Endothelial Cell Turnover Research Articles

Transendothelial macromolecular transport in the aorta of spontaneously hypertensive rats.

Leaky endothelial junctions occurring during cell turnover have been postulated to be a major pathway for enhanced lipoprotein transport across the vascular endothelial layer, which leads to the development of atherosclerosis. Because hypertension has been well documented as one of the major risk factors for atherosclerosis, we explored the possibility that hypertension accelerates atherogenesis by increasing the turnover of endothelial cells and hence the transendothelial macromolecular permeability. The investigations were performed on thoracic aortas of 10 male 3-4-month-old spontaneously hypertensive rats and eight male age-matched Wistar-Kyoto normotensive rats. In en face preparations of aortic specimens, mitotic endothelial cells were identified by hematoxylin nuclear staining; dying or dead endothelial cells containing cytoplasmic immunoglobulin G were detected by indirect immunoperoxidase technique; and endothelial leakage to Evans blue-albumin conjugate was visualized by fluorescence microscopy. The number of leaky foci per unit endothelial surface area in spontaneously hypertensive rats was found to be approximately three times that in Wistar-Kyoto control rats; the frequencies of both endothelial cell mitosis and death in spontaneously hypertensive rats were also approximately three times the corresponding values in Wistar-Kyoto rats. These findings indicate that hypertension in spontaneously hypertensive rats is accompanied by increased endothelial cell turnover and an attendant enhancement of permeability to macromolecules.

Macromolecular transport across arterial and venous endothelium in rats. Studies with Evans blue-albumin and horseradish peroxidase.

Atherosclerotic lesions are characterized by lipid infiltration in regions with high rates of endothelial cell turnover. The present investigation was designed to elucidate the route of macromolecular transport across vascular endothelium. The aorta and vena cava of male Sprague-Dawley rats were perfusion-fixed after the intravenous injection of Evans-blue albumin (EBA) or horseradish peroxidase (HRP). Fluorescence microscopic examination of en face preparation of the aorta stained with hematoxylin allowed the identification of endothelial cells that underwent mitosis, together with the localization and quantification of fluorescent spots for EBA leakage. The HRP specimens were subjected to histochemical treatment, and HRP leakage was seen as brown spots under the light microscope. Silver nitrate stain was added in both EBA and HRP studies to outline cell boundaries and to visualize stigmata, stomata, and dead cells. In the aorta, almost every dividing cell showed junctional leakage to albumin and HRP, with clustering of leaky spots around the branch orifices. Time-dependent studies showed gradual increases in the diameter and number of these heterogeneously sized leaky spots, which finally fused to sizes corresponding to the "blue areas" for EBA or "brown areas" for HRP. Compared with arteries, veins had fewer mitotic cells, but more dead cells and diffuse dye-staining areas, indicating a more rapid transport of macromolecules. The leaky spots in the artery were associated mainly with mitotic cells, dead cells, and stigmata, whereas those in the vein occurred primarily at regions with dead cells. These results suggest that the preferential association of the enhanced transport of macromolecules with mitosis in the arterial as compared to venous endothelium and the differential behavior in transmural transport between arteries and veins may form the basis for the predilection of atherosclerosis in arteries.

Histamine and bradykinin stimulate the phosphoinositide turnover in human umbilical vein endothelial cells via different G-proteins

The G-proteins which regulate hormonal turnover of phosphoinositide (PI) in human umbilical vein endothelial cells have been investigated. A 40–41 kDa doublet present in the membranes of these cells was selectively ADP ribosylated by pertussis toxin (PTx), and this doublet was G α 2 and G α 3 according to immunoblotting with specific antisera. By contrast, a doublet of 24–26 kDa proteins in the same membrane preparations was ADP ribosylated by the C 3 component of botulinum toxin (BoTx). PTx-dependent ADP ribosylation blocked stimulation of PI turnover by histamine, but did not affect stimulation by bradykinin, whereas BoTx (C 2 + C 3 components) had the opposite effect. Thus two different groups of G-proteins may be involved in hormone-dependent stimulation of PI turnover in human umbilical vein endothelial cells.

Regression of white adipose tissue in diabetic rats.

