Glucocorticoid-mediated inhibition of angiogenic changes in human endothelial cells is not caused by reductions in cell proliferation or migration

Abstract

Background: Glucocorticoid-mediated inhibition of angiogenesis is important in physiology, pathophysiology and therapy. However, the mechanisms through which glucocorticoids inhibit growth of new blood vessels have not been established. This study addresses the hypothesis that physiological levels of glucocorticoids inhibit angiogenesis by directly preventing tube formation by endothelial cells.

Methodology/principal findings: Cultured human umbilical vein (HUVEC) and aortic (HAoEC) endothelial cells were used to determine the influence of glucocorticoids on tube-like structure (TLS) formation, and on cellular proliferation (5-bromo-2'-deoxyuridine (BrdU) incorporation), viability (ATP production) and migration (Boyden chambers). Dexamethasone or cortisol (at physiological concentrations) inhibited both basal and prostaglandin F(2α) (PGF(2α))-induced and vascular endothelial growth factor (VEGF) stimulated TLS formation in endothelial cells (ECs) cultured on Matrigel, effects which were blocked with the glucocorticoid receptor antagonist RU38486. Glucocorticoids had no effect on EC viability, migration or proliferation. Time-lapse imaging showed that cortisol blocked VEGF-stimulated cytoskeletal reorganisation and initialisation of tube formation. Real time PCR suggested that increased expression of thrombospodin-1 contributed to glucocorticoid-mediated inhibition of TLS formation.

Conclusions/significance: We conclude that glucocorticoids interact directly with glucocorticoid receptors on vascular ECs to inhibit TLS formation. This action, which was conserved in ECs from two distinct vascular territories, was due to alterations in cell morphology rather than inhibition of EC viability, migration or proliferation and may be mediated in part by induction of thrombospodin-1. These findings provide important insights into the anti-angiogenic action of endogenous glucocorticoids in health and disease.

Figure 1. Formation of tube-like structures (TLS) and expression of glucocorticoid receptors (GR) in human endothelial cells.

(A) Human umbilical vein endothelial cells (HUVECs) cultured on Matrigel formed a network of tube-like structures (TLSs), after approximately 4 hrs, that retained immunoreactivity for the endothelial cell marker CD31 (original magnification ×10). At higher magnification, cell membranes were evident (arrow head), suggesting development of a lumen in the cell-cell connections (original magnification ×40). No staining was observed in negative controls lacking primary antibody. (B) HUVECs cultured on uncoated cover slips showed clearly-defined cytoskeletal components: filamentous (F)-actin (stained with phalloidin-488; green) and α-tubulin (stained with goat anti-mouse IgG Alexa Fluor 594 secondary antibody; red), and nucleus (DNA stained with 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI); blue). Exposure to cortisol (600 nM; 1 hour) had no apparent effect on microtubule staining but induced a more diffuse and homogeneous distribution of F-actin throughout cell. TLSs stained after 10 hours or 22 hours in culture, consisted of adjoining, filopodia-like extensions, containing both F-actin (green) and α-tubulin (red), connecting neighbouring cells (DNA, blue). (C) GR were detected both in first passage (P1) human umbilical vein (HUVECs) and in passaged human aortic (HAoECs) endothelial cells (P2–P4), by RT-PCR (354 bp product). GR expression was maintained in HUVECs 22 hours after TLS formation. L, Liver (positive control). Negative controls included no reverse transcriptase and no cDNA (not shown).

Figure 2. Cortisol, not cortisone, induces glucocorticoid receptor-dependent inhibition of TLS formation by human umbilical endothelial cells.

Compared with controls, cortisol, but not cortisone, (300–1200 nM) reduced tube-like structure (TLS) formation by human umbilical vein endothelial cells (HUVECs) after 5 hours in culture (A), an effect which achieved significance after 22–24 hours (B). The glucocorticoid receptor (GR)-selective steroid dexamethasone (600 nM) produced a similar reduction in TLS formation whilst the response to cortisol was abolished by GR antagonism with RU38486 (1 µM) (C). Data represent mean±standard error of mean (SEM) (n = 3–6, each condition performed in triplicate) and were analysed using one-way analysis of variance (ANOVA) and Dunnett's post hoc test (*p<0.05, **p<0.01).

