The binding characteristics of [3H]quinuclidinyl benzilate ([3H]QNB) to isolated crude membranes of cultured bovine aortic endothelial cells were investigated. [3H]QNB bound to endothelial cell membranes with high affinity (kD = 0.056 nM) and limited capacity (132 fmol/mg DNA). The binding specificity, order of affinity and inhibition constants (Ki) were determined by displacement of bound [3H]QNB with unlabeled ligands. The order of affinity was QNB > atropine > 4-diphenylacetoxy-N-methyl-piperidine methiodide (4-DAMP) > p-fluoro-hexahydro-sila-difenidol (p-F-HHSiD) (M3 antagonist) > pirenzepine (M1 antagonist) > AFDX-116 (M2 antagonist) > (4-hydroxy-2-butynyl) trimethylammonium chloride m-chlorocarbanilate (McN-A-343, M1 agonist). These observations suggest that muscarinic receptors of endothelial cells in culture are likely to be of M3 and M1 subtype. Northern blot analysis of receptor subtypes using cDNA probes did not provide conclusive results due to the low level expression of these receptors in cultured cells. Solubilization of protein bound [3H]QNB with 1% digitonin and 0.02% cholate followed by analysis on sucrose density gradients demonstrated the presence of a specifically bound [3H]QNB-protein complex sedimenting at the 6.2S region of the gradient. These data demonstrate the presence of muscarinic acetylcholine receptor protein in cultured bovine aortic endothelial cells.

The localization, synthesis, and activity of endothelin and the receptor types mediating its effects in penile corpus cavernosum were investigated in whole tissue and in cultured cells derived from this tissue. With immunocytochemistry, utilizing an antiendothelin 1 (ET-1) monoclonal antibody, endothelin-like immunoreactivity was localized intensely in the endothelium and to a lesser degree in the trabecular smooth muscle. Human corpus cavernosum endothelial cells in culture expressed preproendothelin 1 mRNA, as determined by Northern blot analysis. Significant amounts of endothelin-like immunoreactivity were measured by radioimmunoassay in the supernatants of corpus cavernosum endothelial cells in culture. Endothelins are potent constrictors and caused long-lasting contractions of corporeal strips in organ chambers. Equilibrium binding analysis of endothelins to their receptor sites revealed high-affinity, specific, and saturable binding of labeled endothelins to corporeal membranes. Competition binding experiments demonstrated receptors with high affinity for ET-1 and -2 and low affinity for ET-3 and another, less abundant, set of receptors with high affinity for ET-1, -2, and -3. Affinity labeling of endothelins to corporeal membranes, followed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis, revealed that ET-1 and -2 cross-linked specifically to three different molecular mass components (75, 52, and 34 kDa). ET-3 bound only to the 34-kDa component. It is concluded that human corpus cavernosum endothelium has the ability to synthesize and release endothelin, that endothelins contract corporeal smooth muscle, and that at least two distinct endothelin receptors may exist and are differentiated by their affinity for ET-3.

The binding characteristics of the muscarinic acetylcholine receptor antagonist, [3H]quinuclidinyl benzilate, to isolated membranes of human corpus cavernosum and endothelial cells, cultured from this tissue, were investigated. [3H]quinuclidinyl benzilate bound to membranes of human corpus cavernosum and endothelial cells with high affinity and limited capacity. Analysis of the binding data by Scatchard plot revealed the presence of one class of binding sites. The ligand binding specificity was determined by competitive binding assay. The data obtained show that [3H]quinuclidinyl benzilate was displaced by unlabeled competitors in the following order of efficacy in both membrane preparations: quinuclidinyl benzilate greater than scopolamine greater than atropine greater than 4-diphenylacetoxy-N-methyl-piperidine methiodide, M3 antagonist greater than pirenzepine, M1 antagonist greater than oxotremorine greater than (4-hydroxy-2-butynyl)trimethylammonium chloride m-chlorocarbanilate, M1 agonist greater than carbachol greater than hexamethonium. Solubilization of the muscarinic acetylcholine receptors from human corpus cavernosum and endothelial cells, with 1% digitonin and 0.02% cholate and subsequent analysis on sucrose density gradients, demonstrated the presence of a macromolecule specifically bound to [3H]quinuclidinyl benzilate sedimenting at the 6.2 S region of the gradient. These results demonstrate the presence of muscarinic acetylcholine receptors in human corpus cavernosum and in endothelial cells cultured from this tissue.

