Physiological transport properties of cultured retinal microvascular endothelial cell monolayers.

PURPOSE: To characterize baseline transport properties: hydraulic conductivity (Lp), albumin permeability (Pe), and transendothelial electrical resistance (TER) of bovine retinal microvascular endothelial cells (RMEC) in the development of an in vitro model of the blood-retinal barrier (BRB). METHODS: RMEC were grown on porous, polycarbonate filters for determination of the number of days required to achieve minimal transport rates. Lp, Pe, and TER were measured by utilizing a bubble tracking spectrophotometer, by quantifying the diffusional movement of fluorescein isothiocyanate-labeled albumin, and by utilizing a Millipore electrical resistance meter, respectively. RESULTS: Lp decreased significantly from 7.82 +/- 0.85 x 10(-7) (mean +/- SEM) cm/sec/cm H2O at post-plating Day 5 to 1.44 +/- 0.26 x 10(-7) cm/sec/cm H2O at Day 9. Pe of the monolayer also decreased progressively with days post-plating from 3.44 +/- 0.53 x 10(-6) cm/sec at Day 7 to a minimum of 1.95 +/- 0.29 x 10(-6) cm/sec at Day II. Peak TER fluctuated until Day 7, when it began to steadily increase from 17.14 ohm-cm2 to a peak value of 25.42 ohm-cm2 at Day 10, decreasing from then on to 22.24 ohm.cm2 on Day 12. Known disrupters of the BRB, NECA and VEGF, elicited significant increase in RMEC Lp showing the sensitivity of this model to pharmacological alterations. CONCLUSIONS: Our data indicate that RMEC grown on polycarbonate filters form a restrictive monolayer of cells, which exhibit dynamic alterations in response to pharmacological agents, thus demonstrating an in vitro model of the BRB. Future studies with the model may offer insights into the pathogenesis of retinal vascular diseases and allow convenient testing of pharmacological interventions.

Shear stress increases hydraulic conductivity of cultured endothelial monolayers.

To examine the effect of shear stress on hydraulic conductivity (Lp) of bovine aortic endothelial cell monolayers grown on polycarbonate filters, we developed a rotating disk system, which imposed a defined shear stress while Lp was measured. A 10-cmH2O pressure differential was applied to monolayers, and baseline Lp was established between 1.65 +/- 0.85 and 4.94 +/- 1.05 x 10(-7) cm.s-1.cmH2O-1. One-hour exposure to 10 dyn/cm2 shear stress caused a significant (P < 0.05) increase in Lp by 2.16-fold (+/- 0.42), and Lp remained elevated when shear stress was removed. Three-hour exposure to shear stresses between 0.1 and 20.0 dyn/cm2 revealed a threshold for shear-induced increase in Lp of 0.5 dyn/cm2. At 20 dyn/cm2, Lp initially decreased by 30% (+/- 13.4%, P < 0.05) and then increased to a level 3.76-fold (+/- 0.83, P < 0.05) greater than baseline Lp at 3 h. The shear-induced increase in Lp was reversed with dibutyryl adenosine 3',5'-cyclic monophosphate (DBcAMP, 1 mM) and could be significantly (P < 0.05) inhibited when monolayers were preincubated with 0.3 mM DBcAMP, a concentration that did not significantly affect baseline Lp. Furthermore, preincubation with a general phosphodiesterase inhibitor, 3-isobutyl-1-methylxanthine (1 mM), completely blocked the shear-induced increase in Lp. On the basis of these results, we conclude that shear stress alters endothelial Lp through a cellular mechanism involving signal transduction, not by a purely physical mechanism.

Antihistamines reduce blood-retinal barrier permeability in type I (insulin-dependent) diabetic patients with nonproliferative retinopathy. A pilot study.