Figure 3. Cortisol and dexamethasone inhibit basal and stimulated TLS formation in human aortic endothelial cells.

(A) Cortisol (600 nM) and dexamethasone (600 nM) inhibited basal tube-like structure (TLS) formation of human aortic endothelial cells (HAoEC) after (i) 5 hr and (ii) 22–24 hour treatment. (B) Similarly, growth factor (VEGF)-induced TLS formation by HAoECs was also inhibited after 22–24 hours exposure to cortisol or Dex. Data represent mean±standard error of mean (SEM) (n = 5, each exposure performed in triplicate) and were analysed by one-way analysis of variance (ANOVA) and Dunnett's post hoc test (*p<0.05, **p<0.01).

Figure 4. Glucocorticoids reduce formation and stability of TLSs formed by human umbilical vein endothelial cells grown on Matrigel.

TLS formation by human umbilical vein endothelial cells (HUVECs) was quantified using time-lapse video microscopy. (A) Tube-like structures (TLS) formation (solid arrows) occurred rapidly (0–4 hours) after seeding, stabilised (4–8 hours) and was followed by degradation and detachment (8–24 hours). (B) Cortisol (600 nM) reduced TLS formation (0–4 hours) and TLS stability (2–4 hours). Exposure to vascular endothelial growth factor (VEGF, 10 ng/ml) increased and accelerated TLS formation but did not influence TLS stability. This effect of VEGF was abolished by co-incubation with cortisol (600 nM) (**p<0.01 by repeated measures analysis of variance (ANOVA)). (C) The rate to maximum TLS development was increased by VEGF (*p<0.05 by one-way ANOVA and Dunnett's post hoc test) and this effect was blocked by cortisol. Data represent mean±standard error of mean (SEM) (n = 6).

Figure 5. Cortisol does not inhibit human umbilical vein endothelial cell migration, proliferation or viability.

(A) Human umbilical vein endothelial cell (HUVEC) viability was increased by VEGF (25 ng/ml) and this effect was blocked in a concentration-dependent manner by (1–1000 nM) SU51416 but not by (3–1000 nM) cortisol (B) Similarly, VEGF (25 ng/ml)-stimulated proliferation of cultured HUVECs was blocked in a concentration-dependent manner by (1–1000 nM) SU51416 but not by (3–1000 nM) cortisol. (C) Vascular endothelial growth factor (VEGF) (10 ng/ml; 24 hr 37°C)-stimulated migration of (HUVECs) was abolished by (1 µM) SU51416 but not by (600 nM) cortisol. Data represent mean±standard error of mean (SEM) (n = 6, each condition performed in triplicate); *p<0.05 and **p<0.01 analysed by one-way analysis of variance (ANOVA) and Dunnett's post hoc test.

Figure 6. Cortisol increases expression of thrombospondin 1, but not of VEGF or VEGF receptor 2 during tube-like structure formation by HUVECs.

Effects of cortisol (600 nM) on the expression of (A) vascular endothelial growth factor (VEGF), (B) vascular endothelial growth factor receptor 2 (VEGFR2) and (C) thrombospondin-1 (TSP-1) during tube-like structure (TLS) development by human umbilical vein endothelial cells (HUVECs). Cortisol (600 nM) added to the culture medium at time of cell seeding, did not alter expression of VEGF or VEGFR2 but transiently increased (*p<0.05) TSP-1 expression after 8 hours in culture. Quantitative real-time PCR (QrtPCR) measurements; were expressed as a ratio of concentration of mRNA of gene of interest to internal control (cyclophilin A; n = 8–9, with each condition performed in triplicate).