A method for culturing endothelial cells (HCC-EC) from surgical specimens of human corpus cavernosum has been developed. The approach involves selective endothelial outgrowth from explants and may be generally applicable to tissue whose endothelium is not amenable to isolation by routine mechanical or enzymatic methods. The tissue is minced into pieces which are placed onto gelatin-or fibronectin-coated tissue culture plastic, and grown in medium suitable for microvascular endothelial cell growth (Carson and Haudenschild, In Vitro 22:344–354, 1986). By Days 5 to 7 EC colonies are found. Within a day or two after the appearance of the EC colonies, a non-EC cell type appears and, if undisturbed, quickly overgrows the EC. An exploitable temporal separation between the emergence of EC and non-EC is obtained when both conditioned medium (from bovine aortic endothelium) and retinal extract are present during the outgrowth period. Explants are removed by pipetting at the first sign of the emergence of the non-EC cell type. Once isolated, HCC-EC do not require conditioned medium but do require either retinal extract or acidic fibroblast growth factor for survival and growth. Approximately 60% of the first passage cultures are at least 80% EC as judged by DiI-Ac-LDL labeling. One corpus (0.3×0.3×0.5 cm) usually produces 120 cm2 of primary culture within 2 wk. These EC form contact-inhibited monolayers and stain positively for Factor VIII. They have a doubling time at 6th passage of 48 h and a plateau density of 5 to 7×104 cells/cm2. The availability of such cultures should facilitate the study of endothelium-mediated responses which play an important role in the erectile function of human penile corpus cavernosum.

Isolation procedures employed by various laboratories to obtain cerebral microvessels generally utilize meshes to sieve and collect the microvessels from homogenized brain. This is followed in some cases by further purification using density gradients of Percoll or sucrose, or albumin flotation. We have evaluated microvessels prepared by these methods in terms of ATP content and ATPADP ratio, which reflect the cellular energy state, and enrichment of the marker enzymes, alkaline phosphatase and γ-glutamyltransferase. Albumin flotation generally increased the enrichment of marker enzymes; however, preparations using albumin flotation or a Percoll gradient exhibited considerable variability in ATP content and ATPADP ratio with the mean ATPADP ratio significantly lower than that observed in microvessels isolated by sieving through meshes. More uniformly high values for both ATP (∼1.6 nmole ATP/mg protein) and the ATPADP ratio (∼2.3) were obtained with meshes alone. Use of a sucrose gradient consistently resulted in preparations with a much lower ATP content and ATPADP ratio, compared with preparations obtained with the other methods. Values using the other methods were higher than those previously reported, yet were still lower than the ATP content of about 23 and ATPADP ratios of 18 and 7 we found in cultured microvascular endothelium and pericyte, respectively. These low values were not improved by supplying additional fuel to the microvessels during isolation, suggesting they were not the result of fuel deprivations during isolation. Despite the probable damage incurred during isolation, microvessel preparations are a useful in vitro model in which fuel metabolism appears to reflect the prior hormonal/nutritional state of donor animals. However, our data indicate the advisability of measurements of ATP content and ATPADP ratio for quality control of preparations used for metabolic studies, especially after Percoll density gradient or albumin flotation steps.

