Basal and angiopoietin-1-mediated endothelial permeability is regulated by sphingosine kinase-1

Abstract

Endothelial cells (ECs) regulate the barrier function of blood vessels. Here we show that basal and angiopoietin-1 (Ang-1)-regulated control of EC permeability is mediated by 2 different functional states of sphingosine kinase-1 (SK-1). Mice depleted of SK-1 have increased vascular leakiness, whereas mice transgenic for SK-1 in ECs show attenuation of leakiness. Furthermore, Ang-1 rapidly and transiently stimulates SK-1 activity and phosphorylation, and induces an increase in intracellular sphingosine-1-phosphate (S1P) concentration. Overexpression of SK-1 resulted in inhibition of permeability similar to that seen for Ang-1, whereas knockdown of SK-1 by small interfering RNA blocked Ang-1-mediated inhibition of permeability. Transfection with SKS225A, a nonphosphorylatable mutant of SK-1, inhibited basal leakiness, and both SKS225A and a dominant-negative SK-1 mutant removed the capacity of Ang-1 to inhibit permeability. These effects were independent of extracellular S1P as knockdown or inhibition of S1P1, S1P2, or S1P3, did not affect the Ang-1 response. Thus, SK-1 levels in ECs powerfully regulate basal permeability in vitro and in vivo. In addition, the Ang-1-induced inhibition of leakiness is mediated through activation of SK-1, defining a new signaling pathway in the Ang-1 regulation of permeability.

Figures

Figure 1

Ang-1 induces SK-1 activation in HUVECs. (A) HUVECs were either untreated or treated with Ang-1 at 0.2 μg/mL for various times. Cells were lysed and SK-1 activity measured. Shown are pooled data from 3 experiments and expressed as fold change in relation to untreated group, where the SK-1 activity was normalized to 1.0 (*P < .01 vs untreated group). (B) HUVECs were untreated (Nil) or treated with Ang-1 at 0.2 μg/mL for 15 minutes. Intracellular S1P levels were determined. Data are mean plus or minus SEM from 3 independent experiments (*P < .001 vs untreated group).

Figure 2

Ang-1 induces SK-1 activation in HUVECs through ERK1/2. (A) HUVECs were infected with adenovirus carrying hSK-1-FLAG. Cells were lysed after no treatment (Nil), treatment with Ang-1 at 0.2 μg/mL for 30 minutes (Ang-1), or pretreatment with U0126 (2 μM for 20 minutes) followed by treatment with Ang-1 (U0126 + Ang-1). Phospho-SK-1 in the top panel and total SK-1 in the bottom panel are shown. (B) HUVECs were untreated (−), or treated with Ang-1 for various times. Phospho-ERK is shown in the top panel and total ERK in the bottom panel. (C) HUVECs were untreated (Nil), treated with Ang-1 at 0.2 μg/mL for 15 minutes (Ang-1), or pretreated with U0126 (2 μM) for 20 minutes and then treated with Ang-1 for 15 minutes (UO126 + Ang-1). The SK-1 activity was measured. Pooled data from 3 experiments are expressed as in Figure 1A (mean ± SEM; *P < .05 vs untreated cells).

Figure 3

SK-1 regulates EC permeability changes. (A) HUVECs were untreated (Nil) or treated with 5 μM DMS for 15 minutes (DMS), Ang-1 (0.2 μg/mL) for 30 minutes (Ang-1), or Ang-1 and DMS (Ang-1 + DMS). Permeability is given as the FITC-dextran passage in 30 minutes. Pooled data from 3 experiments are shown (*P < .01 vs untreated cells). (B) HUVECs were transfected with control siRNA or siRNA against hSK-1. After 48 hours, cells were lysed and SK-1 activity measured. Pooled data from 3 experiments are shown and are expressed as the fold change in relation to control cells where the SK-1 activity was set to 1.0 (*P < .001 vs control siRNA cells). (C) HUVECs were transfected with control siRNA or siRNA against hSK-1. Permeability was measured 48 hours later. Permeability is given as the FITC-dextran passage in 30 minutes. Pooled data from 3 experiments are shown (*P < .01 vs control siRNA cells). (D) HUVECs were infected with adenoviral carrying EV or human SK-1. After 48 hours, SK-1 activity was measured. Pooled data from 3 experiments are shown and are expressed as the fold change in relation to EV cells, which was normalized to 1.0 (*P < .001 vs EV cells). (E) HUVECs were infected with EV or SK-1 in adenovirus. Permeability was measured 72 hours later. Permeability is given as the FITC-dextran passage in 30 minutes. Shown are pooled data from 3 experiments (*P < .01 vs EV cells). (F) HUVECs were infected with EV or SK-1 in adenovirus. After 48 hours, cell lysates were immunoprecipitated with an anti-PECAM-1 antibody. Western blots for phosphotyrosine (top panel) and PECAM-1 (bottom panel) are shown. (G) HUVECs were transfected with control siRNA or siRNA against hSK-1. After 48 hours, cells were lysed and immunoprecipitated with an anti-PECAM-1 antibody. Western blots for phosphotyrosine (top panel) and PECAM-1 (bottom panel) are shown. (H) HUVECs were infected with EV (EV) or SK-1 (SK-1) in adenovirus. Forty-eight hours after infection, the cells were replated onto LabTek slides and washed and fixed 45 minutes after plating. Cells were stained with anti–VE-cadherin antibody. For imaging information, see “VE-cadherin staining” section in “Methods.” All data are mean plus or minus SEM.

