serum from healthy subjects. We conclude that the altered injury in the perfused pig livers can be attributed to the relative complement deficiency that accompanies liver failure.

Development of an in vitro blood-brain barrier.

The blood-brain barrier represents a significant protective barrier for the passage of pathogens and toxicants, but it is difficult to study in vivo. This unit describes an in vitro system that can be used as a model for the blood-brain barrier to study its functions in an accessible format.

Interactions between HIV-infected monocyte-derived macrophages and human brain microvascular endothelial cells result in increased expression of CC chemokines.

The presence of perivascular monocytic infiltration is a major hallmark of HIV-1-associated dementia. Since CC chemokines are chemoattractant cytokines that are able to attract T cells and monocytes/macrophages to sites of inflammation, and since infiltrating monocytes/macrophages remain in close contact with the brain endothelium, we investigated whether interactions between HIV-1-infected macrophages and brain endothelium result in an altered chemokine production. We found an increased mRNA expression of monocyte chemotactic protein-1 (MCP-1), macrophage inflammatory protein (MIP)-1 alpha and MIP-1 beta, and RANTES by macrophages after HIV-1 infection. Interactions between HIV-infected macrophages and brain microvascular endothelial cells resulted in an additional upregulation of chemokine mRNA expression, during cell-cell contact as well as in a trans-well system. Since IL-1 beta can function as a modulator of chemokine expression we investigated if interleukin-1 beta could be involved in the regulation of chemokine induction. Coculturing of HIV-infected macrophages and endothelial cells resulted in immune-activation as indicated by increased mRNA expression of IL-1 beta. Subsequently, addition of a neutralizing antibody against IL-1 beta resulted in altered chemokine expression by macrophages, but not by endothelial cells. Thus, IL-1 beta appears to play a major role in the regulation of chemokines during cellular interactions in HIV-associated dementia, but other factors may also be involved.

Human endometrial endothelial cells: isolation, characterization, and inflammatory-mediated expression of tissue factor and type 1 plasminogen activator inhibitor.

Binding of Ulex europaeus lectin to microvessels was used to isolate endothelial cells from cycling human endometrium. Cultured human endometrial endothelial cells (HEECs) exhibited endothelial cell-specific characteristics such as tube formation on a basement membrane matrix and sequestration of acetylated low-density lipoprotein. Markers for potentially contaminating epithelial, stromal, smooth muscle, and bone marrow-derived cells were not detected in the HEEC cultures. Basal and proinflammatory-stimulated immunostaining profiles for endothelial cell-specific adhesion markers, as exemplified by Von Willebrand's factor and E-selectin, were similar for cultured HEECs and human umbilical venous cord endothelial cells (HUVECs). However, HUVECs expressed several extracellular matrix proteins that were absent from cultured HEECs. In the latter, the protein kinase C agonist phorbol myristate acetate transiently enhanced tissue factor (TF) mRNA levels and elicited a more prolonged elevation in TF protein levels, but did not affect plasminogen activator inhibitor-1 (PAI-1) mRNA and protein levels. Inappropriate expression of TF, which initiates hemostasis by generating thrombin, and of PAI-1, which regulates hemostasis by acting as the primary inhibitor of fibrinolysis, can each lead to thrombosis. The differential regulation of TF and PAI-1 expression revealed in the current study emphasizes the importance of using HEECs to evaluate mechanisms regulating the hemostatic/thrombotic balance in human endometrium.

Complement activation in discordant hepatic xenotransplantation.

Little is known about hyperacute rejection in hepatic xenotransplantation. Information from clinical xenoperfusions suggests that the liver may be rejected by a mechanism less vigorous than either kidney or heart xenografts. We used the in vitro model of porcine hepatic sinusoidal endothelial cells (PHEC) incubated with either complement replete or deficient human serum to determine the relative roles of the classical and alternate pathways of complement in the immediate response to hepatic xenotransplantation. Our results suggest that either the classical or alternate pathways are capable of independently activating the complement cascade upon exposure to the porcine hepatic sinusoidal endothelium. Our results also imply that either pathway alone is capable of initiating similar degrees of injury as the entire cascade.