PURPOSE: To determine if histamine receptor stimulation mediates increased blood-retinal barrier (BRB) permeability in patients with diabetic retinopathy, as it does in experimental diabetes. METHODS: Fourteen patients with type I (insulin-dependent) diabetes and mild nonproliferative diabetic retinopathy were treated with combined astemizole, 20 mg, and ranitidine, 600 mg, or an identical placebo for 6 months in a double-masked fashion. Blood-retinal barrier permeability was measured by vitreous fluorometry at baseline and at 3 and 6 months. RESULTS: Permeability was significantly reduced in the group treated with antihistamines (P < 0.05) compared with the placebo group. There were no concomitant significant changes in systemic arterial blood pressure or HbA1c values. CONCLUSION: These pilot data suggest that histamine receptors influence permeability of the BRB in human diabetes. Further studies of the effects of antihistamines on diabetic retinopathy are warranted.

Astemizole reduces blood-retinal barrier leakage in experimental diabetes.

We examined the potential of astemizole, a histamine H1-receptor antagonist that does not cross the blood-brain barrier, to reverse blood-retinal barrier leakage to albumin in streptozotocin diabetic rats. Four groups of nondiabetic and four groups of diabetic rats received vehicle or astemizole at dosages of 5, 10, or 20 mg/kg body weight for days 22-28 of a 28-day holding period. There were no significant differences in nondiabetic plasma-vitreous albumin ratios between animals receiving vehicle or any of the three astemizole dosages. Only diabetic rats receiving vehicle showed a significant (p < 0.05) 100% increase in the plasma-vitreous albumin ratio over their nondiabetic counterparts. Diabetic rats receiving either 5, 10, or 20 mg/kg astemizole exhibited total normalization of vitreous albumin accumulation, despite persistence of diabetes. These data indicate that astemizole, an H1-receptor antagonist that does not cross the blood-retinal barrier, is effective in reversing blood-retinal barrier leakage of albumin in experimental diabetes.

Endothelial albumin permeability is shear dependent, time dependent, and reversible.

Altered permeability of vascular endothelium to macromolecules may play a role in vascular disease as well as vascular homeostasis. Because the shear stress of flowing blood on the vascular wall is known to influence many endothelial cell properties, an in vitro system to measure transendothelial permeability (Pe) to fluorescein isothiocyanate conjugated bovine serum albumin under defined physiological levels of steady laminar shear stress was developed. Bovine aortic endothelial cells grown on polycarbonate filters pretreated with gelatin and fibronectin constituted the model system. Onset of 1 dyn/cm2 shear stress resulted in a Pe rise from 5.1 +/- 1.3 x 10(-6) cm/s to 21.9 +/- 4.6 X 10(-6) cm/s at 60 min (n = 6); while 10 dyn/cm2 shear stress increased Pe from 4.8 +/- 1.5 X 10(-6) cm/s to 50.2 +/- 6.8 X 10(-6) cm/s at 30 min and 49.6 +/- 8.9 X 10(-6) cm/s at 60 (n = 9). Pe returned to preshear values within 120 and 60 min after removal of 1 and 10 dyn/cm2 shear stress, respectively. The data show that endothelial cell Pe in vitro is acutely sensitive to shear stress.

The effect of varying albumin concentration and hydrostatic pressure on hydraulic conductivity and albumin permeability of cultured endothelial monolayers.