Figure 4

SK-1 regulates EC permeability and vascular leakage in vivo. (A) The absorbances of the dye eluted from the injected areas (PBS, VEGF, and histamine) of 10 age- and sex-matched WT mice and 10 SK-1 KO mice (SK-1−/−) were read at 620 nm. The pooled results (mean ± SEM) for each group of mice are shown (*P < .05 vs WT mice injected with PBS). (B) The absorbances of the dye eluted from the injected areas of 8 age- and sex-matched WT mice and 8 SK-1 transgenic mice (SK-1+/+) were read at 620 nm. The pooled results (mean ± SEM) for each group of mice are shown (*P < .05 vs the WT mice injected with PBS).

Figure 5

SK-1 mediates the Ang-1 effects on permeability. (A) HUVECs were transfected with control siRNA or siRNA against hSK-1. After 48 hours, cells were treated with or without Ang-1 for 30 minutes and SK-1 activity measured. Pooled data from 3 experiments are shown and are expressed as the fold change in relation to untreated control cells where the SK-1 activity was set to 1.0 (*P < .001 vs control siRNA cells without treatment). NS indicates no significant difference versus SK-1 siRNA cells without treatment. (B) HUVECs were transfected with control siRNA or siRNA against hSK-1. Forty-eight hours later, cells were untreated or treated with Ang-1 (0.2 μg/mL) for 30 minutes, and permeability is measured. Permeability is given as the FITC-dextran passed after 30 minutes. Pooled data from 3 experiments are shown (mean ± SEM; *P < .01 vs control siRNA cells without treatment). NS, vs SK-1 siRNA cells without treatment. (C) HUVECs were transfected with control siRNA (Ci and Cii) or siRNA against hSK-1 (Ciii and Civ). Forty-eight hours later, cells were untreated (Ci and Ciii) or treated (Cii and Civ) with Ang-1 for 1 hour and then stained for VE-cadherin. In panel Ci, arrow shows diffuse, broad VE-cadherin staining with classic zipper-like pattern indicative of immature junctions. In panel Cii, arrow indicates increased VE-cadherin staining at the junctions, with a more linear staining pattern indicative of mature EC junctions. (D) HUVECs were transfected with control siRNA or siRNA against hSK-1. Forty-eight hours later, cells were untreated (−) or treated (+) with Ang-1 (0.2 μg/mL) for 30 minutes then were lysed and immunoprecipitated with an anti-Tie2 antibody. Western blots for phosphotyrosine (top panel) and Tie2 (bottom panel) are shown. For imaging information, see “VE-cadherin staining” section in “Methods.”

Figure 6

Ang-1 inhibition of EC permeability is independent of the S1P receptor. (A) HUVECs were transfected with control siRNA, siRNA against S1P1 or against S1P3. Seventy-two hours after transfection, cells were untreated (Nil), treated with Ang-1 for 30 minutes (Ang-1), or treated with S1P at 1 μM for 15 minutes (S-1-P) and permeability measured. Permeability is given as the FITC-dextran passage in 30 minutes. The pooled data from 3 experiments are shown (*P < .01 vs untreated cells). NS indicates no significant difference versus untreated cells. (B) HUVECs were untreated (Nil), treated with Ang-1 (Ang-1) 0.2 μg/mL for 30 minutes, S1P (S1P) 1 μM for 15 minutes, JET-013 (JET) 1μM, and VPC 23 019 (VPC) 10 μM for 30 minutes (JET + VPC), or pretreated with JET-013 and VPC 23 019 for 30 minutes then treated with Ang-1 (J + V + Ang) or S1P (J + V + S1P). Permeability is given as the FITC-dextran passage in 30 minutes. Pooled data from 3 experiments are shown (*P < .01 vs untreated cells). All data are mean plus or minus SEM. NS indicates no significant difference versus untreated cells.

Figure 7

Mutation of SK-1 blocks the Ang-1 effects on permeability change. (A) HUVECs were infected with adenoviral carrying EV, SK-1, SKG82D, or SKS225A. After 48 hours, SK-1 activity was measured from cells that were either untreated (−) or treated with Ang-1 (+). Pooled data from 3 experiments are shown, expressed as the fold change in relation to EV untreated cells, normalized to 1.0 (*P < .01 versus untreated EV-infected cells or untreated SK-1-infected cells). NS indicates no significant difference versus untreated SKG82D or untreated SKS225A cells. (B) HUVECs were infected with EV, SK-1, SKG82D, or SKS225A in adenovirus. After 72 hours, cells were either untreated (−) or treated with Ang-1 (+) for 30 minutes and permeability measured. Permeability is given as the FITC-dextran passage in 30 minutes. Shown are the pooled data from 3 experiments (*P < .01 vs nontreated EV-infected cells). NS indicates no significant difference versus untreated SK-1, SKG82D, or untreated SKS225A-infected cells. All data are mean plus or minus SEM. (C) HUVECs were infected with SK-1 or SKS225A in adenovirus. Forty-eight hours after infection, cells were lysed after either no treatment (−) or treatment with Ang-1 (+). Phospho-SK-1 is shown in the top and total SK-1 in the bottom panel.