The effects of long-term diabetes (4 wk) on the development of parametrial (PWAT) and retroperitoneal (RWAT) white adipose tissues were studied in young Sprague-Dawley rats (170-200 g) injected with a single dose of streptozotocin (75 mg/kg). Diabetes stopped animal growth and totally abolished the normal increases in the wet weight, total protein content, and cellularity (estimated by DNA content) of PWAT and RWAT. Remarkably, the prolonged lack of insulin induced a progressive decrease of the cellularity of RWAT to levels that were lower than those of the initial controls. It also resulted in a marked reduction of adipocyte size. The tiny adipocytes seen in diabetic animals were characterized by the presence of multilocular triglyceride droplets. In general, the decreases in cell number, cell size, and protein content were more pronounced in RWAT than in PWAT. Quantitative cellular frequency studies revealed that adipocytes, and possibly also endothelial cells, contribute to the decrease in RWAT cellularity. The results demonstrate that 1) diabetes inhibits proliferative activity in adipose tissue, 2) total cell number reduction may occur in adipose depot of young growing rats, 3) this effect is depot dependent, and 4) the turnover of adipocytes and endothelial cells is relatively slow (several weeks).

Transendothelial transport of low density lipoprotein in association with cell mitosis in rat aorta.

Atherosclerosis is characterized by focal areas of lipid accumulation and intimal smooth muscle cell proliferation in large arteries. In vivo studies on rat aorta with Evans blue-albumin conjugate (EBA) have shown that there are preferential sites of increased permeability with an increased uptake of the conjugate. It has been shown that these blue areas are associated with a high endothelial cell turnover rate and an enhanced permeability to lipids. In a previous study, we demonstrated that 99% of endothelial cells in the mitotic (M) phase as identified by hematoxylin staining of the dividing nuclei exhibited EBA leakage and that these dividing cells accounted for 30% of all leakage sites. In the present study, experiments were performed on the thoracic aortas of 10 adult male Sprague-Dawley rats to determine the statistical frequency of isolated leaks to Lucifer yellow-low density lipoprotein conjugate (LY-LDL) at the level of individual cells and to assess the relationship of such leaks to the cell turnover processes. Leakage of LY-LDL around individual endothelial cells was visualized by fluorescence microscopy, and cells in mitosis on the same specimens were identified by hematoxylin staining. Although endothelial cell mitosis is infrequent (0.034%), 80% of dividing cells in the M phase were associated with LY-LDL leakage. These dividing cells accounted for 45% of all leakage spots. These findings lend support to our recent hypothesis that transiently open junctions surrounding the endothelial cells undergoing cell turnover provide pathways through which LDL enters the subendothelial space, resulting in lipid accumulation.

Endothelium as a transducing surface

Endothelial cells responses to a variety of agonists include release of endothelium dependent vasodilators, such as endothelium dependent relaxing factor (EDRF) and prostacyclin (PGI 2). These substances act on vascular smooth muscle to cause relaxation and also have potent anti-aggregatory effects on platelets. A study of the mechanisms of signal transduction involved in these processes was undertaken. An investigation of intracellular calcium using FURA-2 and INDO-1 loaded endothelial cells shows transient elevation in response to vasodilator agonists. The calcium content of endothelial cells calculated using 45Ca flux techniques is increased in response to bradykinin and thrombin. Receptor activation leads to increased phosphoinositide turnover in endothelial cells and activates protein kinase C, the latter may be involved in feedback regulation. Patch clamp studies have demonstrated receptor-operated ionic channels in the endothelial cell membrane. Thus, intracellular calcium concentration is elevated in response to receptor activation, both as a result of liberation of calcium from intracellular stores and calcium entry from extracellular sources. Endothelial cells also respond to particulate stimuli. They can selectively bind and phagocytize bacteria. Phagocytosis leads to generation of superoxide aionin, a process which also seems to be controlled by elevation of intracellular calcium and activation of protein kinase C. In addition phagocytosis activates endothelial cells resulting in increased migration, division and further phagocytosis. All in all, the plethora of different endothelial responses to a variety of stimuli suggests a complex and multipotent cell type. However, the conservatism of the mechanisms of signal transduction may indicate that endothelial cells do not respond to one signal with one response but that “activation” may occur in response to a variety of stimuli.