An in vitro model of the endothelial transport barrier was developed using bovine aortic endothelial cell monolayers cultured on a porous polycarbonate substrate. Hydraulic conductivity (Lp) was measured by a bubble tracking technique at varying pressure differentials and albumin concentrations. The effective albumin permeability (Pc) was determined by measuring the flux of fluorescent-labeled albumin across monolayers at varying hydrostatic pressures. Lp determined at pressure differentials between 5.0 and 10 cm H2O demonstrated a strong dependence on albumin concentration, decreasing approximately 10-fold from 21.3 x 10(-7) +/- 3.18 x 10(-7) cm/sec/cm H2O (mean +/- SEM) at 0.0 g/dl to 2.35 x 10(-7) +/- 0.20 x 10(-7) cm/sec/cm H2O at 1.0 g/dl albumin. Increasing the albumin concentration from 1.0 to 4.0 g/dl reduced Lp by an additional 16% to 1.97 x 10(-7) +/- 0.17 x 10(-7) cm/sec/cm H2O. Furthermore, Lp was moderately dependent on the pressure differential, increasing by about a factor of two with a doubling of the pressure differential. The effective permeability (Pc) was also dependent on the pressure differential. At an albumin concentration of 4.0 g/dl, Pc increased from 1.37 x 10(-6) +/- 0.26 x 10(-6) cm/sec at 0.0 cm H2O to 5.06 x 10(-6) +/- 1.92 x 10(-6) cm/sec at 10 cm H2O. Analysis of Pc and Jv data, however, demonstrates that water and albumin do not share a common pathway in crossing the endothelial monolayer. These data suggest the existence of a large pore pathway for albumin. Thus, the in vitro system has many of the transport characteristics of intact vessels in vivo and should be useful for physiological studies of the endothelial transport barrier.

Histamine H1 receptors mediate increased blood-retinal barrier permeability in experimental diabetes.

To test the hypothesis that histamine receptors mediate increased blood-retinal barrier permeability in experimental diabetes, 51 rats were made diabetic by streptozocin injection (65 mg/kg; jugular vein) and were held for four weeks. The seven animal groups were as follows: untreated controls; untreated diabetic rats; diabetic rats receiving diphenhydramine hydrochloride (Benadryl); diabetic rats receiving cimetidine hydrochloride (Tagamet); diabetic rats receiving diphenhydramine and cimetidine; diabetic rats receiving purified pork insulin (Iletin II); and diabetic rats receiving insulin and diphenhydramine. All treatments were given during the last week. Blood-retinal barrier permeability was assessed through measurement of the vitreous content of fluorescein isothiocyanate conjugated to bovine serum albumin (FITCBSA) after 20 minutes of FITCBSA circulation. Vitreous FITCBSA content of the diabetic group was 64% greater than control content. Diabetic rats treated with either diphenhydramine or diphenhydramine and insulin had respective decreases of 43% and 40% in vitreous FITCBSA content. The vitreous content of the diabetic group receiving insulin was lowered 37% below untreated diabetic values, while the vitreous FITCBSA content of the diabetic group receiving both insulin and diphenhydramine was reduced 63%. These data indicate that retinal histamine H1-receptor activation may be partially responsible for initial blood-retinal barrier leakage of macromolecules into the vitreous and that this abnormal leakage can be prevented both by diphenhydramine and by insulin. Histamine H1 receptors may play an important role in mediating increased blood-retinal barrier permeability in experimental diabetes.

Retinal histamine synthesis is increased in experimental diabetes.

We examined retinal de novo histamine synthesis mediated by retinal histidine decarboxylase in normal and streptozotocin-diabetic male, Sprague Dawley rats that were diabetic for 21 days. We also examined effects of insulin and alpha-hydrazinohistidine (alpha HH) treatments on retinal histamine synthesis in this diabetic model. alpha HH is a specific inhibitor of histidine decarboxylase. Results indicate that the retina contains an active histidine decarboxylase enzyme system, and that in streptozotocin diabetes retinal histamine synthesis is increased 197%. Both insulin and alpha HH independently reverse and normalize retinal histamine synthesis. These data thus indicate that the retinal inducible histamine pool is increased in experimental diabetes, and that insulin is an important modulator of retinal histamine metabolism. This newly described retinal metabolic alteration may be one factor responsible for increased retinal vascular permeability in diabetic retinopathy.

Antihistamines reverse blood-ocular barrier breakdown in experimental diabetes.