Thrombotic thrombocytopenic purpura and sporadic hemolytic-uremic syndrome plasmas induce apoptosis in restricted lineages of human microvascular endothelial cells.

Thrombotic thrombocytopenic purpura (TTP) and sporadic hemolytic-uremic syndrome (HUS) are thrombotic microangiopathies that occur in the absence of an inflammatory response. Ultrastructural features of tissues involved in TTP/sporadic HUS suggest an apoptotic process. Consistent with these findings, we observed that TTP plasmas induce apoptosis in primary human endothelial cells (EC) of dermal microvascular but not umbilical vein origin (Laurence et al, Blood 87:3245, 1996). We now document the ability of plasmas from both TTP and sporadic HUS patients, but not from a patient with childhood/diarrhea-associated HUS, to induce apoptosis and expression of the apoptosis-associated molecule Fas (CD95) in restricted lineages of microvascular EC. EC of small vessel dermal, renal, and cerebral origin were susceptible to induction of Fas and an apoptotic cell death. In contrast, microvascular EC of pulmonary and hepatic origin, as well as EC of a large vessel, coronary artery, were resistant to both processes. This dichotomy parallels the in vivo pathology of TTP/sporadic HUS, with notable sparing of the pulmonary and hepatic microvasculature. Apoptotic EC also had some features of a procoagulant phenotype, including depressed production of prostaglandin I2 (prostacyclin). These phenomena support the pathophysiologic significance of microvascular EC apoptosis in TTP, extend it to a related disorder (sporadic HUS), and suggest consideration of apoptosis inhibitors in the experimental therapeutics of these syndromes.

Mechanisms for the transendothelial migration of HIV-1-infected monocytes into brain.

HIV-1 penetration of the brain is a pivotal event in the neuropathogenesis of AIDS-associated dementia. The establishment of productive viral replication or up-regulation of adhesion molecule expression on brain microvascular endothelial cells (BMVEC) could permit entry of HIV into the central nervous system. To investigate the contribution of both, we inoculated primary human BMVEC with high titer macrophage-tropic HIV-1 or cocultured them with virus-infected monocytes. In both instances, BMVEC failed to demonstrate productive viral replication. Cell to cell contact between monocytes and microvascular endothelium resulted in E-selectin expression on BMVEC. BMVEC. cocultured with LPS-activated HIV-infected monocytes expressed even higher levels of E-selectin and vascular cell adhesion molecule-1 (VCAM-1). Transwell assays supported a role of soluble factors, from virus-infected monocytes, for the induction of adhesion molecules on BMVEC. To verify the in vivo relevance of these findings, levels of adhesion molecules were compared with those of proinflammatory cytokines and HIV-1 gene products in brain tissue of AIDS patients with or without encephalitis and HIV-seronegative controls. E-Selectin, and to a lesser degree VCAM-1, paralleled the levels of HIV-1 gene products and proinflammatory cytokines in brain tissue of subjects with encephalitis. Most importantly, an association between macrophage infiltration and increased endothelial cell adhesion molecules was observed in encephalitic brains. Monocyte binding to encephalitic brain tissue was blocked with Abs to VCAM-1 and E-selectin. These data, taken together, suggest that HIV entry into brain is, in part, a consequence of the ability of virus-infected and immune-activated monocytes to induce adhesion molecules on brain endothelium.

HIV-1 infection alters monocyte interactions with human microvascular endothelial cells.