It has been proposed that the development of diabetic complications may involve a depletion of cellular myo-inositol due to an increase in polyol (sorbitol) formation. We therefore initially examined the effect of diabetes on levels of these metabolites in isolated cerebral microvessels. Compared with microvessels from control rats, microvessels from diabetic animals showed no detectable alteration in myo-inositol levels and a small increase in sorbitol content. To assess whether myo-inositol depletion might occur in only certain microvascular cells, cultured bovine cerebral microvascular pericytes and endothelium were grown for 3 or 18–20 days at 1.1, 5.5, or 22.2 mmol/l glucose. Increased medium glucose concentration resulted in increased sorbitol content in both cell types after both periods of incubation (p<0.05). In contrast, a significant decrease in myo-inositol content (22%, p<0.01) was observed only in pericytes grown for 18–20 days in the high glucose medium. Neither the adenosine 5′-triphosphate content nor the adenosine 5′-triphosphate/adenosine 5′-diphosphate (ATP/ADP) ratio of the pericytes was affected by the medium glucose concentration, indicating that the decrease in myo-inositol was not caused by a deficiency in the cellular energy state affecting the active transport of myo-inositol. These data suggest that myo-inositol depletion occurs selectively in the pericyte, a cell type known to be the site of early morphological changes in diabetes. Furthermore, the depletion apparently requires prolonged exposure to high glucose and is not due to a change in energy state.

We have utilized cultures of bovine brain microvascular endothelial cells (MEC) and pericytes to study two aspects of intercellular relations in the microvasculature. First, the apparent contradiction between the reported demonstration of dye transfer between endothelial cells in capillaries and venules in rat omentum and the lack of ultrastructurally demonstrable interendothelial gap junctions in the same vessels in omentum, brain, and other tissues led us to examine this problem in vitro. MEC showed extensive transfer of both fluorescent dye (Lucifer yellow CH, 96% transfer incidence in primary culture) and radiolabeled uridine nucleotides (97%). Freeze-fracture replicas of MEC revealed both gap and tight junctions. These results demonstrate that MEC are capable of producing gap junctions and engaging in junctional communication in vitro. Second, we have examined the interaction of pericytes with MEC. Cultured pericytes showed gap junctions in freeze-fracture replicas, variable dye transfer (cell density dependent, 19-91%), and extensive nucleotide transfer (94%). While the incidence of dye transfer between MEC and pericytes was low (10-31%), nucleotide transfer between these cells was extensive (86-96%). The demonstration of junctional transfer between MEC and pericytes in vitro may be particularly significant considering the high frequency of junctional contact between these cells in vivo. These cultured cell models should help us to better understand the complex interactions of vessel wall cells in microvascular physiology and pathophysiology.

Although one of the earliest findings of diabetic retinopathy is altered capillary permeability, metabolic factors in diabetes that may increase the permeability of capillaries to fluorescein are unknown. We have studied the effect of a variety of vascular and retinal cells and hyperglycemia on the diffusion rate of fluorescein. These studies were performed with a cell culture system that mimics the cross-section of a capillary by having two chambers separated with a porous membrane that can support the growth of cells on either side of the membrane. The addition of a confluent layer of endothelial cells or retinal pigmented epithelial (RPE) cells inhibited fluorescein diffusion between the two chambers 20- and 300-fold, respectively, after cells were cultured for >5 days. Exposure of endothelial cells to 400 mg/dl glucose for either 3 or 100 days did not affect the barrier function of these cells. The barrier function of capillary endothelial cells isolated from BB rats with chronic diabetes and from nondiabetic animals did not differ. In contrast to endothelial cells and RPE cells, arterial smooth muscle and pericytes, which are not known to form tight junctions, did not inhibit the diffusion of fluorescein more than 2-fold. Surprisingly, the dual culture of endothelial cells with either retinal pericytes or smooth muscle cells resulted in a 50-fold increase in the rate of fluorescein diffusion, showing a disruption of the endothelial barrier. In summary, the intercellular connections between endothelial and epithelial cells that are responsible for the barrier to fluorescein diffusion are not functionally affected by chronic exposure to hyperglycemia or diabetic conditions. The interaction between endothelial cells with capillary pericytes is important in maintaining vascular permeability and may be altered in the diabetic state.