Retinal and other tissue histamine synthesis is increased in experimental diabetes; histamine infusion causes blood-ocular barrier breakdown in nondiabetic rats. We have examined the hypothesis that antihistamines prevent blood-ocular barrier breakdown in streptozotocin diabetes using male Sprague-Dawley rats held 28 days. During the last 7 days they were divided into these treatment groups: control (C), untreated diabetic (D), diabetic rats receiving diphenhydramine-HCl (B), diabetic rats receiving ranitidine (R) and diabetic rats receiving diphenhydramine and ranitidine (BR). Vitreous albumin content was measured 6 hr following fluorescein isothiocyanate bovine serum albumin (FITCBSA) injection. Data show that D had a 98.3% increase in vitreous body FITCBSA over C (p less than 0.05) while B and R showed respective decreases of 34.9% and 51.4% compared to D, R being significantly lower than D (p less than 0.05). BR showed a decrease of 71% (p less than 0.05) compared to D, and R and BR groups were not significantly different from C (p less than 0.05). Leakage into the vitreous was from the retina, not the ciliary body. These data indicate that 1) experimental diabetes results in elevated blood-ocular barrier permeability, which can be reversed by diphenhydramine-HCl and ranitidine; and 2) histamine H1- and H2-receptor activation and interaction by altered endogenous histamine metabolism may mediate blood-ocular barrier breakdown, implicating a pathogenic role of histamine in diabetic retinopathy.

Cultured bovine aortic endothelial cells show increased histamine metabolism when exposed to oscillatory shear stress.

Oscillatory shear stress applied to the lining of blood vessels causes endothelial cell injury, one of the essential postulated prerequisites to the development of atherosclerosis. The purpose of this investigation was to study effects of shear stress on bovine aortic endothelial cells (BAEC), in vitro, for varying lengths of time (6 h, 12 h, 24 h) on BAEC histamine content (HC) and histidine decarboxylase activity (HD). Low intensity stress (1.6 dynes/cm2) as well as intermediate and high intensity shear stresses (3.5 dynes/cm2 and 7.6 dynes/cm2) resulted in an accelerated HD (281%) and elevated HC (144%). These data indicate that oscillatory shear stress produces increases in histamine metabolism.

Density-dependent contraction of the endothelial nascent histamine pool by exogenous heparin.

The aortic endothelial cell nascent histamine pool has been implicated in the control of vessel wall permeability under conditions of stress and injury. We report the contraction of this histamine pool in low density bovine aortic endothelial cell (BAEC) cultures by exogenous heparin. Untreated BAEC exhibit a decline in histamine content in 3-day cultures with increasing plating density between 1000 and 16,000 cells/cm2. Heparin abolished this density-related difference by effecting a 67% contraction of the histamine pool in the lowest density cultures. This effect was reversible and specific to heparin. At a confluent density, endothelial cells secrete heparin-like glycosaminoglycans which affect smooth muscle and endothelial metabolism. We propose that the metabolic effects of exogenous heparin, and perhaps endogenous heparins, extend to specific modulations of the BAEC nascent histamine pool.

Renal histamine increases in the streptozotocin-diabetic rat.

Renal histamine concentration, total histamine and protein contents were measured in rats made diabetic via iv streptozotocin injection and held for 13 weeks following diagnosis of diabetes. Insulin (7 U/day) or alpha-hydrazinohistidine (alpha-HH, 25 mg/kg/day) or both drugs were administered to diabetic subgroups the last 2 weeks of the holding period. Untreated diabetics developed significant increases of renal histamine concentration and total histamine content, up 45 and 46%, respectively. Drug interventions reduced the diabetic increases of histamine concentration and content (in order) as follows: diabetic-insulin, down 7 and 8%; diabetic-alpha-HH down 25 and 26%; diabetic-insulin + alpha-HH, down 35 and 36%. Renal tissue protein content was unchanged and qualitative proteinuria was present in all diabetic subgroups. The data indicate that in experimental diabetes there is an increase of the renal inducible histamine pool which is partially reduced by insulin and/or alpha-HH treatments. In view of the long-recognized actions of histamine upon microvascular permeability, elevated renal histamine may be one pathophysiological mediator of the diabetic functional renal microangiopathy manifest as proteinuria.

