Nitric oxide mediates tightening of the endothelial barrier by ascorbic acid

Abstract

Vitamin C, or ascorbic acid, decreases paracellular endothelial permeability in a process that requires rearrangement of the actin cytoskeleton. To define the proximal mechanism of this effect, we tested whether it might involve enhanced generation and/or sparing of nitric oxide (NO) by the vitamin. EA.hy926 endothelial cells cultured on semi-porous filter supports showed decreased endothelial barrier permeability to radiolabeled inulin in response to exogenous NO provided by the NO donor spermine NONOATE, as well as to activation of the downstream NO pathway by 8-bromo-cyclic GMP, a cell-penetrant cyclic GMP analog. Inhibition of endothelial nitric oxide synthase (eNOS) with N(ω)-nitro-l-arginine methyl ester increased endothelial permeability, indicating a role constitutive NO generation by eNOS in maintaining the permeability barrier. Inhibition of guanylate cyclase by 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one also increased endothelial permeability and blocked barrier tightening by spermine NONOATE. Loading cells with what are likely physiologic concentrations of ascorbate decreased endothelial permeability. This effect was blocked by inhibition of either eNOS or guanylate cyclase, suggesting that it involved generation of NO by eNOS and subsequent NO-dependent activation of guanylate cyclase. These results show that endothelial permeability barrier function depends on constitutive generation of NO and that ascorbate-dependent tightening of this barrier involves maintaining NO through the eNOS/guanylate cyclase pathway.

Copyright © 2010 Elsevier Inc. All rights reserved.

Figures

Figure 1. eNOS-derived NO decreases endothelial barrier permeability in EA.hy926 cells

Panel A. Cells in culture on porous membrane filters were incubated at 37 °C with increasing concentrations of spermine NONOATE (circles) or 8-bromo-cGMP (squares). Spermine NONOATE was added just before the 60 min inulin transfer assay due to its instability, whereas cGMP was added 30 min before the assay to allow it to enter cells. The transfer assay was carried out as described in Materials and methods. Panel B. Cells were treated at 37 °C with increasing concentrations of L-NAME for 30 min followed by the 60 min inulin transfer assay. Results for each panel are shown from 3–5 experiments, with an “*” indicating p < 0.05 compared to the untreated sample.

Figure 2. Inhibition of guanylate cyclase increases basal endothelial permeability and prevents NO-dependent decreases in permeability

Panel A. EA.hy926 cells cultured on filters were treated for 30 min with the indicated concentrations of ODQ, followed by the radioactive inulin transfer assay. Panel B. Cells were incubated with or without ODQ as indicated for 30 min, spermine NONOATE was added as indicated, and the inulin transfer assay was carried out. Results are shown from 4 experiments for Panel A and from 5 experiments from Panel B, with an “*” indicating p

Figure 3. Intracellular ascorbate decreases endothelial permeability

Panel A: EA.hy926 cells cultured on membrane filters were incubated for 90 min at 37 °C with increasing concentrations of DHA and taken for measurement of intracellular ascorbate. Panel B. Cells incubated with DHA as in Panel A for 30 min were then incubated with radiolabeled inulin for an additional 60 min in the transfer assay. Results are shown from 3 experiments in Panel A and from 4 experiments in Panel B, with an “*” indicating p

Figure 4. Prevention of the ascorbate-dependent decrease in endothelial permeability by inhibition of eNOS or guanylate cyclase

Cells cultured on filters were treated with the indicated concentrations of L-NAME and DHA (Panel A) or of ODQ (Panel B) with or without DHA as indicated for 30 min at 37 °C before addition of radiolabeled inulin and the 60 min transfer assay. Results are shown from 6 experiments for Panel A and from 4 experiments for Panel B, with an “*” indicating p < 0.05 compared to the other treatments in the same panel.