Cultured microvascular endothelial cells (MEC) have become a valuable model for studies of microvascular physiology and pathology. Most current techniques involve manual removal of undesirable cell types or cloning, require one to several months, and yield high population doubling level cultures derived from a relatively small sample of the original population. We have devised a technique to more rapidly produce larger numbers of MEC. This method provided primary cultures consisting predominantly of MEC within 1 wk. The technique involves selective aspiration of gray matter from the bovine cerebral cortex followed by homogenization, sieving, enzymatic dissociation, and then dense plating (104 to 105 vessel fragments/cm2) onto gelatin- or fibronectin-coated plastic. Typical yields were 0.1 to 0.5 × 106 fragments/g of aspirated gray matter. The optimal culture medium for these cells was 15% equine plasma derived serum, 20% conditioned medium, 2% retinal extract, 60% fresh medium, and 500 μg/ml heparin. Cells attached within 24 h, well-spread colonies were present within 1 to 2 d, and cultures approached confluence within 2 to 3 d. Alkaline phosphatase staining confirmed the microvascular origin of the material plated. Morphology, Factor VIII-related antigen staining and 1,1′-dioctacecyl-3,3,3′3,-tetramethyl-indocarbocyanine perchlorate acetylated low density lipoprotein uptake suggested that MEC predominated. Cultures could be passaged and additionally purified by sequential exposure to pancreatin and trypsin-EDTA. Pancreatin selectively removed MEC colonies leaving a relatively homogeneous pericyte population. The relative ease with which such cultures can be produced should facilitate the in vitro study of brain microvascular function and may also provide insights useful for growing MEC from other vascular beds.

Cultured bovine aortic endothelial cells bind and internalize[125I]insulin and down regulate insulin receptors. Internalized insulin was not degraded significantly and diffused from the endothelial cells. Neither 5-hydroxytryptamine, methylamine, nor dansyl-cadaverine have any observable effect on insulin binding, internalization, metabolism, or down regulation of insulin receptors. Transglutaminase activity, however, is inhibited by 5-hydroxytryptamine and methylamine. These data indicate that transglutaminase is not required for insulin receptor-mediated endocytosis by bovine aortic endothelial cells in culture.

Endothelial cells isolated from bovine aortas without prior treatment with enzymes were cultured in RPMI 1640 medium containing 17% fetal calf serum and antibiotics. The endothelial cells at confluency (7 days) were similar to endothelium in situ or to freshly isolated endothelial cells from blood vessels as seen by light, scanning, and electron microscopy. Cultured and freshly isolated indothelial cells exposed to labeled serotonin, even in the presence of iproniazid (5 x 10-4M), took up approximately 125 and 250 pmoles 14C-serotonin/mg protein, respectively, in 3 hours. Imipramine (10-4M) reduced uptake for both cell groups. Cold (4 degrees C) and metabolic inhibitors sharply reduced serotonin uptake by both freshly isolated and cultured endothelial cells. Ouabain (10-5M) almost completely blocked serotonin transport. Six analogues of serotonin at concentrations ten times above experimental serotonin concentrations did not affect serotonin transport in the cultured endothelial cells but did reduce it in the freshly isolated endothelial cells by 50%. The data on transport suggest that serotonin uptake is not unique to pulmonary endothelium, as has been suggested previously. In addition, using cultured indothelial cells to study serotonin transport is compatible with using other serotonin model systems such as platelets, lung, or brain. Lastly, serotonin uptake by endothelial cells may involve an active transport mechanism similar to that described for the pulmonary circulation, platelets, and insect salivary glands.

The frequency of clones not permanently resistant to azaguanine (AG) was measured in Chinese hamster ovary cells (CHO) grown in vitro by plating them in 7.5 μg/ml AG and isolating a number of clones in the course of 5 experiments. Such isolated clones were propagated to a point at which their resistance to both AG and the reverse selective medium, HAT, could be determined. Out of a total of 13 clones isolated, 4 of these could not be distinguished from the parent CHO line, either on the basis of their growth in a gradient of AG concentrations or the reverse selective HAT medium or on the basis of their mutation frequency to resistance to 30 μg/ml AG. All four of the apparent phenocopies were isolated from plates in which although lower numbers of cells were seeded, a higher frequency of clones able to grow in AG was yielded. This suggests that the higher “mutation” frequencies obtained at lower cell densities are due to the appearance of phenocopies which occur only under these conditions. It is concluded that under low plating density conditions, the lower levels of AG (7.5 μg/ml) are not satisfactory for mutagenesis and mutation rate studies.