Changes in plasma histamine concentration in the streptozotocin-diabetic rat.

Plasma histamine concentrations were measured in rats made diabetic via jugular vein injection of streptozotocin and held 4 weeks following diabetes diagnosis. At least 15 diabetic animals received insulin (6-8 U/day) or alpha-hydrazinohistidine (alpha-HH) for the last week of the holding period. alpha-HH is a specific inhibitor of histidine decarboxylase (HD), the principle histamine-forming enzyme in mammals. Plasma histamine concentrations, expressed as means and mean standard errors (ng/ml) were as follows: control, 25.5 +/- 2.4; diabetic, 47.1 +/- 5.2; diabetic-insulin, 34.6 +/- 2.9; diabetic-alpha-HH, 28.1 +/- 2.1. These data indicate that in experimental diabetes there is an expansion of the nascent, or inducible histamine pool, an increase which is reflected by increased circulating plasma histamine. This may be one component mediating altered microvessel as well as large vessel permeability characteristics, an underlying component of both diabetic microangiopathy and macroangiopathy.

Aortic histamine synthesis and aortic albumin accumulation in diabetes: activity-uptake relationships.

The relationship between inhibition of aortic histamine synthesis induced by varied dosages of alpha-hydrazinohistidine (alpha HH) and aortic uptake of fluorescein-labeled bovine serum albumin (FITCBSA) was examined in 40 male Wistar rats made diabetic by streptozotocin. Compared to nondiabetic animals, aortic uptake of FITCBSA in untreated diabetic animals was increased 43%. alpha HH administration produced essentially a complete inhibition of this accelerated histamine synthesis in diabetic animals at all dosages examined and likewise prevented an increase in aortic FITCBSA uptake among these animals. In diabetic animals, aortic histamine synthesis and aortic FITCBSA uptake showed a significant, strong positive correlation (r = 0.822, P less than 0.001) when examined in relation to the degree of inhibition of accelerated aortic histamine synthesis achieved among the individual animals. These data support the hypothesis that elevations in de novo aortic histamine formation, and thus elevations in the inducible histamine pool, are responsible, at least in part, for increased aortic uptake of macromolecules in diabetes. Since increases in various permeability characteristics occur early in atherogenesis, these data also indicate that expansion of the nascent histamine pool occurring in diabetes may be an important factor in the predisposition of diabetics to atherosclerotic vascular disease.

Inhibition of aortic histamine synthesis by alpha-hydrazinohistidine inhibits increased aortic albumin accumulation in experimental diabetes in the rat.

We examined the interrelationship between inhibition of aortic histamine synthesis through inhibition of aortic histidine decarboxylase and intra-aortic albumin accumulation in rats made diabetic by a jugular vein injection of 60 mg/kg of streptozotocin under ether anesthesia. Animals were held for 4 weeks following overt manifestation of diabetes. At the end of 3 weeks, at least six animals in each of the diabetic and non-diabetic groups received intra-peritoneal injections of alpha-hydrazinohistidine (25 mg/kg at 12 h) for the last 7 days. Aortic albumin accumulation was measured by quantification of aortic uptake of fluorescein isothiocyanate conjugated to rat serum albumin injected in the jugular vein 1 h before sacrifice. The aortic albumin mass transfer and flux rates of the diabetic group were more than 300% higher than that of the control group; alpha-hydrazinohistidine treated diabetic rats had aortic albumin mass transfer rates equivalent to control values. The aortic albumin content was nearly tenfold higher in untreated diabetic rats, but again treatment with alpha-hydrazinohistidine returned this to control values. These data offer strong support to the premise that accelerated aortic histamine synthesis, which occurs in experimental diabetes, is an important mediator of increased aortic macromolecule uptake, and as such, may be one component of the multitude of factors responsible for increased susceptibility of atherosclerosis among individuals having diabetes mellitus.

