Age-associated stresses induce an anti-inflammatory senescent phenotype in endothelial cells

Abstract

Age is the greatest risk factor for cardiovascular disease. In addition, inflammation and age (senescence) have been linked at both the clinical and molecular levels. In general, senescent cells have been described as pro-inflammatory based on their senescence associated secretory phenotype (SASP). However, we have previously shown that senescence induced by overexpression ofSENEX (or ARHGAP18), in endothelial cells results in an anti-inflammatory phenotype. We have investigated, at the individual cellular level, the senescent phenotype of endothelial cells following three of the chief signals associated with ageing; oxidative stress, disturbed flow and hypoxia. All three stimuli induce senescence and, based on neutrophil adhesion and expression of the adhesion molecules E-selectin and VCAM-1, a population of senescent cells is seen that is resistant to inflammatory stimuli and thus we define as anti-inflammatory. The proportion of anti-inflammatory cells increases with time but remains stable at approximately 50% by eight days after induction of senescence, suggesting that these are stable phenotypes of endothelial cell senescence. Similar to other senescent cell types, p38MAPK blockade inhibits the development of the pro-inflammatory phenotype but unique to EC, there is a corresponding increase in the number of anti-inflammatory senescent cells. Thus stress-induced senescent endothelial cells display a mosaic of inflammatory phenotypes. The anti-inflammatory population suggests that senescent endothelial cells may have an unique protective role, to inhibit uncontrolled proliferation and to limit the local inflammatory response.

Figure 1. Oxidative stress induces senescence in ECs

(A) HUVECs were treated with 0.2mM H₂O₂. After 2 days cells began to demonstrate the morphology of EC senescence (ii-Red outline) compared to untreated cells (i). (B) After a further 2 days the number of senescent cells had increased and the size of the senescent cells had also increased (ii-Red outline), compared to untreated cells (i). This is a representative of 10 HUVEC lines. Bar=220μm. (C) HUVECs were treated with 0.2 mM H₂O₂ and after 4 days stained for SA-β-gal. This is a representative of 10 HUVEC lines. Bar=25μm. (D) HUVECs were untreated (i) or treated with 0.2 mM H₂O₂ and after 4 days stained using immunofluorescence for DAPI (i) and p21 (ii). This is a representative of 5 HUVEC lines. Bar=50μM. (E) and (F) Cells were treated with 0.2 mM H₂O₂ and lysates analysed for levels of γH2A.X (E), p53 and p21 (F). β-Actin was used as a loading control. This is a representative of 8 HUVEC lines.

Figure 2. Induction of EC senescence with disturbed flow

(A) HUVECs were left in static conditions (i) or subjected to 48hrs of flow at 20dyne/cm² (ii) or 2dyne/cm² (iii). Senescent cells are circled in red. This is a representative of 5 HUVEC lines. Bar=100μm (B) Cells exposed to 2dyne/cm² flow for 48hrs were fixed and stained for SA-β-gal. This is a representative of 3 HUVEC lines. Bar=50μM. (C) Cells exposed to 2dyne/cm² flow for 48hrs, fixed and stained for DAPI (blue) and p21 (green). Two representative senescent cells have been highlighted. This is a representative of 3 HUVEC lines. Bar=100μm.

Figure 3. Hypoxia induced senescence

(A) HUVECs were cultured in normoxic conditions (i) or under hypoxic conditions (0.5% oxygen) for 5 days and then stained for SA-β-gal and Eosin counter-staining (ii). Bar=100μm. (B) The number of senescent cells per random field was counted after 5 days treatment at different oxygen tensions. The number of senescent cells at each oxygen tension is normalised to the number of cells in normoxic conditions using matched lines (mean of four experiments per oxygen tension ± SD. * P<0.05, paired Student's t-test). (C) Cell growth was compared at normoxia (squares), 1% oxygen (circles) and 0.5% oxygen (triangle) for 5 days. Cells were seeded 24 hours (Day −1) prior to being placed into the hypoxic incubator on day 0 (mean of four experiments per oxygen tension ± SD. *** P<0.001, paired Student's t-test). (D) HUVECs were untreated (i) or treated with 10μM DFO for 72 hrs (ii) then fixed and stained for SA-β-gal. This is a representative of 3 HUVEC lines. Bar=100μm.

Figure 4. Oxidative stress induced senescence has a unique inflammatory phenotype

(A) Normal HUVECs were stained with cell tracker green for visualization (i and iii) and then neutrophil adhesion was determined by a static adhesion assay. Cells were either untreated (i,ii) or treated with 5ng/ml of TNFα for 5 hours (iii, iv). (B) HUVECs were treated with 0.2 mM H₂O_2. After 4 days the cells were stimulated with 5ng/ml TNFα and stained with cell tracker green for visualization. Senescent cells are highlighted with a yellow line. Representative photos of the pro-inflammatory senescent cells (i) and the anti-inflammatory senescent cells (ii) are shown. This is a representative of 6 HUVEC lines. Bar=25μm. (C) From the photographs taken in (B) the number of anti- and pro-inflammatory senescent cells was determined and given as the % of total senescent cells. This is the representative of the mean +/− SD of 200 senescent cells from 6 HUVEC lines. (D) From videos taken of neutrophil rolling and adhesion (Videos 1-5) the percentage of pro and anti-inflammatory senescent cells after TNFα stimulation was determined. This is a representative of the mean +/− SD of 220 senescent cells from 4 HUVEC lines.