The Human Preproendothelin-1 Gene

Preproendothelin-1 (ppET-1) mRNA has previously been demonstrated to be markedly induced in cultured endothelial cells (EC) by the addition to the medium of thrombin, an agent known to stimulate phosphoinositide turnover in EC. In this study, the mechanism of regulation of ppET-1 mRNA expression was investigated in cultured human umbilical vein EC by RNA blot analysis with cloned ppET-1 gene as a probe. The mRNA for ppET-1 was rapidly upregulated by O-tetradecanoylphorbol-13-acetate (TPA) (0.5 microM) and by ionomycin (5 microM) within 10 min of addition to the medium, but not by forskolin (50 microM). The rapid induction of ppET-1 mRNA by TPA or ionomycin occurred even in the presence of cycloheximide, indicating that the mRNA induction does not require de novo protein synthesis. The ppET-1 mRNA was an extremely unstable species of mRNA with an apparent half-life of about 15 min. However, the half-life of ppET-1 mRNA was not appreciably affected by TPA or ionomycin, suggesting that the mRNA induction by these agents is mostly due to an activation of the transcription of the mRNA. These observations indicate that the production of ET-1 in human EC can be controlled by a transcriptional gene regulation directly coupled to the intracellular signals from the phosphoinositide-turnover pathway, i.e., activation of protein kinase C and increase in intracellular Ca2+. These mechanisms are discussed in relation to information on the primary structure of cloned ppET-1 gene.

Pancreatic islet transplantation ameliorates the effect of diabetes on retinal capillary endothelial cell kinetics

Perturbations in capillary endothelial cell function are early events in the pathogenesis of diabetic retinopathy. We used tritiated thymidine autoradiography to study the turn-over rate of retinal capillary endothelial cells in retinovascular preparations from normal, untreated diabetic and diabetic mice treated with pancreatic islet-cell transplants. Capillary endothelial cell turnover was significantly increased in poorly controlled diabetes and this increase was reduced by islet-cell transplantation. The prevention of excessive endothelial cell replication by tight control of hyperglycemia may help prevent the characteristic structural changes of diabetic microangiopathy, such as basement membrane accumulation.

Endothelial cell denudation, labelling indices and monocyte attachment in advanced swine coronary artery lesions

This study shows that frank endothelial denudation, as measured by scanning electron microscopy, is associated with coronary artery atherosclerotic lesions in swine fed hyperlipidemic diets for 9 or 18 months. The largest areas of endothelial cell denudation were found to be associated with the most advanced atherosclerotic lesions, and lesser areas of denudation were present in less severely atherosclerotic arteries. Overall, the study suggests that frank endothelial denudation in swine coronaries occurs secondarily to advanced lesion development, and is proportionate to the amount of lesion present. In contrast attachment of monocytes to the endothelial surface was not proportionate to the amount of lesion present. While the number of attached monocytes after 9 or 18 months of a hyperlipidemic diet was much greater than in the mash-fed swine, there was no difference between the two hyperlipidemic groups. In both hyperlipidemic groups, the attached monocytes were concentrated over lesion areas. By scanning electron microscopy, most attached monocytes in these perfused coronary arteries showed a slender foot process extending into a gap between endothelial cells. Endothelial cell turnover, as measured by tritiated thymidine labelling indices, increased with the amount of coronary artery atherosclerosis. However, it did not become significantly greater than in mash-fed swine until advanced lesions developed.

The role of arterial endothelial cell mitosis in macromolecular permeability.

The present experiments were performed on twelve male Wistar rats to study the quantitative, topographic correlation between transendothelial permeability of Evans Blue-albumin (EBA) conjugate and endothelial cell replication at the single-cell level. En face preparations of the thoracic aorta were examined by fluorescence microscopy. We found a high degree of correlation between endothelial cell mitosis and EBA leaky spots. Although endothelial cell mitosis is very rare in occurrence, nearly all junctions around the dividing cells were leaky (99%), in contrast to only 0.03% of the non-mitotic cells. In addition, electron microscopic observations showed that the junction around a dividing endothelial cell is leaky, whereas that around a dying cell is not. With the aid of our theoretical model, we were able to analyze the dynamics of macromolecular passage through leaky endothelial junctions. The duration of endothelial cell mitosis was estimated to be 67 min, which constituted 0.01% of the duration of the total cell cycle. The time-dependent change in junctional geometry during endothelial cell turnover leads to an inverse relationship between macromolecular size and duration of junctional leakage. For albumin the duration of leakiness across aortic endothelial cell is approximately 3.7 hr. The present findings lend support to our hypothesis that transiently open junctions surrounding the dividing endothelial cells provide the major pathway through which macromolecules enter the subendothelial space to result in lipid accumulation.