Time-dependent changes in aortic albumin permeability characteristics in experimental diabetes.

The aortic albumin flux, mass transport rates, and aortic albumin-plasma ratios were examined at 20, 30, and 60 min following injection of bovine serum albumin conjugated to fluorescein isothiocyanate (FITCBSA) in rats made diabetic via streptozotocin injection (65 mg/kg, jugular vein, ether anesthesia). These studies were undertaken in order to establish some basic semiquantitative permeability parameters of large arteries in experimental diabetes. Data indicate that under diabetic conditions, both the aortic flux and the mass transfer rates are significantly elevated, changes which correspond to increased tissue accumulation of FITCBSA. In addition, the time needed to saturate the albumin space in aortas of diabetic animals is significantly lower than normal.

Cytochemical correlates of atherosclerosis-resistant and susceptible lesions of the normal rabbit aorta.

Quantitative cytochemical and microfluorimetric techniques were employed to compare mural intermediary metabolism--endothelial macromolecular uptake changes in spontaneous aortic-arteriosclerotic lesions of normolipemic New Zealand White rabbits. Specifically, mural succinic (SDH), lactic (LDH), and glucose-6-phosphate (G-6-PDH) dehydrogenase activities and luminal surface uptake of fluorescein isothiocyanate-conjugated bovine serum albumin (FITC-BSA) were measured in lesion sites abnormally resistant (calcified) and susceptible (proliferative) to dietary hypercholesterolemia. Calcified lesions exhibited severe (55-66%) diminution of SDH, LDH, and G-6-PDH activities within the involved inner mural zone and a comparable (68%) decline in luminal FITC-BSA uptake. Concomitant reductions in FITC-BSA uptake (30%) and marker enzymes of the predominant energy transducing pathways in arterial tissue, i.e., SDH (30%) and LDH (31%), were evidenced in proliferative foci, whereas G-6-PDH was augmented (52%) in comparison to nonlesioned aortic segments. These data lend additional support to the concept that endothelial uptake of plasma-borne macromolecules is coupled to oxidizable substrate requirements of inner avascular compartments of the arterial wall. It is postulated that diminished macromolecular transport in these degenerative lesions stems from reduced mural metabolic demands, and that pharmacologic reduction of vascular smooth muscle metabolism may depress uptake of sclerogenic macromolecules.

Aortic endothelial and smooth muscle histamine metabolism. Relationship to aortic 125I-albumin accumulation in experimental diabetes.

We studied rat aortic endothelial and smooth muscle cell de novo histamine synthesis mediated by histidine decarboxylase (HD) and the effects of its inhibition by alpha-hydrazinohistidine on the intracellular histamine content and intraaortic albumin accumulation in streptozotocin-induced diabetes. Diabetes was induced by a single jugular vein injection of streptozotocin (60 mg/kg, pH 4.5, ether anesthesia), with animals held 4 weeks following the overt manifestation of diabetes. Additional diabetic and nondiabetic rats received alpha-hydrazinohistidine (25 mg/kg, i.p. every 12 hours) during the last week; this had no effect on the severity of diabetes in any animal receiving streptozotocin. Data indicate that the aortic endothelial (EC) HD activity was increased more than 130% in the untreated diabetic group but was similar to control values in the diabetic group receiving alpha-hydrazinohistidine; similarily, the EC histamine content from diabetic aortas increased 127% over control values, but in EC from diabetic animals receiving alpha-hydrazinohistidine it was comparable to control values. Similar trends were observed for the subjacent aortic smooth muscle. In untreated diabetic animals the aortic 125I-albumin mass transfer rate was increased 60% over control values, while in diabetic animals receiving alpha-hydrazinohistidine the 125I-albumin mass transfer rate was essentially identical to controls. These data indicate that in streptozotocin diabetes there is an expansion of the inducible aortic histamine pool, and that this expansion is intimately related to the increased aortic albumin accumulation.