Figure 5. Senescent cells have altered adhesion molecule expression

(A and B) HUVECs were either untreated (A) or stimulated with 5ng/ml TNFα (B) and stained for the surface expression of VCAM-1 (red) (i) and E-selectin (red) (ii) and co-stained with DAPI (blue). This is a representative of 6 HUVEC lines. Bar=100μm. (C) HUVECs were treated with 0.2 mM H₂O₂, for 4 days, stimulated with 5ng/ml TNFα then fixed and stained for the surface expression of VCAM-1 (red) and co-stained with DAPI (blue). Representative pro-inflammatory senescent cells (i) and anti-inflammatory senescent cells (ii) are shown, highlighted by the white outline. This is a representative of 6 HUVEC lines. Bar=100 μm. (D) Cells were treated as in (C) but stained for the surface expression of E-selectin (red) and co-stained with DAPI (blue). This is a representative of 6 HUVEC lines. Bar=100μm. (E) From the photographs taken in (C and D) the number of anti and pro inflammatory senescent cells was determined. This is a representative of the mean percentage +/− SD of 150 senescent cells from 6 HUVEC lines.

Figure 6. The anti-inflammatory phenotype is stable and long lasting

(A) HUVECs were treated with 0.2mM H₂O₂ and after 8 or 16 days fixed and stained for SA-β-gal. This is a representative of 3 HUVEC lines. Bar=100μm. (B) HUVECs were treated with 0.2mM H₂O₂ and after 8 and 16 days fixed and stained for DAPI (blue) and p21 (green). This is a representative of 3 HUVEC lines. Bar=100μm. The percentage of induced pro and anti-inflammatory senescent cells on day 8 (C) and 16 (D) based on E-selectin and VCAM-1 expression were determined. This is a representative of the mean percentage +/− SD of 80 senescent cells from 3 HUVEC lines.

Figure 7. Shear stress and hypoxia induce an anti-inflammatory phenotype

(A) HUVECs were subjected to 48hrs of flow at 2dyne/cm². The cells were then treated with 5ng/ml of TNFα for 5 hours, fixed and stained for E-selectin (red). Induced pro-inflammatory senescent cells (i) and anti-inflammatory senescent cells (ii) are shown. A senescent cell is highlighted by the white outline. This is a representative of 3 HUVEC lines. Bar=100μm. (B) Binding of neutrophils to HUVECs that had been cultured at 0.5% oxygen for five days followed by treatment for six hours with 5ng/ml TNFα (left hand panel). Upper panels show light microscopy of neutrophils upon HUVECs. Lower panels show CMFDA-labelled HUVEC monolayers. Cells with a senescent morphology are indicated with an arrow. Hypoxic treated cells that were not stimulated with TNFα are shown in the left hand panel. Shown are representative images of three experiments using independent HUVEC lines. (C) HUVCs exposed to 10μM DFO for 72 hrs were treated with 5ng/ml of TNFα for 5 hours, fixed and stained for p21 (green), E-selectin (red) and co-stained with DAPI (blue). Representative pro-inflammatory senescent cells (i) and anti-inflammatory senescent cells (ii) are shown, highlighted by the white outline. This is a representative of 3 HUVEC lines. Bar=50μm.

Figure 8. Human coronary artery EC display the anti-inflammatory phenotype

(A) HCAEC were either untreated (i) or treated with 0.2mM H₂O₂ (ii)and after 4 days fixed and stained for SA-β-gal. This is a representative of 3 separate experiments. Bar=100μm. (B) HCAEC were treated with 0.2mM H₂O₂ and after 4 days stimulated with 5ng/ml of TNFα for 5 hours, fixed and stained for p21 (green), E-selectin (red) and co-stained for DAPI (blue). Representative pro-inflammatory (i) and anti-inflammatory senescent cells (ii) are shown, highlighted by the white outline. This is a representative of 3 separate experiments. Bar=100μm.

Figure 9. p38 MAPK activation influences the anti-inflammatory phenotype

(A) HUVECs were treated with either 0.02 mM (+) or 0.2 mM (++) H₂O₂ and lysates analysed for phosphorylated p38 and total p38 levels. β-Actin was used as a loading control. This is a representative of 3 HUVEC lines. (B) HUVECs were untreated or treated with the p38 inhibitor SB203580 (10μM) for 2 hours. The cells are then treated with 0.2 mM H₂O₂ for 4 days, stimulated with TNFα and neutrophil adhesion assessed by the static adhesion assay. The number of anti and inducible pro-inflammatory senescent cells from 300 senescent cells from 6 HUVEC lines is shown. The mean percentage +/−SD is given, * p<0.05 compared to untreated (paired Student's t-test).