Platelet-activating factor effects on bovine pulmonary artery endothelial cells.

Endothelial cells (ECs) were isolated from bovine pulmonary artery and maintained in long-term culture. On reaching confluency, ECs formed a characteristic "cobblestone" monolayer. One hour after addition of 1 nM platelet-activating factor (PAF) to the growth medium, ECs underwent dramatic changes in shape from their normal polygonal morphology to more elongated spindle-shaped forms. More pronounced effects were evident in the presence of 0.1 nM phorbol-12-myristate-13-acetate (PMA), a potent activator of C kinase. It was found that at concentrations from 10(-11)-10(-7) M, PAF stimulates the phosphoinositide turnover in EC. The half-maximal activation in the release of inositol phosphates was at 10(-9) M. This finding suggested that an increase in intracellular Ca2+ concentration and activation of protein kinase C were involved in the mechanism of action of PAF on EC. The metabolic responses of EC were evaluated by measuring the activity of beta-adrenergic receptor-coupled adenylate cyclase (AC) in a crude membrane fraction and by assay of prostacyclin and thromboxane released by cultured EC. AC from control membranes was activated by isoproterenol in a dose-dependent manner (EC50 = 30 nM) from 0.8-5.5 pmol cAMP/min/mg protein. If the membranes were isolated after preincubation of ECs with 1 nM PAF or 0.1 nM PMA, the AC activity was decreased by 70 and 90%, respectively; in both cases, affinity for isoproterenol was lowered threefold (EC50 = 100 nM). Our data suggest that PAF interaction with EC leads to an apparent beta-adrenergic receptor desensitization that probably acts via a phosphorylation mechanism involving C kinase.(ABSTRACT TRUNCATED AT 250 WORDS)

Cholesterol oxygenated derivatives induce erythrocyte accumulation and endothelium alterations in the aorta of rats. Turnover of endothelial cells in the apparently intact areas remains unchanged

In this work we have studied the effect of cholesterol autooxidation derivatives on the aortic endothelium of rat. Endothelial alterations were evident after 24 h treatment. Areas showing many spindle shaped nuclei and focal accumulation of erythrocytes were observed. Moreover, in some preparations areas with destroyed and missing endothelium were observed. Autoradiography using 3H-thymidine does not show differences in the mitotic activity between control and treated animals in those areas showing apparently intact endothelium. Therefore, we conclude that cholesterol derivatives present ability to produce important lesions in aortic endothelium and that this effect is similar to that described when cholesterol is used.

Endothel, Angiogenese und Metastasierung

The turnover of endothelium in a normal adult male is very low. Angiogenesis, the formation of new blood vessels, involves a high turnover of endothelial cells. It takes place during embryogenesis, ovulation and many pathological conditions, e.g. tumor growth. Soluble polypeptide growth factors for endothelial cells have been biochemically characterized. They are able to induce angiogenesis in vivo. Inhibitors of angiogenesis can inhibit tumor vascularization and the growth of solid tumors in mice. Tumor metastasis involves the migration of malignant tumor cells through endothelial cells. The organ and tissue specificity of these and other invasive migrations through vascular endothelium may be related to the heterogeneity and organ specificity of endothelial cells.

Atherosclerotic lesions in coronary arteries of hyperlipidemic swine: Part 2. Endothelial cell kinetics and leukocyte adherence associated with early lesions

The role of endothelial cells (EC) in the development and progression of early swine intimal cell mass (ICM)-derived coronary artery lesions in 55 swine fed either a mash or hyperlipidemic diet for 14, 49 or 90 days was investigated. Characteristics studied were endothelial cell turnover (using tritiated thymidine autoradiography), adhesion of leukocytes (presumably chiefly monocytes) to endothelium, and the presence or absence of endothelial cell denudation. The major findings were: (a) An increased adherence of leucocytes to endothelial cells over ICM-lesions in the HL-90 day group compared to the corresponding mash value at 90 days as well as to that of each of the other HL and mash groups. (b) Significant positive correlation between the labeling indices (LI) of endothelial cells lying over coronary artery lesion cells and the labeling indices of the underlying ICM-lesion cells; also, a significant positive correlation between the labeling indices of endothelial cells over lesions of the abdominal aorta and those of the coronary arteries. (c) At 90 days the endothelial cell LI over lesions in the HL group was significantly higher than the corresponding values in the mash group. Since the EC increase rates by growth in the two groups are also significantly different, the differences in LIs reflect at least in part EC growth differences and no strong conclusion can be made regarding possible cell turnover differences. (d) No frank endothelial denudation was found. The findings suggest that in swine coronary arteries participation by monocytes from circulating blood is a factor in the early progression of the lesion as well as smooth muscle cell proliferation.