HIV infection of monocytes resulted in twofold elevation of adhesion molecule LFA-1 (both alpha L/CD11a and beta 2/CD18 subunits) and LFA-3 (CD58), with no apparent increase in LFA-2 (CD2) or various beta 1-integrins. Homotypic aggregation of monocytes was evident 2 h after exposure to virus and was inhibited by mAbs to both the alpha L- and beta 2-subunits of LFA-1. HIV-infected monocytes also showed a marked increase in adherence to human capillary endothelial cell monolayers derived from brain, lung, and skin. This adherence was inhibited by mAb to either LFA-1 subunit and by mAb to the counter-receptor intercellular adhesion molecule-1. Cocultivation of HIV-infected monocytes with endothelial cells increased permeability of endothelial cell monolayers to 125I albumin in transwell assay systems. The increased endothelial permeability induced by HIV-infected monocytes was associated with a substantial disruption of the endothelial cell monolayer. Morphologic disruption was not a direct toxic effect on endothelial cells, but appeared to be secondary to changes in endothelial cell-cell or cell-matrix interactions. Northern blot analysis showed increased expression of gelatinase B (92-kDa gelatinase), tissue inhibitor of metalloproteinase TIMP-1, and TIMP-2 in the HIV-infected monocytes. Consistent with these Northern analyses, secretion of gelatinase activity in culture fluids of HIV-infected monocytes was also increased and was dependent on the stage of virus replication. Incubation of HIV-infected monocytes with the proteinase inhibitors TIMP-1 and TIMP-2 inhibited the increased permeability of endothelial cell monolayers to 125I albumin. These results suggest possible mechanisms for extravasation of HIV-infected monocytes through vascular endothelium into tissue in early stages of HIV disease.

Group B streptococci (GBS) injure lung endothelium in vitro: GBS invasion and GBS-induced eicosanoid production is greater with microvascular than with pulmonary artery cells.

Neonatal group B streptococcal (GBS) sepsis and pneumonia cause lung endothelial cell injury. GBS invasion of the lung endothelium may be a mechanism for injury and the release of vasoactive eicosanoids. Pulmonary artery endothelial cells (PAEC) and lung microvascular endothelial cells (LMvEC) were isolated from neonatal piglets and were characterized as endothelial on the basis of morphology, uptake of acyl low-density lipoprotein, factor VIII staining, and formation of tube-like structures on Matrigel. PAEC and LMvEC monolayers were infected with COH-1 (parent GBS strain), isogenic mutants of COH-1 devoid of capsular sialic acid or all capsular polysaccharide, or a noninvasive Escherichia coli strain, DH5 alpha. Intracellular GBS were assayed by plate counting of colony-forming units resistant to incubation with extracellular antibiotics. All GBS strains invaded LMvEC significantly more than PAEC, showing that the site of lung endothelial cell origin influences invasion. DH5 alpha was not invasive in either cell type. Both isogenic mutants invaded PAEC and LMvEC more than COH-1 did, showing that GBS capsular polysaccharide attenuates invasion. Live GBS caused both LMvEC and PAEC injury as assessed by lactate dehydrogenase release; heat-killed GBS and DH5 alpha caused no significant injury. Supernatants from PAEC and LMvEC were assayed by radioimmunoassay for prostaglandin E2 (PGE2), the stable metabolite of prostacyclin (6-keto-PGF1 alpha), and the thromboxane metabolite thromoxane B2. At 4 h, live COH-1 caused no significant increases in eicosanoids from both PAEC and LMvEC. At 16 h, live COH-1, but not heat-killed COH-1, caused a significant increase in 6-keto-PGF1 alpha greater than PGE2 from LMvEC, but not PAEC. We conclude that live GBS injure and invade the lung microvascular endothelium and induce release of prostacyclin and PGE2. We postulate that GBS invasion and injury of the lung microvasculature contribute to the pathogenesis of GBS disease.

Group B streptococci invade endothelial cells: type III capsular polysaccharide attenuates invasion.