Turbulent fluid shear stress induces vascular endothelial cell turnover in vitro.

The effects of hemodynamic forces upon vascular endothelial cell turnover were studied by exposing contact-inhibited confluent cell monolayers to shear stresses of varying amplitude in either laminar or turbulent flow. Laminar shear stresses (range, 8-15 dynes/cm2; 24 hr) induced cell alignment in the direction of flow without initiating the cell cycle. In contrast, turbulent shear stresses as low as 1.5 dynes/cm2 for as short a period as 3 hr stimulated substantial endothelial DNA synthesis in the absence of cell alignment, discernible cell retraction, or cell loss. The results of these in vitro experiments suggest that in atherosclerotic lesion-prone regions of the vascular system, unsteady blood flow characteristics, rather than the magnitude of wall shear stress per se, may be the major determinant of hemodynamically induced endothelial cell turnover.

Injury and repair of endothelium at sites of flow disturbances near abdominal aortic coarctations in rabbits.

The acute and chronic effects of flow disturbances on arterial endothelium were studied by locally constricting the diameter of the rabbit abdominal aorta by 62% + 2.4% (mean +/- SE). This procedure produced a region of elevated shear stress immediately upstream from the coarctation. A region of small irregular vortices was formed in the first 5 to 7 mm downstream, whereas an annular vortex was observed in the region from 2.0 to 2.5 cm further downstream. Morphologic changes to the endothelium near these coarctations were assessed by scanning electron microscopy; endothelial cell replication rates as a function of time after coarctation were monitored with 3H-thymidine. These studies established that profound alterations in endothelial cell shape caused by changes in local flow conditions immediately downstream from coarctations are primarily the result of alterations in morphology of pre-existing cells, rather than the proliferation of new cells with altered morphology. Definitive morphologic evidence of injury to endothelium was not seen at any sites after coarctation. Indeed, any early injury to endothelium caused by the flow disturbances was not sufficiently severe to cause a significant increase in endothelial cell turnover rate during the first week of coarctation. On the other hand, we observed a major increase in cell turnover (over 100-fold) by 30 days after coarctation at sites immediately upstream and immediately downstream of the coarctation. This finding suggests that flow-induced trauma can ultimately injure the cell sufficiently to cause cell death if the source of injury is persistent. Finally, we demonstrated that the high shear stress immediately upstream from the coarctation and the secondary flow disturbances immediately downstream can retard migration of endothelial cells into sites of injury, whereas repair was enhanced in the region of the primary annular vortex.(ABSTRACT TRUNCATED AT 250 WORDS)

Participation of the endothelium in the development of the atherosclerotic plaque

In the past decade, initiated by the response-to-injury hypothesis of Ross and Glomset, the endothelium has been implicated in atherogenesis but as a passive participant--more involved through its absence than its presence. The hypothesis stated that endothelial desquamation due to an undefined injury led to platelet adhesion to the exposed basement membrane, and infiltration of serum lipoproteins. The subsequent release from the platelet alpha-granule of a potent smooth muscle cell mitogen and chemoattractant--the platelet-derived growth factor (PDGF)--was postulated to cause the intimal proliferative response that is known to be important in atherosclerotic plaque development. Recent evidence from several laboratories indicates that the endothelium has the potential to play a more active role in plaque development than simply contributing to pathological sequelae resulting from the loss of the nonthrombogenic surface provided by the endothelium. First, the endothelial cell (EC) is the site of attachment, and possibly activation, of blood-borne monocytes which enter the vessel wall as an early event in experimental atherogenesis. We have obtained in vitro evidence that the expression of monocyte binding sites on the surface of EC is a regulatable process and that increased EC turnover and certain exogenous agents acting on EC cause increased monocyte adhesion. Similar events may be responsible for focal adhesion of monocytes to the endothelium in vivo following hypercholesterolemia. Secondly, EC in culture are capable of chemically modifying low density lipoprotein (LDL) by a free radical oxidation process that renders the LDL toxic to proliferating cells and recognizable to the scavenger receptor of monocyte-derived macrophages. Thus, by oxidation of LDL, the EC have the potential to play an active role both in the formation of lipid-laden foam cells and in the accumulation of necrotic tissue which are hallmarks of the atherosclerotic lesion. Thirdly, cultured EC have been recently shown to secrete multiple mitogens for cultured smooth muscle cells. One of these mitogens appears to be closely related, if not identical, to PDGF using the criteria of receptor binding and biochemical and immunological similarity. Production of growth factors by EC is a regulatable process that is stimulated by exogenous agents such as endotoxin and phorbol esters which cause severe injury to cultured EC. Such a regulatory mechanism may participate in the in vivo proliferation of vascular SMC during the atherosclerotic process.(ABSTRACT TRUNCATED AT 400 WORDS)