Group B streptococci (GBS) are the most common cause of neonatal sepsis and pneumonia. The pathogenesis of GBS disease is not completely defined. GBS-induced endothelial cell injury is suggested by histological findings at autopsy and in animal studies. We hypothesized that (i) type III GBS (COH-1) invade and injure human umbilical vein endothelial (HUVE) cells and (ii) isogenic mutations in GBS capsule synthesis would influence HUVE invasion. Confluent HUVE monolayers were infected for 0.5, 2, or 6 h. Media with penicillin plus gentamicin were added and incubated for 2 h to kill extracellular bacteria. Cells were washed and lysed, and the number of live intracellular bacteria was determined by plate counting. COH-1 invaded HUVE cells in a time-dependent manner at levels 1,000-fold higher than those of the noninvasive Escherichia coli strain but significantly lower than those of Staphylococcus aureus. There was no evidence for net intracellular replication of GBS within HUVE cells. COH-1 infection of HUVE cells caused the release of lactate dehydrogenase activity. GBS invasion was inhibited by cytochalasin D in a dose-dependent manner; GBS-induced lactate dehydrogenase release was attenuated by cytochalasin D. The isogenic strains COH 1-11, devoid of capsular sialic acid, and COH 1-13, devoid of all type III capsule, invaded HUVE cells three- to fivefold more than the parent COH-1 strain. The type III capsular polysaccharide and particularly the capsular sialic acid attenuate GBS invasion of HUVE cells. Electron micrographs of lung tissue from a GBS-infected newborn Macaca nemestrina also showed GBS within capillary endothelial cells. We conclude that endothelial cell invasion and injury are potential mechanisms in the pathogenesis of GBS disease.

Isolation, characterization, and propagation in vitro of human glomerular endothelial cells.

Human glomerular endothelial cells have been isolated, cloned, and characterized. They appeared as the first outgrowth from human glomeruli in the presence of platelet-derived growth factor, which was also a requirement for continuous growth. By phase microscopy they appeared as monolayers of polygonal cells. Von Willebrand's factor (VWF) was detected in the cytoplasm of all clones. Their intermediate filaments differed antigenically from that present in human umbilical vein endothelial cells. Like other endothelial cells, they demonstrated high levels of membrane-associated angiotensin-converting enzyme (ACE).

Linked marker analysis of spontaneous HLA variants of somatic cells.

Fourteen independent, spontaneously arising HLA-A1 and -B27 variant clones were isolated from the pseudodiploid B lymphoid line T5-1 by selection using A1 and B27 alloantiserum, respectively, and complement. T5-1 is heterozygous for the tightly linked loci HLA-DR, -B, and -A and probably for -C as well. Following recloning, each of the variants was tested for the HLA specificities of T5-1. None of these spontaneous variants had a genetic lesion which extended to the nearest flanking HLA marker, less than 1 cM distant. On the other hand, a variant isolated from mutagenized cells had a lesion which extended completely through one HLA region in cis, suggesting that haploid expression in the HLA region is compatible with viability, that there are no recessive lethals in the opposite HLA region, and that spontaneously arising variants with extended lesions could have been recovered had they occurred. From these results, we conclude that extended genetic lesions of 0.8 cM or longer contribute less than 10% to the overall rate of spontaneous variation for HLA alleles, which we previously estimated at about 5 x 10(-7) per cell per generation.

HLA variants of cultured human lymphoid cells: evidence for mutational origin and estimation of mutation rate.

Variants of a diploid lymphoid cell line that show a loss of HLA-B27 antigen occur randomly in time and independently of exposure to the alloantiserum used for their isolation. From these and previous findings of variant stability, inducibility by mutagens, and the absence of linked variation, we colclude that most HLA variants arise by mutation. The mutation rate for HLA-B27 loss is 8 x 10(-7) per cell per generation.

Isolation of stable and unstable ("suppressed") Beta2-microglobulin-deficient clones from a lymphoid cell line.

After selection with anti-Beta2m and C against a human lymphoid line, about 10% of surviving clones manifested decreased cytotoxic sensitivity to anti-Beta2m("suppression") which gradually reverted to normal after 35 to 55 cell doublings. Antiserum alone induced suppression. Exposure of suppressed clones to DMSO substantially reversed suppression. One clone, surviving selection, had a stable decrease in anti-Beta2m sensitivity; it possibly represents a Beta2m mutation.