Hyperplastic growth response of vascular smooth muscle cells following induction of acute hypertension in rats by aortic coarctation.

This study examines the growth response of vascular smooth muscle cells following induction of acute hypertension in rats by partial ligation of the abdominal aorta between the renal arteries. Systolic blood pressures proximal to the ligature increased dramatically within 3 days (from 135 +/- 3 to 195 +/- 7 torr) of surgery while pressures distal to the ligature were reduced from control values. The frequency of smooth muscle cells undergoing DNA replication was increased 25-fold in thoracic aortas of coarctation rats compared to sham-operated controls 9 days post-coarctation, while no differences were observed in cells in abdominal aortic segments 1 cm distal to the ligature. A small but significant increase in the frequency of tetraploid smooth muscle cells was observed in thoracic aortas of coarctation rats, but this increase accounted for less than 2% of the increase in medial DNA content. By far, the major growth response was hyperplasia, as evidenced by a 25% increase in thoracic aortic medial smooth muscle cell number without a change in mean cellular volume (micron3/cell) or mass (ng/cell). Whereas no evidence of endothelial denudation was observed in thoracic aortas of coarctation rats by scanning electron microscopy, endothelial cell turnover rates were increased 23-fold compared to controls, indicating that some form of endothelial "injury" or dysfunction was present. Consistent with this, morphological changes indicative of endothelial injury (e.g., subendothelial edema) were observed by light and transmission electron microscopy. The marked contrast between results of this study and our previous studies showing that aortic medial hypertrophy in spontaneously hypertensive and Goldblatt hypertensive rats was due to cellular hypertrophy and hyperploidy without hyperplasia, clearly demonstrates that the growth response of smooth muscle cells within a given blood vessel can be quite different, depending on the model of hypertension. It is suggested that a non-denuding form of endothelial "injury" may play an important role in the proliferative growth response of smooth muscle cells following induction of coarctation hypertension.

Effect of cell turnover and leaky junctions on arterial macromolecular transport.

A new quantitative model is presented to explore the changes in vascular permeability that would result if the intercellular clefts around widely scattered endothelial cells were to become leaky to macromolecules in the range of roughly 4-10 nm during normal cell turnover. Although these open junctions occupy less than 10(-5) of the en face area of the endothelial surface, it is shown that the endothelial permeability can increase by 50-100% due to the experimentally observed regional variations in turnover in the larger arteries, whereas in the thinner walled veins and smaller arteries the subendothelial concentration is not significantly elevated. These results provide a very plausible explanation for the observed focal differences in the uptake of 125I-albumin and 131I-fibrinogen in blue and white areas and the nonselectivity of the local enhancement in uptake for these two molecules as a function of molecular size. The model has important implications for the localization of atherogenesis and the importance of endothelial cell turnover on the transport of proteins in vessels of all sizes.

The inner blood-retinal barrier in diabetes

The inner blood-retinal barrier is highly complex, dependent on a normally functioning retinal vascular endothelium and possibly adjacent perivascular cells. There are undoubtedly many different grades of barrier failure, and in the early stages an inbuilt compensatory mechanism may prevent significant fluid accumulation and delay associated structural and functional changes. In diabetes, early barrier failure may be a function of an increase in the rate of endothelial cell death and turnover, probably exaggerated by the early and widespread loss of the supporting cells and dilatation of the small blood vessels.